Pharmacoepidemiological analysis of chronic conditions and treatment patterns among type 2 diabetes mellitus patients in the South African private health sector

Abstract

This dissertation aimed at investigating the occurrence and medicine treatment of chronic disease list (CDL) conditions in patients with type 2 diabetes mellitus (T2DM) in the South African private health sector. The study consisted of a literature review and an empirical investigation. The literature study provided an overview of T2DM and its comorbidities, which include its prevalence, pathophysiology, risk factors and treatment options. The empirical investigation was performed using medicine claims data for the period 1 January 2008 to 31 December 2016, obtained from a pharmaceutical benefit management (PBM) company in the South African private health sector. The specific research objectives included to: (i) determine the demographic profile of patients with T2DM on the database; (ii) determine the incidence of the other chronic disease list (CD) conditions, as well as the timeline from a baseline year, free of other CDL conditions (2008), to occurrence of these conditions among T2DM patients; (iii) determine the association between the CDL conditions and the T2DM patients’ age and gender; and (iv) determine the medicine prescribing patterns (i.e. number of medicine items, the active substances and the prescribed daily dose) and potentially serious level 1 and 2 drug-drug interactions based on Tatro’s scale for drug-interactions for T2DM patients. Type 2 diabetes mellitus patients and the presence of CDL conditions were identified using International Classification of Diseases and Related Health Problems, 10th Revision (ICD-10) codes on medicine claims data, reimbursed between 2008 and 2016 (i.e. the study period). The timeline to the development of a CDL condition was defined using the receipt of medication in conjunction with a ICD-10 code for the condition as proxy. A total of 114 902 T2DM patients with mean age of 56.95 ± 12.74 years were identified on the database (2008-2016). The majority were males (55.1%, n = 63 326). There was no difference in the mean age across gender groups (p > 0.05; Cohen d <0.2). To determine the demographic profile of patients with T2DM on the database, the incidence rate of other CDL conditions, as well as the timeline from a baseline year, free of other CDL condition (2008) to occurrence of these conditions among T2DM and association between other CDL conditions and the T2DM patients’ age and gender, a closed cohort study design was used. A total 3 390 patients (ICD-10, E11) with mean age 58.42 ± 10.80 years were identified for the study. Of these patients, 2015 (59.6%) were male. Analysis showed that 506 (14.2%) of patients had no other CDL conditions at index year (2008) compared to 340 (67.9%) who had developed at least one additional CDL by 2016. A median of 3 (2-4) other

CDL conditions were identified per patient. In 2016, hypertension was present in 585 per 1 000 persons (n = 297), followed by dyslipidaemia at 454 per 1 000 persons (n = 230). Within the first 500 days from index, 63 (21.2%) of patients had developed hypertension and 44 (19.1%) had developed dyslipidaemia. The association between the other CDL conditions and the T2DM patients’ age and gender was determined using age-adjusted odds ratios. Based on this analysis, only hypothyroidism was significantly associated with female gender in T2DM patients (OR = 3.23 (95% CI, 1.30-8.02), p = 0.012). Age furthermore decreased the odds for women having asthma and epilepsy. In T2DM women, age had a significant effect on the prevalence of dysrithymia, increasing the odds significantly from 0.41 (CI, 0.09-1.93) (p = 0.261) to 1.12 (CI, 1.16-1.18) (p <0.0001). To determine the medicine prescribing patterns (i.e. number of medicine items, the active substances and the prescribed daily dose) and potentially serious level 1 and 2 drug-drug interactions based on Tatro’s scale for drug-interactions for T2DM patients, a longitudinal cross-sectional study, using of 45 179 medicinal active pharmaceutical ingredients (APIs) was conducted. These APIs were identified from the last prescription per patient claimed in 2016, where a prescription was defined as all medicine items claimed during the calender month per patient. A median of 5 (1-10) medicinal API’s were claimed per prescription. The majority of claims were for metformin with 28 961 (64.1%), followed by sulphonylureas at 13 912 (30.3%) and insulin at 1 660 (2.6%). Overall, 223 potentially serious drug interactions based on Tatro’s level 1 and 2 were identified from the 3 390 prescriptions, at an incidence rate of 65.78 per 1000 prescriptions. Statins were identified in 170 (76.2%) of the potential drug-drug interactions. In conclusion, this study provided an estimate of demographic profiles in T2DM patients, the prevalence, incidence and time to development of additional CDL as well as prescribing patterns in T2DM and its CDL conditions in the South African private health sector. Recommendations for future studies were made.