Designing a generic, adaptive protocol resource for the measurement of health impact in cash transfer pilot and feasibility studies and trials in high-income countries

Elliott Aidan Johnson, Matthew Thomas Johnson, Christodoulos Kypridemos, Aase Villadsen, Kate E. Pickett

In the context of the COVID-19 pandemic, upstream interventions that tackle social determinants of health inequalities have never been more important. Evaluations of upstream cash transfer trials have failed to capture comprehensively the impacts that such systems might have on population health through inadequate design of the interventions themselves and failure to implement consistent, thorough research measures that can be used in microsimulations to model long-term impact. In this article, we describe the process of developing a generic, adaptive protocol resource to address this issue and the challenges involved in that process. The resource is designed for use in high-income countries (HIC) but draws on examples from a UK context to illustrate means of development and deployment. The resource is capable of further adaptation for use in low- and middle-income countries (LMIC). It has particular application for trials of Universal Basic Income but can be adapted to those covering other kinds of cash transfer and welfare system changes. We outline two types of prospective intervention based on pilots and trials currently under discussion. In developing the remainder of the resource, we establish six key principles, implement a modular approach based on types of measure and their prospective resource intensity, and source (validated where possible) measures and baseline data primarily from routine collection and large, longitudinal cohort studies. Through these measures, we seek to cover all areas of health impact identified in our theoretical model for use in pilot and feasibility studies. We find that, in general, self-reported measures alongside routinely collected linked respondent data may provide a feasible means of producing data capable of demonstrating comprehensive health impact. However, we also suggest that, where possible, physiological measures should be included to elucidate underlying biological effects that may not be accurately captured through self-reporting alone and can enable modelling of long-term health outcomes. In addition, accurate self-reported objective income data remains a challenge and requires further development and testing. A process of development and implementation of the resource in pilot and feasibility studies will support assessment of whether or not our proposed health outcome measures are acceptable, feasible and can be used with validity and reliability in the target population. We suggest that while Open Access evaluation instruments are available and usable to measure most constructs of interest, there remain some areas for which further development is necessary. This includes self-reported wellbeing measures that require paid licences but are used in a range of nationally important longitudinal studies instead of Open Access alternatives.