Load packages

library(tidyverse)
## ── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
## ✔ dplyr     1.1.2     ✔ readr     2.1.4
## ✔ forcats   1.0.0     ✔ stringr   1.5.0
## ✔ ggplot2   3.4.2     ✔ tibble    3.2.1
## ✔ lubridate 1.9.2     ✔ tidyr     1.3.0
## ✔ purrr     1.0.1     
## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## ✖ dplyr::filter() masks stats::filter()
## ✖ dplyr::lag()    masks stats::lag()
## ℹ Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errors
library(here) 
## here() starts at /Users/sarahdaniels/Desktop/Total_Analyses_4
library(janitor) 
## 
## Attaching package: 'janitor'
## 
## The following objects are masked from 'package:stats':
## 
##     chisq.test, fisher.test
library(haven) 
library(naniar) 
library(ggpubr) 
library(report) 
library(ggplot2)
library(reshape2)
## 
## Attaching package: 'reshape2'
## 
## The following object is masked from 'package:tidyr':
## 
##     smiths
library(lme4)
## Loading required package: Matrix
## 
## Attaching package: 'Matrix'
## 
## The following objects are masked from 'package:tidyr':
## 
##     expand, pack, unpack
library(sjPlot)
## #refugeeswelcome
library(parameters)
library(mediation)
## Loading required package: MASS
## 
## Attaching package: 'MASS'
## 
## The following object is masked from 'package:dplyr':
## 
##     select
## 
## Loading required package: mvtnorm
## Loading required package: sandwich
## mediation: Causal Mediation Analysis
## Version: 4.5.0
library(lavaan)
## This is lavaan 0.6-15
## lavaan is FREE software! Please report any bugs.
library(lmerTest)
## 
## Attaching package: 'lmerTest'
## 
## The following object is masked from 'package:lme4':
## 
##     lmer
## 
## The following object is masked from 'package:stats':
## 
##     step
library(modEvA)
library(report)

Set up

Reading the full data (intervention group, psychoeducation control group, and EC group) and calculating total scores

Full_data_all_t <- read_csv("MI_Data_B1W1M1.csv") %>% 
  rowwise() %>%
  mutate(A_PRE_IUS_total = sum(B_IUS_1, B_IUS_2, B_IUS_3, B_IUS_4, B_IUS_5, B_IUS_6, B_IUS_7, B_IUS_8, B_IUS_9, B_IUS_10, B_IUS_11, B_IUS_12, na.rm = TRUE)) %>%
  mutate(A_PRE_FI_total = sum(B_FI_friends, B_FI_strangers, B_FI_work, B_FI_education, B_FI_hobbies, na.rm = TRUE)) %>% 
  mutate(A_PRE_RTQ_total = sum(B_RTQ_1, B_RTQ_2, B_RTQ_3, B_RTQ_4, B_RTQ_5, B_RTQ_6, B_RTQ_7, B_RTQ_8, B_RTQ_9, B_RTQ_10, na.rm = TRUE)) %>% 
  mutate(A_PRE_ERQ_Rtotal = sum(B_ERQ_1, B_ERQ_3, B_ERQ_5, B_ERQ_7, B_ERQ_8, B_ERQ_10, na.rm = TRUE)) %>% 
  mutate(A_PRE_PHQ_total = sum(B_PHQ_1, B_PHQ_2, B_PHQ_3, B_PHQ_4, B_PHQ_5, B_PHQ_6, B_PHQ_7, B_PHQ_8, na.rm = TRUE)) %>% 
  mutate(A_PRE_GAD_total = sum(B_GAD_1, B_GAD_2, B_GAD_3, B_GAD_4, B_GAD_5, B_GAD_6, B_GAD_7, na.rm = TRUE)) %>% 
  mutate(B_POST_IUS_total = sum(POST_IUS_1, POST_IUS_2, POST_IUS_3, POST_IUS_4, POST_IUS_5, POST_IUS_6, POST_IUS_7, POST_IUS_8, POST_IUS_9, POST_IUS_10, POST_IUS_11, POST_IUS_12, na.rm = TRUE)) %>%
  mutate(C_W1_IUS_total = sum(W1_IUS_1, W1_IUS_2, W1_IUS_3, W1_IUS_4, W1_IUS_5, W1_IUS_6, W1_IUS_7, W1_IUS_8, W1_IUS_9, W1_IUS_10, W1_IUS_11, W1_IUS_12, na.rm = TRUE)) %>%
  mutate(C_W1_FI_total = sum(W1_FI_friends, W1_FI_strangers, W1_FI_work, W1_FI_education, W1_FI_hobbies, na.rm = TRUE)) %>% 
  mutate(C_W1_RTQ_total = sum(W1_RTQ_1, W1_RTQ_2, W1_RTQ_3, W1_RTQ_4, W1_RTQ_5, W1_RTQ_6, W1_RTQ_7, W1_RTQ_8, W1_RTQ_9, W1_RTQ_10, na.rm = TRUE)) %>% 
  mutate(C_W1_ERQ_Rtotal = sum(W1_ERQ_1, W1_ERQ_3, W1_ERQ_5, W1_ERQ_7, W1_ERQ_8, W1_ERQ_10, na.rm = TRUE)) %>% 
  mutate(C_W1_PHQ_total = sum(W1_PHQ_1, W1_PHQ_2, W1_PHQ_3, W1_PHQ_4, W1_PHQ_5, W1_PHQ_6, W1_PHQ_7, W1_PHQ_8, na.rm = TRUE)) %>% 
  mutate(C_W1_GAD_total = sum(W1_GAD_1, W1_GAD_2, W1_GAD_3, W1_GAD_4, W1_GAD_5, W1_GAD_6, W1_GAD_7, na.rm = TRUE)) %>% 
  mutate(D_M1_IUS_total = sum(M1_IUS_1, M1_IUS_2, M1_IUS_3, M1_IUS_4, M1_IUS_5, M1_IUS_6, M1_IUS_7, M1_IUS_8, M1_IUS_9, M1_IUS_10, M1_IUS_11, M1_IUS_12, na.rm = TRUE)) %>%
  mutate(D_M1_FI_total = sum(M1_FI_friends, M1_FI_strangers, M1_FI_work, M1_FI_education, M1_FI_hobbies, na.rm = TRUE)) %>% 
  mutate(D_M1_RTQ_total = sum(M1_RTQ_1, M1_RTQ_2, M1_RTQ_3, M1_RTQ_4, M1_RTQ_5, M1_RTQ_6, M1_RTQ_7, M1_RTQ_8, M1_RTQ_9, M1_RTQ_10, na.rm = TRUE)) %>% 
  mutate(D_M1_ERQ_Rtotal = sum(M1_ERQ_1, M1_ERQ_3, M1_ERQ_5, M1_ERQ_7, M1_ERQ_8, M1_ERQ_10, na.rm = TRUE)) %>% 
  mutate(D_M1_PHQ_total = sum(M1_PHQ_1, M1_PHQ_2, M1_PHQ_3, M1_PHQ_4, M1_PHQ_5, M1_PHQ_6, M1_PHQ_7, M1_PHQ_8, na.rm = TRUE)) %>% 
  mutate(D_M1_GAD_total = sum(M1_GAD_1, M1_GAD_2, M1_GAD_3, M1_GAD_4, M1_GAD_5, M1_GAD_6, M1_GAD_7, na.rm = TRUE)) %>% 
  ungroup()
## New names:
## Rows: 259 Columns: 207
## ── Column specification
## ──────────────────────────────────────────────────────── Delimiter: "," chr
## (2): Prolific_ID, Group dbl (205): ...1, ID, B_IUS_1, B_IUS_2, B_IUS_3,
## B_IUS_4, B_IUS_5, B_IUS_6, B...
## ℹ Use `spec()` to retrieve the full column specification for this data. ℹ
## Specify the column types or set `show_col_types = FALSE` to quiet this message.
## • `` -> `...1`

Calculating mood means

Full_data_all <- mutate(Full_data_all_t, A_PRE_mood_mean = rowMeans(dplyr::select(Full_data_all_t, c(B_distressed_pleasant, B_anxious_relaxed)), na.rm = TRUE)) %>% 
  mutate(Full_data_all_t, B_POST_mood_mean = rowMeans(dplyr::select(Full_data_all_t, c(POST_distressed_pleasant, POST_anxious_relaxed)), na.rm = TRUE)) %>%
  mutate(Full_data_all_t, C_W1_mood_mean = rowMeans(dplyr::select(Full_data_all_t, c(W1_distressed_pleasant, W1_anxious_relaxed)), na.rm = TRUE)) %>% 
 mutate(Full_data_all_t, D_M1_mood_mean = rowMeans(dplyr::select(Full_data_all_t, c(M1_distressed_pleasant, M1_anxious_relaxed)), na.rm = TRUE))

Hypothesis 1

H1a: IUS and Mood association at PRE

#Distressed
PRE_IUS_Distress_lm <- lm(A_PRE_IUS_total ~ B_distressed_pleasant, data = Full_data_all)
summary(PRE_IUS_Distress_lm)
## 
## Call:
## lm(formula = A_PRE_IUS_total ~ B_distressed_pleasant, data = Full_data_all)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -25.8365  -5.8426   0.4123   6.3128  20.1076 
## 
## Coefficients:
##                       Estimate Std. Error t value Pr(>|t|)    
## (Intercept)            44.3702     0.7218  61.471  < 2e-16 ***
## B_distressed_pleasant  -0.0559     0.0129  -4.334 2.11e-05 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 8.571 on 257 degrees of freedom
## Multiple R-squared:  0.0681, Adjusted R-squared:  0.06448 
## F-statistic: 18.78 on 1 and 257 DF,  p-value: 2.105e-05
anova(PRE_IUS_Distress_lm) %>% 
  report()
## The ANOVA suggests that:
## 
##   - The main effect of B_distressed_pleasant is statistically significant and
## medium (F(1, 257) = 18.78, p < .001; Eta2 = 0.07, 95% CI [0.03, 1.00])
## 
## Effect sizes were labelled following Field's (2013) recommendations.
#Anxious
PRE_IUS_Anxiety_lm <- lm(A_PRE_IUS_total ~ B_anxious_relaxed, data = Full_data_all)
summary(PRE_IUS_Anxiety_lm)
## 
## Call:
## lm(formula = A_PRE_IUS_total ~ B_anxious_relaxed, data = Full_data_all)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -26.2289  -5.7679   0.5456   5.9105  20.4901 
## 
## Coefficients:
##                   Estimate Std. Error t value Pr(>|t|)    
## (Intercept)       44.00514    0.61670  71.355  < 2e-16 ***
## B_anxious_relaxed -0.05551    0.01021  -5.437 1.26e-07 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 8.416 on 256 degrees of freedom
##   (1 observation deleted due to missingness)
## Multiple R-squared:  0.1035, Adjusted R-squared:    0.1 
## F-statistic: 29.56 on 1 and 256 DF,  p-value: 1.262e-07
anova(PRE_IUS_Anxiety_lm) %>% 
  report()
## The ANOVA suggests that:
## 
##   - The main effect of B_anxious_relaxed is statistically significant and medium
## (F(1, 256) = 29.56, p < .001; Eta2 = 0.10, 95% CI [0.05, 1.00])
## 
## Effect sizes were labelled following Field's (2013) recommendations.
#Combined
PRE_IUS_mood_lm <- lm(A_PRE_IUS_total ~ A_PRE_mood_mean, data = Full_data_all)
summary(PRE_IUS_mood_lm)
## 
## Call:
## lm(formula = A_PRE_IUS_total ~ A_PRE_mood_mean, data = Full_data_all)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -25.2840  -6.0303   0.1752   6.0332  20.9727 
## 
## Coefficients:
##                 Estimate Std. Error t value Pr(>|t|)    
## (Intercept)     44.57405    0.67547  65.989  < 2e-16 ***
## A_PRE_mood_mean -0.06646    0.01228  -5.412 1.43e-07 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 8.412 on 257 degrees of freedom
## Multiple R-squared:  0.1023, Adjusted R-squared:  0.0988 
## F-statistic: 29.29 on 1 and 257 DF,  p-value: 1.431e-07
anova(PRE_IUS_mood_lm) %>% 
  report()
## The ANOVA suggests that:
## 
##   - The main effect of A_PRE_mood_mean is statistically significant and medium
## (F(1, 257) = 29.29, p < .001; Eta2 = 0.10, 95% CI [0.05, 1.00])
## 
## Effect sizes were labelled following Field's (2013) recommendations.

H1a: IUS and mental health association at PRE

#Depression
PRE_IUS_PHQ_lm <- lm(A_PRE_IUS_total ~ A_PRE_PHQ_total, data = Full_data_all)
summary(PRE_IUS_PHQ_lm)
## 
## Call:
## lm(formula = A_PRE_IUS_total ~ A_PRE_PHQ_total, data = Full_data_all)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -28.9576  -4.5129   0.0901   4.9155  22.1617 
## 
## Coefficients:
##                 Estimate Std. Error t value Pr(>|t|)    
## (Intercept)      35.4890     0.9776  36.303  < 2e-16 ***
## A_PRE_PHQ_total   0.6746     0.0841   8.021 3.72e-14 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 7.94 on 257 degrees of freedom
## Multiple R-squared:  0.2002, Adjusted R-squared:  0.1971 
## F-statistic: 64.34 on 1 and 257 DF,  p-value: 3.716e-14
anova(PRE_IUS_PHQ_lm)
## Analysis of Variance Table
## 
## Response: A_PRE_IUS_total
##                  Df  Sum Sq Mean Sq F value    Pr(>F)    
## A_PRE_PHQ_total   1  4056.3  4056.3  64.344 3.716e-14 ***
## Residuals       257 16201.4    63.0                      
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
#Anxiety
PRE_IUS_GAD_lm <- lm(A_PRE_IUS_total ~ A_PRE_GAD_total,  data = Full_data_all)
summary(PRE_IUS_GAD_lm)
## 
## Call:
## lm(formula = A_PRE_IUS_total ~ A_PRE_GAD_total, data = Full_data_all)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -25.3228  -4.8228   0.7116   4.1772  20.6429 
## 
## Coefficients:
##                 Estimate Std. Error t value Pr(>|t|)    
## (Intercept)     35.04669    0.88936   39.41   <2e-16 ***
## A_PRE_GAD_total  0.82761    0.08639    9.58   <2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 7.621 on 257 degrees of freedom
## Multiple R-squared:  0.2631, Adjusted R-squared:  0.2603 
## F-statistic: 91.78 on 1 and 257 DF,  p-value: < 2.2e-16
anova(PRE_IUS_GAD_lm)
## Analysis of Variance Table
## 
## Response: A_PRE_IUS_total
##                  Df  Sum Sq Mean Sq F value    Pr(>F)    
## A_PRE_GAD_total   1  5330.7  5330.7   91.78 < 2.2e-16 ***
## Residuals       257 14926.9    58.1                      
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

H1b: IUS and BT at PRE

# Adding in groups + excluding participants who only sampled (never made a choice = did not understand task)
BT_PRE_POST <- merge(BT_full,Full_data_all,
                      by=c("ID"),
                      all = TRUE) %>% 
  dplyr::select("ID", "Group", "A_PRE_samples", "B_POST_samples", "A_PRE_IUS_total", "B_POST_IUS_total") %>% 
  filter(ID != "8892522", ID != "8892570", ID != "8892628", ID != "8892668", ID != "8892681", ID != "8892779", ID != "8892794", ID != "8893157", ID != "8893186", ID != "8892873", ID != "9113535", ID != "9113549", ID != "9113550") # excluding those not making a choice

Keeping those who never sample (only make choices)

#Analysis
PRE_IUS_BT_lm <- lm(A_PRE_IUS_total ~ A_PRE_samples, data = BT_PRE_POST)
summary(PRE_IUS_BT_lm)
## 
## Call:
## lm(formula = A_PRE_IUS_total ~ A_PRE_samples, data = BT_PRE_POST)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -29.813  -5.889   1.036   6.280  17.187 
## 
## Coefficients:
##               Estimate Std. Error t value Pr(>|t|)    
## (Intercept)   42.81292    0.61535  69.575  < 2e-16 ***
## A_PRE_samples -0.04991    0.01818  -2.745  0.00651 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 8.754 on 244 degrees of freedom
## Multiple R-squared:  0.02995,    Adjusted R-squared:  0.02598 
## F-statistic: 7.534 on 1 and 244 DF,  p-value: 0.006505
anova(PRE_IUS_BT_lm) %>% 
  report()
## The ANOVA suggests that:
## 
##   - The main effect of A_PRE_samples is statistically significant and small (F(1,
## 244) = 7.53, p = 0.007; Eta2 = 0.03, 95% CI [4.79e-03, 1.00])
## 
## Effect sizes were labelled following Field's (2013) recommendations.

Excluding those who never sample

BT_removed_PRE <- BT_PRE_POST %>% 
  filter(A_PRE_samples != "0") # only excluding from PRE 
#Analysis
PRE_IUS_BT_removed_lm <- lm(A_PRE_IUS_total ~ A_PRE_samples, data = BT_removed_PRE)
summary(PRE_IUS_BT_removed_lm)
## 
## Call:
## lm(formula = A_PRE_IUS_total ~ A_PRE_samples, data = BT_removed_PRE)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -24.606  -5.932   1.014   5.934  15.934 
## 
## Coefficients:
##               Estimate Std. Error t value Pr(>|t|)    
## (Intercept)   43.14794    0.72669  59.376  < 2e-16 ***
## A_PRE_samples -0.05408    0.01801  -3.003  0.00307 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 8.265 on 171 degrees of freedom
## Multiple R-squared:  0.0501, Adjusted R-squared:  0.04454 
## F-statistic: 9.018 on 1 and 171 DF,  p-value: 0.003074
anova(PRE_IUS_BT_removed_lm)
## Analysis of Variance Table
## 
## Response: A_PRE_IUS_total
##                Df  Sum Sq Mean Sq F value   Pr(>F)   
## A_PRE_samples   1   616.1  616.12  9.0184 0.003074 **
## Residuals     171 11682.4   68.32                    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Hypothesis 2

H2a: difference in change in cognitive IUS over time between groups

IUS_alltimepoints <- Full_data_all %>% 
  dplyr::select("ID", "Group", "A_PRE_IUS_total", "B_POST_IUS_total", "C_W1_IUS_total", "D_M1_IUS_total")
## Formatting table as needed
IUS_alltimepoints_long <- IUS_alltimepoints %>%
  pivot_longer(cols = c(A_PRE_IUS_total, B_POST_IUS_total, C_W1_IUS_total, D_M1_IUS_total),
               names_to = "Time",
               values_to = "IUS_Score")
IUS_MEM <- lmer(IUS_Score ~ Group * Time + (1|ID), data = IUS_alltimepoints_long, REML = TRUE)
summary(IUS_MEM)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: IUS_Score ~ Group * Time + (1 | ID)
##    Data: IUS_alltimepoints_long
## 
## REML criterion at convergence: 7807.9
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -4.4802 -0.3254  0.0609  0.4737  2.3749 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev.
##  ID       (Intercept) 72.64    8.523   
##  Residual             77.08    8.779   
## Number of obs: 1036, groups:  ID, 259
## 
## Fixed effects:
##                                          Estimate Std. Error       df t value
## (Intercept)                               41.0800     1.7304 600.1493  23.740
## GroupB_Controls                            0.9483     2.0993 600.1493   0.452
## GroupC_Intervention                        1.9880     2.1090 600.1493   0.943
## TimeB_POST_IUS_total                      -0.2800     1.7559 768.0000  -0.159
## TimeC_W1_IUS_total                        -0.7200     1.7559 768.0000  -0.410
## TimeD_M1_IUS_total                        -2.9000     1.7559 768.0000  -1.652
## GroupB_Controls:TimeB_POST_IUS_total      -3.5879     2.1301 768.0000  -1.684
## GroupC_Intervention:TimeB_POST_IUS_total  -6.7200     2.1400 768.0000  -3.140
## GroupB_Controls:TimeC_W1_IUS_total        -1.8272     2.1301 768.0000  -0.858
## GroupC_Intervention:TimeC_W1_IUS_total    -4.6878     2.1400 768.0000  -2.191
## GroupB_Controls:TimeD_M1_IUS_total        -3.2415     2.1301 768.0000  -1.522
## GroupC_Intervention:TimeD_M1_IUS_total    -6.2553     2.1400 768.0000  -2.923
##                                          Pr(>|t|)    
## (Intercept)                               < 2e-16 ***
## GroupB_Controls                           0.65162    
## GroupC_Intervention                       0.34627    
## TimeB_POST_IUS_total                      0.87334    
## TimeC_W1_IUS_total                        0.68188    
## TimeD_M1_IUS_total                        0.09902 .  
## GroupB_Controls:TimeB_POST_IUS_total      0.09251 .  
## GroupC_Intervention:TimeB_POST_IUS_total  0.00175 ** 
## GroupB_Controls:TimeC_W1_IUS_total        0.39128    
## GroupC_Intervention:TimeC_W1_IUS_total    0.02879 *  
## GroupB_Controls:TimeD_M1_IUS_total        0.12848    
## GroupC_Intervention:TimeD_M1_IUS_total    0.00357 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) GrpB_C GrpC_I TB_POS TC_W1_ TD_M1_ GB_C:TB GC_I:TB GB_C:TC
## GrpB_Cntrls -0.824                                                           
## GrpC_Intrvn -0.820  0.676                                                    
## TB_POST_IUS -0.507  0.418  0.416                                             
## TmC_W1_IUS_ -0.507  0.418  0.416  0.500                                      
## TmD_M1_IUS_ -0.507  0.418  0.416  0.500  0.500                               
## GB_C:TB_POS  0.418 -0.507 -0.343 -0.824 -0.412 -0.412                        
## GC_I:TB_POS  0.416 -0.343 -0.507 -0.820 -0.410 -0.410  0.676                 
## GB_C:TC_W1_  0.418 -0.507 -0.343 -0.412 -0.824 -0.412  0.500   0.338         
## GC_I:TC_W1_  0.416 -0.343 -0.507 -0.410 -0.820 -0.410  0.338   0.500   0.676 
## GB_C:TD_M1_  0.418 -0.507 -0.343 -0.412 -0.412 -0.824  0.500   0.338   0.500 
## GC_I:TD_M1_  0.416 -0.343 -0.507 -0.410 -0.410 -0.820  0.338   0.500   0.338 
##             GC_I:TC GB_C:TD
## GrpB_Cntrls                
## GrpC_Intrvn                
## TB_POST_IUS                
## TmC_W1_IUS_                
## TmD_M1_IUS_                
## GB_C:TB_POS                
## GC_I:TB_POS                
## GB_C:TC_W1_                
## GC_I:TC_W1_                
## GB_C:TD_M1_  0.338         
## GC_I:TD_M1_  0.500   0.676
anova  (IUS_MEM)
## Type III Analysis of Variance Table with Satterthwaite's method
##            Sum Sq Mean Sq NumDF DenDF F value    Pr(>F)    
## Group       175.6   87.79     2   256  1.1390   0.32176    
## Time       4324.7 1441.58     3   768 18.7031 1.022e-11 ***
## Group:Time 1012.0  168.66     6   768  2.1882   0.04221 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
sjPlot::tab_model(IUS_MEM)
  IUS Score
Predictors Estimates CI p
(Intercept) 41.08 37.68 – 44.48 <0.001
Group [B_Controls] 0.95 -3.17 – 5.07 0.652
Group [C_Intervention] 1.99 -2.15 – 6.13 0.346
Time [B_POST_IUS_total] -0.28 -3.73 – 3.17 0.873
Time [C_W1_IUS_total] -0.72 -4.17 – 2.73 0.682
Time [D_M1_IUS_total] -2.90 -6.35 – 0.55 0.099
Group [B_Controls] × Time
[B_POST_IUS_total] 		-3.59 -7.77 – 0.59 0.092
Group [C_Intervention] ×
Time [B_POST_IUS_total] 		-6.72 -10.92 – -2.52 0.002
Group [B_Controls] × Time
[C_W1_IUS_total] 		-1.83 -6.01 – 2.35 0.391
Group [C_Intervention] ×
Time [C_W1_IUS_total] 		-4.69 -8.89 – -0.49 0.029
Group [B_Controls] × Time
[D_M1_IUS_total] 		-3.24 -7.42 – 0.94 0.128
Group [C_Intervention] ×
Time [D_M1_IUS_total] 		-6.26 -10.45 – -2.06 0.004
Random Effects
σ2 77.08
Ï„00 ID 72.64
ICC 0.49
N ID 259
Observations 1036
Marginal R2 / Conditional R2 0.049 / 0.510 
parameters::standardise_parameters(IUS_MEM)
## # Standardization method: refit
## 
## Parameter                                | Std. Coef. |         95% CI
## ----------------------------------------------------------------------
## (Intercept)                              |       0.20 | [-0.08,  0.47]
## GroupB_Controls                          |       0.08 | [-0.25,  0.41]
## GroupC_Intervention                      |       0.16 | [-0.17,  0.49]
## TimeB_POST_IUS_total                     |      -0.02 | [-0.30,  0.25]
## TimeC_W1_IUS_total                       |      -0.06 | [-0.33,  0.22]
## TimeD_M1_IUS_total                       |      -0.23 | [-0.51,  0.04]
## GroupB_Controls:TimeB_POST_IUS_total     |      -0.29 | [-0.62,  0.05]
## GroupC_Intervention:TimeB_POST_IUS_total |      -0.54 | [-0.87, -0.20]
## GroupB_Controls:TimeC_W1_IUS_total       |      -0.15 | [-0.48,  0.19]
## GroupC_Intervention:TimeC_W1_IUS_total   |      -0.38 | [-0.71, -0.04]
## GroupB_Controls:TimeD_M1_IUS_total       |      -0.26 | [-0.59,  0.08]
## GroupC_Intervention:TimeD_M1_IUS_total   |      -0.50 | [-0.84, -0.16]
report(IUS_MEM)
## We fitted a linear mixed model (estimated using REML and nloptwrap optimizer)
## to predict IUS_Score with Group and Time (formula: IUS_Score ~ Group * Time).
## The model included ID as random effect (formula: ~1 | ID). The model's total
## explanatory power is substantial (conditional R2 = 0.51) and the part related
## to the fixed effects alone (marginal R2) is of 0.05. The model's intercept,
## corresponding to Group = A_ECs and Time = A_PRE_IUS_total, is at 41.08 (95% CI
## [37.68, 44.48], t(1022) = 23.74, p < .001). Within this model:
## 
##   - The effect of Group [B_Controls] is statistically non-significant and
## positive (beta = 0.95, 95% CI [-3.17, 5.07], t(1022) = 0.45, p = 0.652; Std.
## beta = 0.08, 95% CI [-0.25, 0.41])
##   - The effect of Group [C_Intervention] is statistically non-significant and
## positive (beta = 1.99, 95% CI [-2.15, 6.13], t(1022) = 0.94, p = 0.346; Std.
## beta = 0.16, 95% CI [-0.17, 0.49])
##   - The effect of Time [B_POST_IUS_total] is statistically non-significant and
## negative (beta = -0.28, 95% CI [-3.73, 3.17], t(1022) = -0.16, p = 0.873; Std.
## beta = -0.02, 95% CI [-0.30, 0.25])
##   - The effect of Time [C_W1_IUS_total] is statistically non-significant and
## negative (beta = -0.72, 95% CI [-4.17, 2.73], t(1022) = -0.41, p = 0.682; Std.
## beta = -0.06, 95% CI [-0.33, 0.22])
##   - The effect of Time [D_M1_IUS_total] is statistically non-significant and
## negative (beta = -2.90, 95% CI [-6.35, 0.55], t(1022) = -1.65, p = 0.099; Std.
## beta = -0.23, 95% CI [-0.51, 0.04])
##   - The effect of Group [B_Controls] × Time [B_POST_IUS_total] is statistically
## non-significant and negative (beta = -3.59, 95% CI [-7.77, 0.59], t(1022) =
## -1.68, p = 0.092; Std. beta = -0.29, 95% CI [-0.62, 0.05])
##   - The effect of Group [C_Intervention] × Time [B_POST_IUS_total] is
## statistically significant and negative (beta = -6.72, 95% CI [-10.92, -2.52],
## t(1022) = -3.14, p = 0.002; Std. beta = -0.54, 95% CI [-0.87, -0.20])
##   - The effect of Group [B_Controls] × Time [C_W1_IUS_total] is statistically
## non-significant and negative (beta = -1.83, 95% CI [-6.01, 2.35], t(1022) =
## -0.86, p = 0.391; Std. beta = -0.15, 95% CI [-0.48, 0.19])
##   - The effect of Group [C_Intervention] × Time [C_W1_IUS_total] is statistically
## significant and negative (beta = -4.69, 95% CI [-8.89, -0.49], t(1022) = -2.19,
## p = 0.029; Std. beta = -0.38, 95% CI [-0.71, -0.04])
##   - The effect of Group [B_Controls] × Time [D_M1_IUS_total] is statistically
## non-significant and negative (beta = -3.24, 95% CI [-7.42, 0.94], t(1022) =
## -1.52, p = 0.128; Std. beta = -0.26, 95% CI [-0.59, 0.08])
##   - The effect of Group [C_Intervention] × Time [D_M1_IUS_total] is statistically
## significant and negative (beta = -6.26, 95% CI [-10.45, -2.06], t(1022) =
## -2.92, p = 0.004; Std. beta = -0.50, 95% CI [-0.84, -0.16])
## 
## Standardized parameters were obtained by fitting the model on a standardized
## version of the dataset. 95% Confidence Intervals (CIs) and p-values were
## computed using a Wald t-distribution approximation.

Baseline to post

IUS_BP <- Full_data_all %>% 
  dplyr::select("ID", "Group", "A_PRE_IUS_total", "B_POST_IUS_total")
## Formatting table as needed
IUS_BP_long <- IUS_BP %>%
  pivot_longer(cols = c(A_PRE_IUS_total, B_POST_IUS_total),
               names_to = "Time",
               values_to = "IUS_Score")
IUS_MEM_BP <- lmer(IUS_Score ~ Group * Time + (1|ID), data = IUS_BP_long, REML = TRUE)
summary(IUS_MEM_BP)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: IUS_Score ~ Group * Time + (1 | ID)
##    Data: IUS_BP_long
## 
## REML criterion at convergence: 3656.7
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -4.2320 -0.4094  0.0034  0.4135  3.2300 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev.
##  ID       (Intercept) 71.59    8.461   
##  Residual             28.73    5.360   
## Number of obs: 518, groups:  ID, 259
## 
## Fixed effects:
##                                          Estimate Std. Error       df t value
## (Intercept)                               41.0800     1.4165 339.2570  29.002
## GroupB_Controls                            0.9483     1.7184 339.2570   0.552
## GroupC_Intervention                        1.9880     1.7264 339.2570   1.152
## TimeB_POST_IUS_total                      -0.2800     1.0721 256.0000  -0.261
## GroupB_Controls:TimeB_POST_IUS_total      -3.5879     1.3006 256.0000  -2.759
## GroupC_Intervention:TimeB_POST_IUS_total  -6.7200     1.3067 256.0000  -5.143
##                                          Pr(>|t|)    
## (Intercept)                               < 2e-16 ***
## GroupB_Controls                           0.58141    
## GroupC_Intervention                       0.25033    
## TimeB_POST_IUS_total                      0.79417    
## GroupB_Controls:TimeB_POST_IUS_total      0.00622 ** 
## GroupC_Intervention:TimeB_POST_IUS_total 5.38e-07 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) GrpB_C GrpC_I TB_POS GB_C:T
## GrpB_Cntrls -0.824                            
## GrpC_Intrvn -0.820  0.676                     
## TB_POST_IUS -0.378  0.312  0.311              
## GB_C:TB_POS  0.312 -0.378 -0.256 -0.824       
## GC_I:TB_POS  0.311 -0.256 -0.378 -0.820  0.676
anova  (IUS_MEM_BP)
## Type III Analysis of Variance Table with Satterthwaite's method
##             Sum Sq Mean Sq NumDF DenDF F value    Pr(>F)    
## Group        21.31   10.66     2   256  0.3708    0.6905    
## Time       1587.48 1587.48     1   256 55.2457 1.595e-12 ***
## Group:Time  787.45  393.72     2   256 13.7020 2.222e-06 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
sjPlot::tab_model(IUS_MEM_BP)
  IUS Score
Predictors Estimates CI p
(Intercept) 41.08 38.30 – 43.86 <0.001
Group [B_Controls] 0.95 -2.43 – 4.32 0.581
Group [C_Intervention] 1.99 -1.40 – 5.38 0.250
Time [B_POST_IUS_total] -0.28 -2.39 – 1.83 0.794
Group [B_Controls] × Time
[B_POST_IUS_total] 		-3.59 -6.14 – -1.03 0.006
Group [C_Intervention] ×
Time [B_POST_IUS_total] 		-6.72 -9.29 – -4.15 <0.001
Random Effects
σ2 28.73
Ï„00 ID 71.59
ICC 0.71
N ID 259
Observations 518
Marginal R2 / Conditional R2 0.062 / 0.731 
parameters::standardise_parameters(IUS_MEM_BP)
## # Standardization method: refit
## 
## Parameter                                | Std. Coef. |         95% CI
## ----------------------------------------------------------------------
## (Intercept)                              |       0.10 | [-0.17,  0.37]
## GroupB_Controls                          |       0.09 | [-0.24,  0.42]
## GroupC_Intervention                      |       0.19 | [-0.14,  0.52]
## TimeB_POST_IUS_total                     |      -0.03 | [-0.23,  0.18]
## GroupB_Controls:TimeB_POST_IUS_total     |      -0.35 | [-0.60, -0.10]
## GroupC_Intervention:TimeB_POST_IUS_total |      -0.65 | [-0.90, -0.40]
plot_model(IUS_MEM_BP, type = "int")

Baseline to 1W

IUS_B1W <- Full_data_all %>% 
  dplyr::select("ID", "Group", "A_PRE_IUS_total", "C_W1_IUS_total")
## Formatting table as needed
IUS_B1W_long <- IUS_B1W %>%
  pivot_longer(cols = c(A_PRE_IUS_total, C_W1_IUS_total),
               names_to = "Time",
               values_to = "IUS_Score")
IUS_MEM_B1W <- lmer(IUS_Score ~ Group * Time + (1|ID), data = IUS_B1W_long, REML = TRUE)
summary(IUS_MEM_B1W)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: IUS_Score ~ Group * Time + (1 | ID)
##    Data: IUS_B1W_long
## 
## REML criterion at convergence: 3746.6
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -3.9845 -0.3439  0.0235  0.4237  2.5991 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev.
##  ID       (Intercept) 69.60    8.342   
##  Residual             39.31    6.270   
## Number of obs: 518, groups:  ID, 259
## 
## Fixed effects:
##                                        Estimate Std. Error       df t value
## (Intercept)                             41.0800     1.4759 363.5479  27.834
## GroupB_Controls                          0.9483     1.7904 363.5479   0.530
## GroupC_Intervention                      1.9880     1.7988 363.5479   1.105
## TimeC_W1_IUS_total                      -0.7200     1.2540 256.0000  -0.574
## GroupB_Controls:TimeC_W1_IUS_total      -1.8272     1.5213 256.0000  -1.201
## GroupC_Intervention:TimeC_W1_IUS_total  -4.6878     1.5284 256.0000  -3.067
##                                        Pr(>|t|)    
## (Intercept)                             < 2e-16 ***
## GroupB_Controls                         0.59668    
## GroupC_Intervention                     0.26981    
## TimeC_W1_IUS_total                      0.56637    
## GroupB_Controls:TimeC_W1_IUS_total      0.23084    
## GroupC_Intervention:TimeC_W1_IUS_total  0.00239 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) GrpB_C GrpC_I TC_W1_ GB_C:T
## GrpB_Cntrls -0.824                            
## GrpC_Intrvn -0.820  0.676                     
## TmC_W1_IUS_ -0.425  0.350  0.349              
## GB_C:TC_W1_  0.350 -0.425 -0.287 -0.824       
## GC_I:TC_W1_  0.349 -0.287 -0.425 -0.820  0.676
anova  (IUS_MEM_B1W)
## Type III Analysis of Variance Table with Satterthwaite's method
##            Sum Sq Mean Sq NumDF DenDF F value    Pr(>F)    
## Group        3.95    1.98     2   256  0.0503  0.950967    
## Time       961.28  961.28     1   256 24.4509 1.381e-06 ***
## Group:Time 425.11  212.56     2   256  5.4065  0.005014 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
sjPlot::tab_model(IUS_MEM_B1W)
  IUS Score
Predictors Estimates CI p
(Intercept) 41.08 38.18 – 43.98 <0.001
Group [B_Controls] 0.95 -2.57 – 4.47 0.597
Group [C_Intervention] 1.99 -1.55 – 5.52 0.270
Time [C_W1_IUS_total] -0.72 -3.18 – 1.74 0.566
Group [B_Controls] × Time
[C_W1_IUS_total] 		-1.83 -4.82 – 1.16 0.230
Group [C_Intervention] ×
Time [C_W1_IUS_total] 		-4.69 -7.69 – -1.69 0.002
Random Effects
σ2 39.31
Ï„00 ID 69.60
ICC 0.64
N ID 259
Observations 518
Marginal R2 / Conditional R2 0.032 / 0.651 
parameters::standardise_parameters(IUS_MEM_B1W)
## # Standardization method: refit
## 
## Parameter                              | Std. Coef. |         95% CI
## --------------------------------------------------------------------
## (Intercept)                            |       0.05 | [-0.23,  0.32]
## GroupB_Controls                        |       0.09 | [-0.24,  0.42]
## GroupC_Intervention                    |       0.19 | [-0.15,  0.52]
## TimeC_W1_IUS_total                     |      -0.07 | [-0.30,  0.17]
## GroupB_Controls:TimeC_W1_IUS_total     |      -0.17 | [-0.46,  0.11]
## GroupC_Intervention:TimeC_W1_IUS_total |      -0.44 | [-0.73, -0.16]
plot_model(IUS_MEM_B1W, type = "int")

Baseline to 1M

IUS_B1M <- Full_data_all %>% 
  dplyr::select("ID", "Group", "A_PRE_IUS_total", "D_M1_IUS_total")
## Formatting table as needed
IUS_B1M_long <- IUS_B1M %>%
  pivot_longer(cols = c(A_PRE_IUS_total, D_M1_IUS_total),
               names_to = "Time",
               values_to = "IUS_Score")
IUS_MEM_B1M <- lmer(IUS_Score ~ Group * Time + (1|ID), data = IUS_B1M_long, REML = TRUE)
summary(IUS_MEM_B1M)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: IUS_Score ~ Group * Time + (1 | ID)
##    Data: IUS_B1M_long
## 
## REML criterion at convergence: 4070.9
## 
## Scaled residuals: 
##      Min       1Q   Median       3Q      Max 
## -2.92417 -0.36721  0.08337  0.54180  1.88810 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev.
##  ID       (Intercept)  59.84    7.735  
##  Residual             108.95   10.438  
## Number of obs: 518, groups:  ID, 259
## 
## Fixed effects:
##                                        Estimate Std. Error       df t value
## (Intercept)                             41.0800     1.8373 454.8351  22.359
## GroupB_Controls                          0.9483     2.2289 454.8351   0.425
## GroupC_Intervention                      1.9880     2.2393 454.8351   0.888
## TimeD_M1_IUS_total                      -2.9000     2.0876 256.0000  -1.389
## GroupB_Controls:TimeD_M1_IUS_total      -3.2415     2.5325 256.0000  -1.280
## GroupC_Intervention:TimeD_M1_IUS_total  -6.2553     2.5443 256.0000  -2.459
##                                        Pr(>|t|)    
## (Intercept)                              <2e-16 ***
## GroupB_Controls                          0.6707    
## GroupC_Intervention                      0.3751    
## TimeD_M1_IUS_total                       0.1660    
## GroupB_Controls:TimeD_M1_IUS_total       0.2017    
## GroupC_Intervention:TimeD_M1_IUS_total   0.0146 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) GrpB_C GrpC_I TD_M1_ GB_C:T
## GrpB_Cntrls -0.824                            
## GrpC_Intrvn -0.820  0.676                     
## TmD_M1_IUS_ -0.568  0.468  0.466              
## GB_C:TD_M1_  0.468 -0.568 -0.384 -0.824       
## GC_I:TD_M1_  0.466 -0.384 -0.568 -0.820  0.676
anova  (IUS_MEM_B1M)
## Type III Analysis of Variance Table with Satterthwaite's method
##            Sum Sq Mean Sq NumDF DenDF F value    Pr(>F)    
## Group        42.2    21.1     2   256  0.1937    0.8240    
## Time       4229.7  4229.7     1   256 38.8236 1.896e-09 ***
## Group:Time  688.0   344.0     2   256  3.1574    0.0442 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
sjPlot::tab_model(IUS_MEM_B1M)
  IUS Score
Predictors Estimates CI p
(Intercept) 41.08 37.47 – 44.69 <0.001
Group [B_Controls] 0.95 -3.43 – 5.33 0.671
Group [C_Intervention] 1.99 -2.41 – 6.39 0.375
Time [D_M1_IUS_total] -2.90 -7.00 – 1.20 0.165
Group [B_Controls] × Time
[D_M1_IUS_total] 		-3.24 -8.22 – 1.73 0.201
Group [C_Intervention] ×
Time [D_M1_IUS_total] 		-6.26 -11.25 – -1.26 0.014
Random Effects
σ2 108.95
Ï„00 ID 59.84
ICC 0.35
N ID 259
Observations 518
Marginal R2 / Conditional R2 0.070 / 0.400 
parameters::standardise_parameters(IUS_MEM_B1M)
## # Standardization method: refit
## 
## Parameter                              | Std. Coef. |         95% CI
## --------------------------------------------------------------------
## (Intercept)                            |       0.16 | [-0.11,  0.43]
## GroupB_Controls                        |       0.07 | [-0.26,  0.40]
## GroupC_Intervention                    |       0.15 | [-0.18,  0.48]
## TimeD_M1_IUS_total                     |      -0.22 | [-0.52,  0.09]
## GroupB_Controls:TimeD_M1_IUS_total     |      -0.24 | [-0.61,  0.13]
## GroupC_Intervention:TimeD_M1_IUS_total |      -0.47 | [-0.84, -0.09]
plot_model(IUS_MEM_B1M, type = "int")

H2a: difference in change in behavioural IUS over time between groups (baseline to post)

Not excluding never samplers

# Excluding them at pre and post
BT_BP <- BT_PRE_POST %>%
  dplyr::select("ID", "Group", "A_PRE_samples", "B_POST_samples")
## Formatting table as needed
BT_BP_long <- BT_BP %>%
  pivot_longer(cols = c(A_PRE_samples, B_POST_samples),
               names_to = "Time",
               values_to = "BT_Score")
BT_MEM_BP <- lmer(BT_Score ~ Group * Time + (1|ID), data = BT_BP_long, REML = TRUE)
summary(BT_MEM_BP)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: BT_Score ~ Group * Time + (1 | ID)
##    Data: BT_BP_long
## 
## REML criterion at convergence: 4395.3
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -4.9264 -0.2398 -0.0766  0.0958  8.1574 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev.
##  ID       (Intercept) 354      18.81   
##  Residual             263      16.22   
## Number of obs: 488, groups:  ID, 246
## 
## Fixed effects:
##                                        Estimate Std. Error       df t value
## (Intercept)                             13.5745     3.6233 364.5272   3.746
## GroupB_Controls                          3.6055     4.3930 364.5272   0.821
## GroupC_Intervention                     -1.9583     4.4001 364.5272  -0.445
## TimeB_POST_samples                      -0.5106     3.3457 240.3829  -0.153
## GroupB_Controls:TimeB_POST_samples      -7.2867     4.0668 240.9058  -1.792
## GroupC_Intervention:TimeB_POST_samples  -2.4873     4.0736 240.9148  -0.611
##                                        Pr(>|t|)    
## (Intercept)                            0.000208 ***
## GroupB_Controls                        0.412330    
## GroupC_Intervention                    0.656542    
## TimeB_POST_samples                     0.878821    
## GroupB_Controls:TimeB_POST_samples     0.074429 .  
## GroupC_Intervention:TimeB_POST_samples 0.542042    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) GrpB_C GrpC_I TB_POS GB_C:T
## GrpB_Cntrls -0.825                            
## GrpC_Intrvn -0.823  0.679                     
## TmB_POST_sm -0.462  0.381  0.380              
## GB_C:TB_POS  0.380 -0.461 -0.313 -0.823       
## GC_I:TB_POS  0.379 -0.313 -0.460 -0.821  0.676
anova  (BT_MEM_BP)
## Type III Analysis of Variance Table with Satterthwaite's method
##             Sum Sq Mean Sq NumDF  DenDF F value  Pr(>F)  
## Group       322.16  161.08     2 243.67  0.6124 0.54290  
## Time       1532.59 1532.59     1 241.18  5.8262 0.01653 *
## Group:Time 1018.76  509.38     2 241.35  1.9364 0.14645  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
sjPlot::tab_model(BT_MEM_BP)
  BT Score
Predictors Estimates CI p
(Intercept) 13.57 6.45 – 20.69 <0.001
Group [B_Controls] 3.61 -5.03 – 12.24 0.412
Group [C_Intervention] -1.96 -10.60 – 6.69 0.656
Time [B_POST_samples] -0.51 -7.08 – 6.06 0.879
Group [B_Controls] × Time
[B_POST_samples] 		-7.29 -15.28 – 0.70 0.074
Group [C_Intervention] ×
Time [B_POST_samples] 		-2.49 -10.49 – 5.52 0.542
Random Effects
σ2 263.05
Ï„00 ID 353.98
ICC 0.57
N ID 246
Observations 488
Marginal R2 / Conditional R2 0.015 / 0.580 
parameters::standardise_parameters(BT_MEM_BP)
## # Standardization method: refit
## 
## Parameter                              | Std. Coef. |        95% CI
## -------------------------------------------------------------------
## (Intercept)                            |       0.06 | [-0.23, 0.35]
## GroupB_Controls                        |       0.14 | [-0.20, 0.49]
## GroupC_Intervention                    |      -0.08 | [-0.43, 0.27]
## TimeB_POST_samples                     |      -0.02 | [-0.28, 0.24]
## GroupB_Controls:TimeB_POST_samples     |      -0.29 | [-0.61, 0.03]
## GroupC_Intervention:TimeB_POST_samples |      -0.10 | [-0.42, 0.22]
plot_model(BT_MEM_BP, type = "int")

Excluding never samplers at PRE

# Excluding them at pre and post
BT_BP_removed <- BT_PRE_POST %>% 
  filter(A_PRE_samples != "0") %>% # only excluding from PRE
  dplyr::select("ID", "Group", "A_PRE_samples", "B_POST_samples")
## Formatting table as needed
BT_BP_long_removed <- BT_BP_removed %>%
  pivot_longer(cols = c(A_PRE_samples, B_POST_samples),
               names_to = "Time",
               values_to = "BT_Score")
BT_MEM_BP_removed <- lmer(BT_Score ~ Group * Time + (1|ID), data = BT_BP_long_removed, REML = TRUE)
summary(BT_MEM_BP_removed)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: BT_Score ~ Group * Time + (1 | ID)
##    Data: BT_BP_long_removed
## 
## REML criterion at convergence: 3169.1
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -4.0812 -0.2761 -0.0683  0.1403  6.8856 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev.
##  ID       (Intercept) 429.4    20.72   
##  Residual             358.4    18.93   
## Number of obs: 343, groups:  ID, 173
## 
## Fixed effects:
##                                        Estimate Std. Error       df t value
## (Intercept)                             16.7895     4.5533 261.0959   3.687
## GroupB_Controls                          9.2408     5.7157 261.0959   1.617
## GroupC_Intervention                     -0.1228     5.6702 261.0959  -0.022
## TimeB_POST_samples                      -1.1053     4.3434 168.0021  -0.254
## GroupB_Controls:TimeB_POST_samples     -11.2692     5.4762 168.6705  -2.058
## GroupC_Intervention:TimeB_POST_samples  -3.4718     5.4196 168.3090  -0.641
##                                        Pr(>|t|)    
## (Intercept)                            0.000276 ***
## GroupB_Controls                        0.107143    
## GroupC_Intervention                    0.982737    
## TimeB_POST_samples                     0.799442    
## GroupB_Controls:TimeB_POST_samples     0.041143 *  
## GroupC_Intervention:TimeB_POST_samples 0.522658    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) GrpB_C GrpC_I TB_POS GB_C:T
## GrpB_Cntrls -0.797                            
## GrpC_Intrvn -0.803  0.640                     
## TmB_POST_sm -0.477  0.380  0.383              
## GB_C:TB_POS  0.378 -0.475 -0.304 -0.793       
## GC_I:TB_POS  0.382 -0.304 -0.476 -0.801  0.636
anova  (BT_MEM_BP_removed)
## Type III Analysis of Variance Table with Satterthwaite's method
##             Sum Sq Mean Sq NumDF  DenDF F value   Pr(>F)   
## Group       601.59  300.79     2 170.47  0.8392 0.433833   
## Time       2885.84 2885.84     1 168.72  8.0513 0.005105 **
## Group:Time 1785.79  892.90     2 168.82  2.4911 0.085856 . 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
sjPlot::tab_model(BT_MEM_BP_removed)
  BT Score
Predictors Estimates CI p
(Intercept) 16.79 7.83 – 25.75 <0.001
Group [B_Controls] 9.24 -2.00 – 20.48 0.107
Group [C_Intervention] -0.12 -11.28 – 11.03 0.983
Time [B_POST_samples] -1.11 -9.65 – 7.44 0.799
Group [B_Controls] × Time
[B_POST_samples] 		-11.27 -22.04 – -0.50 0.040
Group [C_Intervention] ×
Time [B_POST_samples] 		-3.47 -14.13 – 7.19 0.522
Random Effects
σ2 358.43
Ï„00 ID 429.41
ICC 0.55
N ID 173
Observations 343
Marginal R2 / Conditional R2 0.028 / 0.558 
parameters::standardise_parameters(BT_MEM_BP_removed)
## # Standardization method: refit
## 
## Parameter                              | Std. Coef. |         95% CI
## --------------------------------------------------------------------
## (Intercept)                            |  -6.00e-03 | [-0.32,  0.31]
## GroupB_Controls                        |       0.33 | [-0.07,  0.72]
## GroupC_Intervention                    |  -4.34e-03 | [-0.40,  0.39]
## TimeB_POST_samples                     |      -0.04 | [-0.34,  0.26]
## GroupB_Controls:TimeB_POST_samples     |      -0.40 | [-0.78, -0.02]
## GroupC_Intervention:TimeB_POST_samples |      -0.12 | [-0.50,  0.25]
plot_model(BT_MEM_BP_removed, type = "int")

Excluding never samplers at both PRE and POST

# Excluding them at pre and post
BT_BP_removed_both <- BT_PRE_POST %>% 
  filter(A_PRE_samples != "0", B_POST_samples != "0") %>% # only excluding from PRE
  dplyr::select("ID", "Group", "A_PRE_samples", "B_POST_samples")
## Formatting table as needed
BT_BP_long_removed_both <- BT_BP_removed_both %>%
  pivot_longer(cols = c(A_PRE_samples, B_POST_samples),
               names_to = "Time",
               values_to = "BT_Score")
BT_MEM_BP_removed_both <- lmer(BT_Score ~ Group * Time + (1|ID), data = BT_BP_long_removed_both, REML = TRUE)
summary(BT_MEM_BP_removed_both)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: BT_Score ~ Group * Time + (1 | ID)
##    Data: BT_BP_long_removed_both
## 
## REML criterion at convergence: 2294.4
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -3.8069 -0.2375 -0.0513  0.1451  5.7469 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev.
##  ID       (Intercept) 545.0    23.35   
##  Residual             206.9    14.38   
## Number of obs: 256, groups:  ID, 128
## 
## Fixed effects:
##                                        Estimate Std. Error      df t value
## (Intercept)                              19.250      4.848 163.890   3.971
## GroupB_Controls                           4.990      6.208 163.890   0.804
## GroupC_Intervention                       3.685      6.312 163.890   0.584
## TimeB_POST_samples                       -0.625      3.596 125.000  -0.174
## GroupB_Controls:TimeB_POST_samples       -5.895      4.605 125.000  -1.280
## GroupC_Intervention:TimeB_POST_samples   -4.288      4.683 125.000  -0.916
##                                        Pr(>|t|)    
## (Intercept)                            0.000107 ***
## GroupB_Controls                        0.422661    
## GroupC_Intervention                    0.560191    
## TimeB_POST_samples                     0.862299    
## GroupB_Controls:TimeB_POST_samples     0.202877    
## GroupC_Intervention:TimeB_POST_samples 0.361562    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) GrpB_C GrpC_I TB_POS GB_C:T
## GrpB_Cntrls -0.781                            
## GrpC_Intrvn -0.768  0.600                     
## TmB_POST_sm -0.371  0.290  0.285              
## GB_C:TB_POS  0.290 -0.371 -0.222 -0.781       
## GC_I:TB_POS  0.285 -0.222 -0.371 -0.768  0.600
anova  (BT_MEM_BP_removed_both)
## Type III Analysis of Variance Table with Satterthwaite's method
##            Sum Sq Mean Sq NumDF DenDF F value  Pr(>F)  
## Group       26.79   13.40     2   125  0.0647 0.93734  
## Time       996.01  996.01     1   125  4.8141 0.03008 *
## Group:Time 346.12  173.06     2   125  0.8365 0.43565  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
sjPlot::tab_model(BT_MEM_BP_removed_both)
  BT Score
Predictors Estimates CI p
(Intercept) 19.25 9.70 – 28.80 <0.001
Group [B_Controls] 4.99 -7.24 – 17.22 0.422
Group [C_Intervention] 3.68 -8.75 – 16.12 0.560
Time [B_POST_samples] -0.62 -7.71 – 6.46 0.862
Group [B_Controls] × Time
[B_POST_samples] 		-5.90 -14.97 – 3.18 0.202
Group [C_Intervention] ×
Time [B_POST_samples] 		-4.29 -13.51 – 4.93 0.361
Random Effects
σ2 206.89
Ï„00 ID 545.04
ICC 0.72
N ID 128
Observations 256
Marginal R2 / Conditional R2 0.009 / 0.727 
parameters::standardise_parameters(BT_MEM_BP_removed_both)
## # Standardization method: refit
## 
## Parameter                              | Std. Coef. |        95% CI
## -------------------------------------------------------------------
## (Intercept)                            |      -0.04 | [-0.39, 0.31]
## GroupB_Controls                        |       0.18 | [-0.27, 0.63]
## GroupC_Intervention                    |       0.14 | [-0.32, 0.59]
## TimeB_POST_samples                     |      -0.02 | [-0.28, 0.24]
## GroupB_Controls:TimeB_POST_samples     |      -0.22 | [-0.55, 0.12]
## GroupC_Intervention:TimeB_POST_samples |      -0.16 | [-0.50, 0.18]
plot_model(BT_MEM_BP_removed_both, type = "int")

H2a: difference in change in growth mindsets over time between groups

GM_alltimepoints <- Full_data_all %>% 
  dplyr::select("ID", "Group", "A_PRE_GM", "B_POST_GM", "C_W1_GM", "D_M1_GM")
## Formatting table as needed
GM_alltimepoints_long <- GM_alltimepoints %>%
  pivot_longer(cols = c(A_PRE_GM, B_POST_GM, C_W1_GM, D_M1_GM),
               names_to = "Time",
               values_to = "GM_Score")
GM_MEM <- lmer(GM_Score ~ Group * Time + (1|ID), data = GM_alltimepoints_long, REML = TRUE)
summary(GM_MEM)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: GM_Score ~ Group * Time + (1 | ID)
##    Data: GM_alltimepoints_long
## 
## REML criterion at convergence: 3110.4
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -3.1305 -0.5023 -0.0609  0.4554  3.6922 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev.
##  ID       (Intercept) 1.3098   1.1445  
##  Residual             0.7768   0.8814  
## Number of obs: 996, groups:  ID, 259
## 
## Fixed effects:
##                                    Estimate Std. Error        df t value
## (Intercept)                         2.74000    0.20429 463.07671  13.413
## GroupB_Controls                     0.40151    0.24783 463.07671   1.620
## GroupC_Intervention                 0.16291    0.24898 463.07671   0.654
## TimeB_POST_GM                       0.04000    0.17627 728.13424   0.227
## TimeC_W1_GM                        -0.07589    0.17861 729.70994  -0.425
## TimeD_M1_GM                        -0.03987    0.18380 733.00453  -0.217
## GroupB_Controls:TimeB_POST_GM      -0.53057    0.21384 728.13424  -2.481
## GroupC_Intervention:TimeB_POST_GM  -0.71126    0.21528 728.38399  -3.304
## GroupB_Controls:TimeC_W1_GM        -0.23713    0.21621 729.54890  -1.097
## GroupC_Intervention:TimeC_W1_GM    -0.51437    0.21744 729.67920  -2.366
## GroupB_Controls:TimeD_M1_GM        -0.35002    0.22306 733.13394  -1.569
## GroupC_Intervention:TimeD_M1_GM    -0.58867    0.22393 733.05960  -2.629
##                                   Pr(>|t|)    
## (Intercept)                        < 2e-16 ***
## GroupB_Controls                    0.10589    
## GroupC_Intervention                0.51323    
## TimeB_POST_GM                      0.82055    
## TimeC_W1_GM                        0.67105    
## TimeD_M1_GM                        0.82834    
## GroupB_Controls:TimeB_POST_GM      0.01332 *  
## GroupC_Intervention:TimeB_POST_GM  0.00100 ** 
## GroupB_Controls:TimeC_W1_GM        0.27310    
## GroupC_Intervention:TimeC_W1_GM    0.01826 *  
## GroupB_Controls:TimeD_M1_GM        0.11704    
## GroupC_Intervention:TimeD_M1_GM    0.00875 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) GrpB_C GrpC_I TB_POS TC_W1_ TD_M1_ GB_C:TB GC_I:TB GB_C:TC
## GrpB_Cntrls -0.824                                                           
## GrpC_Intrvn -0.820  0.676                                                    
## TmB_POST_GM -0.431  0.356  0.354                                             
## TimeC_W1_GM -0.426  0.351  0.349  0.493                                      
## TimeD_M1_GM -0.414  0.341  0.340  0.480  0.472                               
## GB_C:TB_POS  0.356 -0.431 -0.292 -0.824 -0.407 -0.395                        
## GC_I:TB_POS  0.353 -0.291 -0.431 -0.819 -0.404 -0.393  0.675                 
## GB_C:TC_W1_  0.352 -0.427 -0.289 -0.408 -0.826 -0.390  0.495   0.334         
## GC_I:TC_W1_  0.350 -0.288 -0.426 -0.405 -0.821 -0.388  0.334   0.493   0.679 
## GB_C:TD_M1_  0.341 -0.414 -0.280 -0.395 -0.389 -0.824  0.479   0.324   0.474 
## GC_I:TD_M1_  0.340 -0.280 -0.414 -0.394 -0.388 -0.821  0.324   0.479   0.320 
##             GC_I:TC GB_C:TD
## GrpB_Cntrls                
## GrpC_Intrvn                
## TmB_POST_GM                
## TimeC_W1_GM                
## TimeD_M1_GM                
## GB_C:TB_POS                
## GC_I:TB_POS                
## GB_C:TC_W1_                
## GC_I:TC_W1_                
## GB_C:TD_M1_  0.320         
## GC_I:TD_M1_  0.474   0.676
anova  (GM_MEM)
## Type III Analysis of Variance Table with Satterthwaite's method
##             Sum Sq Mean Sq NumDF  DenDF F value    Pr(>F)    
## Group       4.6999  2.3499     2 256.22  3.0251   0.05029 .  
## Time       21.1646  7.0549     3 731.30  9.0817 6.584e-06 ***
## Group:Time 10.6523  1.7754     6 731.29  2.2855   0.03414 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
sjPlot::tab_model(GM_MEM)
  GM Score
Predictors Estimates CI p
(Intercept) 2.74 2.34 – 3.14 <0.001
Group [B_Controls] 0.40 -0.08 – 0.89 0.106
Group [C_Intervention] 0.16 -0.33 – 0.65 0.513
Time [B_POST_GM] 0.04 -0.31 – 0.39 0.821
Time [C_W1_GM] -0.08 -0.43 – 0.27 0.671
Time [D_M1_GM] -0.04 -0.40 – 0.32 0.828
Group [B_Controls] × Time
[B_POST_GM] 		-0.53 -0.95 – -0.11 0.013
Group [C_Intervention] ×
Time [B_POST_GM] 		-0.71 -1.13 – -0.29 0.001
Group [B_Controls] × Time
[C_W1_GM] 		-0.24 -0.66 – 0.19 0.273
Group [C_Intervention] ×
Time [C_W1_GM] 		-0.51 -0.94 – -0.09 0.018
Group [B_Controls] × Time
[D_M1_GM] 		-0.35 -0.79 – 0.09 0.117
Group [C_Intervention] ×
Time [D_M1_GM] 		-0.59 -1.03 – -0.15 0.009
Random Effects
σ2 0.78
Ï„00 ID 1.31
ICC 0.63
N ID 259
Observations 996
Marginal R2 / Conditional R2 0.037 / 0.641 
parameters::standardise_parameters(GM_MEM)
## # Standardization method: refit
## 
## Parameter                         | Std. Coef. |         95% CI
## ---------------------------------------------------------------
## (Intercept)                       |       0.05 | [-0.22,  0.33]
## GroupB_Controls                   |       0.27 | [-0.06,  0.61]
## GroupC_Intervention               |       0.11 | [-0.22,  0.44]
## TimeB_POST_GM                     |       0.03 | [-0.21,  0.26]
## TimeC_W1_GM                       |      -0.05 | [-0.29,  0.19]
## TimeD_M1_GM                       |      -0.03 | [-0.27,  0.22]
## GroupB_Controls:TimeB_POST_GM     |      -0.36 | [-0.65, -0.08]
## GroupC_Intervention:TimeB_POST_GM |      -0.49 | [-0.77, -0.20]
## GroupB_Controls:TimeC_W1_GM       |      -0.16 | [-0.45,  0.13]
## GroupC_Intervention:TimeC_W1_GM   |      -0.35 | [-0.64, -0.06]
## GroupB_Controls:TimeD_M1_GM       |      -0.24 | [-0.54,  0.06]
## GroupC_Intervention:TimeD_M1_GM   |      -0.40 | [-0.70, -0.10]

Baseline to post

GM_BP <- Full_data_all %>% 
  dplyr::select("ID", "Group", "A_PRE_GM", "B_POST_GM")
## Formatting table as needed
GM_BP_long <- GM_BP %>%
  pivot_longer(cols = c(A_PRE_GM, B_POST_GM),
               names_to = "Time",
               values_to = "GM_Score")
GM_MEM_BP <- lmer(GM_Score ~ Group * Time + (1|ID), data = GM_BP_long, REML = TRUE)
summary(GM_MEM_BP)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: GM_Score ~ Group * Time + (1 | ID)
##    Data: GM_BP_long
## 
## REML criterion at convergence: 1673.5
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -2.5473 -0.4356 -0.0320  0.4087  2.9611 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev.
##  ID       (Intercept) 1.471    1.2128  
##  Residual             0.614    0.7836  
## Number of obs: 516, groups:  ID, 259
## 
## Fixed effects:
##                                   Estimate Std. Error       df t value Pr(>|t|)
## (Intercept)                         2.7400     0.2042 341.6189  13.418  < 2e-16
## GroupB_Controls                     0.4015     0.2477 341.6189   1.621 0.105989
## GroupC_Intervention                 0.1629     0.2489 341.6189   0.655 0.513178
## TimeB_POST_GM                       0.0400     0.1567 254.5431   0.255 0.798745
## GroupB_Controls:TimeB_POST_GM      -0.5306     0.1901 254.5431  -2.791 0.005657
## GroupC_Intervention:TimeB_POST_GM  -0.7237     0.1915 254.9544  -3.779 0.000196
##                                      
## (Intercept)                       ***
## GroupB_Controls                      
## GroupC_Intervention                  
## TimeB_POST_GM                        
## GroupB_Controls:TimeB_POST_GM     ** 
## GroupC_Intervention:TimeB_POST_GM ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) GrpB_C GrpC_I TB_POS GB_C:T
## GrpB_Cntrls -0.824                            
## GrpC_Intrvn -0.820  0.676                     
## TmB_POST_GM -0.384  0.316  0.315              
## GB_C:TB_POS  0.316 -0.384 -0.260 -0.824       
## GC_I:TB_POS  0.314 -0.259 -0.383 -0.818  0.674
anova  (GM_MEM_BP)
## Type III Analysis of Variance Table with Satterthwaite's method
##             Sum Sq Mean Sq NumDF  DenDF F value    Pr(>F)    
## Group       2.0352  1.0176     2 256.34  1.6574 0.1926771    
## Time       16.3667 16.3667     1 254.86 26.6558   4.9e-07 ***
## Group:Time  8.8331  4.4165     2 254.92  7.1931 0.0009142 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
sjPlot::tab_model(GM_MEM_BP)
  GM Score
Predictors Estimates CI p
(Intercept) 2.74 2.34 – 3.14 <0.001
Group [B_Controls] 0.40 -0.09 – 0.89 0.106
Group [C_Intervention] 0.16 -0.33 – 0.65 0.513
Time [B_POST_GM] 0.04 -0.27 – 0.35 0.799
Group [B_Controls] × Time
[B_POST_GM] 		-0.53 -0.90 – -0.16 0.005
Group [C_Intervention] ×
Time [B_POST_GM] 		-0.72 -1.10 – -0.35 <0.001
Random Effects
σ2 0.61
Ï„00 ID 1.47
ICC 0.71
N ID 259
Observations 516
Marginal R2 / Conditional R2 0.043 / 0.718 
parameters::standardise_parameters(GM_MEM_BP)
## # Standardization method: refit
## 
## Parameter                         | Std. Coef. |         95% CI
## ---------------------------------------------------------------
## (Intercept)                       |   1.11e-03 | [-0.27,  0.27]
## GroupB_Controls                   |       0.27 | [-0.06,  0.60]
## GroupC_Intervention               |       0.11 | [-0.22,  0.44]
## TimeB_POST_GM                     |       0.03 | [-0.18,  0.24]
## GroupB_Controls:TimeB_POST_GM     |      -0.36 | [-0.61, -0.11]
## GroupC_Intervention:TimeB_POST_GM |      -0.49 | [-0.75, -0.24]
plot_model(GM_MEM_BP, type = "int")

Baseline to 1W

GM_B1W <- Full_data_all %>% 
  dplyr::select("ID", "Group", "A_PRE_GM", "C_W1_GM")
## Formatting table as needed
GM_B1W_long <- GM_B1W %>%
  pivot_longer(cols = c(A_PRE_GM, C_W1_GM),
               names_to = "Time",
               values_to = "GM_Score")
GM_MEM_B1W <- lmer(GM_Score ~ Group * Time + (1|ID), data = GM_B1W_long, REML = TRUE)
summary(GM_MEM_B1W)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: GM_Score ~ Group * Time + (1 | ID)
##    Data: GM_B1W_long
## 
## REML criterion at convergence: 1747.5
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -2.2664 -0.5109 -0.1518  0.4940  2.7870 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev.
##  ID       (Intercept) 1.038    1.019   
##  Residual             1.010    1.005   
## Number of obs: 511, groups:  ID, 259
## 
## Fixed effects:
##                                 Estimate Std. Error       df t value Pr(>|t|)
## (Intercept)                       2.7400     0.2024 403.7332  13.539   <2e-16
## GroupB_Controls                   0.4015     0.2455 403.7332   1.635   0.1027
## GroupC_Intervention               0.1629     0.2467 403.7332   0.661   0.5093
## TimeC_W1_GM                      -0.0784     0.2041 253.9941  -0.384   0.7012
## GroupB_Controls:TimeC_W1_GM      -0.2376     0.2470 253.3522  -0.962   0.3370
## GroupC_Intervention:TimeC_W1_GM  -0.5133     0.2484 253.6544  -2.066   0.0398
##                                    
## (Intercept)                     ***
## GroupB_Controls                    
## GroupC_Intervention                
## TimeC_W1_GM                        
## GroupB_Controls:TimeC_W1_GM        
## GroupC_Intervention:TimeC_W1_GM *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) GrpB_C GrpC_I TC_W1_ GB_C:T
## GrpB_Cntrls -0.824                            
## GrpC_Intrvn -0.820  0.676                     
## TimeC_W1_GM -0.489  0.403  0.401              
## GB_C:TC_W1_  0.404 -0.490 -0.332 -0.826       
## GC_I:TC_W1_  0.402 -0.331 -0.490 -0.822  0.679
anova  (GM_MEM_B1W)
## Type III Analysis of Variance Table with Satterthwaite's method
##             Sum Sq Mean Sq NumDF  DenDF F value    Pr(>F)    
## Group       5.0659  2.5329     2 256.72  2.5080 0.0834210 .  
## Time       12.1159 12.1159     1 253.25 11.9969 0.0006256 ***
## Group:Time  4.6763  2.3381     2 253.08  2.3152 0.1008383    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
sjPlot::tab_model(GM_MEM_B1W)
  GM Score
Predictors Estimates CI p
(Intercept) 2.74 2.34 – 3.14 <0.001
Group [B_Controls] 0.40 -0.08 – 0.88 0.103
Group [C_Intervention] 0.16 -0.32 – 0.65 0.509
Time [C_W1_GM] -0.08 -0.48 – 0.32 0.701
Group [B_Controls] × Time
[C_W1_GM] 		-0.24 -0.72 – 0.25 0.337
Group [C_Intervention] ×
Time [C_W1_GM] 		-0.51 -1.00 – -0.03 0.039
Random Effects
σ2 1.01
Ï„00 ID 1.04
ICC 0.51
N ID 259
Observations 511
Marginal R2 / Conditional R2 0.035 / 0.524 
parameters::standardise_parameters(GM_MEM_B1W)
## # Standardization method: refit
## 
## Parameter                       | Std. Coef. |         95% CI
## -------------------------------------------------------------
## (Intercept)                     |      -0.03 | [-0.31,  0.24]
## GroupB_Controls                 |       0.28 | [-0.06,  0.61]
## GroupC_Intervention             |       0.11 | [-0.22,  0.45]
## TimeC_W1_GM                     |      -0.05 | [-0.33,  0.22]
## GroupB_Controls:TimeC_W1_GM     |      -0.16 | [-0.50,  0.17]
## GroupC_Intervention:TimeC_W1_GM |      -0.35 | [-0.69, -0.02]
plot_model(GM_MEM_B1W, type = "int")

Baseline to 1M

GM_B1M <- Full_data_all %>% 
  dplyr::select("ID", "Group", "A_PRE_GM", "D_M1_GM")
## Formatting table as needed
GM_B1M_long <- GM_B1M %>%
  pivot_longer(cols = c(A_PRE_GM, D_M1_GM),
               names_to = "Time",
               values_to = "GM_Score")
GM_MEM_B1M <- lmer(GM_Score ~ Group * Time + (1|ID), data = GM_B1M_long, REML = TRUE)
summary(GM_MEM_B1M)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: GM_Score ~ Group * Time + (1 | ID)
##    Data: GM_B1M_long
## 
## REML criterion at convergence: 1639.3
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -2.3875 -0.5367 -0.1142  0.4999  2.8661 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev.
##  ID       (Intercept) 1.0920   1.0450  
##  Residual             0.8918   0.9444  
## Number of obs: 487, groups:  ID, 259
## 
## Fixed effects:
##                                  Estimate Std. Error        df t value Pr(>|t|)
## (Intercept)                       2.74000    0.19919 375.54541  13.756   <2e-16
## GroupB_Controls                   0.40151    0.24164 375.54541   1.662   0.0974
## GroupC_Intervention               0.16291    0.24277 375.54541   0.671   0.5026
## TimeD_M1_GM                      -0.03867    0.19861 235.43449  -0.195   0.8458
## GroupB_Controls:TimeD_M1_GM      -0.30728    0.24103 235.51016  -1.275   0.2036
## GroupC_Intervention:TimeD_M1_GM  -0.59125    0.24193 235.33298  -2.444   0.0153
##                                    
## (Intercept)                     ***
## GroupB_Controls                 .  
## GroupC_Intervention                
## TimeD_M1_GM                        
## GroupB_Controls:TimeD_M1_GM        
## GroupC_Intervention:TimeD_M1_GM *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) GrpB_C GrpC_I TD_M1_ GB_C:T
## GrpB_Cntrls -0.824                            
## GrpC_Intrvn -0.820  0.676                     
## TimeD_M1_GM -0.451  0.372  0.370              
## GB_C:TD_M1_  0.372 -0.451 -0.305 -0.824       
## GC_I:TD_M1_  0.370 -0.305 -0.451 -0.821  0.676
anova  (GM_MEM_B1M)
## Type III Analysis of Variance Table with Satterthwaite's method
##             Sum Sq Mean Sq NumDF  DenDF F value   Pr(>F)    
## Group       4.3400  2.1700     2 254.05  2.4332 0.089803 .  
## Time       11.8930 11.8930     1 235.41 13.3353 0.000321 ***
## Group:Time  5.5615  2.7807     2 235.41  3.1180 0.046079 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
sjPlot::tab_model(GM_MEM_B1M)
  GM Score
Predictors Estimates CI p
(Intercept) 2.74 2.35 – 3.13 <0.001
Group [B_Controls] 0.40 -0.07 – 0.88 0.097
Group [C_Intervention] 0.16 -0.31 – 0.64 0.503
Time [D_M1_GM] -0.04 -0.43 – 0.35 0.846
Group [B_Controls] × Time
[D_M1_GM] 		-0.31 -0.78 – 0.17 0.203
Group [C_Intervention] ×
Time [D_M1_GM] 		-0.59 -1.07 – -0.12 0.015
Random Effects
σ2 0.89
Ï„00 ID 1.09
ICC 0.55
N ID 259
Observations 487
Marginal R2 / Conditional R2 0.039 / 0.568 
parameters::standardise_parameters(GM_MEM_B1M)
## # Standardization method: refit
## 
## Parameter                       | Std. Coef. |         95% CI
## -------------------------------------------------------------
## (Intercept)                     |      -0.03 | [-0.30,  0.25]
## GroupB_Controls                 |       0.28 | [-0.05,  0.61]
## GroupC_Intervention             |       0.11 | [-0.22,  0.45]
## TimeD_M1_GM                     |      -0.03 | [-0.30,  0.25]
## GroupB_Controls:TimeD_M1_GM     |      -0.21 | [-0.55,  0.12]
## GroupC_Intervention:TimeD_M1_GM |      -0.41 | [-0.75, -0.08]
plot_model(GM_MEM_B1M, type = "int")

H2b: difference in change in PHQ over time between groups

PHQ_alltimepoints <- Full_data_all %>% 
  dplyr::select("ID", "Group", "A_PRE_PHQ_total", "C_W1_PHQ_total", "D_M1_PHQ_total")
## Formatting table as needed
PHQ_alltimepoints_long <- PHQ_alltimepoints %>%
  pivot_longer(cols = c(A_PRE_PHQ_total, C_W1_PHQ_total, D_M1_PHQ_total),
               names_to = "Time",
               values_to = "PHQ_Score")
PHQ_MEM <- lmer(PHQ_Score ~ Group * Time + (1|ID), data = PHQ_alltimepoints_long, REML = TRUE)
summary(PHQ_MEM)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: PHQ_Score ~ Group * Time + (1 | ID)
##    Data: PHQ_alltimepoints_long
## 
## REML criterion at convergence: 4789.8
## 
## Scaled residuals: 
##      Min       1Q   Median       3Q      Max 
## -2.90257 -0.51396 -0.08277  0.47380  2.91348 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev.
##  ID       (Intercept) 21.01    4.584   
##  Residual             16.94    4.116   
## Number of obs: 777, groups:  ID, 259
## 
## Fixed effects:
##                                        Estimate Std. Error       df t value
## (Intercept)                              9.9600     0.8713 476.1441  11.431
## GroupB_Controls                         -0.5166     1.0570 476.1441  -0.489
## GroupC_Intervention                      0.7196     1.0619 476.1441   0.678
## TimeC_W1_PHQ_total                      -0.4400     0.8233 512.0000  -0.534
## TimeD_M1_PHQ_total                      -1.0600     0.8233 512.0000  -1.288
## GroupB_Controls:TimeC_W1_PHQ_total      -0.6449     0.9987 512.0000  -0.646
## GroupC_Intervention:TimeC_W1_PHQ_total  -1.2882     1.0034 512.0000  -1.284
## GroupB_Controls:TimeD_M1_PHQ_total      -1.1570     0.9987 512.0000  -1.158
## GroupC_Intervention:TimeD_M1_PHQ_total  -2.2313     1.0034 512.0000  -2.224
##                                        Pr(>|t|)    
## (Intercept)                              <2e-16 ***
## GroupB_Controls                          0.6252    
## GroupC_Intervention                      0.4983    
## TimeC_W1_PHQ_total                       0.5933    
## TimeD_M1_PHQ_total                       0.1985    
## GroupB_Controls:TimeC_W1_PHQ_total       0.5187    
## GroupC_Intervention:TimeC_W1_PHQ_total   0.1998    
## GroupB_Controls:TimeD_M1_PHQ_total       0.2472    
## GroupC_Intervention:TimeD_M1_PHQ_total   0.0266 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) GrpB_C GrpC_I TC_W1_ TD_M1_ GB_C:TC GC_I:TC GB_C:TD
## GrpB_Cntrls -0.824                                                    
## GrpC_Intrvn -0.820  0.676                                             
## TmC_W1_PHQ_ -0.472  0.389  0.388                                      
## TmD_M1_PHQ_ -0.472  0.389  0.388  0.500                               
## GB_C:TC_W1_  0.389 -0.472 -0.320 -0.824 -0.412                        
## GC_I:TC_W1_  0.388 -0.320 -0.472 -0.820 -0.410  0.676                 
## GB_C:TD_M1_  0.389 -0.472 -0.320 -0.412 -0.824  0.500   0.338         
## GC_I:TD_M1_  0.388 -0.320 -0.472 -0.410 -0.820  0.338   0.500   0.676
anova  (PHQ_MEM)
## Type III Analysis of Variance Table with Satterthwaite's method
##            Sum Sq Mean Sq NumDF DenDF F value    Pr(>F)    
## Group       30.63  15.317     2   256  0.9039    0.4063    
## Time       551.10 275.550     2   512 16.2622 1.421e-07 ***
## Group:Time  88.29  22.072     4   512  1.3026    0.2679    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
sjPlot::tab_model(PHQ_MEM)
  PHQ Score
Predictors Estimates CI p
(Intercept) 9.96 8.25 – 11.67 <0.001
Group [B_Controls] -0.52 -2.59 – 1.56 0.625
Group [C_Intervention] 0.72 -1.37 – 2.80 0.498
Time [C_W1_PHQ_total] -0.44 -2.06 – 1.18 0.593
Time [D_M1_PHQ_total] -1.06 -2.68 – 0.56 0.198
Group [B_Controls] × Time
[C_W1_PHQ_total] 		-0.64 -2.61 – 1.32 0.519
Group [C_Intervention] ×
Time [C_W1_PHQ_total] 		-1.29 -3.26 – 0.68 0.200
Group [B_Controls] × Time
[D_M1_PHQ_total] 		-1.16 -3.12 – 0.80 0.247
Group [C_Intervention] ×
Time [D_M1_PHQ_total] 		-2.23 -4.20 – -0.26 0.026
Random Effects
σ2 16.94
Ï„00 ID 21.01
ICC 0.55
N ID 259
Observations 777
Marginal R2 / Conditional R2 0.033 / 0.568 
parameters::standardise_parameters(PHQ_MEM)
## # Standardization method: refit
## 
## Parameter                              | Std. Coef. |         95% CI
## --------------------------------------------------------------------
## (Intercept)                            |       0.18 | [-0.09,  0.46]
## GroupB_Controls                        |      -0.08 | [-0.42,  0.25]
## GroupC_Intervention                    |       0.12 | [-0.22,  0.45]
## TimeC_W1_PHQ_total                     |      -0.07 | [-0.33,  0.19]
## TimeD_M1_PHQ_total                     |      -0.17 | [-0.43,  0.09]
## GroupB_Controls:TimeC_W1_PHQ_total     |      -0.10 | [-0.42,  0.21]
## GroupC_Intervention:TimeC_W1_PHQ_total |      -0.21 | [-0.52,  0.11]
## GroupB_Controls:TimeD_M1_PHQ_total     |      -0.19 | [-0.50,  0.13]
## GroupC_Intervention:TimeD_M1_PHQ_total |      -0.36 | [-0.67, -0.04]

Baseline to 1W

PHQ_B1W <- Full_data_all %>% 
  dplyr::select("ID", "Group", "A_PRE_PHQ_total", "C_W1_PHQ_total")
## Formatting table as needed
PHQ_B1W_long <- PHQ_B1W %>%
  pivot_longer(cols = c(A_PRE_PHQ_total, C_W1_PHQ_total),
               names_to = "Time",
               values_to = "PHQ_Score")
PHQ_MEM_B1W <- lmer(PHQ_Score ~ Group * Time + (1|ID), data = PHQ_B1W_long, REML = TRUE)
summary(PHQ_MEM_B1W)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: PHQ_Score ~ Group * Time + (1 | ID)
##    Data: PHQ_B1W_long
## 
## REML criterion at convergence: 3137.6
## 
## Scaled residuals: 
##      Min       1Q   Median       3Q      Max 
## -2.55240 -0.51079 -0.03476  0.44851  3.00888 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev.
##  ID       (Intercept) 24.72    4.972   
##  Residual             10.79    3.286   
## Number of obs: 518, groups:  ID, 259
## 
## Fixed effects:
##                                        Estimate Std. Error       df t value
## (Intercept)                              9.9600     0.8428 344.8985  11.818
## GroupB_Controls                         -0.5166     1.0224 344.8985  -0.505
## GroupC_Intervention                      0.7196     1.0272 344.8985   0.701
## TimeC_W1_PHQ_total                      -0.4400     0.6571 256.0000  -0.670
## GroupB_Controls:TimeC_W1_PHQ_total      -0.6449     0.7972 256.0000  -0.809
## GroupC_Intervention:TimeC_W1_PHQ_total  -1.2882     0.8009 256.0000  -1.608
##                                        Pr(>|t|)    
## (Intercept)                              <2e-16 ***
## GroupB_Controls                           0.614    
## GroupC_Intervention                       0.484    
## TimeC_W1_PHQ_total                        0.504    
## GroupB_Controls:TimeC_W1_PHQ_total        0.419    
## GroupC_Intervention:TimeC_W1_PHQ_total    0.109    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) GrpB_C GrpC_I TC_W1_ GB_C:T
## GrpB_Cntrls -0.824                            
## GrpC_Intrvn -0.820  0.676                     
## TmC_W1_PHQ_ -0.390  0.321  0.320              
## GB_C:TC_W1_  0.321 -0.390 -0.264 -0.824       
## GC_I:TC_W1_  0.320 -0.264 -0.390 -0.820  0.676
anova  (PHQ_MEM_B1W)
## Type III Analysis of Variance Table with Satterthwaite's method
##             Sum Sq Mean Sq NumDF DenDF F value    Pr(>F)    
## Group       17.842   8.921     2   256  0.8264 0.4387762    
## Time       135.177 135.177     1   256 12.5224 0.0004774 ***
## Group:Time  29.474  14.737     2   256  1.3652 0.2571857    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
sjPlot::tab_model(PHQ_MEM_B1W)
  PHQ Score
Predictors Estimates CI p
(Intercept) 9.96 8.30 – 11.62 <0.001
Group [B_Controls] -0.52 -2.53 – 1.49 0.614
Group [C_Intervention] 0.72 -1.30 – 2.74 0.484
Time [C_W1_PHQ_total] -0.44 -1.73 – 0.85 0.503
Group [B_Controls] × Time
[C_W1_PHQ_total] 		-0.64 -2.21 – 0.92 0.419
Group [C_Intervention] ×
Time [C_W1_PHQ_total] 		-1.29 -2.86 – 0.29 0.108
Random Effects
σ2 10.79
Ï„00 ID 24.72
ICC 0.70
N ID 259
Observations 518
Marginal R2 / Conditional R2 0.017 / 0.701 
parameters::standardise_parameters(PHQ_MEM_B1W)
## # Standardization method: refit
## 
## Parameter                              | Std. Coef. |        95% CI
## -------------------------------------------------------------------
## (Intercept)                            |       0.09 | [-0.19, 0.37]
## GroupB_Controls                        |      -0.09 | [-0.42, 0.25]
## GroupC_Intervention                    |       0.12 | [-0.22, 0.46]
## TimeC_W1_PHQ_total                     |      -0.07 | [-0.29, 0.14]
## GroupB_Controls:TimeC_W1_PHQ_total     |      -0.11 | [-0.37, 0.15]
## GroupC_Intervention:TimeC_W1_PHQ_total |      -0.22 | [-0.48, 0.05]
plot_model(PHQ_MEM_B1W, type = "int")

Baseline to 1M

PHQ_B1M <- Full_data_all %>% 
  dplyr::select("ID", "Group", "A_PRE_PHQ_total", "D_M1_PHQ_total")
## Formatting table as needed
PHQ_B1M_long <- PHQ_B1M %>%
  pivot_longer(cols = c(A_PRE_PHQ_total, D_M1_PHQ_total),
               names_to = "Time",
               values_to = "PHQ_Score")
PHQ_MEM_B1M <- lmer(PHQ_Score ~ Group * Time + (1|ID), data = PHQ_B1M_long, REML = TRUE)
summary(PHQ_MEM_B1M)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: PHQ_Score ~ Group * Time + (1 | ID)
##    Data: PHQ_B1M_long
## 
## REML criterion at convergence: 3284.9
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -2.3373 -0.6131 -0.0873  0.5219  2.5998 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev.
##  ID       (Intercept) 18.42    4.292   
##  Residual             20.25    4.500   
## Number of obs: 518, groups:  ID, 259
## 
## Fixed effects:
##                                        Estimate Std. Error       df t value
## (Intercept)                              9.9600     0.8795 417.3054  11.325
## GroupB_Controls                         -0.5166     1.0669 417.3054  -0.484
## GroupC_Intervention                      0.7196     1.0719 417.3054   0.671
## TimeD_M1_PHQ_total                      -1.0600     0.9000 256.0000  -1.178
## GroupB_Controls:TimeD_M1_PHQ_total      -1.1570     1.0918 256.0000  -1.060
## GroupC_Intervention:TimeD_M1_PHQ_total  -2.2313     1.0969 256.0000  -2.034
##                                        Pr(>|t|)    
## (Intercept)                              <2e-16 ***
## GroupB_Controls                           0.628    
## GroupC_Intervention                       0.502    
## TimeD_M1_PHQ_total                        0.240    
## GroupB_Controls:TimeD_M1_PHQ_total        0.290    
## GroupC_Intervention:TimeD_M1_PHQ_total    0.043 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) GrpB_C GrpC_I TD_M1_ GB_C:T
## GrpB_Cntrls -0.824                            
## GrpC_Intrvn -0.820  0.676                     
## TmD_M1_PHQ_ -0.512  0.422  0.420              
## GB_C:TD_M1_  0.422 -0.512 -0.346 -0.824       
## GC_I:TD_M1_  0.420 -0.346 -0.512 -0.820  0.676
anova  (PHQ_MEM_B1M)
## Type III Analysis of Variance Table with Satterthwaite's method
##            Sum Sq Mean Sq NumDF DenDF F value    Pr(>F)    
## Group       34.23   17.12     2   256  0.8453    0.4306    
## Time       551.08  551.08     1   256 27.2138 3.763e-07 ***
## Group:Time  87.52   43.76     2   256  2.1609    0.1173    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
sjPlot::tab_model(PHQ_MEM_B1M)
  PHQ Score
Predictors Estimates CI p
(Intercept) 9.96 8.23 – 11.69 <0.001
Group [B_Controls] -0.52 -2.61 – 1.58 0.628
Group [C_Intervention] 0.72 -1.39 – 2.83 0.502
Time [D_M1_PHQ_total] -1.06 -2.83 – 0.71 0.239
Group [B_Controls] × Time
[D_M1_PHQ_total] 		-1.16 -3.30 – 0.99 0.290
Group [C_Intervention] ×
Time [D_M1_PHQ_total] 		-2.23 -4.39 – -0.08 0.042
Random Effects
σ2 20.25
Ï„00 ID 18.42
ICC 0.48
N ID 259
Observations 518
Marginal R2 / Conditional R2 0.045 / 0.500 
parameters::standardise_parameters(PHQ_MEM_B1M)
## # Standardization method: refit
## 
## Parameter                              | Std. Coef. |         95% CI
## --------------------------------------------------------------------
## (Intercept)                            |       0.18 | [-0.09,  0.45]
## GroupB_Controls                        |      -0.08 | [-0.41,  0.25]
## GroupC_Intervention                    |       0.11 | [-0.22,  0.45]
## TimeD_M1_PHQ_total                     |      -0.17 | [-0.45,  0.11]
## GroupB_Controls:TimeD_M1_PHQ_total     |      -0.18 | [-0.52,  0.16]
## GroupC_Intervention:TimeD_M1_PHQ_total |      -0.35 | [-0.69, -0.01]
plot_model(PHQ_MEM_B1M, type = "int")

H2b: difference in change in GAD over time between groups

# Merging across timepoints
GAD_alltimepoints <- Full_data_all %>% 
  dplyr::select("ID", "Group", "A_PRE_GAD_total", "C_W1_GAD_total", "D_M1_GAD_total")
## Formatting table as needed
GAD_alltimepoints_long <- GAD_alltimepoints %>%
  pivot_longer(cols = c(A_PRE_GAD_total, C_W1_GAD_total, D_M1_GAD_total),
               names_to = "Time",
               values_to = "GAD_Score")
GAD_MEM <- lmer(GAD_Score ~ Group * Time + (1|ID), data = GAD_alltimepoints_long, REML = TRUE)
summary(GAD_MEM)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: GAD_Score ~ Group * Time + (1 | ID)
##    Data: GAD_alltimepoints_long
## 
## REML criterion at convergence: 4678.2
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -3.3291 -0.4994 -0.0495  0.5366  3.3314 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev.
##  ID       (Intercept) 19.57    4.424   
##  Residual             14.26    3.777   
## Number of obs: 777, groups:  ID, 259
## 
## Fixed effects:
##                                        Estimate Std. Error       df t value
## (Intercept)                              8.0200     0.8226 460.1094   9.749
## GroupB_Controls                          0.4706     0.9980 460.1094   0.472
## GroupC_Intervention                      1.2615     1.0026 460.1094   1.258
## TimeC_W1_GAD_total                       0.0200     0.7554 512.0000   0.026
## TimeD_M1_GAD_total                      -0.1600     0.7554 512.0000  -0.212
## GroupB_Controls:TimeC_W1_GAD_total      -0.8313     0.9164 512.0000  -0.907
## GroupC_Intervention:TimeC_W1_GAD_total  -1.2821     0.9206 512.0000  -1.393
## GroupB_Controls:TimeD_M1_GAD_total      -1.7834     0.9164 512.0000  -1.946
## GroupC_Intervention:TimeD_M1_GAD_total  -2.4031     0.9206 512.0000  -2.610
##                                        Pr(>|t|)    
## (Intercept)                             < 2e-16 ***
## GroupB_Controls                         0.63749    
## GroupC_Intervention                     0.20894    
## TimeC_W1_GAD_total                      0.97889    
## TimeD_M1_GAD_total                      0.83233    
## GroupB_Controls:TimeC_W1_GAD_total      0.36473    
## GroupC_Intervention:TimeC_W1_GAD_total  0.16432    
## GroupB_Controls:TimeD_M1_GAD_total      0.05218 .  
## GroupC_Intervention:TimeD_M1_GAD_total  0.00931 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) GrpB_C GrpC_I TC_W1_ TD_M1_ GB_C:TC GC_I:TC GB_C:TD
## GrpB_Cntrls -0.824                                                    
## GrpC_Intrvn -0.820  0.676                                             
## TmC_W1_GAD_ -0.459  0.378  0.377                                      
## TmD_M1_GAD_ -0.459  0.378  0.377  0.500                               
## GB_C:TC_W1_  0.378 -0.459 -0.311 -0.824 -0.412                        
## GC_I:TC_W1_  0.377 -0.311 -0.459 -0.820 -0.410  0.676                 
## GB_C:TD_M1_  0.378 -0.459 -0.311 -0.412 -0.824  0.500   0.338         
## GC_I:TD_M1_  0.377 -0.311 -0.459 -0.410 -0.820  0.338   0.500   0.676
anova  (GAD_MEM)
## Type III Analysis of Variance Table with Satterthwaite's method
##             Sum Sq Mean Sq NumDF DenDF F value    Pr(>F)    
## Group        6.601   3.301     2   256  0.2314    0.7936    
## Time       279.498 139.749     2   512  9.7971 6.677e-05 ***
## Group:Time  98.745  24.686     4   512  1.7306    0.1418    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
sjPlot::tab_model(GAD_MEM)
  GAD Score
Predictors Estimates CI p
(Intercept) 8.02 6.41 – 9.63 <0.001
Group [B_Controls] 0.47 -1.49 – 2.43 0.637
Group [C_Intervention] 1.26 -0.71 – 3.23 0.209
Time [C_W1_GAD_total] 0.02 -1.46 – 1.50 0.979
Time [D_M1_GAD_total] -0.16 -1.64 – 1.32 0.832
Group [B_Controls] × Time
[C_W1_GAD_total] 		-0.83 -2.63 – 0.97 0.365
Group [C_Intervention] ×
Time [C_W1_GAD_total] 		-1.28 -3.09 – 0.53 0.164
Group [B_Controls] × Time
[D_M1_GAD_total] 		-1.78 -3.58 – 0.02 0.052
Group [C_Intervention] ×
Time [D_M1_GAD_total] 		-2.40 -4.21 – -0.60 0.009
Random Effects
σ2 14.26
Ï„00 ID 19.57
ICC 0.58
N ID 259
Observations 777
Marginal R2 / Conditional R2 0.021 / 0.587 
parameters::standardise_parameters(GAD_MEM)
## # Standardization method: refit
## 
## Parameter                              | Std. Coef. |         95% CI
## --------------------------------------------------------------------
## (Intercept)                            |       0.03 | [-0.24,  0.31]
## GroupB_Controls                        |       0.08 | [-0.25,  0.42]
## GroupC_Intervention                    |       0.22 | [-0.12,  0.55]
## TimeC_W1_GAD_total                     |   3.42e-03 | [-0.25,  0.26]
## TimeD_M1_GAD_total                     |      -0.03 | [-0.28,  0.23]
## GroupB_Controls:TimeC_W1_GAD_total     |      -0.14 | [-0.45,  0.17]
## GroupC_Intervention:TimeC_W1_GAD_total |      -0.22 | [-0.53,  0.09]
## GroupB_Controls:TimeD_M1_GAD_total     |      -0.30 | [-0.61,  0.00]
## GroupC_Intervention:TimeD_M1_GAD_total |      -0.41 | [-0.72, -0.10]

Baseline to 1W

# Merging across timepoints
GAD_B1W <- Full_data_all %>% 
  dplyr::select("ID", "Group", "A_PRE_GAD_total", "C_W1_GAD_total")
## Formatting table as needed
GAD_B1W_long <- GAD_B1W %>%
  pivot_longer(cols = c(A_PRE_GAD_total, C_W1_GAD_total),
               names_to = "Time",
               values_to = "GAD_Score")
GAD_MEM_B1W <- lmer(GAD_Score ~ Group * Time + (1|ID), data = GAD_B1W_long, REML = TRUE)
summary(GAD_MEM_B1W)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: GAD_Score ~ Group * Time + (1 | ID)
##    Data: GAD_B1W_long
## 
## REML criterion at convergence: 3091.9
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -2.8845 -0.4805 -0.0673  0.4776  3.0414 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev.
##  ID       (Intercept) 22.200   4.712   
##  Residual              9.998   3.162   
## Number of obs: 518, groups:  ID, 259
## 
## Fixed effects:
##                                        Estimate Std. Error       df t value
## (Intercept)                              8.0200     0.8025 347.0310   9.994
## GroupB_Controls                          0.4706     0.9735 347.0310   0.483
## GroupC_Intervention                      1.2615     0.9780 347.0310   1.290
## TimeC_W1_GAD_total                       0.0200     0.6324 256.0000   0.032
## GroupB_Controls:TimeC_W1_GAD_total      -0.8313     0.7672 256.0000  -1.084
## GroupC_Intervention:TimeC_W1_GAD_total  -1.2821     0.7708 256.0000  -1.663
##                                        Pr(>|t|)    
## (Intercept)                              <2e-16 ***
## GroupB_Controls                          0.6291    
## GroupC_Intervention                      0.1980    
## TimeC_W1_GAD_total                       0.9748    
## GroupB_Controls:TimeC_W1_GAD_total       0.2796    
## GroupC_Intervention:TimeC_W1_GAD_total   0.0974 .  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) GrpB_C GrpC_I TC_W1_ GB_C:T
## GrpB_Cntrls -0.824                            
## GrpC_Intrvn -0.820  0.676                     
## TmC_W1_GAD_ -0.394  0.325  0.323              
## GB_C:TC_W1_  0.325 -0.394 -0.267 -0.824       
## GC_I:TC_W1_  0.323 -0.267 -0.394 -0.820  0.676
anova  (GAD_MEM_B1W)
## Type III Analysis of Variance Table with Satterthwaite's method
##            Sum Sq Mean Sq NumDF DenDF F value  Pr(>F)  
## Group       7.794   3.897     2   256  0.3898 0.67763  
## Time       53.863  53.863     1   256  5.3872 0.02107 *
## Group:Time 27.698  13.849     2   256  1.3851 0.25216  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
sjPlot::tab_model(GAD_MEM_B1W)
  GAD Score
Predictors Estimates CI p
(Intercept) 8.02 6.44 – 9.60 <0.001
Group [B_Controls] 0.47 -1.44 – 2.38 0.629
Group [C_Intervention] 1.26 -0.66 – 3.18 0.198
Time [C_W1_GAD_total] 0.02 -1.22 – 1.26 0.975
Group [B_Controls] × Time
[C_W1_GAD_total] 		-0.83 -2.34 – 0.68 0.279
Group [C_Intervention] ×
Time [C_W1_GAD_total] 		-1.28 -2.80 – 0.23 0.097
Random Effects
σ2 10.00
Ï„00 ID 22.20
ICC 0.69
N ID 259
Observations 518
Marginal R2 / Conditional R2 0.009 / 0.692 
parameters::standardise_parameters(GAD_MEM_B1W)
## # Standardization method: refit
## 
## Parameter                              | Std. Coef. |        95% CI
## -------------------------------------------------------------------
## (Intercept)                            |      -0.05 | [-0.33, 0.23]
## GroupB_Controls                        |       0.08 | [-0.25, 0.42]
## GroupC_Intervention                    |       0.22 | [-0.12, 0.56]
## TimeC_W1_GAD_total                     |   3.52e-03 | [-0.22, 0.22]
## GroupB_Controls:TimeC_W1_GAD_total     |      -0.15 | [-0.41, 0.12]
## GroupC_Intervention:TimeC_W1_GAD_total |      -0.23 | [-0.49, 0.04]
plot_model(GAD_MEM_B1W, type = "int")

Baseline to 1M

# Merging across timepoints
GAD_B1M <- Full_data_all %>% 
  dplyr::select("ID", "Group", "A_PRE_GAD_total", "D_M1_GAD_total")
## Formatting table as needed
GAD_B1M_long <- GAD_B1M %>%
  pivot_longer(cols = c(A_PRE_GAD_total, D_M1_GAD_total),
               names_to = "Time",
               values_to = "GAD_Score")
GAD_MEM_B1M <- lmer(GAD_Score ~ Group * Time + (1|ID), data = GAD_B1M_long, REML = TRUE)
summary(GAD_MEM_B1M)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: GAD_Score ~ Group * Time + (1 | ID)
##    Data: GAD_B1M_long
## 
## REML criterion at convergence: 3204.6
## 
## Scaled residuals: 
##      Min       1Q   Median       3Q      Max 
## -2.09051 -0.58212 -0.09107  0.54259  2.63822 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev.
##  ID       (Intercept) 16.93    4.115   
##  Residual             16.70    4.087   
## Number of obs: 518, groups:  ID, 259
## 
## Fixed effects:
##                                        Estimate Std. Error       df t value
## (Intercept)                              8.0200     0.8202 408.4768   9.778
## GroupB_Controls                          0.4706     0.9950 408.4768   0.473
## GroupC_Intervention                      1.2616     0.9996 408.4768   1.262
## TimeD_M1_GAD_total                      -0.1600     0.8174 256.0000  -0.196
## GroupB_Controls:TimeD_M1_GAD_total      -1.7834     0.9916 256.0000  -1.799
## GroupC_Intervention:TimeD_M1_GAD_total  -2.4031     0.9962 256.0000  -2.412
##                                        Pr(>|t|)    
## (Intercept)                              <2e-16 ***
## GroupB_Controls                          0.6365    
## GroupC_Intervention                      0.2077    
## TimeD_M1_GAD_total                       0.8450    
## GroupB_Controls:TimeD_M1_GAD_total       0.0733 .  
## GroupC_Intervention:TimeD_M1_GAD_total   0.0166 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) GrpB_C GrpC_I TD_M1_ GB_C:T
## GrpB_Cntrls -0.824                            
## GrpC_Intrvn -0.820  0.676                     
## TmD_M1_GAD_ -0.498  0.411  0.409              
## GB_C:TD_M1_  0.411 -0.498 -0.337 -0.824       
## GC_I:TD_M1_  0.409 -0.337 -0.498 -0.820  0.676
anova  (GAD_MEM_B1M)
## Type III Analysis of Variance Table with Satterthwaite's method
##             Sum Sq Mean Sq NumDF DenDF F value    Pr(>F)    
## Group        8.891   4.445     2   256  0.2662   0.76653    
## Time       278.165 278.165     1   256 16.6540 5.998e-05 ***
## Group:Time  98.051  49.026     2   256  2.9352   0.05491 .  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
sjPlot::tab_model(GAD_MEM_B1M)
  GAD Score
Predictors Estimates CI p
(Intercept) 8.02 6.41 – 9.63 <0.001
Group [B_Controls] 0.47 -1.48 – 2.43 0.636
Group [C_Intervention] 1.26 -0.70 – 3.23 0.208
Time [D_M1_GAD_total] -0.16 -1.77 – 1.45 0.845
Group [B_Controls] × Time
[D_M1_GAD_total] 		-1.78 -3.73 – 0.16 0.073
Group [C_Intervention] ×
Time [D_M1_GAD_total] 		-2.40 -4.36 – -0.45 0.016
Random Effects
σ2 16.70
Ï„00 ID 16.93
ICC 0.50
N ID 259
Observations 518
Marginal R2 / Conditional R2 0.032 / 0.519 
parameters::standardise_parameters(GAD_MEM_B1M)
## # Standardization method: refit
## 
## Parameter                              | Std. Coef. |         95% CI
## --------------------------------------------------------------------
## (Intercept)                            |       0.04 | [-0.24,  0.31]
## GroupB_Controls                        |       0.08 | [-0.25,  0.41]
## GroupC_Intervention                    |       0.22 | [-0.12,  0.55]
## TimeD_M1_GAD_total                     |      -0.03 | [-0.30,  0.25]
## GroupB_Controls:TimeD_M1_GAD_total     |      -0.30 | [-0.64,  0.03]
## GroupC_Intervention:TimeD_M1_GAD_total |      -0.41 | [-0.74, -0.08]
plot_model(GAD_MEM_B1M, type = "int")

H2b: difference in change in mood over time between groups

Mood_alltimepoints <- Full_data_all %>% 
  dplyr::select("ID", "Group", "A_PRE_mood_mean", "B_POST_mood_mean", "C_W1_mood_mean", "D_M1_mood_mean")
## Formatting tables as needed
Mood_alltimepoints_long <- Mood_alltimepoints %>%
  pivot_longer(cols = c("A_PRE_mood_mean", "B_POST_mood_mean", "C_W1_mood_mean", "D_M1_mood_mean"),
               names_to = "Time",
               values_to = "Mood_Score")
Mood_MEM <- lmer(Mood_Score ~ Group * Time + (1|ID), data = Mood_alltimepoints_long, REML = TRUE)
summary(Mood_MEM)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: Mood_Score ~ Group * Time + (1 | ID)
##    Data: Mood_alltimepoints_long
## 
## REML criterion at convergence: 9998.7
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -3.9245 -0.4840  0.0543  0.5791  3.4075 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev.
##  ID       (Intercept) 892.9    29.88   
##  Residual             982.7    31.35   
## Number of obs: 995, groups:  ID, 259
## 
## Fixed effects:
##                                          Estimate Std. Error      df t value
## (Intercept)                                40.300      6.125 599.204   6.580
## GroupB_Controls                            -2.984      7.430 599.204  -0.402
## GroupC_Intervention                       -10.669      7.465 599.204  -1.429
## TimeB_POST_mood_mean                       -0.020      6.270 729.243  -0.003
## TimeC_W1_mood_mean                        -14.945      6.350 731.731  -2.353
## TimeD_M1_mood_mean                        -21.095      6.579 738.302  -3.206
## GroupB_Controls:TimeB_POST_mood_mean       19.992      7.606 729.243   2.628
## GroupC_Intervention:TimeB_POST_mood_mean   29.483      7.649 729.438   3.855
## GroupB_Controls:TimeC_W1_mood_mean          8.695      7.695 731.631   1.130
## GroupC_Intervention:TimeC_W1_mood_mean      9.945      7.739 731.707   1.285
## GroupB_Controls:TimeD_M1_mood_mean         14.488      7.957 737.849   1.821
## GroupC_Intervention:TimeD_M1_mood_mean     14.299      7.996 737.788   1.788
##                                          Pr(>|t|)    
## (Intercept)                              1.03e-10 ***
## GroupB_Controls                          0.688120    
## GroupC_Intervention                      0.153456    
## TimeB_POST_mood_mean                     0.997456    
## TimeC_W1_mood_mean                       0.018866 *  
## TimeD_M1_mood_mean                       0.001403 ** 
## GroupB_Controls:TimeB_POST_mood_mean     0.008758 ** 
## GroupC_Intervention:TimeB_POST_mood_mean 0.000126 ***
## GroupB_Controls:TimeC_W1_mood_mean       0.258872    
## GroupC_Intervention:TimeC_W1_mood_mean   0.199181    
## GroupB_Controls:TimeD_M1_mood_mean       0.069040 .  
## GroupC_Intervention:TimeD_M1_mood_mean   0.074150 .  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) GrpB_C GrpC_I TB_POS TC_W1_ TD_M1_ GB_C:TB GC_I:TB GB_C:TC
## GrpB_Cntrls -0.824                                                           
## GrpC_Intrvn -0.820  0.676                                                    
## TmB_POST_m_ -0.512  0.422  0.420                                             
## TmC_W1_md_m -0.505  0.417  0.415  0.494                                      
## TmD_M1_md_m -0.488  0.402  0.400  0.476  0.469                               
## GB_C:TB_POS  0.422 -0.512 -0.346 -0.824 -0.407 -0.393                        
## GC_I:TB_POS  0.420 -0.346 -0.511 -0.820 -0.405 -0.391  0.676                 
## GB_C:TC_W1_  0.417 -0.506 -0.342 -0.407 -0.825 -0.387  0.494   0.334         
## GC_I:TC_W1_  0.415 -0.342 -0.505 -0.405 -0.821 -0.385  0.334   0.493   0.677 
## GB_C:TD_M1_  0.403 -0.489 -0.331 -0.394 -0.388 -0.827  0.478   0.323   0.472 
## GC_I:TD_M1_  0.401 -0.331 -0.489 -0.392 -0.386 -0.823  0.323   0.477   0.319 
##             GC_I:TC GB_C:TD
## GrpB_Cntrls                
## GrpC_Intrvn                
## TmB_POST_m_                
## TmC_W1_md_m                
## TmD_M1_md_m                
## GB_C:TB_POS                
## GC_I:TB_POS                
## GB_C:TC_W1_                
## GC_I:TC_W1_                
## GB_C:TD_M1_  0.319         
## GC_I:TD_M1_  0.471   0.680
anova  (Mood_MEM)
## Type III Analysis of Variance Table with Satterthwaite's method
##            Sum Sq Mean Sq NumDF  DenDF F value  Pr(>F)    
## Group        2110    1055     2 258.46  1.0734 0.34336    
## Time       104490   34830     3 734.51 35.4434 < 2e-16 ***
## Group:Time  15983    2664     6 734.42  2.7107 0.01307 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
sjPlot::tab_model(Mood_MEM)
  Mood Score
Predictors Estimates CI p
(Intercept) 40.30 28.28 – 52.32 <0.001
Group [B_Controls] -2.98 -17.56 – 11.60 0.688
Group [C_Intervention] -10.67 -25.32 – 3.98 0.153
Time [B_POST_mood_mean] -0.02 -12.32 – 12.28 0.997
Time [C_W1_mood_mean] -14.94 -27.41 – -2.48 0.019
Time [D_M1_mood_mean] -21.09 -34.01 – -8.18 0.001
Group [B_Controls] × Time
[B_POST_mood_mean] 		19.99 5.07 – 34.92 0.009
Group [C_Intervention] ×
Time [B_POST_mood_mean] 		29.48 14.47 – 44.49 <0.001
Group [B_Controls] × Time
[C_W1_mood_mean] 		8.69 -6.41 – 23.80 0.259
Group [C_Intervention] ×
Time [C_W1_mood_mean] 		9.94 -5.24 – 25.13 0.199
Group [B_Controls] × Time
[D_M1_mood_mean] 		14.49 -1.13 – 30.10 0.069
Group [C_Intervention] ×
Time [D_M1_mood_mean] 		14.30 -1.39 – 29.99 0.074
Random Effects
σ2 982.69
Ï„00 ID 892.93
ICC 0.48
N ID 259
Observations 995
Marginal R2 / Conditional R2 0.079 / 0.517 
parameters::standardise_parameters(Mood_MEM)
## # Standardization method: refit
## 
## Parameter                                | Std. Coef. |         95% CI
## ----------------------------------------------------------------------
## (Intercept)                              |       0.10 | [-0.17,  0.36]
## GroupB_Controls                          |      -0.07 | [-0.39,  0.26]
## GroupC_Intervention                      |      -0.24 | [-0.56,  0.09]
## TimeB_POST_mood_mean                     |  -4.45e-04 | [-0.27,  0.27]
## TimeC_W1_mood_mean                       |      -0.33 | [-0.61, -0.06]
## TimeD_M1_mood_mean                       |      -0.47 | [-0.76, -0.18]
## GroupB_Controls:TimeB_POST_mood_mean     |       0.44 | [ 0.11,  0.78]
## GroupC_Intervention:TimeB_POST_mood_mean |       0.66 | [ 0.32,  0.99]
## GroupB_Controls:TimeC_W1_mood_mean       |       0.19 | [-0.14,  0.53]
## GroupC_Intervention:TimeC_W1_mood_mean   |       0.22 | [-0.12,  0.56]
## GroupB_Controls:TimeD_M1_mood_mean       |       0.32 | [-0.03,  0.67]
## GroupC_Intervention:TimeD_M1_mood_mean   |       0.32 | [-0.03,  0.67]

Baseline to post

Mood_BP <- Full_data_all %>% 
  dplyr::select("ID", "Group", "A_PRE_mood_mean", "B_POST_mood_mean")
## Formatting tables as needed
Mood_BP_long <- Mood_BP %>%
  pivot_longer(cols = c("A_PRE_mood_mean", "B_POST_mood_mean"),
               names_to = "Time",
               values_to = "Mood_Score")
Mood_MEM_BP <- lmer(Mood_Score ~ Group * Time + (1|ID), data = Mood_BP_long, REML = TRUE)
summary(Mood_MEM_BP)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: Mood_Score ~ Group * Time + (1 | ID)
##    Data: Mood_BP_long
## 
## REML criterion at convergence: 5079
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -4.1547 -0.3855  0.0476  0.4451  3.3781 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev.
##  ID       (Intercept) 1035.6   32.18   
##  Residual              513.4   22.66   
## Number of obs: 517, groups:  ID, 259
## 
## Fixed effects:
##                                          Estimate Std. Error      df t value
## (Intercept)                                40.300      5.566 353.535   7.240
## GroupB_Controls                            -2.984      6.752 353.535  -0.442
## GroupC_Intervention                       -10.669      6.784 353.535  -1.573
## TimeB_POST_mood_mean                       -0.020      4.532 255.115  -0.004
## GroupB_Controls:TimeB_POST_mood_mean       19.992      5.498 255.115   3.637
## GroupC_Intervention:TimeB_POST_mood_mean   29.341      5.530 255.341   5.305
##                                          Pr(>|t|)    
## (Intercept)                              2.81e-12 ***
## GroupB_Controls                          0.658816    
## GroupC_Intervention                      0.116674    
## TimeB_POST_mood_mean                     0.996482    
## GroupB_Controls:TimeB_POST_mood_mean     0.000334 ***
## GroupC_Intervention:TimeB_POST_mood_mean 2.44e-07 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) GrpB_C GrpC_I TB_POS GB_C:T
## GrpB_Cntrls -0.824                            
## GrpC_Intrvn -0.820  0.676                     
## TmB_POST_m_ -0.407  0.336  0.334              
## GB_C:TB_POS  0.336 -0.407 -0.275 -0.824       
## GC_I:TB_POS  0.334 -0.275 -0.407 -0.819  0.675
anova  (Mood_MEM_BP)
## Type III Analysis of Variance Table with Satterthwaite's method
##             Sum Sq Mean Sq NumDF  DenDF F value    Pr(>F)    
## Group        677.9   339.0     2 255.99  0.6603    0.5176    
## Time       30949.7 30949.7     1 255.29 60.2843 1.989e-13 ***
## Group:Time 14453.2  7226.6     2 255.32 14.0761 1.588e-06 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
sjPlot::tab_model(Mood_MEM_BP)
  Mood Score
Predictors Estimates CI p
(Intercept) 40.30 29.36 – 51.24 <0.001
Group [B_Controls] -2.98 -16.25 – 10.28 0.659
Group [C_Intervention] -10.67 -24.00 – 2.66 0.116
Time [B_POST_mood_mean] -0.02 -8.92 – 8.88 0.996
Group [B_Controls] × Time
[B_POST_mood_mean] 		19.99 9.19 – 30.79 <0.001
Group [C_Intervention] ×
Time [B_POST_mood_mean] 		29.34 18.48 – 40.21 <0.001
Random Effects
σ2 513.40
Ï„00 ID 1035.58
ICC 0.67
N ID 259
Observations 517
Marginal R2 / Conditional R2 0.079 / 0.695 
parameters::standardise_parameters(Mood_MEM_BP)
## # Standardization method: refit
## 
## Parameter                                | Std. Coef. |        95% CI
## ---------------------------------------------------------------------
## (Intercept)                              |      -0.11 | [-0.38, 0.16]
## GroupB_Controls                          |      -0.07 | [-0.40, 0.25]
## GroupC_Intervention                      |      -0.26 | [-0.59, 0.07]
## TimeB_POST_mood_mean                     |  -4.90e-04 | [-0.22, 0.22]
## GroupB_Controls:TimeB_POST_mood_mean     |       0.49 | [ 0.23, 0.75]
## GroupC_Intervention:TimeB_POST_mood_mean |       0.72 | [ 0.45, 0.98]
plot_model(Mood_MEM_BP, type = "int")

Baseline to 1W

Mood_B1W <- Full_data_all %>% 
  dplyr::select("ID", "Group", "A_PRE_mood_mean", "C_W1_mood_mean")
## Formatting tables as needed
Mood_B1W_long <- Mood_B1W %>%
  pivot_longer(cols = c("A_PRE_mood_mean", "C_W1_mood_mean"),
               names_to = "Time",
               values_to = "Mood_Score")
Mood_MEM_B1W <- lmer(Mood_Score ~ Group * Time + (1|ID), data = Mood_B1W_long, REML = TRUE)
summary(Mood_MEM_B1W)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: Mood_Score ~ Group * Time + (1 | ID)
##    Data: Mood_B1W_long
## 
## REML criterion at convergence: 5220.7
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -3.1570 -0.4429  0.1159  0.5899  2.4716 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev.
##  ID       (Intercept)  866.2   29.43   
##  Residual             1117.7   33.43   
## Number of obs: 509, groups:  ID, 259
## 
## Fixed effects:
##                                        Estimate Std. Error      df t value
## (Intercept)                              40.300      6.299 424.920   6.398
## GroupB_Controls                          -2.984      7.641 424.920  -0.390
## GroupC_Intervention                     -10.669      7.677 424.920  -1.390
## TimeC_W1_mood_mean                      -14.705      6.786 253.748  -2.167
## GroupB_Controls:TimeC_W1_mood_mean        8.265      8.220 253.359   1.005
## GroupC_Intervention:TimeC_W1_mood_mean    9.377      8.269 253.708   1.134
##                                        Pr(>|t|)    
## (Intercept)                            4.16e-10 ***
## GroupB_Controls                          0.6964    
## GroupC_Intervention                      0.1653    
## TimeC_W1_mood_mean                       0.0312 *  
## GroupB_Controls:TimeC_W1_mood_mean       0.3157    
## GroupC_Intervention:TimeC_W1_mood_mean   0.2579    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) GrpB_C GrpC_I TC_W1_ GB_C:T
## GrpB_Cntrls -0.824                            
## GrpC_Intrvn -0.820  0.676                     
## TmC_W1_md_m -0.523  0.431  0.429              
## GB_C:TC_W1_  0.432 -0.524 -0.354 -0.825       
## GC_I:TC_W1_  0.429 -0.354 -0.523 -0.821  0.677
anova  (Mood_MEM_B1W)
## Type III Analysis of Variance Table with Satterthwaite's method
##            Sum Sq Mean Sq NumDF  DenDF F value  Pr(>F)   
## Group      2236.2  1118.1     2 257.40  1.0004 0.36916   
## Time       8711.4  8711.4     1 253.43  7.7944 0.00564 **
## Group:Time 1553.5   776.7     2 253.35  0.6950 0.50003   
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
sjPlot::tab_model(Mood_MEM_B1W)
  Mood Score
Predictors Estimates CI p
(Intercept) 40.30 27.92 – 52.68 <0.001
Group [B_Controls] -2.98 -18.00 – 12.03 0.696
Group [C_Intervention] -10.67 -25.75 – 4.41 0.165
Time [C_W1_mood_mean] -14.70 -28.04 – -1.37 0.031
Group [B_Controls] × Time
[C_W1_mood_mean] 		8.26 -7.89 – 24.42 0.315
Group [C_Intervention] ×
Time [C_W1_mood_mean] 		9.38 -6.87 – 25.62 0.257
Random Effects
σ2 1117.65
Ï„00 ID 866.19
ICC 0.44
N ID 259
Observations 509
Marginal R2 / Conditional R2 0.014 / 0.445 
parameters::standardise_parameters(Mood_MEM_B1W)
## # Standardization method: refit
## 
## Parameter                              | Std. Coef. |         95% CI
## --------------------------------------------------------------------
## (Intercept)                            |       0.20 | [-0.07,  0.48]
## GroupB_Controls                        |      -0.07 | [-0.40,  0.27]
## GroupC_Intervention                    |      -0.24 | [-0.58,  0.10]
## TimeC_W1_mood_mean                     |      -0.33 | [-0.63, -0.03]
## GroupB_Controls:TimeC_W1_mood_mean     |       0.19 | [-0.18,  0.55]
## GroupC_Intervention:TimeC_W1_mood_mean |       0.21 | [-0.15,  0.57]
plot_model(Mood_MEM_B1W, type = "int")

Baseline to 1M

Mood_B1M <- Full_data_all %>% 
  dplyr::select("ID", "Group", "A_PRE_mood_mean", "D_M1_mood_mean")
## Formatting tables as needed
Mood_B1M_long <- Mood_B1M %>%
  pivot_longer(cols = c("A_PRE_mood_mean", "D_M1_mood_mean"),
               names_to = "Time",
               values_to = "Mood_Score")
Mood_MEM_B1M <- lmer(Mood_Score ~ Group * Time + (1|ID), data = Mood_B1M_long, REML = TRUE)
summary(Mood_MEM_B1M)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: Mood_Score ~ Group * Time + (1 | ID)
##    Data: Mood_B1M_long
## 
## REML criterion at convergence: 5011.6
## 
## Scaled residuals: 
##      Min       1Q   Median       3Q      Max 
## -2.91923 -0.51813  0.07265  0.59296  2.42299 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev.
##  ID       (Intercept)  893.6   29.89   
##  Residual             1156.5   34.01   
## Number of obs: 487, groups:  ID, 259
## 
## Fixed effects:
##                                        Estimate Std. Error      df t value
## (Intercept)                              40.300      6.403 411.735   6.294
## GroupB_Controls                          -2.984      7.768 411.735  -0.384
## GroupC_Intervention                     -10.669      7.804 411.735  -1.367
## TimeD_M1_mood_mean                      -20.875      7.188 246.670  -2.904
## GroupB_Controls:TimeD_M1_mood_mean       14.198      8.688 245.758   1.634
## GroupC_Intervention:TimeD_M1_mood_mean   14.331      8.732 245.852   1.641
##                                        Pr(>|t|)    
## (Intercept)                            7.95e-10 ***
## GroupB_Controls                         0.70108    
## GroupC_Intervention                     0.17236    
## TimeD_M1_mood_mean                      0.00402 ** 
## GroupB_Controls:TimeD_M1_mood_mean      0.10351    
## GroupC_Intervention:TimeD_M1_mood_mean  0.10205    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) GrpB_C GrpC_I TD_M1_ GB_C:T
## GrpB_Cntrls -0.824                            
## GrpC_Intrvn -0.820  0.676                     
## TmD_M1_md_m -0.503  0.414  0.412              
## GB_C:TD_M1_  0.416 -0.504 -0.341 -0.827       
## GC_I:TD_M1_  0.414 -0.341 -0.504 -0.823  0.681
anova  (Mood_MEM_B1M)
## Type III Analysis of Variance Table with Satterthwaite's method
##             Sum Sq Mean Sq NumDF  DenDF F value    Pr(>F)    
## Group       2302.4  1151.2     2 260.11  0.9954 0.3709746    
## Time       13435.3 13435.3     1 245.36 11.6170 0.0007637 ***
## Group:Time  3690.3  1845.2     2 245.06  1.5954 0.2049198    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
sjPlot::tab_model(Mood_MEM_B1M)
  Mood Score
Predictors Estimates CI p
(Intercept) 40.30 27.72 – 52.88 <0.001
Group [B_Controls] -2.98 -18.25 – 12.28 0.701
Group [C_Intervention] -10.67 -26.00 – 4.67 0.172
Time [D_M1_mood_mean] -20.87 -35.00 – -6.75 0.004
Group [B_Controls] × Time
[D_M1_mood_mean] 		14.20 -2.87 – 31.27 0.103
Group [C_Intervention] ×
Time [D_M1_mood_mean] 		14.33 -2.83 – 31.49 0.101
Random Effects
σ2 1156.52
Ï„00 ID 893.64
ICC 0.44
N ID 259
Observations 487
Marginal R2 / Conditional R2 0.020 / 0.447 
parameters::standardise_parameters(Mood_MEM_B1M)
## # Standardization method: refit
## 
## Parameter                              | Std. Coef. |         95% CI
## --------------------------------------------------------------------
## (Intercept)                            |       0.22 | [-0.06,  0.49]
## GroupB_Controls                        |      -0.07 | [-0.40,  0.27]
## GroupC_Intervention                    |      -0.23 | [-0.57,  0.10]
## TimeD_M1_mood_mean                     |      -0.46 | [-0.77, -0.15]
## GroupB_Controls:TimeD_M1_mood_mean     |       0.31 | [-0.06,  0.69]
## GroupC_Intervention:TimeD_M1_mood_mean |       0.31 | [-0.06,  0.69]
plot_model(Mood_MEM_B1M, type = "int")

Hypothesis 3

# Baseline to 1W/1M changes (creating new columns)
changeinvariables <- mutate(Full_data_all,
                      IUS_BP_change = B_POST_IUS_total - A_PRE_IUS_total,
                      IUS_B1W_change = C_W1_IUS_total - A_PRE_IUS_total,
                      IUS_B1M_change = D_M1_IUS_total - A_PRE_IUS_total,
                      PHQ_B1W_change = C_W1_PHQ_total - A_PRE_PHQ_total,
                      PHQ_B1M_change = D_M1_PHQ_total - A_PRE_PHQ_total,
                      GAD_B1W_change = C_W1_GAD_total - A_PRE_GAD_total,
                      GAD_B1M_change = D_M1_GAD_total - A_PRE_GAD_total,
                      Mood_BP_change = B_POST_mood_mean - A_PRE_mood_mean,
                      Mood_B1W_change = C_W1_mood_mean - A_PRE_mood_mean,
                      Mood_B1M_change = D_M1_mood_mean - A_PRE_mood_mean)
# Separating out each group

Intervention_group <- changeinvariables %>% 
  filter(Group == "C_Intervention")

Psychoed_group <- changeinvariables %>% 
  filter(Group == "B_Controls")

ECs_group <- changeinvariables %>% 
  filter(Group == "A_ECs")

H3: IUS mediating change in PHQ over time between groups

Full sample at 1 week: Group to change in PHQ via change in IUS

Mediation.PHQchange.1W <-
  '#regressions
PHQ_B1W_change ~ c1 * Group  
IUS_B1W_change ~ a1 * Group 
PHQ_B1W_change ~ b1*IUS_B1W_change
indirect1 := a1 * b1
direct := c1
total := c1 + (a1 * b1)
'
group.IUS.PHQ.1W <- sem(Mediation.PHQchange.1W, data=changeinvariables, std.lv=T, std.ov=T, missing='fiml', se='robust', estimator='mlr', auto.var = T)

summary(group.IUS.PHQ.1W, standardized=T, rsquare=T)
## lavaan 0.6.15 ended normally after 17 iterations
## 
##   Estimator                                         ML
##   Optimization method                           NLMINB
##   Number of model parameters                         7
## 
##   Number of observations                           259
##   Number of missing patterns                         1
## 
## Model Test User Model:
##                                               Standard      Scaled
##   Test Statistic                                 0.000       0.000
##   Degrees of freedom                                 0           0
## 
## Parameter Estimates:
## 
##   Standard errors                             Sandwich
##   Information bread                           Observed
##   Observed information based on                Hessian
## 
## Regressions:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##   PHQ_B1W_change ~                                                      
##     Group     (c1)   -0.063    0.079   -0.803    0.422   -0.063   -0.047
##   IUS_B1W_change ~                                                      
##     Group     (a1)   -0.269    0.085   -3.152    0.002   -0.269   -0.199
##   PHQ_B1W_change ~                                                      
##     IUS_B1W_c (b1)    0.279    0.068    4.118    0.000    0.279    0.279
## 
## Intercepts:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .PHQ_B1W_change    0.140    0.171    0.818    0.413    0.140    0.140
##    .IUS_B1W_change    0.592    0.196    3.018    0.003    0.592    0.593
## 
## Variances:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .PHQ_B1W_change    0.911    0.111    8.233    0.000    0.911    0.915
##    .IUS_B1W_change    0.957    0.158    6.071    0.000    0.957    0.960
## 
## R-Square:
##                    Estimate
##     PHQ_B1W_change    0.085
##     IUS_B1W_change    0.040
## 
## Defined Parameters:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##     indirect1        -0.075    0.029   -2.567    0.010   -0.075   -0.056
##     direct           -0.063    0.079   -0.803    0.422   -0.063   -0.047
##     total            -0.138    0.082   -1.692    0.091   -0.138   -0.103

Intervention group at 1 week: pre PHQ to 1 week PHQ via change in IU

Mediation.PHQ.intervention.1W <-
  '#regressions
C_W1_PHQ_total ~ c1 * A_PRE_PHQ_total  
IUS_B1W_change ~ a1 * A_PRE_PHQ_total 
C_W1_PHQ_total ~ b1*IUS_B1W_change
indirect1 := a1 * b1
direct := c1
total := c1 + (a1 * b1)
'
PHQ.IUS.intervention.1W <- sem(Mediation.PHQ.intervention.1W, data=Intervention_group, std.lv=T, std.ov=T, missing='fiml', se='robust', estimator='mlr', auto.var = T)

summary(PHQ.IUS.intervention.1W, standardized=T, rsquare=T)
## lavaan 0.6.15 ended normally after 21 iterations
## 
##   Estimator                                         ML
##   Optimization method                           NLMINB
##   Number of model parameters                         7
## 
##   Number of observations                           103
##   Number of missing patterns                         1
## 
## Model Test User Model:
##                                               Standard      Scaled
##   Test Statistic                                 0.000       0.000
##   Degrees of freedom                                 0           0
## 
## Parameter Estimates:
## 
##   Standard errors                             Sandwich
##   Information bread                           Observed
##   Observed information based on                Hessian
## 
## Regressions:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##   C_W1_PHQ_total ~                                                      
##     A_PRE_PHQ (c1)    0.106    0.015    6.960    0.000    0.106    0.594
##   IUS_B1W_change ~                                                      
##     A_PRE_PHQ (a1)    0.009    0.016    0.550    0.582    0.009    0.048
##   C_W1_PHQ_total ~                                                      
##     IUS_B1W_c (b1)    0.188    0.093    2.020    0.043    0.188    0.188
## 
## Intercepts:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .C_W1_PHQ_total   -1.130    0.143   -7.924    0.000   -1.130   -1.135
##    .IUS_B1W_change   -0.092    0.200   -0.458    0.647   -0.092   -0.092
## 
## Variances:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .C_W1_PHQ_total    0.595    0.094    6.317    0.000    0.595    0.601
##    .IUS_B1W_change    0.988    0.223    4.429    0.000    0.988    0.998
## 
## R-Square:
##                    Estimate
##     C_W1_PHQ_total    0.399
##     IUS_B1W_change    0.002
## 
## Defined Parameters:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##     indirect1         0.002    0.003    0.569    0.569    0.002    0.009
##     direct            0.106    0.015    6.960    0.000    0.106    0.594
##     total             0.107    0.015    6.975    0.000    0.107    0.603

Psychoeducation group at 1 week: pre PHQ to 1 week PHQ via change in IU

Mediation.PHQ.psychoed.1W <-
  '#regressions
C_W1_PHQ_total ~ c1 * A_PRE_PHQ_total  
IUS_B1W_change ~ a1 * A_PRE_PHQ_total 
C_W1_PHQ_total ~ b1*IUS_B1W_change
indirect1 := a1 * b1
direct := c1
total := c1 + (a1 * b1)
'
PHQ.IUS.psychoed.1W <- sem(Mediation.PHQ.psychoed.1W, data=Psychoed_group, std.lv=T, std.ov=T, missing='fiml', se='robust', estimator='mlr', auto.var = T)

summary(PHQ.IUS.psychoed.1W, standardized=T, rsquare=T)
## lavaan 0.6.15 ended normally after 29 iterations
## 
##   Estimator                                         ML
##   Optimization method                           NLMINB
##   Number of model parameters                         7
## 
##   Number of observations                           106
##   Number of missing patterns                         1
## 
## Model Test User Model:
##                                               Standard      Scaled
##   Test Statistic                                 0.000       0.000
##   Degrees of freedom                                 0           0
## 
## Parameter Estimates:
## 
##   Standard errors                             Sandwich
##   Information bread                           Observed
##   Observed information based on                Hessian
## 
## Regressions:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##   C_W1_PHQ_total ~                                                      
##     A_PRE_PHQ (c1)    0.122    0.011   10.762    0.000    0.122    0.754
##   IUS_B1W_change ~                                                      
##     A_PRE_PHQ (a1)   -0.007    0.013   -0.560    0.576   -0.007   -0.043
##   C_W1_PHQ_total ~                                                      
##     IUS_B1W_c (b1)    0.212    0.060    3.523    0.000    0.212    0.212
## 
## Intercepts:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .C_W1_PHQ_total   -1.156    0.095  -12.147    0.000   -1.156   -1.161
##    .IUS_B1W_change    0.067    0.155    0.429    0.668    0.067    0.067
## 
## Variances:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .C_W1_PHQ_total    0.397    0.067    5.959    0.000    0.397    0.401
##    .IUS_B1W_change    0.989    0.233    4.249    0.000    0.989    0.998
## 
## R-Square:
##                    Estimate
##     C_W1_PHQ_total    0.599
##     IUS_B1W_change    0.002
## 
## Defined Parameters:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##     indirect1        -0.001    0.003   -0.554    0.580   -0.001   -0.009
##     direct            0.122    0.011   10.762    0.000    0.122    0.754
##     total             0.121    0.012   10.215    0.000    0.121    0.744

EC group at 1 week: pre PHQ to 1 week PHQ via change in IU

Mediation.PHQ.ECs.1W <-
  '#regressions
C_W1_PHQ_total ~ c1 * A_PRE_PHQ_total  
IUS_B1W_change ~ a1 * A_PRE_PHQ_total 
C_W1_PHQ_total ~ b1*IUS_B1W_change
indirect1 := a1 * b1
direct := c1
total := c1 + (a1 * b1)
'
PHQ.IUS.ECs.1W <- sem(Mediation.PHQ.ECs.1W, data=ECs_group, std.lv=T, std.ov=T, missing='fiml', se='robust', estimator='mlr', auto.var = T)

summary(PHQ.IUS.ECs.1W, standardized=T, rsquare=T)
## lavaan 0.6.15 ended normally after 22 iterations
## 
##   Estimator                                         ML
##   Optimization method                           NLMINB
##   Number of model parameters                         7
## 
##   Number of observations                            50
##   Number of missing patterns                         1
## 
## Model Test User Model:
##                                               Standard      Scaled
##   Test Statistic                                 0.000       0.000
##   Degrees of freedom                                 0           0
## 
## Parameter Estimates:
## 
##   Standard errors                             Sandwich
##   Information bread                           Observed
##   Observed information based on                Hessian
## 
## Regressions:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##   C_W1_PHQ_total ~                                                      
##     A_PRE_PHQ (c1)    0.122    0.009   13.936    0.000    0.122    0.701
##   IUS_B1W_change ~                                                      
##     A_PRE_PHQ (a1)    0.036    0.013    2.654    0.008    0.036    0.206
##   C_W1_PHQ_total ~                                                      
##     IUS_B1W_c (b1)    0.330    0.039    8.532    0.000    0.330    0.330
## 
## Intercepts:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .C_W1_PHQ_total   -1.210    0.106  -11.467    0.000   -1.210   -1.222
##    .IUS_B1W_change   -0.356    0.212   -1.681    0.093   -0.356   -0.360
## 
## Variances:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .C_W1_PHQ_total    0.298    0.083    3.575    0.000    0.298    0.304
##    .IUS_B1W_change    0.938    0.464    2.024    0.043    0.938    0.957
## 
## R-Square:
##                    Estimate
##     C_W1_PHQ_total    0.696
##     IUS_B1W_change    0.043
## 
## Defined Parameters:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##     indirect1         0.012    0.004    2.821    0.005    0.012    0.068
##     direct            0.122    0.009   13.936    0.000    0.122    0.701
##     total             0.133    0.009   14.943    0.000    0.133    0.769

Full sample at 1 month: Group to change in PHQ via change in IUS

Mediation.PHQchange.1M <-
  '#regressions
PHQ_B1M_change ~ c1 * Group  
IUS_B1M_change ~ a1 * Group 
PHQ_B1M_change ~ b1*IUS_B1M_change
indirect1 := a1 * b1
direct := c1
total := c1 + (a1 * b1)
'
group.IUS.PHQ.1M <- sem(Mediation.PHQchange.1M, data=changeinvariables, std.lv=T, std.ov=T, missing='fiml', se='robust', estimator='mlr', auto.var = T)

summary(group.IUS.PHQ.1M, standardized=T, rsquare=T)
## lavaan 0.6.15 ended normally after 16 iterations
## 
##   Estimator                                         ML
##   Optimization method                           NLMINB
##   Number of model parameters                         7
## 
##   Number of observations                           259
##   Number of missing patterns                         1
## 
## Model Test User Model:
##                                               Standard      Scaled
##   Test Statistic                                 0.000       0.000
##   Degrees of freedom                                 0           0
## 
## Parameter Estimates:
## 
##   Standard errors                             Sandwich
##   Information bread                           Observed
##   Observed information based on                Hessian
## 
## Regressions:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##   PHQ_B1M_change ~                                                      
##     Group     (c1)   -0.067    0.069   -0.974    0.330   -0.067   -0.050
##   IUS_B1M_change ~                                                      
##     Group     (a1)   -0.209    0.083   -2.526    0.012   -0.209   -0.155
##   PHQ_B1M_change ~                                                      
##     IUS_B1M_c (b1)    0.508    0.056    9.103    0.000    0.508    0.508
## 
## Intercepts:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .PHQ_B1M_change    0.149    0.161    0.923    0.356    0.149    0.149
##    .IUS_B1M_change    0.461    0.192    2.396    0.017    0.461    0.462
## 
## Variances:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .PHQ_B1M_change    0.729    0.076    9.631    0.000    0.729    0.732
##    .IUS_B1M_change    0.972    0.122    7.988    0.000    0.972    0.976
## 
## R-Square:
##                    Estimate
##     PHQ_B1M_change    0.268
##     IUS_B1M_change    0.024
## 
## Defined Parameters:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##     indirect1        -0.106    0.042   -2.536    0.011   -0.106   -0.079
##     direct           -0.067    0.069   -0.974    0.330   -0.067   -0.050
##     total            -0.174    0.081   -2.150    0.032   -0.174   -0.129

Intervention group at 1 month: pre PHQ to 1 month PHQ via change in IU

Mediation.PHQ.intervention.1M <-
  '#regressions
D_M1_PHQ_total ~ c1 * A_PRE_PHQ_total  
IUS_B1M_change ~ a1 * A_PRE_PHQ_total 
D_M1_PHQ_total ~ b1*IUS_B1M_change
indirect1 := a1 * b1
direct := c1
total := c1 + (a1 * b1)
'
PHQ.IUS.intervention.1M <- sem(Mediation.PHQ.intervention.1M, data=Intervention_group, std.lv=T, std.ov=T, missing='fiml', se='robust', estimator='mlr', auto.var = T)

summary(PHQ.IUS.intervention.1M, standardized=T, rsquare=T)
## lavaan 0.6.15 ended normally after 21 iterations
## 
##   Estimator                                         ML
##   Optimization method                           NLMINB
##   Number of model parameters                         7
## 
##   Number of observations                           103
##   Number of missing patterns                         1
## 
## Model Test User Model:
##                                               Standard      Scaled
##   Test Statistic                                 0.000       0.000
##   Degrees of freedom                                 0           0
## 
## Parameter Estimates:
## 
##   Standard errors                             Sandwich
##   Information bread                           Observed
##   Observed information based on                Hessian
## 
## Regressions:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##   D_M1_PHQ_total ~                                                      
##     A_PRE_PHQ (c1)    0.098    0.011    8.872    0.000    0.098    0.553
##   IUS_B1M_change ~                                                      
##     A_PRE_PHQ (a1)   -0.005    0.015   -0.321    0.748   -0.005   -0.027
##   D_M1_PHQ_total ~                                                      
##     IUS_B1M_c (b1)    0.388    0.076    5.081    0.000    0.388    0.388
## 
## Intercepts:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .D_M1_PHQ_total   -1.050    0.127   -8.255    0.000   -1.050   -1.055
##    .IUS_B1M_change    0.052    0.183    0.282    0.778    0.052    0.052
## 
## Variances:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .D_M1_PHQ_total    0.551    0.087    6.322    0.000    0.551    0.556
##    .IUS_B1M_change    0.990    0.184    5.380    0.000    0.990    0.999
## 
## R-Square:
##                    Estimate
##     D_M1_PHQ_total    0.444
##     IUS_B1M_change    0.001
## 
## Defined Parameters:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##     indirect1        -0.002    0.006   -0.315    0.753   -0.002   -0.011
##     direct            0.098    0.011    8.872    0.000    0.098    0.553
##     total             0.096    0.013    7.336    0.000    0.096    0.542

Psychoeducation group at 1 month: pre PHQ to 1 month PHQ via change in IU

Mediation.PHQ.psychoed.1M <-
  '#regressions
D_M1_PHQ_total ~ c1 * A_PRE_PHQ_total  
IUS_B1M_change ~ a1 * A_PRE_PHQ_total 
D_M1_PHQ_total ~ b1*IUS_B1M_change
indirect1 := a1 * b1
direct := c1
total := c1 + (a1 * b1)
'
PHQ.IUS.psychoed.1M <- sem(Mediation.PHQ.psychoed.1M, data=Psychoed_group, std.lv=T, std.ov=T, missing='fiml', se='robust', estimator='mlr', auto.var = T)

summary(PHQ.IUS.psychoed.1M, standardized=T, rsquare=T)
## lavaan 0.6.15 ended normally after 18 iterations
## 
##   Estimator                                         ML
##   Optimization method                           NLMINB
##   Number of model parameters                         7
## 
##   Number of observations                           106
##   Number of missing patterns                         1
## 
## Model Test User Model:
##                                               Standard      Scaled
##   Test Statistic                                 0.000       0.000
##   Degrees of freedom                                 0           0
## 
## Parameter Estimates:
## 
##   Standard errors                             Sandwich
##   Information bread                           Observed
##   Observed information based on                Hessian
## 
## Regressions:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##   D_M1_PHQ_total ~                                                      
##     A_PRE_PHQ (c1)    0.079    0.013    6.054    0.000    0.079    0.484
##   IUS_B1M_change ~                                                      
##     A_PRE_PHQ (a1)   -0.024    0.017   -1.417    0.157   -0.024   -0.148
##   D_M1_PHQ_total ~                                                      
##     IUS_B1M_c (b1)    0.467    0.065    7.219    0.000    0.467    0.467
## 
## Intercepts:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .D_M1_PHQ_total   -0.742    0.131   -5.681    0.000   -0.742   -0.746
##    .IUS_B1M_change    0.226    0.153    1.478    0.139    0.226    0.227
## 
## Variances:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .D_M1_PHQ_total    0.609    0.083    7.294    0.000    0.609    0.615
##    .IUS_B1M_change    0.969    0.182    5.319    0.000    0.969    0.978
## 
## R-Square:
##                    Estimate
##     D_M1_PHQ_total    0.385
##     IUS_B1M_change    0.022
## 
## Defined Parameters:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##     indirect1        -0.011    0.009   -1.310    0.190   -0.011   -0.069
##     direct            0.079    0.013    6.054    0.000    0.079    0.484
##     total             0.067    0.015    4.572    0.000    0.067    0.415

EC group at 1 month: pre PHQ to 1 month PHQ via change in IU

Mediation.PHQ.ECs.1M <-
  '#regressions
D_M1_PHQ_total ~ c1 * A_PRE_PHQ_total  
IUS_B1M_change ~ a1 * A_PRE_PHQ_total 
D_M1_PHQ_total ~ b1*IUS_B1M_change
indirect1 := a1 * b1
direct := c1
total := c1 + (a1 * b1)
'
PHQ.IUS.ECs.1M <- sem(Mediation.PHQ.ECs.1M, data=ECs_group, std.lv=T, std.ov=T, missing='fiml', se='robust', estimator='mlr', auto.var = T)

summary(PHQ.IUS.ECs.1M, standardized=T, rsquare=T)
## lavaan 0.6.15 ended normally after 23 iterations
## 
##   Estimator                                         ML
##   Optimization method                           NLMINB
##   Number of model parameters                         7
## 
##   Number of observations                            50
##   Number of missing patterns                         1
## 
## Model Test User Model:
##                                               Standard      Scaled
##   Test Statistic                                 0.000       0.000
##   Degrees of freedom                                 0           0
## 
## Parameter Estimates:
## 
##   Standard errors                             Sandwich
##   Information bread                           Observed
##   Observed information based on                Hessian
## 
## Regressions:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##   D_M1_PHQ_total ~                                                      
##     A_PRE_PHQ (c1)    0.092    0.016    5.603    0.000    0.092    0.533
##   IUS_B1M_change ~                                                      
##     A_PRE_PHQ (a1)   -0.009    0.023   -0.419    0.676   -0.009   -0.055
##   D_M1_PHQ_total ~                                                      
##     IUS_B1M_c (b1)    0.572    0.067    8.471    0.000    0.572    0.572
## 
## Intercepts:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .D_M1_PHQ_total   -0.920    0.176   -5.226    0.000   -0.920   -0.930
##    .IUS_B1M_change    0.094    0.221    0.427    0.670    0.094    0.095
## 
## Variances:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .D_M1_PHQ_total    0.414    0.097    4.255    0.000    0.414    0.423
##    .IUS_B1M_change    0.977    0.297    3.293    0.001    0.977    0.997
## 
## R-Square:
##                    Estimate
##     D_M1_PHQ_total    0.577
##     IUS_B1M_change    0.003
## 
## Defined Parameters:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##     indirect1        -0.005    0.013   -0.410    0.682   -0.005   -0.031
##     direct            0.092    0.016    5.603    0.000    0.092    0.533
##     total             0.087    0.022    4.038    0.000    0.087    0.502

H3: IUS mediating change in GAD over time between groups

Full sample at 1 week: Group to change in GAD via change in IUS

Mediation.GADchange.1W <-
  '#regressions
GAD_B1W_change ~ c1 * Group  
IUS_B1W_change ~ a1 * Group 
GAD_B1W_change ~ b1*IUS_B1W_change
indirect1 := a1 * b1
direct := c1
total := c1 + (a1 * b1)
'
group.IUS.GAD.1W <- sem(Mediation.GADchange.1W, data=changeinvariables, std.lv=T, std.ov=T, missing='fiml', se='robust', estimator='mlr', auto.var = T)

summary(group.IUS.GAD.1W, standardized=T, rsquare=T)
## lavaan 0.6.15 ended normally after 15 iterations
## 
##   Estimator                                         ML
##   Optimization method                           NLMINB
##   Number of model parameters                         7
## 
##   Number of observations                           259
##   Number of missing patterns                         1
## 
## Model Test User Model:
##                                               Standard      Scaled
##   Test Statistic                                 0.000       0.000
##   Degrees of freedom                                 0           0
## 
## Parameter Estimates:
## 
##   Standard errors                             Sandwich
##   Information bread                           Observed
##   Observed information based on                Hessian
## 
## Regressions:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##   GAD_B1W_change ~                                                      
##     Group     (c1)   -0.046    0.075   -0.615    0.538   -0.046   -0.034
##   IUS_B1W_change ~                                                      
##     Group     (a1)   -0.269    0.085   -3.152    0.002   -0.269   -0.199
##   GAD_B1W_change ~                                                      
##     IUS_B1W_c (b1)    0.337    0.068    4.954    0.000    0.337    0.337
## 
## Intercepts:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .GAD_B1W_change    0.102    0.157    0.648    0.517    0.102    0.102
##    .IUS_B1W_change    0.592    0.196    3.018    0.003    0.592    0.593
## 
## Variances:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .GAD_B1W_change    0.878    0.126    6.973    0.000    0.878    0.881
##    .IUS_B1W_change    0.957    0.158    6.071    0.000    0.957    0.960
## 
## R-Square:
##                    Estimate
##     GAD_B1W_change    0.119
##     IUS_B1W_change    0.040
## 
## Defined Parameters:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##     indirect1        -0.090    0.036   -2.502    0.012   -0.090   -0.067
##     direct           -0.046    0.075   -0.615    0.538   -0.046   -0.034
##     total            -0.137    0.079   -1.734    0.083   -0.137   -0.101

Intervention group at 1 week: pre GAD to 1 week GAD via change in IU

Mediation.GAD.intervention.1W <-
  '#regressions
C_W1_GAD_total ~ c1 * A_PRE_GAD_total  
IUS_B1W_change ~ a1 * A_PRE_GAD_total 
C_W1_GAD_total ~ b1*IUS_B1W_change
indirect1 := a1 * b1
direct := c1
total := c1 + (a1 * b1)
'
GAD.IUS.intervention.1W <- sem(Mediation.GAD.intervention.1W, data=Intervention_group, std.lv=T, std.ov=T, missing='fiml', se='robust', estimator='mlr', auto.var = T)

summary(GAD.IUS.intervention.1W, standardized=T, rsquare=T)
## lavaan 0.6.15 ended normally after 22 iterations
## 
##   Estimator                                         ML
##   Optimization method                           NLMINB
##   Number of model parameters                         7
## 
##   Number of observations                           103
##   Number of missing patterns                         1
## 
## Model Test User Model:
##                                               Standard      Scaled
##   Test Statistic                                 0.000       0.000
##   Degrees of freedom                                 0           0
## 
## Parameter Estimates:
## 
##   Standard errors                             Sandwich
##   Information bread                           Observed
##   Observed information based on                Hessian
## 
## Regressions:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##   C_W1_GAD_total ~                                                      
##     A_PRE_GAD (c1)    0.116    0.015    7.919    0.000    0.116    0.632
##   IUS_B1W_change ~                                                      
##     A_PRE_GAD (a1)   -0.004    0.017   -0.212    0.832   -0.004   -0.020
##   C_W1_GAD_total ~                                                      
##     IUS_B1W_c (b1)    0.202    0.097    2.089    0.037    0.202    0.202
## 
## Intercepts:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .C_W1_GAD_total   -1.073    0.117   -9.195    0.000   -1.073   -1.078
##    .IUS_B1W_change    0.034    0.191    0.177    0.860    0.034    0.034
## 
## Variances:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .C_W1_GAD_total    0.559    0.108    5.189    0.000    0.559    0.564
##    .IUS_B1W_change    0.990    0.223    4.435    0.000    0.990    1.000
## 
## R-Square:
##                    Estimate
##     C_W1_GAD_total    0.436
##     IUS_B1W_change    0.000
## 
## Defined Parameters:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##     indirect1        -0.001    0.004   -0.204    0.839   -0.001   -0.004
##     direct            0.116    0.015    7.919    0.000    0.116    0.632
##     total             0.115    0.015    7.784    0.000    0.115    0.628

Psychoeducation group at 1 week: pre GAD to 1 week GAD via change in IU

Mediation.GAD.psychoed.1W <-
  '#regressions
C_W1_GAD_total ~ c1 * A_PRE_GAD_total  
IUS_B1W_change ~ a1 * A_PRE_GAD_total 
C_W1_GAD_total ~ b1*IUS_B1W_change
indirect1 := a1 * b1
direct := c1
total := c1 + (a1 * b1)
'
GAD.IUS.psychoed.1W <- sem(Mediation.GAD.psychoed.1W, data=Psychoed_group, std.lv=T, std.ov=T, missing='fiml', se='robust', estimator='mlr', auto.var = T)

summary(GAD.IUS.psychoed.1W, standardized=T, rsquare=T)
## lavaan 0.6.15 ended normally after 24 iterations
## 
##   Estimator                                         ML
##   Optimization method                           NLMINB
##   Number of model parameters                         7
## 
##   Number of observations                           106
##   Number of missing patterns                         1
## 
## Model Test User Model:
##                                               Standard      Scaled
##   Test Statistic                                 0.000       0.000
##   Degrees of freedom                                 0           0
## 
## Parameter Estimates:
## 
##   Standard errors                             Sandwich
##   Information bread                           Observed
##   Observed information based on                Hessian
## 
## Regressions:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##   C_W1_GAD_total ~                                                      
##     A_PRE_GAD (c1)    0.127    0.011   11.850    0.000    0.127    0.732
##   IUS_B1W_change ~                                                      
##     A_PRE_GAD (a1)   -0.009    0.013   -0.695    0.487   -0.009   -0.053
##   C_W1_GAD_total ~                                                      
##     IUS_B1W_c (b1)    0.263    0.046    5.779    0.000    0.263    0.263
## 
## Intercepts:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .C_W1_GAD_total   -1.074    0.101  -10.685    0.000   -1.074   -1.079
##    .IUS_B1W_change    0.078    0.152    0.512    0.609    0.078    0.078
## 
## Variances:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .C_W1_GAD_total    0.411    0.083    4.965    0.000    0.411    0.415
##    .IUS_B1W_change    0.988    0.232    4.253    0.000    0.988    0.997
## 
## R-Square:
##                    Estimate
##     C_W1_GAD_total    0.585
##     IUS_B1W_change    0.003
## 
## Defined Parameters:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##     indirect1        -0.002    0.003   -0.698    0.485   -0.002   -0.014
##     direct            0.127    0.011   11.850    0.000    0.127    0.732
##     total             0.124    0.011   10.906    0.000    0.124    0.718

EC group at 1 week: pre GAD to 1 week GAD via change in IU

Mediation.GAD.ECs.1W <-
  '#regressions
C_W1_GAD_total ~ c1 * A_PRE_GAD_total  
IUS_B1W_change ~ a1 * A_PRE_GAD_total 
C_W1_GAD_total ~ b1*IUS_B1W_change
indirect1 := a1 * b1
direct := c1
total := c1 + (a1 * b1)
'
GAD.IUS.ECs.1W <- sem(Mediation.GAD.ECs.1W, data=ECs_group, std.lv=T, std.ov=T, missing='fiml', se='robust', estimator='mlr', auto.var = T)

summary(GAD.IUS.ECs.1W, standardized=T, rsquare=T)
## lavaan 0.6.15 ended normally after 23 iterations
## 
##   Estimator                                         ML
##   Optimization method                           NLMINB
##   Number of model parameters                         7
## 
##   Number of observations                            50
##   Number of missing patterns                         1
## 
## Model Test User Model:
##                                               Standard      Scaled
##   Test Statistic                                 0.000       0.000
##   Degrees of freedom                                 0           0
## 
## Parameter Estimates:
## 
##   Standard errors                             Sandwich
##   Information bread                           Observed
##   Observed information based on                Hessian
## 
## Regressions:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##   C_W1_GAD_total ~                                                      
##     A_PRE_GAD (c1)    0.152    0.013   11.427    0.000    0.152    0.739
##   IUS_B1W_change ~                                                      
##     A_PRE_GAD (a1)    0.027    0.017    1.622    0.105    0.027    0.131
##   C_W1_GAD_total ~                                                      
##     IUS_B1W_c (b1)    0.337    0.040    8.535    0.000    0.337    0.337
## 
## Intercepts:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .C_W1_GAD_total   -1.221    0.105  -11.594    0.000   -1.221   -1.234
##    .IUS_B1W_change   -0.216    0.213   -1.014    0.311   -0.216   -0.218
## 
## Variances:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .C_W1_GAD_total    0.270    0.057    4.701    0.000    0.270    0.275
##    .IUS_B1W_change    0.963    0.469    2.055    0.040    0.963    0.983
## 
## R-Square:
##                    Estimate
##     C_W1_GAD_total    0.725
##     IUS_B1W_change    0.017
## 
## Defined Parameters:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##     indirect1         0.009    0.005    1.717    0.086    0.009    0.044
##     direct            0.152    0.013   11.427    0.000    0.152    0.739
##     total             0.161    0.014   11.719    0.000    0.161    0.783

Full sample at 1 month: Group to change in GAD via change in IUS

Mediation.GADchange.1M <-
  '#regressions
GAD_B1M_change ~ c1 * Group  
IUS_B1M_change ~ a1 * Group 
GAD_B1M_change ~ b1*IUS_B1M_change
indirect1 := a1 * b1
direct := c1
total := c1 + (a1 * b1)
'
group.IUS.GAD.1M <- sem(Mediation.GADchange.1M, data=changeinvariables, std.lv=T, std.ov=T, missing='fiml', se='robust', estimator='mlr', auto.var = T)

summary(group.IUS.GAD.1M, standardized=T, rsquare=T)
## lavaan 0.6.15 ended normally after 17 iterations
## 
##   Estimator                                         ML
##   Optimization method                           NLMINB
##   Number of model parameters                         7
## 
##   Number of observations                           259
##   Number of missing patterns                         1
## 
## Model Test User Model:
##                                               Standard      Scaled
##   Test Statistic                                 0.000       0.000
##   Degrees of freedom                                 0           0
## 
## Parameter Estimates:
## 
##   Standard errors                             Sandwich
##   Information bread                           Observed
##   Observed information based on                Hessian
## 
## Regressions:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##   GAD_B1M_change ~                                                      
##     Group     (c1)   -0.079    0.070   -1.135    0.257   -0.079   -0.059
##   IUS_B1M_change ~                                                      
##     Group     (a1)   -0.209    0.083   -2.526    0.012   -0.209   -0.155
##   GAD_B1M_change ~                                                      
##     IUS_B1M_c (b1)    0.535    0.054    9.865    0.000    0.535    0.535
## 
## Intercepts:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .GAD_B1M_change    0.175    0.168    1.042    0.297    0.175    0.175
##    .IUS_B1M_change    0.461    0.192    2.396    0.017    0.461    0.462
## 
## Variances:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .GAD_B1M_change    0.698    0.080    8.748    0.000    0.698    0.701
##    .IUS_B1M_change    0.972    0.122    7.988    0.000    0.972    0.976
## 
## R-Square:
##                    Estimate
##     GAD_B1M_change    0.299
##     IUS_B1M_change    0.024
## 
## Defined Parameters:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##     indirect1        -0.112    0.045   -2.490    0.013   -0.112   -0.083
##     direct           -0.079    0.070   -1.135    0.257   -0.079   -0.059
##     total            -0.191    0.081   -2.369    0.018   -0.191   -0.142

Intervention group at 1 month: pre GAD to 1 month GAD via change in IU

Mediation.GAD.intervention.1M <-
  '#regressions
D_M1_GAD_total ~ c1 * A_PRE_GAD_total  
IUS_B1M_change ~ a1 * A_PRE_GAD_total 
D_M1_GAD_total ~ b1*IUS_B1M_change
indirect1 := a1 * b1
direct := c1
total := c1 + (a1 * b1)
'
GAD.IUS.intervention.1M <- sem(Mediation.GAD.intervention.1M, data=Intervention_group, std.lv=T, std.ov=T, missing='fiml', se='robust', estimator='mlr', auto.var = T)

summary(GAD.IUS.intervention.1M, standardized=T, rsquare=T)
## lavaan 0.6.15 ended normally after 25 iterations
## 
##   Estimator                                         ML
##   Optimization method                           NLMINB
##   Number of model parameters                         7
## 
##   Number of observations                           103
##   Number of missing patterns                         1
## 
## Model Test User Model:
##                                               Standard      Scaled
##   Test Statistic                                 0.000       0.000
##   Degrees of freedom                                 0           0
## 
## Parameter Estimates:
## 
##   Standard errors                             Sandwich
##   Information bread                           Observed
##   Observed information based on                Hessian
## 
## Regressions:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##   D_M1_GAD_total ~                                                      
##     A_PRE_GAD (c1)    0.114    0.012    9.726    0.000    0.114    0.624
##   IUS_B1M_change ~                                                      
##     A_PRE_GAD (a1)   -0.004    0.014   -0.249    0.803   -0.004   -0.020
##   D_M1_GAD_total ~                                                      
##     IUS_B1M_c (b1)    0.403    0.071    5.693    0.000    0.403    0.403
## 
## Intercepts:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .D_M1_GAD_total   -1.059    0.104  -10.159    0.000   -1.059   -1.064
##    .IUS_B1M_change    0.033    0.159    0.210    0.833    0.033    0.034
## 
## Variances:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .D_M1_GAD_total    0.453    0.072    6.283    0.000    0.453    0.458
##    .IUS_B1M_change    0.990    0.184    5.378    0.000    0.990    1.000
## 
## R-Square:
##                    Estimate
##     D_M1_GAD_total    0.542
##     IUS_B1M_change    0.000
## 
## Defined Parameters:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##     indirect1        -0.001    0.006   -0.246    0.806   -0.001   -0.008
##     direct            0.114    0.012    9.726    0.000    0.114    0.624
##     total             0.113    0.013    8.766    0.000    0.113    0.616

Psychoeducation group at 1 month: pre GAD to 1 month GAD via change in IU

Mediation.GAD.psychoed.1M <-
  '#regressions
D_M1_GAD_total ~ c1 * A_PRE_GAD_total  
IUS_B1M_change ~ a1 * A_PRE_GAD_total 
D_M1_GAD_total ~ b1*IUS_B1M_change
indirect1 := a1 * b1
direct := c1
total := c1 + (a1 * b1)
'
GAD.IUS.psychoed.1M <- sem(Mediation.GAD.psychoed.1M, data=Psychoed_group, std.lv=T, std.ov=T, missing='fiml', se='robust', estimator='mlr', auto.var = T)

summary(GAD.IUS.psychoed.1M, standardized=T, rsquare=T)
## lavaan 0.6.15 ended normally after 19 iterations
## 
##   Estimator                                         ML
##   Optimization method                           NLMINB
##   Number of model parameters                         7
## 
##   Number of observations                           106
##   Number of missing patterns                         1
## 
## Model Test User Model:
##                                               Standard      Scaled
##   Test Statistic                                 0.000       0.000
##   Degrees of freedom                                 0           0
## 
## Parameter Estimates:
## 
##   Standard errors                             Sandwich
##   Information bread                           Observed
##   Observed information based on                Hessian
## 
## Regressions:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##   D_M1_GAD_total ~                                                      
##     A_PRE_GAD (c1)    0.089    0.012    7.614    0.000    0.089    0.512
##   IUS_B1M_change ~                                                      
##     A_PRE_GAD (a1)   -0.026    0.019   -1.347    0.178   -0.026   -0.149
##   D_M1_GAD_total ~                                                      
##     IUS_B1M_c (b1)    0.539    0.063    8.591    0.000    0.539    0.539
## 
## Intercepts:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .D_M1_GAD_total   -0.752    0.104   -7.211    0.000   -0.752   -0.756
##    .IUS_B1M_change    0.218    0.153    1.427    0.154    0.218    0.219
## 
## Variances:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .D_M1_GAD_total    0.524    0.073    7.162    0.000    0.524    0.529
##    .IUS_B1M_change    0.969    0.179    5.407    0.000    0.969    0.978
## 
## R-Square:
##                    Estimate
##     D_M1_GAD_total    0.471
##     IUS_B1M_change    0.022
## 
## Defined Parameters:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##     indirect1        -0.014    0.011   -1.279    0.201   -0.014   -0.080
##     direct            0.089    0.012    7.614    0.000    0.089    0.512
##     total             0.075    0.016    4.631    0.000    0.075    0.432

EC group at 1 month: pre GAD to 1 month GAD via change in IU

Mediation.GAD.ECs.1M <-
  '#regressions
D_M1_GAD_total ~ c1 * A_PRE_GAD_total  
IUS_B1M_change ~ a1 * A_PRE_GAD_total 
D_M1_GAD_total ~ b1*IUS_B1M_change
indirect1 := a1 * b1
direct := c1
total := c1 + (a1 * b1)
'
GAD.IUS.ECs.1M <- sem(Mediation.GAD.ECs.1M, data=ECs_group, std.lv=T, std.ov=T, missing='fiml', se='robust', estimator='mlr', auto.var = T)

summary(GAD.IUS.ECs.1M, standardized=T, rsquare=T)
## lavaan 0.6.15 ended normally after 28 iterations
## 
##   Estimator                                         ML
##   Optimization method                           NLMINB
##   Number of model parameters                         7
## 
##   Number of observations                            50
##   Number of missing patterns                         1
## 
## Model Test User Model:
##                                               Standard      Scaled
##   Test Statistic                                 0.000       0.000
##   Degrees of freedom                                 0           0
## 
## Parameter Estimates:
## 
##   Standard errors                             Sandwich
##   Information bread                           Observed
##   Observed information based on                Hessian
## 
## Regressions:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##   D_M1_GAD_total ~                                                      
##     A_PRE_GAD (c1)    0.093    0.025    3.772    0.000    0.093    0.450
##   IUS_B1M_change ~                                                      
##     A_PRE_GAD (a1)    0.002    0.018    0.122    0.903    0.002    0.011
##   D_M1_GAD_total ~                                                      
##     IUS_B1M_c (b1)    0.503    0.053    9.546    0.000    0.503    0.503
## 
## Intercepts:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .D_M1_GAD_total   -0.745    0.236   -3.156    0.002   -0.745   -0.752
##    .IUS_B1M_change   -0.018    0.192   -0.093    0.926   -0.018   -0.018
## 
## Variances:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .D_M1_GAD_total    0.529    0.153    3.444    0.001    0.529    0.539
##    .IUS_B1M_change    0.980    0.301    3.256    0.001    0.980    1.000
## 
## R-Square:
##                    Estimate
##     D_M1_GAD_total    0.461
##     IUS_B1M_change    0.000
## 
## Defined Parameters:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##     indirect1         0.001    0.009    0.122    0.903    0.001    0.005
##     direct            0.093    0.025    3.772    0.000    0.093    0.450
##     total             0.094    0.026    3.575    0.000    0.094    0.456

H3: IUS mediating change in mood over time between groups

Full sample at post: Group to change in mood via change in IUS

Mediation.Moodchange.post <-
  '#regressions
Mood_BP_change ~ c1 * Group  
IUS_BP_change ~ a1 * Group 
Mood_BP_change ~ b1*IUS_BP_change
indirect1 := a1 * b1
direct := c1
total := c1 + (a1 * b1)
'
group.IUS.Mood.post <- sem(Mediation.Moodchange.post, data=changeinvariables, std.lv=T, std.ov=T, missing='fiml', se='robust', estimator='mlr', auto.var = T)

summary(group.IUS.Mood.post, standardized=T, rsquare=T)
## lavaan 0.6.15 ended normally after 17 iterations
## 
##   Estimator                                         ML
##   Optimization method                           NLMINB
##   Number of model parameters                         7
## 
##   Number of observations                           259
##   Number of missing patterns                         2
## 
## Model Test User Model:
##                                               Standard      Scaled
##   Test Statistic                                 0.000       0.000
##   Degrees of freedom                                 0           0
## 
## Parameter Estimates:
## 
##   Standard errors                             Sandwich
##   Information bread                           Observed
##   Observed information based on                Hessian
## 
## Regressions:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##   Mood_BP_change ~                                                      
##     Group     (c1)    0.320    0.077    4.160    0.000    0.320    0.236
##   IUS_BP_change ~                                                       
##     Group     (a1)   -0.419    0.079   -5.274    0.000   -0.419   -0.311
##   Mood_BP_change ~                                                      
##     IUS_BP_ch (b1)   -0.241    0.077   -3.154    0.002   -0.241   -0.240
## 
## Intercepts:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .Mood_BP_change   -0.699    0.166   -4.212    0.000   -0.699   -0.697
##    .IUS_BP_change     0.923    0.159    5.816    0.000    0.923    0.924
## 
## Variances:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .Mood_BP_change    0.855    0.144    5.952    0.000    0.855    0.851
##    .IUS_BP_change     0.900    0.139    6.490    0.000    0.900    0.903
## 
## R-Square:
##                    Estimate
##     Mood_BP_change    0.149
##     IUS_BP_change     0.097
## 
## Defined Parameters:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##     indirect1         0.101    0.034    2.984    0.003    0.101    0.075
##     direct            0.320    0.077    4.160    0.000    0.320    0.236
##     total             0.421    0.072    5.811    0.000    0.421    0.311

Intervention group at post: pre mood to post mood via change in IU

Mediation.Mood.intervention.post <-
  '#regressions
B_POST_mood_mean ~ c1 * A_PRE_mood_mean  
IUS_BP_change ~ a1 * A_PRE_mood_mean 
B_POST_mood_mean ~ b1*IUS_BP_change
indirect1 := a1 * b1
direct := c1
total := c1 + (a1 * b1)
'
Mood.IUS.intervention.post <- sem(Mediation.Mood.intervention.post, data=Intervention_group, std.lv=T, std.ov=T, missing='fiml', se='robust', estimator='mlr', auto.var = T)

summary(Mood.IUS.intervention.post, standardized=T, rsquare=T)
## lavaan 0.6.15 ended normally after 17 iterations
## 
##   Estimator                                         ML
##   Optimization method                           NLMINB
##   Number of model parameters                         7
## 
##   Number of observations                           103
##   Number of missing patterns                         2
## 
## Model Test User Model:
##                                               Standard      Scaled
##   Test Statistic                                 0.000       0.000
##   Degrees of freedom                                 0           0
## 
## Parameter Estimates:
## 
##   Standard errors                             Sandwich
##   Information bread                           Observed
##   Observed information based on                Hessian
## 
## Regressions:
##                      Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##   B_POST_mood_mean ~                                                      
##     A_PRE_md_ (c1)      0.014    0.003    4.715    0.000    0.014    0.600
##   IUS_BP_change ~                                                         
##     A_PRE_md_ (a1)      0.001    0.002    0.530    0.596    0.001    0.051
##   B_POST_mood_mean ~                                                      
##     IUS_BP_ch (b1)     -0.162    0.093   -1.751    0.080   -0.162   -0.162
## 
## Intercepts:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .B_POST_mood_mn   -0.407    0.147   -2.767    0.006   -0.407   -0.407
##    .IUS_BP_change    -0.035    0.099   -0.357    0.721   -0.035   -0.035
## 
## Variances:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .B_POST_mood_mn    0.624    0.128    4.889    0.000    0.624    0.624
##    .IUS_BP_change     0.988    0.224    4.416    0.000    0.988    0.997
## 
## R-Square:
##                    Estimate
##     B_POST_mood_mn    0.376
##     IUS_BP_change     0.003
## 
## Defined Parameters:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##     indirect1        -0.000    0.000   -0.495    0.621   -0.000   -0.008
##     direct            0.014    0.003    4.715    0.000    0.014    0.600
##     total             0.014    0.003    4.724    0.000    0.014    0.591

Psychoeducation group at post: pre mood to post mood via change in IU

Mediation.Mood.psychoed.post <-
  '#regressions
B_POST_mood_mean ~ c1 * A_PRE_mood_mean  
IUS_BP_change ~ a1 * A_PRE_mood_mean 
B_POST_mood_mean ~ b1*IUS_BP_change
indirect1 := a1 * b1
direct := c1
total := c1 + (a1 * b1)
'
Mood.IUS.psychoed.post <- sem(Mediation.Mood.psychoed.post, data=Psychoed_group, std.lv=T, std.ov=T, missing='fiml', se='robust', estimator='mlr', auto.var = T)

summary(Mood.IUS.psychoed.post, standardized=T, rsquare=T)
## lavaan 0.6.15 ended normally after 24 iterations
## 
##   Estimator                                         ML
##   Optimization method                           NLMINB
##   Number of model parameters                         7
## 
##   Number of observations                           106
##   Number of missing patterns                         1
## 
## Model Test User Model:
##                                               Standard      Scaled
##   Test Statistic                                 0.000       0.000
##   Degrees of freedom                                 0           0
## 
## Parameter Estimates:
## 
##   Standard errors                             Sandwich
##   Information bread                           Observed
##   Observed information based on                Hessian
## 
## Regressions:
##                      Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##   B_POST_mood_mean ~                                                      
##     A_PRE_md_ (c1)      0.016    0.002    7.911    0.000    0.016    0.695
##   IUS_BP_change ~                                                         
##     A_PRE_md_ (a1)      0.002    0.002    0.698    0.485    0.002    0.067
##   B_POST_mood_mean ~                                                      
##     IUS_BP_ch (b1)     -0.241    0.069   -3.506    0.000   -0.241   -0.241
## 
## Intercepts:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .B_POST_mood_mn   -0.589    0.119   -4.939    0.000   -0.589   -0.592
##    .IUS_BP_change    -0.057    0.138   -0.412    0.681   -0.057   -0.057
## 
## Variances:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .B_POST_mood_mn    0.477    0.083    5.756    0.000    0.477    0.482
##    .IUS_BP_change     0.986    0.196    5.026    0.000    0.986    0.996
## 
## R-Square:
##                    Estimate
##     B_POST_mood_mn    0.518
##     IUS_BP_change     0.004
## 
## Defined Parameters:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##     indirect1        -0.000    0.001   -0.647    0.518   -0.000   -0.016
##     direct            0.016    0.002    7.911    0.000    0.016    0.695
##     total             0.015    0.002    7.039    0.000    0.015    0.679

EC group at post: pre mood to post mood via change in IU

Mediation.Mood.ECs.post <-
  '#regressions
B_POST_mood_mean ~ c1 * A_PRE_mood_mean  
IUS_BP_change ~ a1 * A_PRE_mood_mean 
B_POST_mood_mean ~ b1*IUS_BP_change
indirect1 := a1 * b1
direct := c1
total := c1 + (a1 * b1)
'
Mood.IUS.ECs.post <- sem(Mediation.Mood.ECs.post, data=ECs_group, std.lv=T, std.ov=T, missing='fiml', se='robust', estimator='mlr', auto.var = T)

summary(Mood.IUS.ECs.post, standardized=T, rsquare=T)
## lavaan 0.6.15 ended normally after 25 iterations
## 
##   Estimator                                         ML
##   Optimization method                           NLMINB
##   Number of model parameters                         7
## 
##   Number of observations                            50
##   Number of missing patterns                         1
## 
## Model Test User Model:
##                                               Standard      Scaled
##   Test Statistic                                 0.000       0.000
##   Degrees of freedom                                 0           0
## 
## Parameter Estimates:
## 
##   Standard errors                             Sandwich
##   Information bread                           Observed
##   Observed information based on                Hessian
## 
## Regressions:
##                      Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##   B_POST_mood_mean ~                                                      
##     A_PRE_md_ (c1)      0.022    0.002   13.400    0.000    0.022    0.868
##   IUS_BP_change ~                                                         
##     A_PRE_md_ (a1)      0.001    0.003    0.236    0.814    0.001    0.025
##   B_POST_mood_mean ~                                                      
##     IUS_BP_ch (b1)     -0.063    0.060   -1.052    0.293   -0.063   -0.063
## 
## Intercepts:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .B_POST_mood_mn   -0.895    0.104   -8.601    0.000   -0.895   -0.904
##    .IUS_BP_change    -0.026    0.178   -0.147    0.883   -0.026   -0.026
## 
## Variances:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .B_POST_mood_mn    0.240    0.063    3.801    0.000    0.240    0.245
##    .IUS_BP_change     0.979    0.226    4.336    0.000    0.979    0.999
## 
## R-Square:
##                    Estimate
##     B_POST_mood_mn    0.755
##     IUS_BP_change     0.001
## 
## Defined Parameters:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##     indirect1        -0.000    0.000   -0.231    0.817   -0.000   -0.002
##     direct            0.022    0.002   13.400    0.000    0.022    0.868
##     total             0.022    0.002   13.364    0.000    0.022    0.866

Full sample at 1 week: Group to change in mood via change in IUS

Mediation.Moodchange.1W <-
  '#regressions
Mood_B1W_change ~ c1 * Group  
IUS_B1W_change ~ a1 * Group 
Mood_B1W_change ~ b1*IUS_B1W_change
indirect1 := a1 * b1
direct := c1
total := c1 + (a1 * b1)
'
group.IUS.Mood.1W <- sem(Mediation.Moodchange.1W, data=changeinvariables, std.lv=T, std.ov=T, missing='fiml', se='robust', estimator='mlr', auto.var = T)

summary(group.IUS.Mood.1W, standardized=T, rsquare=T)
## lavaan 0.6.15 ended normally after 17 iterations
## 
##   Estimator                                         ML
##   Optimization method                           NLMINB
##   Number of model parameters                         7
## 
##   Number of observations                           259
##   Number of missing patterns                         2
## 
## Model Test User Model:
##                                               Standard      Scaled
##   Test Statistic                                 0.000       0.000
##   Degrees of freedom                                 0           0
## 
## Parameter Estimates:
## 
##   Standard errors                             Sandwich
##   Information bread                           Observed
##   Observed information based on                Hessian
## 
## Regressions:
##                     Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##   Mood_B1W_change ~                                                      
##     Group     (c1)     0.020    0.088    0.224    0.823    0.020    0.015
##   IUS_B1W_change ~                                                       
##     Group     (a1)    -0.269    0.085   -3.152    0.002   -0.269   -0.199
##   Mood_B1W_change ~                                                      
##     IUS_B1W_c (b1)    -0.211    0.094   -2.246    0.025   -0.211   -0.209
## 
## Intercepts:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .Mood_B1W_chang   -0.021    0.207   -0.102    0.919   -0.021   -0.021
##    .IUS_B1W_change    0.592    0.196    3.018    0.003    0.592    0.593
## 
## Variances:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .Mood_B1W_chang    0.966    0.121    7.987    0.000    0.966    0.955
##    .IUS_B1W_change    0.957    0.158    6.071    0.000    0.957    0.960
## 
## R-Square:
##                    Estimate
##     Mood_B1W_chang    0.045
##     IUS_B1W_change    0.040
## 
## Defined Parameters:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##     indirect1         0.057    0.030    1.914    0.056    0.057    0.042
##     direct            0.020    0.088    0.224    0.823    0.020    0.015
##     total             0.076    0.082    0.929    0.353    0.076    0.056

Intervention group at 1 week: pre mood to 1 week mood via change in IU

Mediation.Mood.intervention.1W <-
  '#regressions
C_W1_mood_mean ~ c1 * A_PRE_mood_mean  
IUS_B1W_change ~ a1 * A_PRE_mood_mean 
C_W1_mood_mean ~ b1*IUS_B1W_change
indirect1 := a1 * b1
direct := c1
total := c1 + (a1 * b1)
'
Mood.IUS.intervention.1W <- sem(Mediation.Mood.intervention.1W, data=Intervention_group, std.lv=T, std.ov=T, missing='fiml', se='robust', estimator='mlr', auto.var = T)

summary(Mood.IUS.intervention.1W, standardized=T, rsquare=T)
## lavaan 0.6.15 ended normally after 17 iterations
## 
##   Estimator                                         ML
##   Optimization method                           NLMINB
##   Number of model parameters                         7
## 
##   Number of observations                           103
##   Number of missing patterns                         2
## 
## Model Test User Model:
##                                               Standard      Scaled
##   Test Statistic                                 0.000       0.000
##   Degrees of freedom                                 0           0
## 
## Parameter Estimates:
## 
##   Standard errors                             Sandwich
##   Information bread                           Observed
##   Observed information based on                Hessian
## 
## Regressions:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##   C_W1_mood_mean ~                                                      
##     A_PRE_md_ (c1)    0.011    0.002    4.421    0.000    0.011    0.457
##   IUS_B1W_change ~                                                      
##     A_PRE_md_ (a1)    0.001    0.002    0.492    0.623    0.001    0.051
##   C_W1_mood_mean ~                                                      
##     IUS_B1W_c (b1)   -0.133    0.114   -1.173    0.241   -0.133   -0.134
## 
## Intercepts:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .C_W1_mood_mean   -0.311    0.110   -2.833    0.005   -0.311   -0.313
##    .IUS_B1W_change   -0.036    0.114   -0.311    0.756   -0.036   -0.036
## 
## Variances:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .C_W1_mood_mean    0.768    0.155    4.947    0.000    0.768    0.779
##    .IUS_B1W_change    0.988    0.220    4.496    0.000    0.988    0.997
## 
## R-Square:
##                    Estimate
##     C_W1_mood_mean    0.221
##     IUS_B1W_change    0.003
## 
## Defined Parameters:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##     indirect1        -0.000    0.000   -0.478    0.633   -0.000   -0.007
##     direct            0.011    0.002    4.421    0.000    0.011    0.457
##     total             0.011    0.003    4.150    0.000    0.011    0.450

Psychoeducation group at 1 week: pre mood to 1 week mood via change in IU

Mediation.Mood.psychoed.1W <-
  '#regressions
C_W1_mood_mean ~ c1 * A_PRE_mood_mean  
IUS_B1W_change ~ a1 * A_PRE_mood_mean 
C_W1_mood_mean ~ b1*IUS_B1W_change
indirect1 := a1 * b1
direct := c1
total := c1 + (a1 * b1)
'
Mood.IUS.psychoed.1W <- sem(Mediation.Mood.psychoed.1W, data=Psychoed_group, std.lv=T, std.ov=T, missing='fiml', se='robust', estimator='mlr', auto.var = T)

summary(Mood.IUS.psychoed.1W, standardized=T, rsquare=T)
## lavaan 0.6.15 ended normally after 18 iterations
## 
##   Estimator                                         ML
##   Optimization method                           NLMINB
##   Number of model parameters                         7
## 
##   Number of observations                           106
##   Number of missing patterns                         2
## 
## Model Test User Model:
##                                               Standard      Scaled
##   Test Statistic                                 0.000       0.000
##   Degrees of freedom                                 0           0
## 
## Parameter Estimates:
## 
##   Standard errors                             Sandwich
##   Information bread                           Observed
##   Observed information based on                Hessian
## 
## Regressions:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##   C_W1_mood_mean ~                                                      
##     A_PRE_md_ (c1)    0.010    0.002    4.478    0.000    0.010    0.444
##   IUS_B1W_change ~                                                      
##     A_PRE_md_ (a1)    0.001    0.002    0.633    0.527    0.001    0.062
##   C_W1_mood_mean ~                                                      
##     IUS_B1W_c (b1)   -0.198    0.120   -1.648    0.099   -0.198   -0.196
## 
## Intercepts:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .C_W1_mood_mean   -0.361    0.135   -2.670    0.008   -0.361   -0.359
##    .IUS_B1W_change   -0.053    0.133   -0.397    0.691   -0.053   -0.053
## 
## Variances:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .C_W1_mood_mean    0.781    0.126    6.223    0.000    0.781    0.776
##    .IUS_B1W_change    0.987    0.235    4.206    0.000    0.987    0.996
## 
## R-Square:
##                    Estimate
##     C_W1_mood_mean    0.224
##     IUS_B1W_change    0.004
## 
## Defined Parameters:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##     indirect1        -0.000    0.001   -0.554    0.579   -0.000   -0.012
##     direct            0.010    0.002    4.478    0.000    0.010    0.444
##     total             0.010    0.002    4.164    0.000    0.010    0.431

EC group at 1 week: pre mood to 1 week mood via change in IU

Mediation.Mood.ECs.1W <-
  '#regressions
C_W1_mood_mean ~ c1 * A_PRE_mood_mean  
IUS_B1W_change ~ a1 * A_PRE_mood_mean 
C_W1_mood_mean ~ b1*IUS_B1W_change
indirect1 := a1 * b1
direct := c1
total := c1 + (a1 * b1)
'
Mood.IUS.ECs.1W <- sem(Mediation.Mood.ECs.1W, data=ECs_group, std.lv=T, std.ov=T, missing='fiml', se='robust', estimator='mlr', auto.var = T)

summary(Mood.IUS.ECs.1W, standardized=T, rsquare=T)
## lavaan 0.6.15 ended normally after 22 iterations
## 
##   Estimator                                         ML
##   Optimization method                           NLMINB
##   Number of model parameters                         7
## 
##   Number of observations                            50
##   Number of missing patterns                         2
## 
## Model Test User Model:
##                                               Standard      Scaled
##   Test Statistic                                 0.000       0.000
##   Degrees of freedom                                 0           0
## 
## Parameter Estimates:
## 
##   Standard errors                             Sandwich
##   Information bread                           Observed
##   Observed information based on                Hessian
## 
## Regressions:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##   C_W1_mood_mean ~                                                      
##     A_PRE_md_ (c1)    0.011    0.003    3.198    0.001    0.011    0.434
##   IUS_B1W_change ~                                                      
##     A_PRE_md_ (a1)    0.003    0.002    2.046    0.041    0.003    0.136
##   C_W1_mood_mean ~                                                      
##     IUS_B1W_c (b1)   -0.078    0.244   -0.320    0.749   -0.078   -0.078
## 
## Intercepts:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .C_W1_mood_mean   -0.434    0.178   -2.433    0.015   -0.434   -0.439
##    .IUS_B1W_change   -0.140    0.154   -0.906    0.365   -0.140   -0.141
## 
## Variances:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .C_W1_mood_mean    0.795    0.129    6.163    0.000    0.795    0.815
##    .IUS_B1W_change    0.962    0.476    2.021    0.043    0.962    0.982
## 
## R-Square:
##                    Estimate
##     C_W1_mood_mean    0.185
##     IUS_B1W_change    0.018
## 
## Defined Parameters:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##     indirect1        -0.000    0.001   -0.314    0.754   -0.000   -0.011
##     direct            0.011    0.003    3.198    0.001    0.011    0.434
##     total             0.011    0.003    3.239    0.001    0.011    0.424

Full sample at 1 month: Group to change in mood via change in IUS

Mediation.Moodchange.1M <-
  '#regressions
Mood_B1M_change ~ c1 * Group  
IUS_B1M_change ~ a1 * Group 
Mood_B1M_change ~ b1*IUS_B1M_change
indirect1 := a1 * b1
direct := c1
total := c1 + (a1 * b1)
'
group.IUS.Mood.1M <- sem(Mediation.Moodchange.1M, data=changeinvariables, std.lv=T, std.ov=T, missing='fiml', se='robust', estimator='mlr', auto.var = T)

summary(group.IUS.Mood.1M, standardized=T, rsquare=T)
## lavaan 0.6.15 ended normally after 17 iterations
## 
##   Estimator                                         ML
##   Optimization method                           NLMINB
##   Number of model parameters                         7
## 
##   Number of observations                           259
##   Number of missing patterns                         2
## 
## Model Test User Model:
##                                               Standard      Scaled
##   Test Statistic                                 0.000       0.000
##   Degrees of freedom                                 0           0
## 
## Parameter Estimates:
## 
##   Standard errors                             Sandwich
##   Information bread                           Observed
##   Observed information based on                Hessian
## 
## Regressions:
##                     Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##   Mood_B1M_change ~                                                      
##     Group     (c1)     0.025    0.089    0.281    0.779    0.025    0.017
##   IUS_B1M_change ~                                                       
##     Group     (a1)    -0.209    0.083   -2.526    0.012   -0.209   -0.155
##   Mood_B1M_change ~                                                      
##     IUS_B1M_c (b1)    -0.433    0.126   -3.450    0.001   -0.433   -0.409
## 
## Intercepts:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .Mood_B1M_chang    0.064    0.219    0.293    0.770    0.064    0.061
##    .IUS_B1M_change    0.461    0.192    2.396    0.017    0.461    0.462
## 
## Variances:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .Mood_B1M_chang    0.926    0.104    8.930    0.000    0.926    0.830
##    .IUS_B1M_change    0.972    0.122    7.988    0.000    0.972    0.976
## 
## R-Square:
##                    Estimate
##     Mood_B1M_chang    0.170
##     IUS_B1M_change    0.024
## 
## Defined Parameters:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##     indirect1         0.090    0.044    2.067    0.039    0.090    0.063
##     direct            0.025    0.089    0.281    0.779    0.025    0.017
##     total             0.115    0.089    1.296    0.195    0.115    0.081

Intervention group at 1 month: pre mood to 1 month mood via change in IU

Mediation.Mood.intervention.1M <-
  '#regressions
D_M1_mood_mean ~ c1 * A_PRE_mood_mean  
IUS_B1M_change ~ a1 * A_PRE_mood_mean 
D_M1_mood_mean ~ b1*IUS_B1M_change
indirect1 := a1 * b1
direct := c1
total := c1 + (a1 * b1)
'
Mood.IUS.intervention.1M <- sem(Mediation.Mood.intervention.1M, data=Intervention_group, std.lv=T, std.ov=T, missing='fiml', se='robust', estimator='mlr', auto.var = T)

summary(Mood.IUS.intervention.1M, standardized=T, rsquare=T)
## lavaan 0.6.15 ended normally after 13 iterations
## 
##   Estimator                                         ML
##   Optimization method                           NLMINB
##   Number of model parameters                         7
## 
##   Number of observations                           103
##   Number of missing patterns                         2
## 
## Model Test User Model:
##                                               Standard      Scaled
##   Test Statistic                                 0.000       0.000
##   Degrees of freedom                                 0           0
## 
## Parameter Estimates:
## 
##   Standard errors                             Sandwich
##   Information bread                           Observed
##   Observed information based on                Hessian
## 
## Regressions:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##   D_M1_mood_mean ~                                                      
##     A_PRE_md_ (c1)    0.010    0.002    4.288    0.000    0.010    0.417
##   IUS_B1M_change ~                                                      
##     A_PRE_md_ (a1)    0.003    0.002    1.160    0.246    0.003    0.111
##   D_M1_mood_mean ~                                                      
##     IUS_B1M_c (b1)   -0.329    0.135   -2.428    0.015   -0.329   -0.320
## 
## Intercepts:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .D_M1_mood_mean   -0.209    0.115   -1.806    0.071   -0.209   -0.204
##    .IUS_B1M_change   -0.077    0.120   -0.645    0.519   -0.077   -0.078
## 
## Variances:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .D_M1_mood_mean    0.785    0.140    5.613    0.000    0.785    0.753
##    .IUS_B1M_change    0.978    0.183    5.338    0.000    0.978    0.988
## 
## R-Square:
##                    Estimate
##     D_M1_mood_mean    0.247
##     IUS_B1M_change    0.012
## 
## Defined Parameters:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##     indirect1        -0.001    0.001   -1.014    0.311   -0.001   -0.036
##     direct            0.010    0.002    4.288    0.000    0.010    0.417
##     total             0.009    0.003    3.594    0.000    0.009    0.381

Psychoeducation group at 1 month: pre mood to 1 month mood via change in IU

Mediation.Mood.psychoed.1M <-
  '#regressions
D_M1_mood_mean ~ c1 * A_PRE_mood_mean  
IUS_B1M_change ~ a1 * A_PRE_mood_mean 
D_M1_mood_mean ~ b1*IUS_B1M_change
indirect1 := a1 * b1
direct := c1
total := c1 + (a1 * b1)
'
Mood.IUS.psychoed.1M <- sem(Mediation.Mood.psychoed.1M, data=Psychoed_group, std.lv=T, std.ov=T, missing='fiml', se='robust', estimator='mlr', auto.var = T)

summary(Mood.IUS.psychoed.1M, standardized=T, rsquare=T)
## lavaan 0.6.15 ended normally after 23 iterations
## 
##   Estimator                                         ML
##   Optimization method                           NLMINB
##   Number of model parameters                         7
## 
##   Number of observations                           106
##   Number of missing patterns                         2
## 
## Model Test User Model:
##                                               Standard      Scaled
##   Test Statistic                                 0.000       0.000
##   Degrees of freedom                                 0           0
## 
## Parameter Estimates:
## 
##   Standard errors                             Sandwich
##   Information bread                           Observed
##   Observed information based on                Hessian
## 
## Regressions:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##   D_M1_mood_mean ~                                                      
##     A_PRE_md_ (c1)    0.010    0.002    4.816    0.000    0.010    0.422
##   IUS_B1M_change ~                                                      
##     A_PRE_md_ (a1)    0.001    0.002    0.786    0.432    0.001    0.060
##   D_M1_mood_mean ~                                                      
##     IUS_B1M_c (b1)   -0.342    0.145   -2.365    0.018   -0.342   -0.332
## 
## Intercepts:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .D_M1_mood_mean   -0.274    0.131   -2.092    0.036   -0.274   -0.268
##    .IUS_B1M_change   -0.051    0.121   -0.423    0.672   -0.051   -0.051
## 
## Variances:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .D_M1_mood_mean    0.764    0.093    8.253    0.000    0.764    0.728
##    .IUS_B1M_change    0.987    0.197    5.013    0.000    0.987    0.996
## 
## R-Square:
##                    Estimate
##     D_M1_mood_mean    0.272
##     IUS_B1M_change    0.004
## 
## Defined Parameters:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##     indirect1        -0.000    0.001   -0.698    0.485   -0.000   -0.020
##     direct            0.010    0.002    4.816    0.000    0.010    0.422
##     total             0.009    0.002    4.193    0.000    0.009    0.403

EC group at 1 month: pre mood to 1 month mood via change in IU

Mediation.Mood.EC.1M <-
  '#regressions
D_M1_mood_mean ~ c1 * A_PRE_mood_mean  
IUS_B1M_change ~ a1 * A_PRE_mood_mean 
D_M1_mood_mean ~ b1*IUS_B1M_change
indirect1 := a1 * b1
direct := c1
total := c1 + (a1 * b1)
'
Mood.IUS.EC.1M <- sem(Mediation.Mood.EC.1M, data=ECs_group, std.lv=T, std.ov=T, missing='fiml', se='robust', estimator='mlr', auto.var = T)

summary(Mood.IUS.EC.1M, standardized=T, rsquare=T)
## lavaan 0.6.15 ended normally after 26 iterations
## 
##   Estimator                                         ML
##   Optimization method                           NLMINB
##   Number of model parameters                         7
## 
##   Number of observations                            50
##   Number of missing patterns                         2
## 
## Model Test User Model:
##                                               Standard      Scaled
##   Test Statistic                                 0.000       0.000
##   Degrees of freedom                                 0           0
## 
## Parameter Estimates:
## 
##   Standard errors                             Sandwich
##   Information bread                           Observed
##   Observed information based on                Hessian
## 
## Regressions:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##   D_M1_mood_mean ~                                                      
##     A_PRE_md_ (c1)    0.016    0.002    7.150    0.000    0.016    0.594
##   IUS_B1M_change ~                                                      
##     A_PRE_md_ (a1)   -0.001    0.003   -0.189    0.850   -0.001   -0.024
##   D_M1_mood_mean ~                                                      
##     IUS_B1M_c (b1)   -0.294    0.292   -1.007    0.314   -0.294   -0.278
## 
## Intercepts:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .D_M1_mood_mean   -0.562    0.206   -2.725    0.006   -0.562   -0.538
##    .IUS_B1M_change    0.025    0.177    0.142    0.887    0.025    0.025
## 
## Variances:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##    .D_M1_mood_mean    0.616    0.122    5.039    0.000    0.616    0.562
##    .IUS_B1M_change    0.979    0.300    3.266    0.001    0.979    0.999
## 
## R-Square:
##                    Estimate
##     D_M1_mood_mean    0.438
##     IUS_B1M_change    0.001
## 
## Defined Parameters:
##                    Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
##     indirect1         0.000    0.001    0.187    0.852    0.000    0.007
##     direct            0.016    0.002    7.150    0.000    0.016    0.594
##     total             0.016    0.002    6.726    0.000    0.016    0.600

Exploratory analyses

E1: Growth mindsets moderating the association between group and PHQ changes across time

# 1 week
moderation_GM_PHQ_1W <- lm(PHQ_B1W_change ~ Group*A_PRE_GM, data = changeinvariables)
summary(moderation_GM_PHQ_1W)
## 
## Call:
## lm(formula = PHQ_B1W_change ~ Group * A_PRE_GM, data = changeinvariables)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -18.0497  -2.3842   0.3762   2.7917  12.7478 
## 
## Coefficients:
##                              Estimate Std. Error t value Pr(>|t|)  
## (Intercept)                   -2.5685     1.4565  -1.763   0.0790 .
## GroupB_Controls                0.7047     1.8204   0.387   0.6990  
## GroupC_Intervention            1.4282     1.8140   0.787   0.4318  
## A_PRE_GM                       0.7768     0.4746   1.637   0.1029  
## GroupB_Controls:A_PRE_GM      -0.5289     0.5705  -0.927   0.3547  
## GroupC_Intervention:A_PRE_GM  -0.9793     0.5824  -1.682   0.0939 .
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 4.641 on 253 degrees of freedom
## Multiple R-squared:  0.02464,    Adjusted R-squared:  0.005359 
## F-statistic: 1.278 on 5 and 253 DF,  p-value: 0.2738
anova(moderation_GM_PHQ_1W)
## Analysis of Variance Table
## 
## Response: PHQ_B1W_change
##                 Df Sum Sq Mean Sq F value Pr(>F)
## Group            2   58.9  29.474  1.3687 0.2563
## A_PRE_GM         1   16.0  15.954  0.7409 0.3902
## Group:A_PRE_GM   2   62.7  31.355  1.4560 0.2351
## Residuals      253 5448.3  21.535
# 1 month
moderation_GM_PHQ_1M <- lm(PHQ_B1M_change ~ Group*A_PRE_GM, data = changeinvariables)
summary(moderation_GM_PHQ_1M)
## 
## Call:
## lm(formula = PHQ_B1M_change ~ Group * A_PRE_GM, data = changeinvariables)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -18.625  -3.325   0.553   3.680  21.375 
## 
## Coefficients:
##                              Estimate Std. Error t value Pr(>|t|)
## (Intercept)                   -0.7511     2.0051  -0.375    0.708
## GroupB_Controls               -2.3907     2.5059  -0.954    0.341
## GroupC_Intervention           -1.5126     2.4971  -0.606    0.545
## A_PRE_GM                      -0.1127     0.6533  -0.173    0.863
## GroupB_Controls:A_PRE_GM       0.4071     0.7853   0.518    0.605
## GroupC_Intervention:A_PRE_GM  -0.2412     0.8017  -0.301    0.764
## 
## Residual standard error: 6.388 on 253 degrees of freedom
## Multiple R-squared:  0.02073,    Adjusted R-squared:  0.001377 
## F-statistic: 1.071 on 5 and 253 DF,  p-value: 0.3768
anova(moderation_GM_PHQ_1M)
## Analysis of Variance Table
## 
## Response: PHQ_B1M_change
##                 Df  Sum Sq Mean Sq F value Pr(>F)
## Group            2   175.0  87.518  2.1446 0.1192
## A_PRE_GM         1     0.4   0.417  0.0102 0.9195
## Group:A_PRE_GM   2    43.1  21.552  0.5281 0.5904
## Residuals      253 10324.6  40.809

E1: Growth mindsets moderating the association between group and GAD changes across time

# 1 week
moderation_GM_GAD_1W <- lm(GAD_B1W_change ~ Group*A_PRE_GM, data = changeinvariables)
summary(moderation_GM_GAD_1W)
## 
## Call:
## lm(formula = GAD_B1W_change ~ Group * A_PRE_GM, data = changeinvariables)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -18.6066  -2.0403   0.2757   2.2738  15.1541 
## 
## Coefficients:
##                              Estimate Std. Error t value Pr(>|t|)  
## (Intercept)                   -1.7641     1.3993  -1.261    0.209  
## GroupB_Controls                2.4269     1.7488   1.388    0.166  
## GroupC_Intervention            0.8493     1.7427   0.487    0.626  
## A_PRE_GM                       0.6511     0.4559   1.428    0.154  
## GroupB_Controls:A_PRE_GM      -1.1204     0.5481  -2.044    0.042 *
## GroupC_Intervention:A_PRE_GM  -0.7708     0.5595  -1.378    0.170  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 4.458 on 253 degrees of freedom
## Multiple R-squared:  0.02821,    Adjusted R-squared:  0.009001 
## F-statistic: 1.469 on 5 and 253 DF,  p-value: 0.2005
anova(moderation_GM_GAD_1W)
## Analysis of Variance Table
## 
## Response: GAD_B1W_change
##                 Df Sum Sq Mean Sq F value Pr(>F)
## Group            2   55.4  27.698  1.3936 0.2501
## A_PRE_GM         1    7.5   7.496  0.3771 0.5397
## Group:A_PRE_GM   2   83.1  41.532  2.0896 0.1259
## Residuals      253 5028.6  19.876
# 1 month
moderation_GM_GAD_1M <- lm(GAD_B1M_change ~ Group*A_PRE_GM, data = changeinvariables)
summary(moderation_GM_GAD_1M)
## 
## Call:
## lm(formula = GAD_B1M_change ~ Group * A_PRE_GM, data = changeinvariables)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -18.9788  -2.7062   0.6174   3.0740  18.6317 
## 
## Coefficients:
##                              Estimate Std. Error t value Pr(>|t|)
## (Intercept)                  -0.90274    1.81773  -0.497    0.620
## GroupB_Controls              -0.95514    2.27178  -0.420    0.675
## GroupC_Intervention          -0.03819    2.26384  -0.017    0.987
## A_PRE_GM                      0.27107    0.59225   0.458    0.648
## GroupB_Controls:A_PRE_GM     -0.29829    0.71193  -0.419    0.676
## GroupC_Intervention:A_PRE_GM -0.82988    0.72677  -1.142    0.255
## 
## Residual standard error: 5.791 on 253 degrees of freedom
## Multiple R-squared:  0.02999,    Adjusted R-squared:  0.01082 
## F-statistic: 1.564 on 5 and 253 DF,  p-value: 0.1707
anova(moderation_GM_GAD_1M)
## Analysis of Variance Table
## 
## Response: GAD_B1M_change
##                 Df Sum Sq Mean Sq F value  Pr(>F)  
## Group            2  196.1  98.051  2.9234 0.05557 .
## A_PRE_GM         1   14.7  14.655  0.4369 0.50920  
## Group:A_PRE_GM   2   51.6  25.779  0.7686 0.46473  
## Residuals      253 8485.5  33.540                  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

E1: Growth mindsets moderating the association between group and mood changes across time

# post
moderation_GM_mood_BP <- lm(Mood_BP_change ~ Group*A_PRE_GM, data = changeinvariables)
summary(moderation_GM_mood_BP)
## 
## Call:
## lm(formula = Mood_BP_change ~ Group * A_PRE_GM, data = changeinvariables)
## 
## Residuals:
##    Min     1Q Median     3Q    Max 
## -72.57 -19.14  -3.97  15.23 171.32 
## 
## Coefficients:
##                              Estimate Std. Error t value Pr(>|t|)   
## (Intercept)                    -8.380     10.096  -0.830  0.40732   
## GroupB_Controls                31.642     12.618   2.508  0.01278 * 
## GroupC_Intervention            38.846     12.577   3.089  0.00223 **
## A_PRE_GM                        3.051      3.289   0.928  0.35455   
## GroupB_Controls:A_PRE_GM       -4.098      3.954  -1.036  0.30097   
## GroupC_Intervention:A_PRE_GM   -3.409      4.037  -0.845  0.39913   
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 32.17 on 252 degrees of freedom
##   (1 observation deleted due to missingness)
## Multiple R-squared:  0.1039, Adjusted R-squared:  0.08617 
## F-statistic: 5.847 on 5 and 252 DF,  p-value: 3.945e-05
anova(moderation_GM_mood_BP)
## Analysis of Variance Table
## 
## Response: Mood_BP_change
##                 Df Sum Sq Mean Sq F value    Pr(>F)    
## Group            2  29097 14548.3 14.0609 1.625e-06 ***
## A_PRE_GM         1      0     0.0  0.0000    0.9988    
## Group:A_PRE_GM   2   1150   575.1  0.5558    0.5743    
## Residuals      252 260736  1034.7                      
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
# 1 week
moderation_GM_mood_1W <- lm(Mood_B1W_change ~ Group*A_PRE_GM, data = changeinvariables)
summary(moderation_GM_mood_1W)
## 
## Call:
## lm(formula = Mood_B1W_change ~ Group * A_PRE_GM, data = changeinvariables)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -185.320  -23.963    2.028   25.016  185.316 
## 
## Coefficients:
##                              Estimate Std. Error t value Pr(>|t|)
## (Intercept)                  -15.2285    15.0894  -1.009    0.314
## GroupB_Controls               15.5416    18.8441   0.825    0.410
## GroupC_Intervention           16.4329    18.9320   0.868    0.386
## A_PRE_GM                       0.1816     4.8849   0.037    0.970
## GroupB_Controls:A_PRE_GM      -2.3073     5.8749  -0.393    0.695
## GroupC_Intervention:A_PRE_GM  -2.5278     6.0285  -0.419    0.675
## 
## Residual standard error: 47.61 on 244 degrees of freedom
##   (9 observations deleted due to missingness)
## Multiple R-squared:  0.00881,    Adjusted R-squared:  -0.0115 
## F-statistic: 0.4338 on 5 and 244 DF,  p-value: 0.8248
anova(moderation_GM_mood_1W)
## Analysis of Variance Table
## 
## Response: Mood_B1W_change
##                 Df Sum Sq Mean Sq F value Pr(>F)
## Group            2   2952  1476.0  0.6511 0.5224
## A_PRE_GM         1   1516  1515.6  0.6686 0.4143
## Group:A_PRE_GM   2    449   224.5  0.0990 0.9057
## Residuals      244 553135  2266.9
# 1 month
moderation_GM_mood_1W <- lm(Mood_B1M_change ~ Group*A_PRE_GM, data = changeinvariables)
summary(moderation_GM_mood_1W)
## 
## Call:
## lm(formula = Mood_B1M_change ~ Group * A_PRE_GM, data = changeinvariables)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -150.984  -28.262    4.266   26.821  183.538 
## 
## Coefficients:
##                              Estimate Std. Error t value Pr(>|t|)  
## (Intercept)                  -29.7643    16.1860  -1.839   0.0673 .
## GroupB_Controls               27.3570    20.3907   1.342   0.1811  
## GroupC_Intervention           13.0556    20.0891   0.650   0.5164  
## A_PRE_GM                       3.0544     5.2872   0.578   0.5641  
## GroupB_Controls:A_PRE_GM      -4.2805     6.4042  -0.668   0.5046  
## GroupC_Intervention:A_PRE_GM   0.4938     6.4979   0.076   0.9395  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 48.65 on 222 degrees of freedom
##   (31 observations deleted due to missingness)
## Multiple R-squared:  0.02057,    Adjusted R-squared:  -0.001493 
## F-statistic: 0.9323 on 5 and 222 DF,  p-value: 0.4608
anova(moderation_GM_mood_1W)
## Analysis of Variance Table
## 
## Response: Mood_B1M_change
##                 Df Sum Sq Mean Sq F value Pr(>F)
## Group            2   7881  3940.5  1.6652 0.1915
## A_PRE_GM         1    904   904.4  0.3822 0.5371
## Group:A_PRE_GM   2   2246  1122.9  0.4745 0.6228
## Residuals      222 525329  2366.3

E2: Association between IUS and baseline functional impairment

# Total FI scale
PRE_IUS_FI_lm <- lm(A_PRE_IUS_total ~ A_PRE_FI_total, data = Full_data_all)
summary(PRE_IUS_FI_lm)
## 
## Call:
## lm(formula = A_PRE_IUS_total ~ A_PRE_FI_total, data = Full_data_all)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -20.1561  -4.1561   0.3806   5.0026  15.3806 
## 
## Coefficients:
##                Estimate Std. Error t value Pr(>|t|)    
## (Intercept)     29.1074     1.2521   23.25   <2e-16 ***
## A_PRE_FI_total   1.2927     0.1148   11.26   <2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 7.265 on 257 degrees of freedom
## Multiple R-squared:  0.3304, Adjusted R-squared:  0.3278 
## F-statistic: 126.8 on 1 and 257 DF,  p-value: < 2.2e-16
anova(PRE_IUS_FI_lm)
## Analysis of Variance Table
## 
## Response: A_PRE_IUS_total
##                 Df  Sum Sq Mean Sq F value    Pr(>F)    
## A_PRE_FI_total   1  6692.6  6692.6   126.8 < 2.2e-16 ***
## Residuals      257 13565.1    52.8                      
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Plot_lm_fi <- ggplot(data = Full_data_all, aes(x = A_PRE_IUS_total, y = A_PRE_FI_total)) +
  geom_point() +
  geom_smooth(method = "lm", color = "black")
print(Plot_lm_fi)
## `geom_smooth()` using formula = 'y ~ x'

# Friends item
PRE_IUS_friends_lm <- lm(A_PRE_IUS_total ~ B_FI_friends, data = Full_data_all)
summary(PRE_IUS_friends_lm)
## 
## Call:
## lm(formula = A_PRE_IUS_total ~ B_FI_friends, data = Full_data_all)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -26.7419  -5.1496   0.6416   6.0252  19.2581 
## 
## Coefficients:
##              Estimate Std. Error t value Pr(>|t|)    
## (Intercept)   36.5090     0.9703  37.627  < 2e-16 ***
## B_FI_friends   3.2329     0.4657   6.942 3.16e-11 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 8.159 on 256 degrees of freedom
##   (1 observation deleted due to missingness)
## Multiple R-squared:  0.1584, Adjusted R-squared:  0.1551 
## F-statistic:  48.2 on 1 and 256 DF,  p-value: 3.163e-11
anova(PRE_IUS_friends_lm)
## Analysis of Variance Table
## 
## Response: A_PRE_IUS_total
##               Df  Sum Sq Mean Sq F value    Pr(>F)    
## B_FI_friends   1  3208.3  3208.3  48.195 3.163e-11 ***
## Residuals    256 17041.8    66.6                      
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
# Strangers item
PRE_IUS_strangers_lm <- lm(A_PRE_IUS_total ~ B_FI_strangers, data = Full_data_all)
summary(PRE_IUS_strangers_lm)
## 
## Call:
## lm(formula = A_PRE_IUS_total ~ B_FI_strangers, data = Full_data_all)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -27.857  -4.948   1.075   6.143  20.189 
## 
## Coefficients:
##                Estimate Std. Error t value Pr(>|t|)    
## (Intercept)     36.7655     1.3132  27.997  < 2e-16 ***
## B_FI_strangers   2.0457     0.4489   4.557 8.03e-06 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 8.548 on 256 degrees of freedom
##   (1 observation deleted due to missingness)
## Multiple R-squared:  0.07504,    Adjusted R-squared:  0.07142 
## F-statistic: 20.77 on 1 and 256 DF,  p-value: 8.034e-06
anova(PRE_IUS_strangers_lm)
## Analysis of Variance Table
## 
## Response: A_PRE_IUS_total
##                 Df  Sum Sq Mean Sq F value    Pr(>F)    
## B_FI_strangers   1  1517.6 1517.59  20.768 8.034e-06 ***
## Residuals      256 18707.0   73.07                      
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
# Work item
PRE_IUS_work_lm <- lm(A_PRE_IUS_total ~ B_FI_work, data = Full_data_all)
summary(PRE_IUS_work_lm)
## 
## Call:
## lm(formula = A_PRE_IUS_total ~ B_FI_work, data = Full_data_all)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -20.9966  -4.3259   0.0034   5.8388  18.9211 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)   33.997      1.037  32.772   <2e-16 ***
## B_FI_work      4.082      0.451   9.051   <2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 7.705 on 243 degrees of freedom
##   (14 observations deleted due to missingness)
## Multiple R-squared:  0.2521, Adjusted R-squared:  0.249 
## F-statistic: 81.92 on 1 and 243 DF,  p-value: < 2.2e-16
anova(PRE_IUS_work_lm)
## Analysis of Variance Table
## 
## Response: A_PRE_IUS_total
##            Df  Sum Sq Mean Sq F value    Pr(>F)    
## B_FI_work   1  4863.8  4863.8   81.92 < 2.2e-16 ***
## Residuals 243 14427.5    59.4                      
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
# Education item
PRE_IUS_education_lm <- lm(A_PRE_IUS_total ~ B_FI_education, data = Full_data_all)
summary(PRE_IUS_education_lm)
## 
## Call:
## lm(formula = A_PRE_IUS_total ~ B_FI_education, data = Full_data_all)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -22.9221  -4.9864   0.0136   5.9976  18.0457 
## 
## Coefficients:
##                Estimate Std. Error t value Pr(>|t|)    
## (Intercept)     35.9221     1.0023  35.840  < 2e-16 ***
## B_FI_education   3.0321     0.4192   7.233 5.79e-12 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 8.153 on 249 degrees of freedom
##   (8 observations deleted due to missingness)
## Multiple R-squared:  0.1736, Adjusted R-squared:  0.1703 
## F-statistic: 52.32 on 1 and 249 DF,  p-value: 5.793e-12
anova(PRE_IUS_education_lm)
## Analysis of Variance Table
## 
## Response: A_PRE_IUS_total
##                 Df  Sum Sq Mean Sq F value    Pr(>F)    
## B_FI_education   1  3478.3  3478.3  52.323 5.793e-12 ***
## Residuals      249 16552.6    66.5                      
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
# Hobbies item
PRE_IUS_hobbies_lm <- lm(A_PRE_IUS_total ~ B_FI_hobbies, data = Full_data_all)
summary(PRE_IUS_hobbies_lm)
## 
## Call:
## lm(formula = A_PRE_IUS_total ~ B_FI_hobbies, data = Full_data_all)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -23.3848  -5.5183   0.0813   5.4372  17.6152 
## 
## Coefficients:
##              Estimate Std. Error t value Pr(>|t|)    
## (Intercept)   36.3848     0.9140  39.807  < 2e-16 ***
## B_FI_hobbies   3.1780     0.4149   7.659 3.85e-13 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 8.022 on 256 degrees of freedom
##   (1 observation deleted due to missingness)
## Multiple R-squared:  0.1864, Adjusted R-squared:  0.1832 
## F-statistic: 58.66 on 1 and 256 DF,  p-value: 3.853e-13
anova(PRE_IUS_hobbies_lm)
## Analysis of Variance Table
## 
## Response: A_PRE_IUS_total
##               Df  Sum Sq Mean Sq F value    Pr(>F)    
## B_FI_hobbies   1  3775.2  3775.2  58.662 3.853e-13 ***
## Residuals    256 16474.9    64.4                      
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

E2: Change in functional impairment over time across groups

# Merging across timepoints
FI_alltimepoints <- Full_data_all %>% 
  dplyr::select("ID", "Group", "A_PRE_FI_total", "C_W1_FI_total", "D_M1_FI_total")
## Formatting table as needed
FI_alltimepoints_long <- FI_alltimepoints %>%
  pivot_longer(cols = c(A_PRE_FI_total, C_W1_FI_total, D_M1_FI_total),
               names_to = "Time",
               values_to = "FI_Score")
FI_MEM <- lmer(FI_Score ~ Group * Time + (1|ID), data = FI_alltimepoints_long, REML = TRUE)
summary(FI_MEM)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: FI_Score ~ Group * Time + (1 | ID)
##    Data: FI_alltimepoints_long
## 
## REML criterion at convergence: 4331.6
## 
## Scaled residuals: 
##      Min       1Q   Median       3Q      Max 
## -3.13881 -0.46729  0.01098  0.49621  2.78553 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev.
##  ID       (Intercept)  9.265   3.044   
##  Residual             10.064   3.172   
## Number of obs: 777, groups:  ID, 259
## 
## Fixed effects:
##                                       Estimate Std. Error       df t value
## (Intercept)                             9.8600     0.6218 526.1943  15.858
## GroupB_Controls                         0.1683     0.7543 526.1943   0.223
## GroupC_Intervention                     0.6157     0.7578 526.1943   0.813
## TimeC_W1_FI_total                      -0.2800     0.6345 512.0000  -0.441
## TimeD_M1_FI_total                      -1.3000     0.6345 512.0000  -2.049
## GroupB_Controls:TimeC_W1_FI_total      -0.1540     0.7697 512.0000  -0.200
## GroupC_Intervention:TimeC_W1_FI_total  -0.9724     0.7733 512.0000  -1.258
## GroupB_Controls:TimeD_M1_FI_total      -0.4170     0.7697 512.0000  -0.542
## GroupC_Intervention:TimeD_M1_FI_total  -0.8748     0.7733 512.0000  -1.131
##                                       Pr(>|t|)    
## (Intercept)                             <2e-16 ***
## GroupB_Controls                          0.824    
## GroupC_Intervention                      0.417    
## TimeC_W1_FI_total                        0.659    
## TimeD_M1_FI_total                        0.041 *  
## GroupB_Controls:TimeC_W1_FI_total        0.842    
## GroupC_Intervention:TimeC_W1_FI_total    0.209    
## GroupB_Controls:TimeD_M1_FI_total        0.588    
## GroupC_Intervention:TimeD_M1_FI_total    0.258    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) GrpB_C GrpC_I TC_W1_ TD_M1_ GB_C:TC GC_I:TC GB_C:TD
## GrpB_Cntrls -0.824                                                    
## GrpC_Intrvn -0.820  0.676                                             
## TmC_W1_FI_t -0.510  0.421  0.419                                      
## TmD_M1_FI_t -0.510  0.421  0.419  0.500                               
## GB_C:TC_W1_  0.421 -0.510 -0.345 -0.824 -0.412                        
## GC_I:TC_W1_  0.419 -0.345 -0.510 -0.820 -0.410  0.676                 
## GB_C:TD_M1_  0.421 -0.510 -0.345 -0.412 -0.824  0.500   0.338         
## GC_I:TD_M1_  0.419 -0.345 -0.510 -0.410 -0.820  0.338   0.500   0.676
anova  (FI_MEM)
## Type III Analysis of Variance Table with Satterthwaite's method
##            Sum Sq Mean Sq NumDF DenDF F value    Pr(>F)    
## Group        0.02   0.012     2   256  0.0012    0.9988    
## Time       351.05 175.527     2   512 17.4411 4.702e-08 ***
## Group:Time  27.41   6.852     4   512  0.6808    0.6055    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
sjPlot::tab_model(FI_MEM)
  FI Score
Predictors Estimates CI p
(Intercept) 9.86 8.64 – 11.08 <0.001
Group [B_Controls] 0.17 -1.31 – 1.65 0.823
Group [C_Intervention] 0.62 -0.87 – 2.10 0.417
Time [C_W1_FI_total] -0.28 -1.53 – 0.97 0.659
Time [D_M1_FI_total] -1.30 -2.55 – -0.05 0.041
Group [B_Controls] × Time
[C_W1_FI_total] 		-0.15 -1.66 – 1.36 0.842
Group [C_Intervention] ×
Time [C_W1_FI_total] 		-0.97 -2.49 – 0.55 0.209
Group [B_Controls] × Time
[D_M1_FI_total] 		-0.42 -1.93 – 1.09 0.588
Group [C_Intervention] ×
Time [D_M1_FI_total] 		-0.87 -2.39 – 0.64 0.258
Random Effects
σ2 10.06
Ï„00 ID 9.27
ICC 0.48
N ID 259
Observations 777
Marginal R2 / Conditional R2 0.030 / 0.495 
parameters::standardise_parameters(FI_MEM)
## # Standardization method: refit
## 
## Parameter                             | Std. Coef. |         95% CI
## -------------------------------------------------------------------
## (Intercept)                           |       0.12 | [-0.15,  0.40]
## GroupB_Controls                       |       0.04 | [-0.30,  0.37]
## GroupC_Intervention                   |       0.14 | [-0.20,  0.47]
## TimeC_W1_FI_total                     |      -0.06 | [-0.34,  0.22]
## TimeD_M1_FI_total                     |      -0.29 | [-0.57, -0.01]
## GroupB_Controls:TimeC_W1_FI_total     |      -0.03 | [-0.37,  0.31]
## GroupC_Intervention:TimeC_W1_FI_total |      -0.22 | [-0.56,  0.12]
## GroupB_Controls:TimeD_M1_FI_total     |      -0.09 | [-0.43,  0.25]
## GroupC_Intervention:TimeD_M1_FI_total |      -0.20 | [-0.54,  0.14]

Baseline to 1W

# Merging across timepoints
FI_B1W <- Full_data_all %>% 
  dplyr::select("ID", "Group", "A_PRE_FI_total", "C_W1_FI_total")
## Formatting table as needed
FI_B1W_long <- FI_B1W %>%
  pivot_longer(cols = c(A_PRE_FI_total, C_W1_FI_total),
               names_to = "Time",
               values_to = "FI_Score")
FI_MEM_B1W <- lmer(FI_Score ~ Group * Time + (1|ID), data = FI_B1W_long, REML = TRUE)
summary(FI_MEM_B1W)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: FI_Score ~ Group * Time + (1 | ID)
##    Data: FI_B1W_long
## 
## REML criterion at convergence: 2835.9
## 
## Scaled residuals: 
##      Min       1Q   Median       3Q      Max 
## -2.85506 -0.47819  0.01114  0.51077  2.57159 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev.
##  ID       (Intercept) 9.146    3.024   
##  Residual             7.698    2.775   
## Number of obs: 518, groups:  ID, 259
## 
## Fixed effects:
##                                       Estimate Std. Error       df t value
## (Intercept)                             9.8600     0.5804 395.4238  16.988
## GroupB_Controls                         0.1683     0.7041 395.4238   0.239
## GroupC_Intervention                     0.6157     0.7074 395.4238   0.870
## TimeC_W1_FI_total                      -0.2800     0.5549 256.0000  -0.505
## GroupB_Controls:TimeC_W1_FI_total      -0.1540     0.6732 256.0000  -0.229
## GroupC_Intervention:TimeC_W1_FI_total  -0.9724     0.6763 256.0000  -1.438
##                                       Pr(>|t|)    
## (Intercept)                             <2e-16 ***
## GroupB_Controls                          0.811    
## GroupC_Intervention                      0.385    
## TimeC_W1_FI_total                        0.614    
## GroupB_Controls:TimeC_W1_FI_total        0.819    
## GroupC_Intervention:TimeC_W1_FI_total    0.152    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) GrpB_C GrpC_I TC_W1_ GB_C:T
## GrpB_Cntrls -0.824                            
## GrpC_Intrvn -0.820  0.676                     
## TmC_W1_FI_t -0.478  0.394  0.392              
## GB_C:TC_W1_  0.394 -0.478 -0.323 -0.824       
## GC_I:TC_W1_  0.392 -0.323 -0.478 -0.820  0.676
anova  (FI_MEM_B1W)
## Type III Analysis of Variance Table with Satterthwaite's method
##            Sum Sq Mean Sq NumDF DenDF F value  Pr(>F)  
## Group       0.335   0.168     2   256  0.0218 0.97847  
## Time       49.392  49.392     1   256  6.4159 0.01191 *
## Group:Time 23.763  11.882     2   256  1.5434 0.21564  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
sjPlot::tab_model(FI_MEM_B1W)
  FI Score
Predictors Estimates CI p
(Intercept) 9.86 8.72 – 11.00 <0.001
Group [B_Controls] 0.17 -1.22 – 1.55 0.811
Group [C_Intervention] 0.62 -0.77 – 2.01 0.384
Time [C_W1_FI_total] -0.28 -1.37 – 0.81 0.614
Group [B_Controls] × Time
[C_W1_FI_total] 		-0.15 -1.48 – 1.17 0.819
Group [C_Intervention] ×
Time [C_W1_FI_total] 		-0.97 -2.30 – 0.36 0.151
Random Effects
σ2 7.70
Ï„00 ID 9.15
ICC 0.54
N ID 259
Observations 518
Marginal R2 / Conditional R2 0.011 / 0.548 
parameters::standardise_parameters(FI_MEM_B1W)
## # Standardization method: refit
## 
## Parameter                             | Std. Coef. |        95% CI
## ------------------------------------------------------------------
## (Intercept)                           |       0.01 | [-0.27, 0.29]
## GroupB_Controls                       |       0.04 | [-0.30, 0.38]
## GroupC_Intervention                   |       0.15 | [-0.19, 0.49]
## TimeC_W1_FI_total                     |      -0.07 | [-0.33, 0.20]
## GroupB_Controls:TimeC_W1_FI_total     |      -0.04 | [-0.36, 0.28]
## GroupC_Intervention:TimeC_W1_FI_total |      -0.24 | [-0.56, 0.09]
plot_model(FI_MEM_B1W, type = "int")

Baseline to 1M

# Merging across timepoints
FI_B1M <- Full_data_all %>% 
  dplyr::select("ID", "Group", "A_PRE_FI_total", "D_M1_FI_total")
## Formatting table as needed
FI_B1M_long <- FI_B1M %>%
  pivot_longer(cols = c(A_PRE_FI_total, D_M1_FI_total),
               names_to = "Time",
               values_to = "FI_Score")
FI_MEM_B1M <- lmer(FI_Score ~ Group * Time + (1|ID), data = FI_B1M_long, REML = TRUE)
summary(FI_MEM_B1M)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: FI_Score ~ Group * Time + (1 | ID)
##    Data: FI_B1M_long
## 
## REML criterion at convergence: 2948
## 
## Scaled residuals: 
##      Min       1Q   Median       3Q      Max 
## -2.42367 -0.48652  0.03745  0.52434  2.27495 
## 
## Random effects:
##  Groups   Name        Variance Std.Dev.
##  ID       (Intercept)  9.36    3.059   
##  Residual             10.58    3.253   
## Number of obs: 518, groups:  ID, 259
## 
## Fixed effects:
##                                       Estimate Std. Error       df t value
## (Intercept)                             9.8600     0.6315 419.5497  15.613
## GroupB_Controls                         0.1683     0.7661 419.5497   0.220
## GroupC_Intervention                     0.6157     0.7697 419.5497   0.800
## TimeD_M1_FI_total                      -1.3000     0.6505 256.0000  -1.998
## GroupB_Controls:TimeD_M1_FI_total      -0.4170     0.7892 256.0000  -0.528
## GroupC_Intervention:TimeD_M1_FI_total  -0.8748     0.7929 256.0000  -1.103
##                                       Pr(>|t|)    
## (Intercept)                             <2e-16 ***
## GroupB_Controls                         0.8262    
## GroupC_Intervention                     0.4242    
## TimeD_M1_FI_total                       0.0467 *  
## GroupB_Controls:TimeD_M1_FI_total       0.5977    
## GroupC_Intervention:TimeD_M1_FI_total   0.2709    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) GrpB_C GrpC_I TD_M1_ GB_C:T
## GrpB_Cntrls -0.824                            
## GrpC_Intrvn -0.820  0.676                     
## TmD_M1_FI_t -0.515  0.425  0.423              
## GB_C:TD_M1_  0.425 -0.515 -0.348 -0.824       
## GC_I:TD_M1_  0.423 -0.348 -0.515 -0.820  0.676
anova  (FI_MEM_B1M)
## Type III Analysis of Variance Table with Satterthwaite's method
##            Sum Sq Mean Sq NumDF DenDF F value   Pr(>F)    
## Group        1.93    0.97     2   256  0.0914   0.9127    
## Time       344.31  344.31     1   256 32.5437 3.21e-08 ***
## Group:Time  13.80    6.90     2   256  0.6524   0.5217    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
sjPlot::tab_model(FI_MEM_B1M)
  FI Score
Predictors Estimates CI p
(Intercept) 9.86 8.62 – 11.10 <0.001
Group [B_Controls] 0.17 -1.34 – 1.67 0.826
Group [C_Intervention] 0.62 -0.90 – 2.13 0.424
Time [D_M1_FI_total] -1.30 -2.58 – -0.02 0.046
Group [B_Controls] × Time
[D_M1_FI_total] 		-0.42 -1.97 – 1.13 0.597
Group [C_Intervention] ×
Time [D_M1_FI_total] 		-0.87 -2.43 – 0.68 0.270
Random Effects
σ2 10.58
Ï„00 ID 9.36
ICC 0.47
N ID 259
Observations 518
Marginal R2 / Conditional R2 0.042 / 0.491 
parameters::standardise_parameters(FI_MEM_B1M)
## # Standardization method: refit
## 
## Parameter                             | Std. Coef. |         95% CI
## -------------------------------------------------------------------
## (Intercept)                           |       0.13 | [-0.14,  0.40]
## GroupB_Controls                       |       0.04 | [-0.29,  0.37]
## GroupC_Intervention                   |       0.14 | [-0.20,  0.47]
## TimeD_M1_FI_total                     |      -0.29 | [-0.57,  0.00]
## GroupB_Controls:TimeD_M1_FI_total     |      -0.09 | [-0.43,  0.25]
## GroupC_Intervention:TimeD_M1_FI_total |      -0.19 | [-0.54,  0.15]
plot_model(FI_MEM_B1M, type = "int")

Extra exploratory: IUS and GM association at PRE

PRE_IUS_GM_lm <- lm(A_PRE_IUS_total ~ A_PRE_GM, data = Full_data_all)
summary(PRE_IUS_GM_lm)
## 
## Call:
## lm(formula = A_PRE_IUS_total ~ A_PRE_GM, data = Full_data_all)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -32.939  -5.868   0.918   6.097  18.132 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  38.6537     1.2714  30.402  < 2e-16 ***
## A_PRE_GM      1.2142     0.3875   3.134  0.00193 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 8.713 on 257 degrees of freedom
## Multiple R-squared:  0.03681,    Adjusted R-squared:  0.03306 
## F-statistic:  9.82 on 1 and 257 DF,  p-value: 0.001926
anova(PRE_IUS_GM_lm) %>%
  report()
## The ANOVA suggests that:
## 
##   - The main effect of A_PRE_GM is statistically significant and small (F(1, 257)
## = 9.82, p = 0.002; Eta2 = 0.04, 95% CI [8.34e-03, 1.00])
## 
## Effect sizes were labelled following Field's (2013) recommendations.
Plot_lm_fi <- ggplot(data = Full_data_all, aes(x = A_PRE_IUS_total, y = A_PRE_GM)) +
  geom_point() +
  geom_smooth(method = "lm", color = "black")
print(Plot_lm_fi)
## `geom_smooth()` using formula = 'y ~ x'