Predictors Of Enrollment Into Community Based Health Insurance Schemes Among Patients with Diabetes Mellitus in Wakiso District – Central Uganda

Abstract

Background: This study was done to assess predictors of enrollment into community based health insurance schemes and patients with type II diabetes mellitus in Wakiso district – central Uganda. Diabetes is linked with 1.5 million deaths annually, with all deaths related to poorly controlled diabetes that arise, in part due to non-adherence to medication. Such non-adherence in most cases arises from financial limitations and so, diabetic patients need to have health insurance cover. Currently, CBHI schemes are the most common in developing countries and some have been customized for diabetes patients, as is the case in Wakiso district. However, enrollment into those schemes remains a challenge.

Method: The study was cross sectional, targeting 291 diabetic patients in Wakiso district, in which all the three facilities with the CBHI scheme were purposively sampled and patients randomly sampled, following which structured interviews were used to collect data. Data was analyzed in SPSS 25 using the log-binomial model.

Results: The proportion of patients with diabetes mellitus in Wakiso district that is enrolled into community based health insurance schemes is 81.4% (237). The prevalence of CBHI enrollment was higher among patients who were Catholics (aPR = 3.982 [CI = 1.190 - 13.318], P = 0.025), who had been educated to secondary school (aPR =12.749 [CI = 3.716 - 43.735], P = 0.000), who rated their health status as being somehow okay had 16 times the odds of being enrolled (aPR = 16.526 [CI = 5.001 - 54.611], P = 0.000). It was however less among patients who had lived with diabetes for less than 5 years (aPR = 0.269 [CI= 0.127 - 0.568], P = 0.001), although it was higher among patients who reported that diabetes medication was always available at their respective health facilities (aPR = 1.979 [CI = 1.066 -3.671], P = 0.030). No bivariate analysis could be conducted because of the null integers in the cells of all cross tabulations between the independent variables and the dependent variable. The null integers notwithstanding, it is evident that there could have been significantly higher proportions of enrolled patients who agreed to the institutional assertions than, those who disagreed.

Conclusion: Enrollment into the CBHI among diabetic patients in Wakiso district is high, but not universal, contrary to ideal standards. It is mainly predicted by intrapersonal characteristics and to some extent by institutional characteristics, although it was possible that scheme characteristics are also been of conceivable importance. Behavior change communication will go a long way in increasing enrollment among the patients