Self-management definitions

To date, self-management has been variably defined in the literature with no commonly accepted or consensus definition . Self-management typically appears to be referred to either in terms of the process or content of self-management (i.e. what must be done or delivered) or in terms of the aims and behavioural outcomes. This lack of consensus hampers the assessment and comparison of studies as it is likely that these different definitions will lead to different interventions and targets for treatment. Future studies should therefore consider adopting a definition that includes the aim, components and outcomes. As an example, we refer readers to a detailed definition by Hurley et al. and Toomey et al. that incorporates each of these elements based on a consolidation of existing definitions from self-management reviews.

Use of theoretical rationale

Few self-management interventions for LBP have reported developing their intervention using theory. For example, Keogh et al.’s review of behaviour change theories and techniques in group-based self-management interventions for persistent LBP and osteoarthritis found that only three of their 22 included studies were classified as ‘theory informed’. Lack of theoretical rationale has been recognised as an important setback for developing effective interventions and may explain some of the heterogeneity in the effectiveness. Interventions developed using a theoretical rationale facilitate an understanding of how and why the intervention influenced outcomes (or not), thereby enabling researchers to focus on understanding the mechanisms of change and allowing theories to be empirically tested and ultimately improved . There are many theories that explain behaviour in the context of self-management . We provide an overview of two psychological theories, the fear-avoidance model (FAM) and social cognitive theory (SCT), which are commonly used in the field of LBP research. Recent evidence from self-management intervention studies that have specifically targeted elements included in these models shows some promise (e.g. Refs. ), but more studies are needed to provide strong evidence for causal associations between these factors and outcomes. Several key reviews of theories relevant to persistent pain are available, and we refer readers to those reviews for further information in this area (e.g. Refs. ).

How to apply theory

While the Medical Research Council (MRC) among others recommends the basing of interventions on a theoretical model, it provides little information on what theories to use and how to apply them. Michie and colleagues have drawn together different theoretical models and focus on the application of theory in changing behaviour through a series of interacting frameworks: the Capability, Opportunity, Motivation – Behaviour (COM-B), the Theoretical Domains Framework (TDF) and the Behaviour Change Technique (BCT) Taxonomy . The COM-B components are designed to bring together factors necessary for Behaviour change (B in COM-B) to occur . Capability (C) is defined as a person’s physical and psychological capacity to engage in the behaviour, Opportunity (O) as factors external to the individual that make the behaviour possible (such as their environment), and Motivation (M) as brain processes that “energise and direct” behaviour, including habits and emotional responses. The TDF consists of 14 domains, which form associations between 36 theories of behaviour change and the constructs that are represented within them . Examples of domains include knowledge, skills, beliefs about capabilities, social influences, emotion and intentions . The BCT Taxonomy provides a list of 93 techniques that can be used to change behaviour such as information about outcomes , modelling, rehearsal, monitoring, feedback , and credible source . Michie et al. have linked the BCTs to each of the 14 domains in the TDF on the basis of available evidence or expert consensus. The BCTs can be used as guidance for coding of all elements of a behaviour change intervention, so they can be properly recorded and addressed.

These frameworks provide researchers with a path from a broad theoretical perspective on what might influence a specific outcome (COM-B) to specific behaviour change components that can be included in an intervention (TDF) and then specific ways in which that intervention components can target the behaviour to improve the outcome of interest (BCT). This requires postulating hypothesised causal pathways and provides a guide for the development of behaviour change interventions. When designing an intervention, the domains most likely to be effective in changing behaviour were selected from the TDF. These domains can then be linked to specific BCTs to help achieve the behaviour required to affect the desired outcome. While these frameworks are designed for behaviour change interventions generally, they can be readily applied to self-management interventions, as the example in Fig. 2 illustrates.

Using a theory-informed stepped approach for developing a complex self-management intervention.

Self-management intervention content

Content of self-management interventions for persistent LBP is varied. Common content of the four aforementioned systematic reviews includes pain education, activity in relation to pain, and specific strategies or resources to use in flare-ups. However, some interventions also included strategies such as goal-setting and practical exercise or problem-solving sessions or included cognitive-behavioural strategies such as cognitive restructuring, pacing and relaxation. Variation in content is possibly due to differences in the definition of self-management, where definitions that include different aims may likely include different strategies. Additionally, other aspects of the intervention, such as frequency, duration, mode of delivery and provider, varied substantially between the studies included in the reviews. For example, most intervention sessions were delivered face-to-face, but others were either delivered online or through a self-help booklet. Some interventions involved a single lay-provider; others included interventions delivered by multiple health professionals from different disciplines. The modes of delivery included group sessions ( n = 12), individual sessions ( n = 5) or a combination of the two ( n = 3), with a wide range in frequency in terms of number of sessions (1–13 sessions) and the duration of the intervention (over 1–12 weeks). It is to be expected that the wide variations in intervention content, frequency, duration and delivery methods likely influence the variations in effectiveness on outcome.

An issue related to this is that of intervention fidelity or the degree to which the content of these interventions are implemented as intended , both in relation to the providers and participants. Fidelity of treatment delivery specifically refers to the degree that intervention components and delivery modes/methods were implemented as intended (e.g. all components such as pain education, coping skills, and strategies to increase self-efficacy for physical activity were actually delivered by the treatment provider), whereas fidelity of treatment receipt refers to the participants receiving the intervention content and being able to understand and apply it . By evaluating the fidelity of an intervention, results may be interpreted with respect to how well it was delivered and received, increasing confidence that the results were indeed due to the intervention . An important precursor to fidelity assessment is the use of strategies to enhance or improve fidelity, such as producing intervention manuals or protocols for providers to improve fidelity of delivery . However, fidelity of self-management interventions for persistent LBP has been poorly enhanced, assessed and reported to date, potentially contributing towards explaining the variability in effectiveness outcomes (e.g. Refs. ). We refer readers to Borrelli’s framework as a guide to understanding the different components of fidelity within behaviour change research and to recent studies by Toomey et al. that provide an example of how to improve and assess fidelity within a self-management intervention for people with persistent LBP and osteoarthritis.

The choice of outcome

Following on from specifying a theoretical basis for a self-management intervention is the need to choose a relevant outcome. There are two broad categories of outcomes that are important to studies of self-management interventions; general clinical outcomes (e.g. pain, disability, health status) and behaviours specific to the aims of the intervention (e.g. physical activity, learning new coping/problem solving strategies, relaxation techniques). Many studies of self-management for persistent LBP include outcomes in the categories of clinical improvements (pain, disability) and, to a lesser degree, health status improvements (global improvement, satisfaction, quality of life). However, measurement of learning (of knowledge and self-management skills) or behaviour does not appear to be common practice within self-management interventions to date . Assessing both categories of outcomes would appear to be beneficial as clinical outcomes can tell us if the intervention worked, while the behaviour outcomes can tell us how it worked . We suggest that future self-management interventions report a logic model as presented in Fig. 2 to illustrate the proposed process of how the intervention should change behaviour and clinical outcomes.

Reporting of self-management interventions

Poor reporting is a common problem, particularly in complex interventions that include multiple components and techniques . For example, in the four aforementioned reviews, under-reporting of intervention content and procedural information was evident in most studies of self-management for persistent LBP. This is problematic as without adequate intervention descriptions, replication is not possible. Guidance such as the Template for Intervention Description and Replication (TIDieR) , which includes 12 categories for reporting including name, why (rationale), what (materials, procedures), who (providers), how (modes of delivery), where (delivery location), when and how much (dose, duration), tailoring, modifications, and intervention fidelity, may help to enhance reporting within these interventions. The specific intervention content can be reported using the BCT Taxonomy to help with transparency and replication. This process for reporting has been used by several behaviour change interventions and, more recently, by two study protocols describing trials of self-management interventions for persistent musculoskeletal pain including LBP, SOLAS and COPERS .