The observer effect in a hospital setting - Experiences from the observed and the observers

Observation studies are used in health care research, e.g. to explore behaviors of patients or health care professionals in hospitals. A methodological challenge in observation studies is the observer effect, as it can jeopardize the quality of a study. To capture different dimensions of the observer effect through health care professionals' and patients' experiences, and their reactions to being observed in a hospital setting, and in addition, observers' experiences from performing an observation study. Four focus group interviews (health care professionals and observers) and 10 individual interviews (patients) were conducted with participants from a Norwegian observation study focusing on medication communication at a hospital ward. In all 26 persons were interviewed, whereof 3 were observers (pharmacist, pharmacy students). Data were collected between September 2019 and January 2020 and analyzed by an inductive, thematic analysis approach. Five main themes were identified; Experiencing being observed; Temporarily adapting medication communication behavior; Consequences for the patients; To interact or not - reflections on the relations and Observing the observers. Respondents reported some observer effects, but also that these diminished with time. Even though minimal interaction was used as a strategy, observers and the observed still built rapport. The observer effect in relation to medication communication seemed to be small and temporary in this specific hospital setting, among other things as staff and patients were used to extra persons (e.g. students) being around. Medication communication in hospital settings is a complex behavior, and appears to not be strongly impacted by the presence of observers, especially with a long observation time. It is important for researchers to monitor and record the observer effect in the specific setting of the study. This can be done by interviews with the observed and the observers by someone not connected to the observation study.