Exploring drivers and challenges in implementation of health promotion in community mental health services: a qualitative multi-site case study using Normalization Process Theory
BACKGROUND: There is an increased interest in improving the physical health of people with mental illness. Little is known about implementing health promotion interventions in adult mental health organisations where many users also have physical health problems. The literature suggests that contextual factors are important for implementation in community settings. This study focused on the change process and analysed the implementation of a structural health promotion intervention in community mental health organisations in different contexts in Denmark. METHODS: The study was based on a qualitative multiple-case design and included two municipal and two regional provider organisations. Data were various written sources and 13 semi-structured interviews with 22 key managers and frontline staff. The analysis was organised around the four main constructs of Normalization Process Theory: Coherence, Cognitive Participation, Collective Action, and Reflexive Monitoring. RESULTS: Coherence: Most respondents found the intervention to be meaningful in that the intervention fitted well into existing goals, practices and treatment approaches. Cognitive Participation: Management engagement varied across providers and low engagement impeded implementation. Engaging all staff was a general problem although some of the initial resistance was apparently overcome. Collective Action: Daily enactment depended on staff being attentive and flexible enough to manage the complex needs and varying capacities of users. Reflexive Monitoring: During implementation, staff evaluations of the progress and impact of the intervention were mostly informal and ad hoc and staff used these to make on-going adjustments to activities. Overall, characteristics of context common to all providers (work force and user groups) seemed to be more important for implementation than differences in the external political-administrative context. CONCLUSIONS: In terms of research, future studies should adopt a more bottom-up, grounded description of context and pay closer attention to the interplay between different dimensions of implementation. In terms of practice, future interventions need to better facilitate the translation of the initial sense of general meaning into daily practice by active local management support that occurs throughout the implementation process and that systematically connects the intervention to existing practices.