Advancing ‘real-world’ trials that take account of social context and human volition

The dissemination over a decade ago of the Medical Research Council’s (MRC) Framework for the Development and Evaluation of Complex Healthcare Interventions [1], and appeals by researchers for more research to be conducted on interventions in complex systems [2] has contributed to the proliferation of research in more open systems, i.e. in ‘real-world’ settings (for an example, see the INCLUSIVE study – [3]). The MRC framework advocates a mixed-method approach, including: qualitative research – such as process evaluation of implementation of an intervention and of contextual factors that could lead to variability in outcomes [4]; causal modelling approaches, that can provide information on the process and outcome of an intervention and that may lead to design changes prior to implementation, e.g. [5]; and a range of experimental methods for evaluating effectiveness, with preference given to randomised designs (randomised controlled trials – RCTs), wherever possible or appropriate, for optimisation of internal validity [6]. The mixed-method approach is an attempt to supplement the randomised trial with procedures embedded in realism that take the complexity of social context and human volition into account, and to shift the focus from identifying causality to explaining causality, or in other words, to explaining how intervention components interact with different populations and in different contexts [7]. Moore and Evans [8] demonstrate the importance of context by describing how an intervention nearly two decades ago, that effectively reduced smoking in young people by prohibiting smoking on school premises and thereby challenging norms surrounding smoking behaviour, may not prove as effective today. Norms surrounding smoking that existed 15–20 years ago (e.g. high prevalence rates and social acceptance) have changed, and consequently, so have the mechanisms that predict smoking behaviour in young people [8].

The combination of positivist and realist approaches found in the MRC framework, which Porter et al. [7] and others [9] refer to as realist RCTs, has given rise to debate [7, 9‐13]. Central to the debate is whether randomised and realist evaluation can be consolidated. While critics argue that the realist RCT design found in the MRC framework does not fully take complexity into account [12], the MRC framework does offer a pragmatic step towards a more open systems (real-world) approach that bridges, to a degree, randomised evaluation with social context and human agency [14]. The recent paper in Trials by Porter and colleagues highlighting the utility of applying a critical realism approach [7] and the ongoing debate exploring the utility between realist and randomised evaluations [9, 12] are welcome and important steps towards the development and refinement of a usable open systems approach.

Since the introduction of the MRC framework, complex interventions are increasingly being developed in an attempt to solve the continually expanding burden of health-related issues. This commentary is, therefore, a plea to policy-makers, and those who make the strategic decisions regarding allocation of research funding, to support to a greater degree all phases of real-world trials that take into account social context and human volition as advocated by the MRC framework, rather than to automatically allocate research funds on the basis of received wisdom.