1 Literature review
1.3 Intervention development
1.3.1 Intervention development models
The literature reviewed thus far identifies the postnatal period as an opportune time for interventions to increase PA levels. Developing health behaviour interventions, defined as those that alter or affect the course of action taken by an individual relating to a health outcome, is a naturally complex process (Craig, Dieppe et al. 2008), and PA is a multi-component behaviour with multiple influencing factors. The complex nature of interventions and problems developing and evaluating them relate to the standardisation of intervention design and delivery, sensitivity to local features, the organisational and logistical difficulty of applying evaluation methods and the length of the causal chains linking intervention with outcome (Craig, Dieppe et al. 2008). Their complex nature demands a systematic approach to development to ensure the effective allocation of resources, to determine the mechanism for change and to enhance the likelihood of effectiveness. Below, I present models of intervention development used throughout this thesis.
1.3.1.1 Medical Research Council guidance: Developing and evaluating complex interventions In 2008, the MRC published a framework of the intervention development cycle (Craig, Dieppe et al.
2008). The guidance places emphasis on the development stages and piloting the intervention and evaluation methods prior to a main efficacy trial. The guidance presents four key stages (Figure 1.4).
These stages are cyclical, and users move between the stages as required by the research findings.
32 Figure 1.4 - MRC guidance stages of intervention development (Craig, Dieppe et al. 2008 p8)
a) Development
The increased focus on intervention development was a key feature of the updated guidance published in 2008. Taking a systematic and thorough approach to intervention development can create an intervention expected to have a worthwhile effect. The guidance proposes three key stages to development. First, identify the existing evidence, through a recently published or conducting an original systematic review. Second, identify or develop appropriate theory because theory-based interventions are likely to be more effective than atheoretical interventions (Michie, Abraham et al. 2009). Additionally, a theory will help understand the likely mechanisms or processes of change in an intervention.
b) Feasibility/piloting
The purpose of this stage is to test the intervention using a phased approach targeting the key uncertainties of the intervention. This stage identifies potential problems ahead of a larger trial and provides the opportunity to implement strategies that address these. The guidance proposes three key elements in this stage: the acceptability of procedures, the likely recruitment and retention rates of participants and a sample size calculation. Both qualitative and quantitative methods are encouraged in this stage and several studies may be required to refine the study design.
c) Evaluation
The third step of the guidance is a large-scale evaluation of intervention efficacy with three key aims: to assess effectiveness, understand the process and assess cost effectiveness. Assessing intervention effectiveness involves two key decisions: the study design and choice of outcomes
33 measured, usually a primary outcome and some secondary measures, guided by the intervention development work. The second aim of evaluation is to understand the processes.
A process evaluation explores the reasons that a successful intervention is successful or an intervention fails. The components of a process evaluation can include fidelity and quality of implementation, clarify causal mechanisms and identify contextual factors associated with variation in outcomes. The third aim of an evaluation is to assess cost-effectiveness, which are useful for decision makers to justify the cost of implementing an intervention.
d) Implementation
Beyond publication, the guidance provides two additional steps to encourage the uptake of the results. The first is to get research into practice through active dissemination of the results in accessible and attractive formats. The second stage is surveillance; monitoring and long-term outcomes because there may be differences in outcomes in a long-term, widely disseminated intervention compared to the research trial.
1.3.1.2 Behaviour Change Wheel
The BCW is an intervention development method that is linked to a model of behaviour (Michie, Atkins et al. 2014). It was developed from a synthesis of nineteen frameworks of behaviour change and has three layers (Figure 1.5). The COM-B model (described in section 1.3.2.6) is the hub, which identifies the sources of behaviour to target in an intervention. Briefly, for a behaviour to occur, individuals must have the capability, opportunity and motivation.
The second layer presents intervention functions and the third layer presents potential policy options. The intervention functions and policy options present a wide choice so that developers can consider the potential of each option. Using these as a guide, intervention developers choose the appropriate intervention function and policy option using the APEASE Criteria. APEASE includes six criteria (affordability, practicability, effectiveness/cost-effectiveness, acceptability, side effects/safety and equity) against which each option is appraised and those meeting all APEASE criteria should be considered.
34 Figure 1.5 – The Behaviour Change Wheel (Michie, Van Stralen et al. 2011 p1)
The BCW has six steps (Table 1.1) to intervention development working through the three layers.
Stage 1 culminates in a behavioural analysis, which identifies the factors that need to change to enable behaviour, categorised according to the COM-B components. During the second stage, the BCW links each behavioural component to the relevant intervention function and policy options for appraisal using the APEASE criteria resulting in the selection of intervention functions and/or policy options for the intervention. Stage three requires designers to identify appropriate BCTs for the chosen intervention functions or policy options. A BCT is an observable, replicable and irreducible component of an intervention designed to alter or redirect causal processes that regulate behaviour (Michie, Richardson et al. 2013). All BCTs are coded using a 93-item taxonomy to identify and report intervention content using a common language to enhance our ability to replicate and compare findings. The BCW identifies the BCTs used ‘more frequently’ and ‘less frequently’ for each intervention function and policy option. Intervention designers should consider the use of each BCT, using the APEASE criteria. The final stage of the BCW is to choose the intervention delivery method, using a taxonomy of delivery mode to consider the most appropriate delivery option.
35 Table 1.1 – Description of each stage of the BCW
Specify the population and the behaviour, eg, what is the behaviour, where does the behaviour occur 2. Select the target
behaviour
Identify all behaviours that contribute to the problem and select the target behaviours of the intervention.
3. Specify the target behaviour
Specify who needs to perform the behaviour, what they need to do differently, when will they do it, how often and with whom
4. Identify what needs to change
What factors within individuals’ capability, opportunity and motivation need to change to enable behaviour
An intervention function is a broad category of means by which an intervention changes behaviour. Using results from the previous step identifying what needs to change, the BCW identifies the intervention functions that are likely to be effective for bringing about the changes.
There are nine intervention functions. Intervention designers appraise each intervention function to choose those likely to be effective.
6. Identify policy categories
For designers who have access to policy options, identify what policy options would support the delivery of the chosen intervention functions 3. Identify content
and
implementation options
7. Identify BCTs Identify BCTs appropriate for the chosen
intervention options from a list of ‘most frequently’
and ‘less frequently’ used, appraising each one 8. Identify mode of
delivery
Using a taxonomy of modes for delivering interventions, decide initially on a face-to-face or distance intervention.
36 The BCW is a relatively new intervention development model. It has been applied in promoting hearing aid use (Barker, Atkins et al. 2016), medication adherence (Jackson, Eliasson et al. 2014), increasing the frequency of very brief PA advice by healthcare professionals to cancer patients (Webb, Hall et al. 2016) and health coaching programme for low income Latina mothers with recent GDM (Handley, Harleman et al. 2015). One systematic review of PA interventions in postnatal women identified ‘self-monitoring of behaviour’ and ‘goal setting’ as the most common BCTs among efficacious interventions (Gilinsky, Dale et al. 2015). In a qualitative study of a postnatal weight management intervention, participants reported using ‘self-monitoring of behaviour’,
‘prompts/cues’ and ‘social support (unspecified)’ (Smith, Taylor et al. 2016).