1 General introduction 9 existing or hypothetical trials and after-the-fact analysis of trial data sets13 and both approaches will be covered in this thesis. Inadequate generalizability can arise from under-representation of important subgroups among people with heart failure such as those of older ages, women, minority ethnic groups. Questions arise to whether meaningful differences exist for outcomes within these subgroups as patient characteristics are increasingly shown to be modifiers of treatment effect or safety. Racial differences in incidence and outcomes are well-established in studies in the U.S., in which Black and Hispanic/Latinx persons with heart failure are known to fare worse than White patients.4 This is attributable to a disproportionate burden of CV risk factors that leads to earlier onset of atherosclerotic disease and shorter life expectancy. Risk differences between racial groups are linked to social and system/institutional determinants of health, which cannot be ignored when considering the effectiveness of treatments in real-world heart failure patients. With that said, a broad racial category such as Asians can be heterogenous in terms of culture, ethnic, language and biology. For instance, people of South Asian and East Asian origins have marked differences in the prevalence of ischemic heart disease and its subsequent disease outcomes.4 For heart failure, inter-ethnic differences in prognosis among diverse South East Asian communities is much less understood compared to those on racial disparities in the U.S. On a similar note, people with multimorbid conditions are often excluded from heart failure trials. And yet it is likely that they are part of the population treated with an approved drug, despite a disproportionately small amount of data on safety. Given that extensive exclusion criteria limits target population representativeness and trial accrual rates (57% of terminated trials were terminated because of poor accrual14), knowing how individual criteria affects eligibility in a collective manner would be of added value at the design stage of trials. Among the challenges to assessing generalizability in trials for heart failure, is that it requires access to trial datasets but this is limited particularly for pharmaceutical industry-sponsored trials. Public-private partnerships such as the BigData@Heart project15,16, has enabled sharing of individual patient data for direct comparison of trial populations and heart failure registries. From here, we examine generalizability in terms of heart failure outcomes among enrolled trial participants relative to observational registry patients, quantify age, sex and ethnic differences in
RkJQdWJsaXNoZXIy MjY0ODMw