How incremental video training did not guarantee implementation due to fluctuating population prevalence.

Patients with stroke admitted at the neurology/neurosurgery ward of the Academic Medical Centre in Amsterdam, The Netherlands, may experience problems in communication, such as aphasia, severe confusion/delirium or severe language barriers. This may prevent self-reported pain assessment; therefore, pain behaviour observation scales are needed. In this project, we therefore aimed to implement the Rotterdam Elderly Pain Observation Scale (REPOS) by video training. We used a stepped-wedge cluster design with clusters of four to five nurses with intervals of 2 weeks, for a total study duration of 34 weeks. Primary endpoint was the proportion of shifts in which nurses used the REPOS when caring for an eligible patient. A questionnaire was send biweekly to assess self-perceived competence and attitude on pain measurement in patients able or unable to self-report pain intensity. No other strategies were used to promote the use of the REPOS. Though the proportion of shifts in which trained nurses cared for eligible patients increased from 0% at baseline to 83% at the end of the study, the proportion of cumulative shifts where the REPOS was used decreased from 14% to 6%, respectively. Process evaluation suggests that this decrease can (in part) be attributed to low and varying prevalence of eligible patients and opportunities for practice. In total, 24 (45.3%) nurses had used the REPOS at least once after 34 weeks, with a median of two times (1-33). Nurses perceived themselves 'competent' to 'very competent' in pain behaviour observation. There was no negative attitude towards pain measurement. This study shows that education alone may not be effective when implementing a pain behaviour observation scale for non-communicative patients with Acquired Brain Injury. Individual motivation of health professionals and individual patient factors may be of influence for the use of the REPOS.