PUCC_CSQ_Analysis

## [1] 1305   11

##  [1] "Internal ID"              "Encounter Date"          
##  [3] "Age"                      "Sex"                     
##  [5] "Race"                     "Ethnicity"               
##  [7] "Coverage"                 "GAD-7"                   
##  [9] "Last Question 9 Response" "PHQ-9"                   
## [11] "Patient Disposition"

## [1] 792  22

##  [1] "ZID"                                   
##  [2] "Visit_Date"                            
##  [3] "Encounter_Date"                        
##  [4] "Visit_Type"                            
##  [5] "new_CSQ1"                              
##  [6] "new_CSQ2"                              
##  [7] "new_CSQ3"                              
##  [8] "new_CSQ4"                              
##  [9] "new_CSQ5"                              
## [10] "new_CSQ6"                              
## [11] "new_CSQ7"                              
## [12] "new_CSQ8"                              
## [13] "new_CSQ9"                              
## [14] "new_CSQ10"                             
## [15] "new_CSQ11"                             
## [16] "new_CSQ12"                             
## [17] "new_CSQ13"                             
## [18] "new_CSQ14"                             
## [19] "new_CSQ_Healthrelatedstressors"        
## [20] "new_CSQ_nonpsychhealthrelatedstressors"
## [21] "new_CSQ15"                             
## [22] "new_CSQ15 reorganized"

## [1] 792  23

#Factors vs. numeric variables

PUCC_final[,c("Sex", "Race", "Ethnicity")] <- lapply(PUCC_final[,c("Sex", "Race", "Ethnicity")], as.factor)
PUCC_final[,c("Age", "PHQ9_tot", "GAD7_tot")] <- lapply(PUCC_final[,c("Age", "PHQ9_tot", "GAD7_tot")], as.numeric)

## Warning in lapply(PUCC_final[, c("Age", "PHQ9_tot", "GAD7_tot")], as.numeric):
## NAs introduced by coercion

#PHQ9 and GAD7 by cut scores

PUCC_final$PHQ9_cut <- cut(PUCC_final$PHQ9_tot, breaks = c(0, 5, 10, 15, 20, 27))
levels(PUCC_final$PHQ9_cut) <- c("Minimal depression", "Mild depression", "Moderate depression", "Moderately severe depression", "Severe depression")

PUCC_final$GAD7_cut <- cut(PUCC_final$GAD7_tot, breaks = c(0, 5, 10, 15, 21))
levels(PUCC_final$GAD7_cut) <- c("Minimal anxiety", "Mild anxiety", "Moderate anxiety", "Severe anxiety")

#PHQ9 and GAD7 by dichotomous (Severe vs. not severe)

PUCC_final$PHQ9_severe <- cut(PUCC_final$PHQ9_tot, breaks = c(0, 15,27))
levels(PUCC_final$PHQ9_severe) <- c("Minimal to moderate depression", "Severe depression")

PUCC_final$GAD7_severe<- cut(PUCC_final$GAD7_tot, breaks = c(0, 15,21))
levels(PUCC_final$GAD7_severe) <- c("Minimal to moderate anxiety", "Severe anxiety")


#PHQ9 and GAD7 by dichotomous (Moderate to severe vs. mild)

PUCC_final$PHQ9_moderate <- cut(PUCC_final$PHQ9_tot, breaks = c(0, 10,27))
levels(PUCC_final$PHQ9_moderate) <- c("Mild depression", "Moderate to severe depression")

PUCC_final$GAD7_moderate<- cut(PUCC_final$GAD7_tot, breaks = c(0, 10,21))
levels(PUCC_final$GAD7_moderate) <- c("Mild anxiety", "Moderate to severe anxiety")

#Sex variable - assigning reference group as female

PUCC_final$Sex<-relevel(PUCC_final$Sex, "Female")

#Age bracket variable - assigning ages to 5 brackets

PUCC_final$Age_cut <- cut(PUCC_final$Age, breaks = c(18, 25, 49 , 65, 80, 95))
levels(PUCC_final$Age_cut) <- c("18-25 years", "26-49 years", "50-65 years", "66-80 years", "81-95 years")
table(PUCC_final$Age_cut  )

## 
## 18-25 years 26-49 years 50-65 years 66-80 years 81-95 years 
##         211         401         101          44          11

#Race and ethnicity variable - assigning dichotomous levels and reference groups

PUCC_final$Ethnicity<-ifelse(PUCC_final$Ethnicity=="Not Hispanic, Latino/a, or Spanish origin [1]", "Not Hispanic", ifelse(PUCC_final$Ethnicity=="Unknown [3]" | PUCC_final$Ethnicity=="Decline to Answer [4] ", "Other", "Hispanic"))
PUCC_final$Ethnicity<-as.factor(PUCC_final$Ethnicity)
PUCC_final$Ethnicity<-relevel(PUCC_final$Ethnicity, "Not Hispanic")
table(PUCC_final$Ethnicity  )

## 
## Not Hispanic     Hispanic        Other 
##          571           50          145

PUCC_final$Race<-ifelse(PUCC_final$Race=="White", "White", ifelse(PUCC_final$Race=="Black or African American", "Black", "Other"))
PUCC_final$Race<-as.factor(PUCC_final$Race)
table(PUCC_final$Race)

## 
## Black Other White 
##    47   152   570

PUCC_final$Race<-relevel(PUCC_final$Race, "White")
table(PUCC_final$Race  )

## 
## White Black Other 
##   570    47   152

#CSQ_tot variable

CSQ_data <- c("new_CSQ1", "new_CSQ2", "new_CSQ3", "new_CSQ4", "new_CSQ5", "new_CSQ6",
              "new_CSQ7","new_CSQ8","new_CSQ9", "new_CSQ10", "new_CSQ11", "new_CSQ12",
              "new_CSQ13","new_CSQ_Healthrelatedstressors","new_CSQ_nonpsychhealthrelatedstressors")

PUCC_final[,CSQ_data] <- lapply(PUCC_final[,CSQ_data], as.numeric)
PUCC_final$CSQ_tot<-rowSums(PUCC_final[ ,CSQ_data]) 
PUCC_final[,CSQ_data] <- lapply(PUCC_final[,CSQ_data], as.factor)

#reorder vars in df

PUCC_FIN<-PUCC_final %>% dplyr::select("ZID", "Encounter_Date", "Sex", "Age", "Race", "Ethnicity", "new_CSQ1", "new_CSQ2", "new_CSQ3", "new_CSQ4", "new_CSQ5", "new_CSQ6","new_CSQ7","new_CSQ8","new_CSQ9", "new_CSQ10", "new_CSQ11", "new_CSQ12",        "new_CSQ13","new_CSQ_Healthrelatedstressors","new_CSQ_nonpsychhealthrelatedstressors", "CSQ_tot", "PHQ9_tot", "PHQ9_cut", "PHQ9_severe", "PHQ9_moderate", "GAD7_tot", "GAD7_cut", "GAD7_severe", "GAD7_moderate")

# % Missingness across all variables

purrr::map(PUCC_FIN, ~mean(is.na(.)))

## $ZID
## [1] 0
## 
## $Encounter_Date
## [1] 0
## 
## $Sex
## [1] 0.0290404
## 
## $Age
## [1] 0.03030303
## 
## $Race
## [1] 0.0290404
## 
## $Ethnicity
## [1] 0.03282828
## 
## $new_CSQ1
## [1] 0
## 
## $new_CSQ2
## [1] 0
## 
## $new_CSQ3
## [1] 0
## 
## $new_CSQ4
## [1] 0
## 
## $new_CSQ5
## [1] 0
## 
## $new_CSQ6
## [1] 0
## 
## $new_CSQ7
## [1] 0
## 
## $new_CSQ8
## [1] 0
## 
## $new_CSQ9
## [1] 0
## 
## $new_CSQ10
## [1] 0
## 
## $new_CSQ11
## [1] 0
## 
## $new_CSQ12
## [1] 0
## 
## $new_CSQ13
## [1] 0
## 
## $new_CSQ_Healthrelatedstressors
## [1] 0
## 
## $new_CSQ_nonpsychhealthrelatedstressors
## [1] 0
## 
## $CSQ_tot
## [1] 0
## 
## $PHQ9_tot
## [1] 0.1401515
## 
## $PHQ9_cut
## [1] 0.1502525
## 
## $PHQ9_severe
## [1] 0.1502525
## 
## $PHQ9_moderate
## [1] 0.1502525
## 
## $GAD7_tot
## [1] 0.1363636
## 
## $GAD7_cut
## [1] 0.1464646
## 
## $GAD7_severe
## [1] 0.1464646
## 
## $GAD7_moderate
## [1] 0.1464646

COMMENTARY: here we see the percent missingness per variable

PUCC_FIN %>% dplyr::select(-ZID, -Encounter_Date) %>% gtsummary::tbl_summary()

Characteristic	N = 792¹
Sex
Female	437 (57%)
Male	332 (43%)
Unknown	23
Age	32 (25, 46)
Unknown	24
Race
White	570 (74%)
Black	47 (6.1%)
Other	152 (20%)
Unknown	23
Ethnicity
Not Hispanic	571 (75%)
Hispanic	50 (6.5%)
Other	145 (19%)
Unknown	26
new_CSQ1
0	522 (66%)
1	270 (34%)
new_CSQ2
0	734 (93%)
1	58 (7.3%)
new_CSQ3
0	528 (67%)
1	264 (33%)
new_CSQ4
0	481 (61%)
1	311 (39%)
new_CSQ5
0	705 (89%)
1	87 (11%)
new_CSQ6
0	631 (80%)
1	161 (20%)
new_CSQ7
0	622 (79%)
1	170 (21%)
new_CSQ8
0	729 (92%)
1	63 (8.0%)
new_CSQ9
0	590 (74%)
1	202 (26%)
new_CSQ10
0	649 (82%)
1	143 (18%)
new_CSQ11
0	737 (93%)
1	55 (6.9%)
new_CSQ12
0	595 (75%)
1	197 (25%)
new_CSQ13
0	473 (60%)
1	319 (40%)
new_CSQ_Healthrelatedstressors
0	720 (91%)
1	72 (9.1%)
new_CSQ_nonpsychhealthrelatedstressors
0	767 (97%)
1	25 (3.2%)
CSQ_tot	3.00 (1.00, 4.00)
PHQ9_tot	18 (12, 22)
Unknown	111
PHQ9_cut
Minimal depression	39 (5.8%)
Mild depression	76 (11%)
Moderate depression	141 (21%)
Moderately severe depression	196 (29%)
Severe depression	221 (33%)
Unknown	119
PHQ9_severe
Minimal to moderate depression	256 (38%)
Severe depression	417 (62%)
Unknown	119
PHQ9_moderate
Mild depression	115 (17%)
Moderate to severe depression	558 (83%)
Unknown	119
GAD7_tot	17.0 (12.0, 20.0)
Unknown	108
GAD7_cut
Minimal anxiety	39 (5.8%)
Mild anxiety	75 (11%)
Moderate anxiety	149 (22%)
Severe anxiety	413 (61%)
Unknown	116
GAD7_severe
Minimal to moderate anxiety	263 (39%)
Severe anxiety	413 (61%)
Unknown	116
GAD7_moderate
Mild anxiety	114 (17%)
Moderate to severe anxiety	562 (83%)
Unknown	116
¹ n (%); Median (IQR)

#Distribution of PHQ9_tot variable
hist(PUCC_FIN$PHQ9_tot) #L skew

  shapiro.test(PUCC_FIN$PHQ9_tot) #p<0.001 - not normally distributed

## 
##  Shapiro-Wilk normality test
## 
## data:  PUCC_FIN$PHQ9_tot
## W = 0.96355, p-value = 5.683e-12

#Distribution of PHQ9_tot residuals
mod_PHQ9<-lm(PHQ9_tot~Age, data=PUCC_FIN)
plot(mod_PHQ9)

COMMENTARY: the Q-Q plot for PHQ9 looks a lot better than GAD7. That’s okay since I dont think we’ll be using GAD7 as an outcome for this analysis.

#Covariate screen by PHQ_tot (continuous)

PUCC_FIN %>%
  dplyr::select(-ZID, -Encounter_Date, -PHQ9_cut, -PHQ9_severe) %>%
  gtsummary::tbl_uvregression(method = lm, y= PHQ9_tot) %>%
  add_global_p() %>%
  bold_p()

## Warning in checkMatrixPackageVersion(): Package version inconsistency detected.
## TMB was built with Matrix version 1.3.2
## Current Matrix version is 1.2.18
## Please re-install 'TMB' from source using install.packages('TMB', type = 'source') or ask CRAN for a binary version of 'TMB' matching CRAN's 'Matrix' package

## add_global_p: Global p-values for variable(s) `add_global_p(include = c("Sex",
## "Age", "Race", "Ethnicity", "new_CSQ1", "new_CSQ2", "new_CSQ3", "new_CSQ4",
## "new_CSQ5", "new_CSQ6", "new_CSQ7", "new_CSQ8", "new_CSQ9", "new_CSQ10",
## "new_CSQ11", "new_CSQ12", "new_CSQ13", "new_CSQ_Healthrelatedstressors",
## "new_CSQ_nonpsychhealthrelatedstressors", "CSQ_tot", "PHQ9_moderate",
## "GAD7_tot", "GAD7_cut", "GAD7_severe", "GAD7_moderate"))` were calculated with
##   `car::Anova(mod = x$model_obj, type = "III")`

Characteristic	N	Beta	95% CI¹	p-value
Sex	681			0.2
Female		—	—
Male		-0.68	-1.7, 0.32
Age	680	-0.01	-0.05, 0.02	0.4
Race	681			0.8
White		—	—
Black		-0.73	-2.8, 1.4
Other		0.02	-1.2, 1.3
Ethnicity	678			0.4
Not Hispanic		—	—
Hispanic		1.0	-1.1, 3.1
Other		0.64	-0.64, 1.9
new_CSQ1	681			0.050
0		—	—
1		1.0	0.00, 2.1
new_CSQ2	681			0.3
0		—	—
1		1.1	-0.88, 3.0
new_CSQ3	681			<0.001
0		—	—
1		2.0	1.0, 3.0
new_CSQ4	681			0.002
0		—	—
1		1.6	0.59, 2.6
new_CSQ5	681			0.4
0		—	—
1		0.73	-0.87, 2.3
new_CSQ6	681			0.004
0		—	—
1		1.8	0.57, 3.0
new_CSQ7	681			0.024
0		—	—
1		1.4	0.18, 2.6
new_CSQ8	681			0.012
0		—	—
1		2.3	0.51, 4.2
new_CSQ9	681			0.011
0		—	—
1		1.5	0.34, 2.6
new_CSQ10	681			0.042
0		—	—
1		1.4	0.05, 2.7
new_CSQ11	681			0.4
0		—	—
1		0.90	-1.1, 2.9
new_CSQ12	681			0.8
0		—	—
1		-0.16	-1.3, 1.0
new_CSQ13	681			<0.001
0		—	—
1		2.1	1.1, 3.0
new_CSQ_Healthrelatedstressors	681			0.7
0		—	—
1		-0.32	-2.0, 1.4
new_CSQ_nonpsychhealthrelatedstressors	681			>0.9
0		—	—
1		0.06	-2.7, 2.9
CSQ_tot	681	0.66	0.43, 0.89	<0.001
PHQ9_moderate	673			<0.001
Mild depression		—	—
Moderate to severe depression		12	12, 13
GAD7_tot	680	0.75	0.67, 0.82	<0.001
GAD7_cut	672			<0.001
Minimal anxiety		—	—
Mild anxiety		3.0	1.0, 5.1
Moderate anxiety		6.2	4.4, 8.1
Severe anxiety		11	9.6, 13
GAD7_severe	672			<0.001
Minimal to moderate anxiety		—	—
Severe anxiety		6.9	6.0, 7.7
GAD7_moderate	672			<0.001
Mild anxiety		—	—
Moderate to severe anxiety		8.0	6.8, 9.1
¹ CI = Confidence Interval

COMMENTARYL: Significant covariates by PHQ9_tot: new_CSQ3 new_CSQ4 new_CSQ6 new_CSQ7 new_CSQ8 new_CSQ9 new_CSQ10 new_CSQ13 CSQ_tot GAD_tot, GAD_cut, GAD_severe

PUCC_FIN %>% dplyr::select(-ZID, -Encounter_Date) %>% 
  gtsummary::tbl_summary(by=PHQ9_severe, statistic = list(all_continuous() ~ "{mean} ({sd})", 
                                                          all_categorical() ~ "{n} / {N} ({p}%)"),
                                                          digits = all_continuous() ~ 2,
                                                          missing_text = "(Missing)") %>%
                                                         add_p()

## 119 observations missing `PHQ9_severe` have been removed. To include these observations, use `forcats::fct_explicit_na()` on `PHQ9_severe` column before passing to `tbl_summary()`.

Characteristic	Minimal to moderate depression, N = 256¹	Severe depression, N = 417¹	p-value²
Sex			0.048
Female	138 / 256 (54%)	257 / 417 (62%)
Male	118 / 256 (46%)	160 / 417 (38%)
Age	36.88 (15.42)	35.88 (14.66)	0.6
(Missing)	1	0
Race			0.4
White	191 / 256 (75%)	310 / 417 (74%)
Black	19 / 256 (7.4%)	21 / 417 (5.0%)
Other	46 / 256 (18%)	86 / 417 (21%)
Ethnicity			0.3
Not Hispanic	199 / 254 (78%)	305 / 416 (73%)
Hispanic	13 / 254 (5.1%)	28 / 416 (6.7%)
Other	42 / 254 (17%)	83 / 416 (20%)
(Missing)	2	1
new_CSQ1			0.3
0	174 / 256 (68%)	267 / 417 (64%)
1	82 / 256 (32%)	150 / 417 (36%)
new_CSQ2			0.4
0	241 / 256 (94%)	385 / 417 (92%)
1	15 / 256 (5.9%)	32 / 417 (7.7%)
new_CSQ3			<0.001
0	190 / 256 (74%)	254 / 417 (61%)
1	66 / 256 (26%)	163 / 417 (39%)
new_CSQ4			0.090
0	163 / 256 (64%)	238 / 417 (57%)
1	93 / 256 (36%)	179 / 417 (43%)
new_CSQ5			0.2
0	234 / 256 (91%)	368 / 417 (88%)
1	22 / 256 (8.6%)	49 / 417 (12%)
new_CSQ6			0.014
0	216 / 256 (84%)	319 / 417 (76%)
1	40 / 256 (16%)	98 / 417 (24%)
new_CSQ7			0.14
0	211 / 256 (82%)	324 / 417 (78%)
1	45 / 256 (18%)	93 / 417 (22%)
new_CSQ8			0.003
0	246 / 256 (96%)	374 / 417 (90%)
1	10 / 256 (3.9%)	43 / 417 (10%)
new_CSQ9			0.13
0	199 / 256 (78%)	302 / 417 (72%)
1	57 / 256 (22%)	115 / 417 (28%)
new_CSQ10			0.4
0	217 / 256 (85%)	344 / 417 (82%)
1	39 / 256 (15%)	73 / 417 (18%)
new_CSQ11			0.5
0	241 / 256 (94%)	387 / 417 (93%)
1	15 / 256 (5.9%)	30 / 417 (7.2%)
new_CSQ12			0.9
0	191 / 256 (75%)	313 / 417 (75%)
1	65 / 256 (25%)	104 / 417 (25%)
new_CSQ13			<0.001
0	172 / 256 (67%)	226 / 417 (54%)
1	84 / 256 (33%)	191 / 417 (46%)
new_CSQ_Healthrelatedstressors			0.7
0	230 / 256 (90%)	379 / 417 (91%)
1	26 / 256 (10%)	38 / 417 (9.1%)
new_CSQ_nonpsychhealthrelatedstressors			0.8
0	247 / 256 (96%)	404 / 417 (97%)
1	9 / 256 (3.5%)	13 / 417 (3.1%)
CSQ_tot	2.61 (1.96)	3.29 (2.10)	<0.001
PHQ9_tot	10.21 (3.87)	21.02 (3.28)	<0.001
PHQ9_cut			<0.001
Minimal depression	39 / 256 (15%)	0 / 417 (0%)
Mild depression	76 / 256 (30%)	0 / 417 (0%)
Moderate depression	141 / 256 (55%)	0 / 417 (0%)
Moderately severe depression	0 / 256 (0%)	196 / 417 (47%)
Severe depression	0 / 256 (0%)	221 / 417 (53%)
PHQ9_moderate			<0.001
Mild depression	115 / 256 (45%)	0 / 417 (0%)
Moderate to severe depression	141 / 256 (55%)	417 / 417 (100%)
GAD7_tot	12.42 (5.47)	17.67 (3.82)	<0.001
(Missing)	1	0
GAD7_cut			<0.001
Minimal anxiety	33 / 254 (13%)	4 / 415 (1.0%)
Mild anxiety	58 / 254 (23%)	14 / 415 (3.4%)
Moderate anxiety	76 / 254 (30%)	73 / 415 (18%)
Severe anxiety	87 / 254 (34%)	324 / 415 (78%)
(Missing)	2	2
GAD7_severe			<0.001
Minimal to moderate anxiety	167 / 254 (66%)	91 / 415 (22%)
Severe anxiety	87 / 254 (34%)	324 / 415 (78%)
(Missing)	2	2
GAD7_moderate			<0.001
Mild anxiety	91 / 254 (36%)	18 / 415 (4.3%)
Moderate to severe anxiety	163 / 254 (64%)	397 / 415 (96%)
(Missing)	2	2
¹ n / N (%); Mean (SD) ² Pearson's Chi-squared test; Wilcoxon rank sum test

Significant covariates according to PHQ9_severe (dichotomous): Sex new_CSQ3 new_CSQ6 new_CSQ8 new_CSQ13 CSQ_tot GAD7 variables

#Group comparison by PHQ9_cut (severe or not severe depression)

PUCC_FIN %>% dplyr::select(-ZID, -Encounter_Date) %>% 
  gtsummary::tbl_summary(by=PHQ9_moderate, statistic = list(all_continuous() ~ "{mean} ({sd})", 
                                                          all_categorical() ~ "{n} / {N} ({p}%)"),
                                                          digits = all_continuous() ~ 2,
                                                          missing_text = "(Missing)") %>%
                                                         add_p()

## 119 observations missing `PHQ9_moderate` have been removed. To include these observations, use `forcats::fct_explicit_na()` on `PHQ9_moderate` column before passing to `tbl_summary()`.

Characteristic	Mild depression, N = 115¹	Moderate to severe depression, N = 558¹	p-value²
Sex			0.6
Female	65 / 115 (57%)	330 / 558 (59%)
Male	50 / 115 (43%)	228 / 558 (41%)
Age	36.89 (15.98)	36.13 (14.75)	>0.9
(Missing)	1	0
Race			0.3
White	90 / 115 (78%)	411 / 558 (74%)
Black	8 / 115 (7.0%)	32 / 558 (5.7%)
Other	17 / 115 (15%)	115 / 558 (21%)
Ethnicity			0.3
Not Hispanic	92 / 114 (81%)	412 / 556 (74%)
Hispanic	6 / 114 (5.3%)	35 / 556 (6.3%)
Other	16 / 114 (14%)	109 / 556 (20%)
(Missing)	1	2
new_CSQ1			0.063
0	84 / 115 (73%)	357 / 558 (64%)
1	31 / 115 (27%)	201 / 558 (36%)
new_CSQ2			0.7
0	106 / 115 (92%)	520 / 558 (93%)
1	9 / 115 (7.8%)	38 / 558 (6.8%)
new_CSQ3			0.029
0	86 / 115 (75%)	358 / 558 (64%)
1	29 / 115 (25%)	200 / 558 (36%)
new_CSQ4			0.005
0	82 / 115 (71%)	319 / 558 (57%)
1	33 / 115 (29%)	239 / 558 (43%)
new_CSQ5			>0.9
0	103 / 115 (90%)	499 / 558 (89%)
1	12 / 115 (10%)	59 / 558 (11%)
new_CSQ6			0.015
0	101 / 115 (88%)	434 / 558 (78%)
1	14 / 115 (12%)	124 / 558 (22%)
new_CSQ7			0.001
0	104 / 115 (90%)	431 / 558 (77%)
1	11 / 115 (9.6%)	127 / 558 (23%)
new_CSQ8			0.12
0	110 / 115 (96%)	510 / 558 (91%)
1	5 / 115 (4.3%)	48 / 558 (8.6%)
new_CSQ9			0.083
0	93 / 115 (81%)	408 / 558 (73%)
1	22 / 115 (19%)	150 / 558 (27%)
new_CSQ10			0.3
0	100 / 115 (87%)	461 / 558 (83%)
1	15 / 115 (13%)	97 / 558 (17%)
new_CSQ11			0.6
0	106 / 115 (92%)	522 / 558 (94%)
1	9 / 115 (7.8%)	36 / 558 (6.5%)
new_CSQ12			0.8
0	87 / 115 (76%)	417 / 558 (75%)
1	28 / 115 (24%)	141 / 558 (25%)
new_CSQ13			0.012
0	80 / 115 (70%)	318 / 558 (57%)
1	35 / 115 (30%)	240 / 558 (43%)
new_CSQ_Healthrelatedstressors			0.7
0	103 / 115 (90%)	506 / 558 (91%)
1	12 / 115 (10%)	52 / 558 (9.3%)
new_CSQ_nonpsychhealthrelatedstressors			0.6
0	110 / 115 (96%)	541 / 558 (97%)
1	5 / 115 (4.3%)	17 / 558 (3.0%)
CSQ_tot	2.35 (2.06)	3.17 (2.05)	<0.001
PHQ9_tot	6.59 (2.68)	19.04 (4.50)	<0.001
PHQ9_cut			<0.001
Minimal depression	39 / 115 (34%)	0 / 558 (0%)
Mild depression	76 / 115 (66%)	0 / 558 (0%)
Moderate depression	0 / 115 (0%)	141 / 558 (25%)
Moderately severe depression	0 / 115 (0%)	196 / 558 (35%)
Severe depression	0 / 115 (0%)	221 / 558 (40%)
PHQ9_severe			<0.001
Minimal to moderate depression	115 / 115 (100%)	141 / 558 (25%)
Severe depression	0 / 115 (0%)	417 / 558 (75%)
GAD7_tot	10.09 (5.18)	16.82 (4.39)	<0.001
(Missing)	1	0
GAD7_cut			<0.001
Minimal anxiety	24 / 113 (21%)	13 / 556 (2.3%)
Mild anxiety	34 / 113 (30%)	38 / 556 (6.8%)
Moderate anxiety	36 / 113 (32%)	113 / 556 (20%)
Severe anxiety	19 / 113 (17%)	392 / 556 (71%)
(Missing)	2	2
GAD7_severe			<0.001
Minimal to moderate anxiety	94 / 113 (83%)	164 / 556 (29%)
Severe anxiety	19 / 113 (17%)	392 / 556 (71%)
(Missing)	2	2
GAD7_moderate			<0.001
Mild anxiety	58 / 113 (51%)	51 / 556 (9.2%)
Moderate to severe anxiety	55 / 113 (49%)	505 / 556 (91%)
(Missing)	2	2
¹ n / N (%); Mean (SD) ² Pearson's Chi-squared test; Wilcoxon rank sum test; Fisher's exact test

COMMENTARY: Positive screens for variables: CSQ 3, 4, 6, 7, 13, CSQ_tot, GAD7 variables

mod_CSQ_tot<-lm(PHQ9_tot~Sex+Age+Race+GAD7_tot+CSQ_tot, data=PUCC_FIN)
summary(mod_CSQ_tot)

## 
## Call:
## lm(formula = PHQ9_tot ~ Sex + Age + Race + GAD7_tot + CSQ_tot, 
##     data = PUCC_FIN)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -17.5069  -3.3519   0.3315   3.5237  20.6467 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  4.77639    0.84410   5.659 2.26e-08 ***
## SexMale     -0.62135    0.40071  -1.551   0.1215    
## Age          0.01114    0.01330   0.838   0.4024    
## RaceBlack   -0.64300    0.83479  -0.770   0.4414    
## RaceOther   -0.34213    0.49927  -0.685   0.4934    
## GAD7_tot     0.72781    0.03776  19.277  < 2e-16 ***
## CSQ_tot      0.21082    0.09798   2.152   0.0318 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 5.13 on 672 degrees of freedom
##   (113 observations deleted due to missingness)
## Multiple R-squared:  0.3862, Adjusted R-squared:  0.3808 
## F-statistic: 70.48 on 6 and 672 DF,  p-value: < 2.2e-16

#Cohen's d effect size function

reg_cohend <- function (glm_output) {
  beta <- summary(glm_output)$coefficients[,1]
  SD <- sqrt(dim(glm_output$model)[1])*summary(glm_output)$coefficients[,2]
  cohend <- round(beta/SD,3)
  return(cohend)
}

reg_cohend(mod_CSQ_tot)

## (Intercept)     SexMale         Age   RaceBlack   RaceOther    GAD7_tot 
##       0.217      -0.060       0.032      -0.030      -0.026       0.740 
##     CSQ_tot 
##       0.083

COMMENTARY: Here we see CSQ_tot (sum tally of all CSQ variables) is significantly associated with depressive severity when measured continuously (PHQ9_tot), but small effect size compared to GAD7.

mod_CSQ_tot<-lm(CSQ_tot~Sex+Age+Race+PHQ9_cut+GAD7_tot, data=PUCC_FIN)
summary(mod_CSQ_tot)

## 
## Call:
## lm(formula = CSQ_tot ~ Sex + Age + Race + PHQ9_cut + GAD7_tot, 
##     data = PUCC_FIN)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -3.8157 -1.4678 -0.2378  1.2620  9.9556 
## 
## Coefficients:
##                                       Estimate Std. Error t value Pr(>|t|)    
## (Intercept)                           1.758762   0.431803   4.073  5.2e-05 ***
## SexMale                               0.078603   0.159965   0.491 0.623323    
## Age                                  -0.006990   0.005306  -1.318 0.188117    
## RaceBlack                            -0.106042   0.334056  -0.317 0.751011    
## RaceOther                             0.173053   0.199343   0.868 0.385645    
## PHQ9_cutMild depression               0.131029   0.404226   0.324 0.745929    
## PHQ9_cutModerate depression           0.245146   0.380698   0.644 0.519838    
## PHQ9_cutModerately severe depression  0.453570   0.381361   1.189 0.234731    
## PHQ9_cutSevere depression             0.484630   0.395012   1.227 0.220307    
## GAD7_tot                              0.070693   0.018552   3.811 0.000152 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 2.025 on 661 degrees of freedom
##   (121 observations deleted due to missingness)
## Multiple R-squared:  0.05781,    Adjusted R-squared:  0.04498 
## F-statistic: 4.506 on 9 and 661 DF,  p-value: 9.105e-06

reg_cohend(mod_CSQ_tot)

##                          (Intercept)                              SexMale 
##                                0.157                                0.019 
##                                  Age                            RaceBlack 
##                               -0.051                               -0.012 
##                            RaceOther              PHQ9_cutMild depression 
##                                0.034                                0.013 
##          PHQ9_cutModerate depression PHQ9_cutModerately severe depression 
##                                0.025                                0.046 
##            PHQ9_cutSevere depression                             GAD7_tot 
##                                0.047                                0.147

COMMENTARY: Here we are demonstrating that CSQ_tot does not distinguish between 4 levels of depressive severity, when PHQ9 is an categorical (ordinal) variable, when adjusted for demographics and anxiety. This suggests that the sum burden of CSQ stressors (CSQ_tot) is not sensitive enough to distinguish depressive severity in a clinically meaningful way. We have get more refined and ask about specific stressors.

mod_CSQ_tot<-glm(PHQ9_severe~Sex+Age+Race+GAD7_tot+CSQ_tot, family=binomial, data=PUCC_FIN)
summary(mod_CSQ_tot)

## 
## Call:
## glm(formula = PHQ9_severe ~ Sex + Age + Race + GAD7_tot + CSQ_tot, 
##     family = binomial, data = PUCC_FIN)
## 
## Deviance Residuals: 
##     Min       1Q   Median       3Q      Max  
## -2.1370  -0.8484   0.5425   0.7739   2.5866  
## 
## Coefficients:
##              Estimate Std. Error z value Pr(>|z|)    
## (Intercept) -3.119717   0.436004  -7.155 8.35e-13 ***
## SexMale     -0.400732   0.187111  -2.142   0.0322 *  
## Age          0.002851   0.006155   0.463   0.6433    
## RaceBlack   -0.503480   0.385531  -1.306   0.1916    
## RaceOther    0.028264   0.238639   0.118   0.9057    
## GAD7_tot     0.226880   0.021038  10.784  < 2e-16 ***
## CSQ_tot      0.069969   0.047093   1.486   0.1373    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 890.21  on 670  degrees of freedom
## Residual deviance: 709.81  on 664  degrees of freedom
##   (121 observations deleted due to missingness)
## AIC: 723.81
## 
## Number of Fisher Scoring iterations: 4

reg_cohend(mod_CSQ_tot)

## (Intercept)     SexMale         Age   RaceBlack   RaceOther    GAD7_tot 
##      -0.276      -0.083       0.018      -0.050       0.005       0.416 
##     CSQ_tot 
##       0.057

Just to be sure, let’s see if CSQ_tot can distinguish between severe vs. not severe depression (dichotomous variable). We find that CSQ_tot is still not significantly associated with depression (severe vs. not severe) when adjusted for sex, age, race and anxiety. In other words, there’s no added clinical benefit to summing total CSQ items for distinguishing severerly depressed patients at the PUCC, that is added value above and beyond asking for GAD7.

mod1_PHQ9_severe<-glm(PHQ9_severe ~ 
                    Sex+
                    Age+
                    Race+
                    GAD7_severe+
                    new_CSQ3+
                    new_CSQ6+
                    new_CSQ8+
                    new_CSQ13, 
                    family=binomial, data=PUCC_FIN)


reg_cohend(mod1_PHQ9_severe)

##               (Intercept)                   SexMale                       Age 
##                    -0.111                    -0.099                     0.021 
##                 RaceBlack                 RaceOther GAD7_severeSevere anxiety 
##                    -0.042                     0.016                     0.397 
##                 new_CSQ31                 new_CSQ61                 new_CSQ81 
##                     0.062                     0.048                     0.059 
##                new_CSQ131 
##                     0.081

summary(mod1_PHQ9_severe)

## 
## Call:
## glm(formula = PHQ9_severe ~ Sex + Age + Race + GAD7_severe + 
##     new_CSQ3 + new_CSQ6 + new_CSQ8 + new_CSQ13, family = binomial, 
##     data = PUCC_FIN)
## 
## Deviance Residuals: 
##     Min       1Q   Median       3Q      Max  
## -2.2470  -0.8932   0.5963   0.7522   1.7335  
## 
## Coefficients:
##                            Estimate Std. Error z value Pr(>|z|)    
## (Intercept)               -0.846506   0.295338  -2.866  0.00415 ** 
## SexMale                   -0.471830   0.184642  -2.555  0.01061 *  
## Age                        0.003221   0.006026   0.534  0.59301    
## RaceBlack                 -0.423229   0.385401  -1.098  0.27214    
## RaceOther                  0.094817   0.233762   0.406  0.68503    
## GAD7_severeSevere anxiety  1.878136   0.183232  10.250  < 2e-16 ***
## new_CSQ31                  0.321904   0.200810   1.603  0.10893    
## new_CSQ61                  0.288340   0.234838   1.228  0.21951    
## new_CSQ81                  0.599698   0.395113   1.518  0.12907    
## new_CSQ131                 0.397496   0.189516   2.097  0.03596 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 886.36  on 667  degrees of freedom
## Residual deviance: 735.87  on 658  degrees of freedom
##   (124 observations deleted due to missingness)
## AIC: 755.87
## 
## Number of Fisher Scoring iterations: 4

#Table

mod1_PHQ9_severe %>%
  tbl_regression(
    exponentiate = TRUE, 
    pvalue_fun = ~style_pvalue(.x, digits = 2),
  ) %>% 
  add_global_p() %>%
  bold_p(t = 0.05) %>%
  bold_labels() %>%
  italicize_levels()

## add_global_p: Global p-values for variable(s) `add_global_p(include = c("Sex",
## "Age", "Race", "GAD7_severe", "new_CSQ3", "new_CSQ6", "new_CSQ8", "new_CSQ13"))`
## were calculated with
##   `car::Anova(x$model_obj, type = "III")`

Characteristic	OR¹	95% CI¹	p-value
Sex			0.010
Female	—	—
Male	0.62	0.43, 0.89
Age	1.00	0.99, 1.02	0.59
Race			0.48
White	—	—
Black	0.65	0.31, 1.40
Other	1.10	0.70, 1.75
GAD7_severe			<0.001
Minimal to moderate anxiety	—	—
Severe anxiety	6.54	4.59, 9.41
new_CSQ3			0.11
0	—	—
1	1.38	0.93, 2.05
new_CSQ6			0.22
0	—	—
1	1.33	0.85, 2.13
new_CSQ8			0.12
0	—	—
1	1.82	0.87, 4.14
new_CSQ13			0.035
0	—	—
1	1.49	1.03, 2.16
¹ OR = Odds Ratio, CI = Confidence Interval

COMMENTARY: this table is a the same as the txt version above

#Removal of CSQ6, then CSQ8

mod2_PHQ9_severe<-glm(PHQ9_severe ~ 
                    Sex+
                    Age+
                    Race+
                    GAD7_severe+
                    new_CSQ3+
                    new_CSQ13, 
                    family=binomial, data=PUCC_FIN)


summary(mod2_PHQ9_severe)

## 
## Call:
## glm(formula = PHQ9_severe ~ Sex + Age + Race + GAD7_severe + 
##     new_CSQ3 + new_CSQ13, family = binomial, data = PUCC_FIN)
## 
## Deviance Residuals: 
##     Min       1Q   Median       3Q      Max  
## -2.1236  -0.9029   0.5954   0.7386   1.7157  
## 
## Coefficients:
##                            Estimate Std. Error z value Pr(>|z|)    
## (Intercept)               -0.848895   0.294967  -2.878   0.0040 ** 
## SexMale                   -0.447705   0.183148  -2.445   0.0145 *  
## Age                        0.004076   0.005976   0.682   0.4951    
## RaceBlack                 -0.379683   0.379580  -1.000   0.3172    
## RaceOther                  0.094091   0.232575   0.405   0.6858    
## GAD7_severeSevere anxiety  1.908618   0.182455  10.461   <2e-16 ***
## new_CSQ31                  0.384722   0.197769   1.945   0.0517 .  
## new_CSQ131                 0.442485   0.187827   2.356   0.0185 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 886.36  on 667  degrees of freedom
## Residual deviance: 740.04  on 660  degrees of freedom
##   (124 observations deleted due to missingness)
## AIC: 756.04
## 
## Number of Fisher Scoring iterations: 4

reg_cohend(mod2_PHQ9_severe)

##               (Intercept)                   SexMale                       Age 
##                    -0.111                    -0.095                     0.026 
##                 RaceBlack                 RaceOther GAD7_severeSevere anxiety 
##                    -0.039                     0.016                     0.405 
##                 new_CSQ31                new_CSQ131 
##                     0.075                     0.091

#Table

mod2_PHQ9_severe %>%
  tbl_regression(
    exponentiate = TRUE, 
    pvalue_fun = ~style_pvalue(.x, digits = 2),
  ) %>% 
  add_global_p() %>%
  bold_p(t = 0.05) %>%
  bold_labels() %>%
  italicize_levels()

## add_global_p: Global p-values for variable(s) `add_global_p(include = c("Sex",
## "Age", "Race", "GAD7_severe", "new_CSQ3", "new_CSQ13"))` were calculated with
##   `car::Anova(x$model_obj, type = "III")`

Characteristic	OR¹	95% CI¹	p-value
Sex			0.014
Female	—	—
Male	0.64	0.45, 0.91
Age	1.00	0.99, 1.02	0.49
Race			0.53
White	—	—
Black	0.68	0.33, 1.45
Other	1.10	0.70, 1.74
GAD7_severe			<0.001
Minimal to moderate anxiety	—	—
Severe anxiety	6.74	4.74, 9.69
new_CSQ3			0.051
0	—	—
1	1.47	1.00, 2.17
new_CSQ13			0.018
0	—	—
1	1.56	1.08, 2.25
¹ OR = Odds Ratio, CI = Confidence Interval

#Removal of CSQ3

mod3_PHQ9_severe<-glm(PHQ9_severe ~ 
                    Sex+
                    Age+
                    Race+
                    GAD7_severe+
                    new_CSQ13, 
                    family=binomial, data=PUCC_FIN)

summary(mod3_PHQ9_severe)

## 
## Call:
## glm(formula = PHQ9_severe ~ Sex + Age + Race + GAD7_severe + 
##     new_CSQ13, family = binomial, data = PUCC_FIN)
## 
## Deviance Residuals: 
##     Min       1Q   Median       3Q      Max  
## -2.0326  -0.9306   0.5631   0.7065   1.6706  
## 
## Coefficients:
##                            Estimate Std. Error z value Pr(>|z|)    
## (Intercept)               -0.737627   0.287756  -2.563  0.01037 *  
## SexMale                   -0.440615   0.182595  -2.413  0.01582 *  
## Age                        0.003210   0.005939   0.540  0.58886    
## RaceBlack                 -0.340629   0.372915  -0.913  0.36102    
## RaceOther                  0.106274   0.231749   0.459  0.64654    
## GAD7_severeSevere anxiety  1.934949   0.181724  10.648  < 2e-16 ***
## new_CSQ131                 0.498170   0.185276   2.689  0.00717 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 886.36  on 667  degrees of freedom
## Residual deviance: 743.86  on 661  degrees of freedom
##   (124 observations deleted due to missingness)
## AIC: 757.86
## 
## Number of Fisher Scoring iterations: 4

reg_cohend(mod3_PHQ9_severe)

##               (Intercept)                   SexMale                       Age 
##                    -0.099                    -0.093                     0.021 
##                 RaceBlack                 RaceOther GAD7_severeSevere anxiety 
##                    -0.035                     0.018                     0.412 
##                new_CSQ131 
##                     0.104

#Table

mod3_PHQ9_severe %>%
  tbl_regression(
    exponentiate = TRUE, 
    pvalue_fun = ~style_pvalue(.x, digits = 2),
  ) %>% 
  add_global_p() %>%
  bold_p(t = 0.05) %>%
  bold_labels() %>%
  italicize_levels()

## add_global_p: Global p-values for variable(s) `add_global_p(include = c("Sex",
## "Age", "Race", "GAD7_severe", "new_CSQ13"))` were calculated with
##   `car::Anova(x$model_obj, type = "III")`

Characteristic	OR¹	95% CI¹	p-value
Sex			0.016
Female	—	—
Male	0.64	0.45, 0.92
Age	1.00	0.99, 1.02	0.59
Race			0.56
White	—	—
Black	0.71	0.34, 1.49
Other	1.11	0.71, 1.76
GAD7_severe			<0.001
Minimal to moderate anxiety	—	—
Severe anxiety	6.92	4.87, 9.94
new_CSQ13			0.007
0	—	—
1	1.65	1.15, 2.37
¹ OR = Odds Ratio, CI = Confidence Interval

COMMENTARY: This is likely the final model: severe depression PUCC presentation is independently associated with increased reported loneliness. This effect persisted after adjusting for sex, age, race, anxiety severity (GAD7, dichotomously measured by severe vs. less severe), and interactions. Of note, financial stressors (CSQ3) trended towards significance in the prior model.

mod4_PHQ9_moderate<-glm(PHQ9_moderate~
                          Sex+
                          Age+
                          Race+
                          GAD7_severe+
                          new_CSQ3+
                          new_CSQ4+
                          new_CSQ6+
                          new_CSQ7+
                          new_CSQ13,
                          family=binomial, data=PUCC_FIN)
summary(mod4_PHQ9_moderate)

## 
## Call:
## glm(formula = PHQ9_moderate ~ Sex + Age + Race + GAD7_severe + 
##     new_CSQ3 + new_CSQ4 + new_CSQ6 + new_CSQ7 + new_CSQ13, family = binomial, 
##     data = PUCC_FIN)
## 
## Deviance Residuals: 
##     Min       1Q   Median       3Q      Max  
## -3.0404   0.2104   0.3118   0.4575   1.2565  
## 
## Coefficients:
##                            Estimate Std. Error z value Pr(>|z|)    
## (Intercept)               -0.058007   0.358836  -0.162   0.8716    
## SexMale                   -0.238601   0.235730  -1.012   0.3115    
## Age                        0.005902   0.007346   0.803   0.4217    
## RaceBlack                 -0.031823   0.485506  -0.066   0.9477    
## RaceOther                  0.415600   0.317185   1.310   0.1901    
## GAD7_severeSevere anxiety  2.383427   0.272594   8.744   <2e-16 ***
## new_CSQ31                 -0.023231   0.270241  -0.086   0.9315    
## new_CSQ41                  0.419891   0.251836   1.667   0.0955 .  
## new_CSQ61                  0.475191   0.336535   1.412   0.1579    
## new_CSQ71                  0.702721   0.368121   1.909   0.0563 .  
## new_CSQ131                 0.393446   0.249004   1.580   0.1141    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 604.09  on 667  degrees of freedom
## Residual deviance: 474.55  on 657  degrees of freedom
##   (124 observations deleted due to missingness)
## AIC: 496.55
## 
## Number of Fisher Scoring iterations: 5

#Removing variables starting with least significant: CSQ6 --> CSQ3 --> CSQ4 --> CSQ13

mod4_PHQ9_moderate<-glm(PHQ9_moderate~
                          Sex+
                          Age+
                          Race+
                          GAD7_severe+
                          new_CSQ7+
                          new_CSQ13,
                          family=binomial, data=PUCC_FIN)
summary(mod4_PHQ9_moderate)

## 
## Call:
## glm(formula = PHQ9_moderate ~ Sex + Age + Race + GAD7_severe + 
##     new_CSQ7 + new_CSQ13, family = binomial, data = PUCC_FIN)
## 
## Deviance Residuals: 
##     Min       1Q   Median       3Q      Max  
## -2.8552   0.2279   0.3182   0.4084   1.1679  
## 
## Coefficients:
##                             Estimate Std. Error z value Pr(>|z|)    
## (Intercept)                0.1107832  0.3373149   0.328   0.7426    
## SexMale                   -0.2087586  0.2334934  -0.894   0.3713    
## Age                        0.0060160  0.0071899   0.837   0.4027    
## RaceBlack                  0.0006439  0.4820406   0.001   0.9989    
## RaceOther                  0.4344841  0.3149547   1.380   0.1677    
## GAD7_severeSevere anxiety  2.4155156  0.2712183   8.906   <2e-16 ***
## new_CSQ71                  0.7440086  0.3578971   2.079   0.0376 *  
## new_CSQ131                 0.4185383  0.2443613   1.713   0.0868 .  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 604.09  on 667  degrees of freedom
## Residual deviance: 479.48  on 660  degrees of freedom
##   (124 observations deleted due to missingness)
## AIC: 495.48
## 
## Number of Fisher Scoring iterations: 5

reg_cohend(mod4_PHQ9_moderate)

##               (Intercept)                   SexMale                       Age 
##                     0.013                    -0.035                     0.032 
##                 RaceBlack                 RaceOther GAD7_severeSevere anxiety 
##                     0.000                     0.053                     0.345 
##                 new_CSQ71                new_CSQ131 
##                     0.080                     0.066

#Table 
mod4_PHQ9_moderate %>%
  tbl_regression(
    exponentiate = TRUE, 
    pvalue_fun = ~style_pvalue(.x, digits = 2),
  ) %>% 
  add_global_p() %>%
  bold_p(t = 0.05) %>%
  bold_labels() %>%
  italicize_levels()

## add_global_p: Global p-values for variable(s) `add_global_p(include = c("Sex",
## "Age", "Race", "GAD7_severe", "new_CSQ7", "new_CSQ13"))` were calculated with
##   `car::Anova(x$model_obj, type = "III")`

Characteristic	OR¹	95% CI¹	p-value
Sex			0.37
Female	—	—
Male	0.81	0.51, 1.28
Age	1.01	0.99, 1.02	0.40
Race			0.36
White	—	—
Black	1.00	0.40, 2.74
Other	1.54	0.85, 2.93
GAD7_severe			<0.001
Minimal to moderate anxiety	—	—
Severe anxiety	11.2	6.72, 19.6
new_CSQ7			0.028
0	—	—
1	2.10	1.08, 4.44
new_CSQ13			0.083
0	—	—
1	1.52	0.95, 2.47
¹ OR = Odds Ratio, CI = Confidence Interval

COMMENTARY: In this model, we show that lowering the threshold of depressive severity from severe to moderate (still dichotomously measured) diminishes the significance of CSQ13 (loneliness) to a trend. It would be interesting to see if the association between loneliness and depressive severity is dose-dependent (quantitative) or if severely depressed patients tend to form a more lonely group (qualitative), but out of scope for this poster. Of note, in this model job instability (CSQ7) is more strongly associated to moderate to severe depression, compared to loneliness.

PUCC_CSQ_Analysis

S. Murata

9/15/2021