The data are from a survey of primary care patients in Cantabrian, Spain, to estimate the prevalence of psychiatric illness. The prevalence for each of the sexes is estimated from data provided by the general practitioner and from completing the general health questionnaire.
Column 1: Subject ID
Column 2: Illness = 1, Otherwise = 0
Column 3: Gender ID
Column 4: General practitioner = GP, General health questionnaire = GPQ
## ID Health Gender Report
## 1 1 0 M GP
## 2 1 0 M GHQ
## 3 2 0 M GP
## 4 2 0 M GHQ
## 5 3 0 M GP
## 6 3 0 M GHQ
## 'data.frame': 1646 obs. of 4 variables:
## $ ID : int 1 1 2 2 3 3 4 4 5 5 ...
## $ Health: int 0 0 0 0 0 0 0 0 0 0 ...
## $ Gender: chr "M" "M" "M" "M" ...
## $ Report: chr "GP" "GHQ" "GP" "GHQ" ...
## Report GHQ GP
## Gender Health
## M Illness 244 287
## Otherwise 80 37
## F Illness 306 420
## Otherwise 193 79
summary(m0 <- geeglm(Health ~ Gender + Report + Gender*Report, data=dta2, id = ID, corstr="exchangeable", family = binomial))##
## Call:
## geeglm(formula = Health ~ Gender + Report + Gender * Report,
## family = binomial, data = dta2, id = ID, corstr = "exchangeable")
##
## Coefficients:
## Estimate Std.err Wald Pr(>|W|)
## (Intercept) -0.46089 0.09192 25.141 5.33e-07 ***
## GenderM -0.65425 0.15826 17.089 3.57e-05 ***
## ReportGP -1.20991 0.12651 91.462 < 2e-16 ***
## GenderM:ReportGP 0.27649 0.22828 1.467 0.226
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation structure = exchangeable
## Estimated Scale Parameters:
##
## Estimate Std.err
## (Intercept) 1 0.07865
## Link = identity
##
## Estimated Correlation Parameters:
## Estimate Std.err
## alpha 0.2975 0.04744
## Number of clusters: 823 Maximum cluster size: 2
| Health | |||
|---|---|---|---|
| Predictors | Odds Ratios | CI | p |
| (Intercept) | 0.63 | 0.53 – 0.76 | <0.001 |
| Gender [M] | 0.52 | 0.38 – 0.71 | <0.001 |
| Report [GP] | 0.30 | 0.23 – 0.38 | <0.001 |
| Gender [M] * Report [GP] | 1.32 | 0.84 – 2.06 | 0.226 |