The authors concluded that case management strategies seemed to be useful for improving clinical and social outcomes among frequent users of emergency departments. These tentative conclusions appear reliable but limitations in the review search and the evidence base should be kept in mind.

To evaluate the effectiveness of case management strategies in populations that use the emergency department at disproportionately high rates.

MEDLINE and EMBASE were searched for publications in English. It appeared that the databases were searched from 1990 to April 2010. Search terms were reported. References of retrieved articles were handsearched for further studies.

Eligible studies compared use of case management interventions with their non-use in adult populations (over 18 years old) deemed to be frequent users of hospital emergency department services. The primary outcome of interest was emergency department utilisation; secondary outcomes included in-patient hospitalisation rates and psychosocial outcomes (such as homelessness and history of substance abuse). Eligible study designs included prospective and retrospective studies, randomised and non-randomised controlled trials, case-control studies and pre- and post-intervention analysis using historical controls.

Where reported, the mean age of participants ranged from 38 to 48 years and 33% to 87% of the participants were men. Study populations varied and included homeless patients, employed and unemployed patients, insured and uninsured patients, patients with or without primary care doctors and patients with psychiatric or substance abuse disorders. Where reported, the definition of frequent use of emergency departments ranged from more than four visits a year to more than five visits a month. Care management interventions varied; more than half of the studies reported use of substance abuse counselling or referral services, assistance with financial entitlements and/or use of individualised care plans. Emergency department utilisation was the primary outcome in all but one of the included studies.

Two reviewers independently selected studies for inclusion.

Risk of bias and methodological limitations of the included studies were reportedly assessed according to criteria for randomisation, sample size, identification and selection of control groups, retrospective data collection, selection bias and follow-up.

Two reviewers independently assessed study quality; disagreements were resolved by discussion and involvement of a third party.

Data were extracted on patient characteristics, care management interventions and outcomes (emergency department utilisation, in-patient hospitalisation rates, psychosocial outcomes).

Two reviewers independently extracted these data.

Data were presented in a narrative synthesis with separate sections for different outcomes.

Twelve studies (960 patients) were included in the review: two randomised controlled trials (RCTs) (322 patients), eight pre-/post-intervention analyses (530 patients) and two studies with age-matched controls (108 patients). Methodological limitations included retrospective design (five studies), no randomisation (seven studies), short follow-up period (four studies), small sample size (three studies), limitations with data and analysis (two studies) and baseline differences between groups at baseline (one study). Length of follow-up ranged from two months to two years; half of the studies had 12 month follow-up periods.

Emergency department use (11 studies): Eight studies (one RCT, six pre-/post intervention analyses, one study with age-matched controls) reported reductions in emergency department use that ranged from 31% to 83%. Two studies (one RCT and one pre-/post intervention analysis) reported no significant reduction. One pre-/post intervention analysis reported a non-significant increase in use. In studies that reported reductions in emergency department use, intervention components included individual and group supportive therapy, financial entitlements, substance abuse referral, assistance with housing, extensive and persistent outreach, individualised care plans and referral to a primary care health professional.

Disposition (four studies): No significant differences were observed between intervention and control groups for hospital admission rates (one RCT and three pre-/post intervention analyses), medical in-patient days, psychiatric emergency visits, psychiatric in-patient admissions or psychiatric in-patient days (one RCT, one pre-/post intervention analysis). One pre-/post intervention analysis demonstrated a significant increase in rates of emergency department overnight observation in those exposed to care management intervention in hospital-based, community and primary health care settings.

Psychosocial variables (four studies): Care management interventions were associated with improvements in mean housing status score (one pre-/post intervention analysis) and reductions in rates of homelessness, lack of health insurance, lack of social security income and unmet financial needs (one RCT, two pre-/post intervention analyses). One RCT and one pre-/post intervention analysis each showed a significant reduction in rates of drug and alcohol use with use of case management but another pre-/post intervention analysis reported no change in rates.

Results for other outcomes were reported.

Greater reductions in emergency department cost were found among patients enrolled in case management interventions compared with those who were not enrolled; this finding was consistent across all four studies that reported cost outcomes. Further details of these findings were reported in the paper.

Case management strategies seemed to be useful for improving both clinical and social outcomes among frequent users of emergency departments.

The review question and inclusion criteria were clearly defined. Relevant databases were searched but conflicting dates for the search were reported in different parts of the paper and the restricted inclusion of papers published in English increased risks of language and publication biases. Efforts were taken to avoid reviewer error and bias at all stages of the review process. Quality assessment revealed some limitations in the studies but no formal assessment tool was used and this made it difficult to assess the true extent of bias in the individual studies.

Study details showed that the studies were diverse so the narrative method of synthesis was appropriate. The authors noted that the included studies had several limitations such as over-reliance on retrospective design or pre-/post intervention analyses that used historical controls, potential misinterpretation of positive intervention effects with the use of people serving as their own controls, small sample sizes, possible selection bias and differences in the degree of detail used to describe care management interventions.

The authors' tentative conclusions appear reliable but limitations in the review search and the evidence base should be kept in mind.

Practice: The authors did not state any implications for practice.

Research: The authors stated that future research should determine the most effective aspects of case management interventions. They stated that further research should be conducted to find out which interventions were most effective in challenging populations of high users such as people with psychiatric and substance abuse disorders and those in whom use of emergency departments was especially high.

Not stated.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.