Predictors of frequent emergency department use among patients with psychiatric illness

Abstract

Objective

To identify the patient characteristics associated with frequent emergency department (ED) use and develop a tool to predict risk for returning in the next month.

Method

Prospective cohort study of 863 adults with psychiatric illness presenting to one of four general hospital EDs. ED visits and relevant clinical information in the year before and one month after the index visit were abstracted.

Results

One hundred sixty-seven of the patients (19%) were considered frequent users. Characteristics associated with frequent user status were homelessness, cocaine-positive toxicology screen, Medicare insurance, a personality disorder and hepatobiliary disease (all P< .05). Patients scoring in the highest risk category had nearly five times the odds of returning to the ED in the month subsequent to the index visit.

Conclusions

Psychiatric patients with frequent ED use are a heterogeneous group, but there are specific target conditions which, if confirmed, may facilitate reduced ED use and be replaced by more appropriate treatment.

Introduction

Overcrowding within our Nation’s emergency departments (EDs) is a major public health concern, as crowded EDs [and associated long lengths of stay (LOS)] lead to high degree of patient and provider stress, greater risk for adverse events, high costs and lower levels of patient satisfaction [1], [2]. These issues appear to be even more significant for patients with mental health conditions, with LOS commonly over eight hours, and not infrequently longer than a full day [3].

All three components of ED patient flow (input, throughput and output) have been implicated in the long ED LOS experienced by psychiatric patients [4]. Greater demand for ED services (input) coupled with a reduction in the number of available inpatient psychiatric beds (output) have resulted in a high congestion queue with slow outflow (throughput) [5], [6], [7], [8]. Patient characteristics such as homelessness, public insurance, behavioral loss of control requiring restraints or sitters, and recent substance use have been associated with prolonged ED stays [1], [2].

Some patients account for a disproportionate volume of ED activity by making frequent visits and therefore may contribute to the phenomenon of ED overcrowding [9]. The definition of frequent ED usage is varied; a number of studies have used a cut-off of four or more ED visits within a 12-month period [10], [11], [12], [13]. Several factors have been identified as being predictive of being a frequent ED user, including low socioeconomic status, public insurance and poor physical health. Mental health and substance abuse conditions have also been recognized as key risk factors for frequent use of psychiatric emergency services [14], [15]. Despite a fairly robust literature, most prior work in this area has focused on a single ED, or utilization of psychiatric emergency services, or relied on large administrative databases, making specific application to broader networks of care more difficult [11], [14], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25]. In addition, relatively few studies have attempted to use the identified risk factors to prospectively predict future general hospital ED use.

The purpose of this study was to identify risk factors associated with frequent ED use in a sample of 863 patients with psychiatric conditions seen across four general hospitals in the Greater Boston area by merging administratively available data on ED utilization with detailed chart review. These data were used to create a simple predictive tool for future ED use which was then tested on a subsequent one-month sample from these hospitals.