Pragmatic trials and implementation science: grounds for divorce?

It will be recalled that the PEMS sub-study [31] labels itself as a ‘theory-based process evaluation’. At the risk of complicating matters, some difference between the authors’ strategy and the approach advocated in this paper are worthy of note. Grimshaw et al. utilise the ‘theory of planned behaviour’ (TPB) to explain the paradoxical outcomes of the main trial. TPB is an example of what Nilsen [32] refers to as ‘classical theory’, that is to say a theory that originate from fields external to an intervention under study. In this case TPB is a generic psychological theory on decision-making, which explains why peoples’ best intentions do not always lead to action. Put simply, the theory is that the motivated individual will only act on their intentions if they have belief in their own ability to perform the requisite actions. It turns out many practitioners in PEMS study were positively inclined to increase retinopathy referrals and confident in their ability to do so – and yet still do not change their behaviour. Interesting and carefully researched as this finding is, it is arguably a refutation of TPB, rather than an exemplification. A standard charge against this psychological theory is that it cannot account for, as in this instance, environmental and organisational aspects of decision making. In general terms, my preference is to start with more mundane programme theories, which more closely reflect the thinking of planners and practitioners, and which render these stakeholders more likely to act on any research which probes into them.

Note that the claim here is not that case study design is inherently superior to any other research design. The four examples here all have their methodological flaws and limitations. One important instance is that all four rely on self-report on the matter of whether practitioners make use the guideline. In this respect PRCTs, like [28], are superior because they test whether the guideline produce changed behaviour. Individual case studies are always open to improvement by optimising the balance of quantitative and qualitative methods. But collectively they are paramount because they provide multiple tests of the same programme theory.