A recent article emphasised the power of using theory to explain why improvement programmes work. When key components of an intervention can be identified and their theoretical contribution to final outcomes understood, causal associations can be identified and tested. Although it has been argued that complex interventions require complex solutions, psychological theories have an important function of explaining human behaviours that are often the focus of change. So how can we isolate and describe theories that may be used formally or informally within complex interventions?
Logic models are a useful tool. They are often described as process maps, theories of change, or logical frameworks. Similarly they are depicted in many different ways and formats.
At their core, they are simplified and graphical descriptions of what happens in an intervention, described in a particular sequence, with clear and practical benefits. It is important to represent the key actions or components that need to take place to drive the change that will generate the expected outcomes.
They usually describe the mechanisms of change, in terms of the reasons that all activities are thought to lead to the intended outcomes. In designing a logic model, it is also useful to ask IF (activities are implemented) THEN (what) behaviours or outcomes are expected? Psychological theory can be used to explain how the actual activity encourages the specific behaviours (and why this is likely to happen).
This is not just good practice, as there is evidence to suggest that multifaceted interventions often work better, when they incorporate interventions components that are known to be effective.
So when should we bother to develop a logic model of what we are doing in a complex intervention? Often they are seen as a useful tool to design evaluation schedules.
However, the process of developing a logic model can be very beneficial for intervention teams, to discuss and clarify hidden assumptions about how the intervention is expected to work. Most people will have mental models or ‘common sense’ explanations of how the active ingredients of an intervention work together to achieve the expected outcomes. However, internal assumptions and explanations may not be aligned and therefore, different practices could be used without being theoretically compatible. Therefore, a logic model can guide the key outcomes to be assessed, the measurement tools, analytical approaches and standards to be used.
Further, when a logic model is developed early in a complex intervention, it can highlight the need for certain types of contextual analysis, through the identification of core components that are necessary for success. It may help in prioritising which barriers are most important to address. Logic models can also influence building efficient implementation plans that can also highlight which key processes need to be monitored and evaluated. This may ensure that evaluation is designed into the intervention from the outset. Finally, it makes the evaluation of these programmes much more usable for future use; when you can affirm that a particular component influenced behaviour to achieved its expected success. It can be even more exciting when Qualitative interviews can confirm the theoretical explanation of why this happened!