Impact of care management on the highest utilizers of Camden NJ's hospitals

Background: The highest utilizers of hospital based care are typically patients who have complex medical conditions compounded by an array of social issues. Programs that improve the outpatient management of these complex patients are likely to reduce hospital utilization. In this study we examined the effect of a citywide care management (CM) project on reducing the probability of experiencing a subsequent emergency department (ED) visit or hospitalization, compared to a matched control group.

Methods: We conducted a retrospective cohort study of 33 patients who met the following CM project enrollment criteria: 1. five or more ED visits during a one year period and 2. a Camden resident. CM patients were recruited from 11/1/2007-4/30/2008 and followed until 6/30/2008. CM patients were matched into a citywide database of all area hospital visits to determine both utilization rates before and after project enrollment. Based on each individual CM patient's age, gender, and baseline utilization, three matched control patients were selected from the database to form a comparison group of 99 patients. Time to event analysis was performed using multivariable Cox regression. The two events of interest were defined as: 1. a subsequent ED visit and 2. a subsequent hospital admission. Censoring took place in the event of death, loss to follow-up, and at study completion on 6/30/2008. The study period was up to 224 days from program enrollment.

Results: 83 patients (62.9%) experienced a returning emergency visit and 84 patients (63.6%) experienced a returning hospital admission. Multivariable Cox regression found that CM project enrollment reduces the risk of ED utilization by 66.7% (hazard ratio 0.343 [CI, 0.175 to 0.67]; P = 0.002). The mean time to next ED visit was found to be 76.9 days longer for CM patients compared to controls. CM project enrollment was also associated with a trend toward lower risk of hospital admission, however this association failed to reach statistical significance (hazard ratio 0.587 [CI, 0.292 to 1.180]; P = 0.135).

Conclusion: In the early evaluation of this intervention, participation in the CM project was associated with a significantly lower risk of a subsequent ED visit. This suggests that providing primary medical care and social support, over a relatively short period of time, is effective in decreasing ED visits. The inability of the CM project to impact hospital admissions is likely due to this initial analysis' short intervention period.