## Warning: package 'labelled' was built under R version 4.3.3
## Warning: package 'gtsummary' was built under R version 4.3.3
##
## Attaching package: 'dplyr'
## The following objects are masked from 'package:stats':
##
## filter, lag
## The following objects are masked from 'package:base':
##
## intersect, setdiff, setequal, union
Import Dataset and isolate selected columns
library(agricolae)
temp1 <- read.csv('C:/Users/User/Desktop/Academics/Practicum and Thesis/Completed Forms/Practicum_dataset.csv')
temp2 = temp1[,c('subjectid', 'wk2_epds_score','wk6_epds_score','mat.age','gest_age_num','nasal_count','text_count','wk2_hh_ttl18to64','race','mat_educ_cat','income_cat','married_cat','ins_type_cat','wk6_blood','wks_active')]
dim(temp2)
## [1] 245 15
Remove rows with incomplete/missing data
all_data <- na.omit(temp2)
dim(all_data)
## [1] 224 15
Create categories for EPDS scores
EPDS Categories: None: 0-6, Mild: 7-13, Moderate: 14-19, Severe:
20-30
all_data$wk2_epds_cat = cut(all_data$wk2_epds_score, breaks=c(-1,6,30), labels= c("Depressed", "Not Depressed"))
hist(all_data$wk2_epds_score)
table(all_data$wk2_epds_cat)
##
## Depressed Not Depressed
## 179 45
Setting compliance (PREVAIL definition)
0=No, 1=Yes
## Set Weeks Active level
all_data$weekly_comp <- ifelse(all_data$wks_active >= 73, 1, 0)
## Set text compliance level
tcomp_scrap <- 100*(all_data$text_count/102)
all_data$tcomp_cat <- ifelse(tcomp_scrap >= 70, 1, 0)
## Create is.comp.2 column
all_data$is.comp.2 <- ifelse(all_data$weekly_comp == 1, ifelse(all_data$tcomp_cat == 1, 1, ifelse(all_data$wk6_blood == 1, 1, 0)), 0)
table(all_data$is.comp.2)
##
## 0 1
## 30 194
Setting categories for factors
## Race
## Binomial
all_data$race_cat = ifelse(all_data$race=="White", 0, 1)
## Category
all_data$race_cat2 = ifelse(all_data$race=="White", "White", "POC")
## Income (Category)
all_data$income_cat_bin <- ifelse(all_data$income_cat == "<25k", "Less than 25k", "25k or more")
GLM with all factors
model <- glm(is.comp.2 ~ wk2_epds_cat + race_cat2 + mat.age + gest_age_num + mat_educ_cat + income_cat_bin + married_cat + wk2_hh_ttl18to64 + ins_type_cat, data = all_data, family = "binomial")
summary(model)
##
## Call:
## glm(formula = is.comp.2 ~ wk2_epds_cat + race_cat2 + mat.age +
## gest_age_num + mat_educ_cat + income_cat_bin + married_cat +
## wk2_hh_ttl18to64 + ins_type_cat, family = "binomial", data = all_data)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 7.60294 7.61630 0.998 0.3182
## wk2_epds_catNot Depressed 1.35042 0.79238 1.704 0.0883 .
## race_cat2White 0.08472 0.61345 0.138 0.8902
## mat.age 0.08059 0.05193 1.552 0.1207
## gest_age_num -0.20234 0.18966 -1.067 0.2860
## mat_educ_cat>HS 0.61223 0.63528 0.964 0.3352
## income_cat_binLess than 25k 0.29262 0.51728 0.566 0.5716
## married_catSingle 0.07922 0.48196 0.164 0.8694
## wk2_hh_ttl18to64 -0.20777 0.22005 -0.944 0.3451
## ins_type_catPublic -0.56882 0.76088 -0.748 0.4547
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 176.42 on 223 degrees of freedom
## Residual deviance: 155.90 on 214 degrees of freedom
## AIC: 175.9
##
## Number of Fisher Scoring iterations: 6
The following are the Fisher’s Exact for categorical factors:
Fisher’s Exact Test - Definition 2 (PREVAIL)
## EPDS categories
Fdata2 <- data.frame("comp_yes" = c(151, 43),
"comp_no" = c(28, 2),
row.names = c("Depressed", "Not Depressed"),
stringsAsFactors = FALSE
)
colnames(Fdata2) <- c("Compliant", "Noncompliant")
print(Fdata2)
## Compliant Noncompliant
## Depressed 151 28
## Not Depressed 43 2
fisher.test(Fdata2)
##
## Fisher's Exact Test for Count Data
##
## data: Fdata2
## p-value = 0.05154
## alternative hypothesis: true odds ratio is not equal to 1
## 95 percent confidence interval:
## 0.02798151 1.06836348
## sample estimates:
## odds ratio
## 0.2519048
## Race
X2 <- all_data[,c("race_cat", "is.comp.2")]
table(X2)
## is.comp.2
## race_cat 0 1
## 0 11 105
## 1 19 89
fisher.test(X2)
##
## Fisher's Exact Test for Count Data
##
## data: X2
## p-value = 1
## alternative hypothesis: two.sided
## Maternal Education
all_data$edu_temp <- ifelse(all_data$mat_educ_cat == "<=HS", 0, 1)
X3 <- all_data[,c("edu_temp", "is.comp.2")]
table(X3)
## is.comp.2
## edu_temp 0 1
## 0 21 80
## 1 9 114
fisher.test(X3)
##
## Fisher's Exact Test for Count Data
##
## data: X3
## p-value = 1
## alternative hypothesis: two.sided
## Income
all_data$inc_temp <- ifelse(all_data$income_cat == "<25k", 0, 1)
all_data$income_cat_bin <- ifelse(all_data$income_cat == "<25k", "Less than 25k", "25k or more")
X4 <- all_data[,c("inc_temp", "is.comp.2")]
table(X4)
## is.comp.2
## inc_temp 0 1
## 0 12 59
## 1 18 135
fisher.test(X4)
##
## Fisher's Exact Test for Count Data
##
## data: X4
## p-value = 1
## alternative hypothesis: two.sided
## Marriage Status
all_data$mar_temp <- ifelse(all_data$married_cat == "Single", 0, 1)
X5 <- all_data[,c("mar_temp", "is.comp.2")]
table(X5)
## is.comp.2
## mar_temp 0 1
## 0 13 58
## 1 17 136
fisher.test(X5)
##
## Fisher's Exact Test for Count Data
##
## data: X5
## p-value = 1
## alternative hypothesis: two.sided
## Insurance Type
all_data$ins_temp <- ifelse(all_data$ins_type_cat == "Private", 0, 1)
X6 <- all_data[,c("ins_temp", "is.comp.2")]
table(X6)
## is.comp.2
## ins_temp 0 1
## 0 7 91
## 1 23 103
fisher.test(X6)
##
## Fisher's Exact Test for Count Data
##
## data: X6
## p-value = 1
## alternative hypothesis: two.sided
Tables and Graphs
Print Table
print(table1)
Table 1: Demographics vs. Compliance
|
Variable
|
Overall, N = 224
|
Compliant, N = 194
|
Not Compliant, N = 30
|
p-value
|
|
Race
|
|
|
|
0.11
|
|
White
|
116 (52%)
|
105 (54%)
|
11 (37%)
|
|
|
Black
|
97 (43%)
|
81 (42%)
|
16 (53%)
|
|
|
Other
|
11 (4.9%)
|
8 (4.1%)
|
3 (10%)
|
|
|
Maternal age
|
|
|
|
0.002
|
|
18-24
|
36 (16%)
|
24 (12%)
|
12 (40%)
|
|
|
25-34
|
139 (62%)
|
125 (64%)
|
14 (47%)
|
|
|
>34
|
49 (22%)
|
45 (23%)
|
4 (13%)
|
|
|
Gestational age
|
|
|
|
0.9
|
|
37 weeks or less
|
9 (4.0%)
|
8 (4.1%)
|
1 (3.3%)
|
|
|
More than 37 weeks
|
215 (96%)
|
186 (96%)
|
29 (97%)
|
|
|
Maternal Education
|
|
|
|
0.003
|
|
High School/GED or lower
|
101 (45%)
|
80 (41%)
|
21 (70%)
|
|
|
Higher than High School/GED
|
123 (55%)
|
114 (59%)
|
9 (30%)
|
|
|
Family Income
|
|
|
|
0.069
|
|
Less than $25k
|
71 (32%)
|
59 (30%)
|
12 (40%)
|
|
|
$25k-$50k
|
44 (20%)
|
35 (18%)
|
9 (30%)
|
|
|
More than $50k
|
109 (49%)
|
100 (52%)
|
9 (30%)
|
|
|
Unknown
|
0 (0%)
|
0 (0%)
|
0 (0%)
|
|
|
Marital status
|
|
|
|
0.14
|
|
Married/Parter
|
153 (68%)
|
136 (70%)
|
17 (57%)
|
|
|
Single
|
71 (32%)
|
58 (30%)
|
13 (43%)
|
|
|
People in household
|
|
|
|
0.9
|
|
One person
|
51 (23%)
|
44 (23%)
|
7 (23%)
|
|
|
Two or more people
|
173 (77%)
|
150 (77%)
|
23 (77%)
|
|
|
Insurance type
|
|
|
|
0.015
|
|
Private
|
98 (44%)
|
91 (47%)
|
7 (23%)
|
|
|
Public
|
126 (56%)
|
103 (53%)
|
23 (77%)
|
|
Print Table
print(table2)
Table 2: EPDS Category vs. Compliance
|
EPDS Category
|
N = 224
|
|
Depressed
|
179 (80%)
|
|
Not Depressed
|
45 (20%)
|
Print Table
print(table3)
Table 2: EPDS Category vs. Compliance
|
|
Overall, N = 224
|
Compliant, N = 194
|
Not Compliant, N = 30
|
|
EPDS Category
|
|
|
|
|
Depressed
|
179 (80%)
|
151 (78%)
|
28 (93%)
|
|
Not Depressed
|
45 (20%)
|
43 (22%)
|
2 (6.7%)
|
#Table 4