Identifying Brief Message Content for Interventions Delivered Via Mobile Devices to Improve Medication Adherence in People with Type 2 Diabetes Mellitus: A Rapid Systematic Review

Background: Current interventions to support medication adherence in people with type 2 diabetes are generally resource-intensive and ineffective. Brief messages, such as those delivered via short message service (SMS) systems, are increasingly used in digital health interventions to support adherence because they can be delivered on a wide-scale and at low cost. The content of messages is a crucial intervention feature for promoting behaviour change, but it is often unclear what the rationale is for chosen wording or any underlying mechanisms targeted for behavioural change. There is little guidance for developing and optimising brief message content for use in mobile-device delivered interventions.

Objective: (1) To identify theoretical constructs (i.e. the targets that interventions aim to change) and behavioural strategies (i.e. features of intervention content) found to be associated with medication adherence in patients with type 2 diabetes. Additionally, (2) to map these onto a standard taxonomy for behaviour change techniques (BCTs), i.e. ‘active ingredients’ of interventions used to promote behavioural change, to produce an evidence-based set of approaches that have shown promise of improving adherence in previous studies and which could be further tested in digital-health interventions.

Methods: A rapid systematic review of existing relevant systematic reviews was conducted. Medline and PsycINFO databases were searched from inception to 10th April 2017. Inclusion criteria: (A) systematic reviews of quantitative data if the studies reviewed (i) identified predictors of or correlates with medication adherence and/or evaluated medication adherence-enhancing interventions and (ii) included adult participants taking medication to manage a chronic physical health condition; and (B) systematic reviews of qualitative studies of experiences of medication adherence for adult participants with type 2 diabetes. Data were extracted on review characteristics and BCTs, theoretical constructs or behavioural strategies associated with improved medication adherence. Constructs and strategies were mapped onto the BCT v1 taxonomy.

Results: A total of 1701 articles were identified; 25 systematic reviews (19 quantitative reviews, 3 qualitative reviews and 3 mixed-method reviews) were included. 20 theoretical constructs (e.g. self-efficacy) and 19 behavioural strategies (e.g. habit analysis) were identified in the included reviews. In total, 46 BCTs were identified as being related to medication adherence in type 2 diabetes (e.g. habit formation, prompts/cues, information about health consequences).

Conclusions: We have identified 46 promising BCTs related to medication adherence in type 2 diabetes, upon which the content of brief messages delivered through mobile devices to improve adherence could be based. By using explicit systematic review methods and linking our findings to a standardised taxonomy of BCTs, we have described a novel approach for the development of digital message content. Future brief message interventions that aim to support medication adherence could incorporate the BCTs identified in this review.