Factors Associated with Patient Satisfaction of Community Mental Health Services: A Multilevel Approach

Abstract

Community care is increasingly the mainstay of mental healthcare provision in many countries and patient satisfaction is an important barometer of quality of patient care. This paper explores the key factors associated with patient satisfaction with community mental health services in England and then compares providers’ performance on patient satisfaction. Our analysis is based on patient-level responses from the community mental health survey, which is run annually by the Care Quality Commission (CQC) for the years 2010 to 2013. We perform a repeated cross-section analysis, identifying factors associated with patient satisfaction via a multi-level ordered probit model, including both patient- and provider-level variables. We identify hospital-specific effects via empirical Bayes estimation. Our analysis identifies a number of novel results. First, patient characteristics such as older age, being employed, and being able to work, are associated with higher satisfaction, while being female is associated with lower satisfaction. Service contact length, time since last visit, condition severity and admission to a mental health institution, are all associated with lower satisfaction. Second, treatment type affects satisfaction, with patients receiving talking therapies or being prescribed medications being more satisfied. Third, care continuity and involvement, as proxied by having a care plan, is associated with higher satisfaction. Fourth, seeing a health professional closer to the community improves satisfaction, with patients seeing a community-psychiatric nurse, a social worker or a mental-health support worker being more satisfied. Finally, our study identifies the need for service integration, with patients experiencing financial, accommodation, or physical health needs being less satisfied. At a provider level, we find a negative association between the percentage of occupied beds and satisfaction. We further identify significant provider-specific effects after accounting for observable differences in patient and provider characteristics which suggests significant differences in provider quality of care.