Conducting cost-effectiveness analyses of type 2 diabetes in low- and middle-income countries: Can locally generated observational study data overcome methodological limitations?

In low- and middle-income countries, the high personal and economic burden of type 2 diabetes is further compounded by inadequate resources for diabetes care when compared with high-income countries. Health technology assessments (HTAs) aim to inform policy decision makers in their efforts to achieve more effective allocation of resources by providing evidence-based input on new technologies. Within the hierarchy of evidence, randomized controlled trials (RCTs) remain the 'gold standard' used to inform HTAs, but are limited by poor external validity (ie, generalizability to real-world populations). Unlike RCTs, observational studies are able to enrol broader patient populations, but their design renders such studies vulnerable to confounding factors and selection bias. However, it is increasingly recognized that observational studies can complement RCTs by supporting and extending efficacy findings from RCTs to real-world clinical practice, particularly across geographical populations. They can also provide locally relevant baseline and disease natural history data to populate health economic models. Thus, observational data are likely to be of considerable informative value to policy makers in developing countries reaching decisions on diabetes care within an environment of scarce resources.