Introduction
In recent years, there has been an increasing interest in the application of systems thinking to exploring, understanding and acting on complex public health issues.1 The interest in systems thinking arises from the growing body of evidence acknowledging the multiple, systemic and complex causes of many public health issues, and that to address them, actions are required at both the individual and societal level.2 A complex system is best considered as a collection of elements (eg, subsystems, sectors), with interconnections between those elements.2 A systems approach to public health conceptualises poor health and health inequalities as outcomes of a multitude of interdependent elements within a connected whole.3 By developing an understanding of these interdependencies, we can develop interventions or approaches that are more sensitive to the complexities of human behaviour. This is consistent with guidance from the recent MRC framework for developing and evaluating complex interventions.1
System mapping has emerged as one of the first steps to exploring the nature of any complex system. The most prominent example of this approach in the UK was the Foresight report, published in 2007.4 This emphasised the complex nature of obesity and argued for it to be addressed using systems approaches, with accompanying system maps. These ‘maps’ are diagrams that show the factors that influence an outcome (in this case, obesity) and how they are—or might be—connected. Guidance from Public Health England now recommends system mapping as part of any approach to exploring and intervening in complex systems.5
A recent scoping review of systems approaches for increasing physical activity in populations found that system mapping has been applied to physical activity in three principal ways.6 First, system maps have been used as a theorising tool, identifying and comparing stakeholder understanding of a complex system. For example, in Derby UK, system maps were developed to help stakeholders explore the influences on physical activity in order to plan and modify an existing town-wide physical activity programme with many complex and interacting components.7 Second, system maps—predominantly using system dynamic modelling (SDM)—have been used to hypothesise and simulate how an intervention might impact on and interact in a complex system.8 For example, SDM was used to compare the effects of policies to increase bicycle commuting, based on participatory methods including interviews and workshops with policy, community and academic stakeholders.9 Third, system maps developed through consensus-based methods such as participatory system mapping or group model-building have been used to inform intervention development. For example, in the UK, participatory systems mapping was used with stakeholders to produce a theoretical framework and intervention plan for the Healthy New Town Programme.10
A further application of system mapping that has been used less frequently is as part of a process evaluation of an intervention or programme. The scoping review cited above6 found 155 examples of systems approaches, of which 11 were part of process evaluations. However, only one of these used a version of system mapping (using an ‘Action Mapping Tool’ to systematically record information related to local authority actions on obesity).11 In this paper, we describe an innovative approach to using system mapping to explore the wider factors that might influence the implementation or effectiveness of an intervention delivered in multiple sites. We will explore the use of system mapping as part of the ongoing process evaluation of a community-based active ageing intervention: the Active, Connected, Engaged (ACE) randomised controlled trial.12 In this paper, we will refer to the trial by the acronym ‘ACE’ although it is known as ‘ACTIF’ in Wales to avoid any confusion with a Welsh national programme called ACE that focuses on adverse childhood experiences.
Background to the ACE trial
The ACE trial aims to evaluate the effectiveness and cost-effectiveness of the ACE intervention, a community-based active ageing programme designed to improve physical function. ACE is a low-cost programme in which older volunteers (55 years+) support older people (65 years+) to improve their mobility by becoming more active within their communities. During the ACE programme, volunteers (n=150) meet with participants twice to get to know each other, find out about and discuss local community-based activities that the participant would like to join and identify and address any barriers to taking part. Then, over a 3-month period, the volunteer–participant pair attend at least three local activities chosen by the participant, together. Over the following 3 months, volunteers support the participant to continue attending these activities independently through regular phone calls, with further joint visits to activities scheduled if needed. Outcome measures include lower limb function; detailed analysis of objectively measured physical activity; muscle-strengthening exercise and a wide range of psychological and social outcomes.12 This is accompanied by an economic evaluation to test cost-effectiveness. ACE was based on the findings of an earlier feasibility study13–15 which led to the development of this definitive effectiveness and cost-effectiveness study, targeting diverse and deprived communities. These communities were identified by demographic data and deprivation index scores. The deprivation index is a measure of relative deprivation at a small local area level produced by the UK Government.16 It is based on seven different domains of deprivation including: income; employment; education; health and disability; crime; housing and living environment. This aims to capture a wide range of aspects of an individual’s living conditions. Based on these criteria, we selected four diverse areas in terms of deprivation including Stoke-on-Trent, Manchester and Bristol in England and Cardiff in Wales. This diversity increases the complexity of the trial and underscores the need for a comprehensive process evaluation to explore the workings of the intervention and appreciate the breadth of contextual factors influencing its effectiveness.
The process evaluation runs alongside the main trial and aims to explore the actual operation of the trial; to understand the mechanisms of change; and to test and, if necessary, refine the trial’s logic model.17 The process evaluation includes the following components (see online supplemental material file for the timing of each component):
Longitudinal, semistructured participant interviews to track the experiences of participants throughout the study.
Semistructured interviews to explore the experiences of peer volunteers and provider organisations.
Assessment of the quality of delivery of the ACE intervention by peer-volunteers, through audio-recordings of a purposive sample of participant-volunteer consultation meetings.
Quantitative data collection among participants and peer-volunteers to explore possible mechanisms of change of the ACE intervention.
System-level process evaluation using a system mapping protocol (the focus of this paper) before the launch of the ACE intervention and repeated after its completion.
System mapping
The ACE intervention has to operate within a complex system with many interacting factors that influence whether or not an individual may be physically active. These operate at a variety of levels, from individual factors such as mobility limitations; social factors such as social isolation and community norms; and environmental factors such as the existence of safe places to be active. Alongside these factors are a multitude of ‘stakeholders’: individuals and agencies who have an interest in helping older people to be active and improve their health and well-being. This part of the process evaluation sets out to explore the system around physical activity and volunteering and place the ACE intervention within this system. We developed the system maps based on primary participatory research with participants at the ACE delivery sites, as it was essential to explore locality-specific issues that might have been missed had we relied on the findings of existing reviews without appreciation for local characteristics. The system mapping component of the process evaluation had the following objectives:
To gather area-specific information on participants’ barriers and facilitators to physical activity and the needs of peer volunteers and service providers prior to the start of the intervention. This can be seen as ‘formative evaluation’: using system mapping to propose refinements to the intervention’s design and delivery and the overall logic model for the intervention, which describes the assumptions and processes by which ACE aims to influence short-term, medium-term and long-term outcomes (see online supplemental material file).
To create ‘baseline’ system maps before the intervention takes place in three ACE delivery sites. This process will be repeated after the completion of the intervention in each area in order to provide a system-level context for the longitudinal process and impact evaluations, helping to identify individual-level and systems-level factors that might influence the effectiveness of the ACE intervention.