Where next for randomised controlled trials in global health?

Abstract

The 2019 Nobel Memorial Prize in Economic Sciences was awarded to three economists—Esther Duflo, Abhijit Bannerjee, and Michael Kremer—for their experiment-based research to mitigate global poverty. The award was notable for several reasons. Esther Duflo was only the second woman to win the prize since it began in 1969, and the youngest ever winner. Previous prizes were awarded largely for contributions to theory—for example, by observing and interpreting the behaviour of markets. But this year, the Sveriges Riksbank committee's recognition of the randomised controlled trial (RCT) elevates a method long used in medicine, but much less so in human development. Although based in the USA, the three laureates have worked with some of the poorest people to understand their lives and the constraints that keep them poor. They made important discoveries. Those living in poverty often spend their meagre disposable income on activities to relieve tedium or bring small pleasures. In India, they found that people spent their budgets on religious festivals. In Nicaragua, it was owning a radio or television. The laureates concluded that simply providing money to the poor to alleviate poverty was not enough because there are too many competing expenditures. They decided to break down the problems into smaller manageable questions and use empirical data to examine which interventions work and which do not, and what motivates people to make the choices they make. They used RCTs to assess the causal effects of an intervention. Together, their work has made important contributions to health care, education, agriculture, and gender issues. Understanding how demand for de-worming medicines to treat parasitic infections in Kenya is affected by price provided the case for why governments should subsidise health care. In India, vaccination uptake increased by improving service quality and providing families with small incentives. The laureates have used their findings to develop new anti-poverty programmes and influence policy. Despite their success in producing evidence for social change, many notable economists and social scientists have criticised RCTs on philosophical, epistemological, political, and methodological grounds. Of course, other study designs and approaches, such as more qualitative analyses, aid our understanding of health and development too, but the RCT remains the best means of discovering whether any proposed intervention may work. While development economics has drawn lessons from medicine, what can medicine learn from this experimentalist turn in economics? The laureates have shown that RCTs can be done in some of the most challenging human circumstances. Importantly, the design of interventions must be based on a detailed understanding of context. Too often, a policy shown to work in one setting is transplanted to another, with scant regard for whether the situation is at all similar. This scenario is especially true for health policy, in which a community of highly paid international consultants travel business class from country to country peddling their favourite idea. A good example of where these lessons have been learnt is the HOPE-4 trial, published last month in The Lancet. This cluster RCT of hypertension management in Malaysia and Colombia achieved impressive results. The intervention had multiple components—task shifting, peer support, free medicines, and simplified guidelines. But what was particularly important was that the intervention in each country was designed following a detailed study of the lived experience of patients and after interrogation of policy makers. This knowledge was integrated with systematic reviews of experiences elsewhere. Why would someone be expected to take a tablet for high blood pressure for life when they feel perfectly well? Only by answering this question and others might interventions be successful. RCTs in global health must evolve to become more meaningful. Too often, trials are severely restricted, with little ability to plan for changes across the study (adaptive) and being ready for unforeseen decision making (simulation modelling) at a huge cost and effort. An upcoming Lancet Global Health Series on improving efficiency in global health clinical trials aimed both at researchers and funding bodies will focus on innovative designs and avoiding research waste. The lesson from this year's Nobel Prizes is that one size does not fit all. In conceiving and doing rigorous experiments to find out what really works, we need to listen to the voices of the poor and design interventions that respond to their beliefs, needs, and expectations.