Predictors of frequent visits to a psychiatric emergency room: a large-scale register study combined with a small-scale interview study.

BACKGROUND

The role of the psychiatric emergency services has undergone extensive changes following a significant downsizing of the number of psychiatric hospital beds during the past decades. A relatively small number of "frequent visitors" accounts for a disproportionately large amount of visits to psychiatric emergency services.

OBJECTIVES

To identify predictors of frequent use of a psychiatric emergency room at a Danish University Psychiatric Hospital through a 12-year period (1995-2007) and to speculate on how changes in the mental healthcare services affect predictors of frequent use through time.

DESIGN

A large-scale register based logistic regression analysis combined with a small-scale explorative, interpretative interview study. Register data were drawn from the Danish Central Psychiatric Research Register. Four-year cohorts (1995, 1998, 2001 and 2004) of patients with at least one visit to the psychiatric emergency room were followed for 3 years to identify general trends of predictors throughout the period. A purposeful sample of 15 frequent visitors were interviewed about their personal motives for visiting the psychiatric emergency room, their pathways to care, and their social network and social support.

RESULTS

The study identified two overall trends of predictors of frequent use of the psychiatric emergency room. High use of psychiatric services: ≥5 visits to the psychiatric emergency room, ≥3 admissions or ≥60 bed days during the year, was and continued to be predictive of high use of the psychiatric emergency room throughout the whole period. Furthermore, the emergence and continual presence of the predictors: severe mental illness (1999-onwards), substance abuse (2002-onwards) and sheltered housing (2002-2003-2005-onwards) indicated changes in the general profile of frequent visitors to the psychiatric emergency room, where predictors related to illness behaviour were supplemented by predictors related to disease.

CONCLUSION

The changing profile of the conditions predicting frequent visits at the psychiatric emergency room was most probably related to the adverse effects of the continuous deinstitutionalising of the Danish mental health services and a radical health care reform. A basic multiplicative model was designed for the early detection of individual frequent visitors.