Thus, it makes it difficult to reach a conclusion with regards to the impact of each variable on enrolment.
The dichotomous dependent variable was enrolment in the government scheme (combining the 'previously-insured' and 'newly-insured') and the explanatory variables were categorical and included socio-demographics, salary level and health status.
Despite the steps taken to reduce endogeneity, the main limitation of this study is the cross-sectional nature of the data, which makes it difficult to infer causation between an independent variable and enrolment.
We also examined the association between interruption in care or death and sociodemographic variables at enrolment, as well as clinical/biological variables six months after cART initiation, using a multivariate Cox proportional hazard regression model.
To compare characteristics of survey respondents and non-respondents, we used χ tests for categorical variables (author employer and location, trial funder, and journal) and the Kruskal-Wallis test for continuous variables (trial enrolment), using two-sided tests with a type I error level of 0.05.
H is a vector of dummy variables of enrolment in the student health insurance and free health insurance programs (it is equal one for the insured and zero for the uninsured).
We chose to control for age and gender during analysis rather than matching for these variables during enrolment, because of considerable logistic difficulties that we would have encountered in identifying appropriate matched controls.
Candidate variables at enrolment selected by the spline Cox model included systolic/diastolic blood pressure (BP), HbA1c, BMI, waist circumference, blood haemoglobin (Hb), white blood cells (WBC) count, HDL-C, LDL-C, triglyceride (TG), total cholesterol and drug usage (Table 1).
We tested for endogeneity using household size as an instrumental variable for MHO enrolment.
*Adjusted for age (time variable), year of enrolment, sex, marital status, education, occupation, smoking status, smoking duration, smoking intensity.
