Objective: To explore the factors perceived to affect rehabilitation assessment and referral practices for patients with stroke. Design: Qualitative study using data from focus groups analysed thematically and then mapped to the Theoretical Domains Framework. Setting: Eight acute stroke units in two states of Australia. Subjects: Health professionals working in acute stroke units. Interventions: Health professionals at all sites had participated in interventions to improve rehabilitation assessment and referral practices, which included provision of copies of an evidence-based decision-making rehabilitation Assessment Tool and pathway. Results: Eight focus groups were conducted (32 total participants). Reported rehabilitation assessment and referral practices varied markedly between units. Continence and mood were not routinely assessed (4 units), and people with stroke symptoms were not consistently referred to rehabilitation (4 units). Key factors influencing practice were identified and included whether health professionals perceived that use of the Assessment Tool would improve rehabilitation assessment practices (theoretical domain 'social and professional role'); beliefs about outcomes from changing practice such as increased equity for patients or conversely that changing rehabilitation referral patterns would not affect access to rehabilitation ('belief about consequences'); the influence of the unit's relationships with other groups including rehabilitation teams ('social influences' domain) and understanding within the acute stroke unit team of the purpose of changing assessment practices ('knowledge' domain). Conclusion: This study has identified that health professionals' perceived roles, beliefs about consequences from changing practice and relationships with rehabilitation service providers were perceived to influence rehabilitation assessment and referral practices on Australian acute stroke units.
- acute stroke unit
- assessment for rehabilitation
- qualitative study
- Theoretical Domains Framework