Program targeting superutilizers did not reduce readmission rates in randomized trial

The results suggest that approaches to care management designed to connect patients with existing resources may be insufficient for a medically and socially complex superutilizer population, study authors said.

An intervention targeting patients with very high use of health care services, or “superutilizers,” had no significant effect on readmission rates compared to usual care in a recent randomized controlled trial.

The superutilizer program, created by the Camden Coalition of Healthcare Providers, enrolled eligible patients while in the hospital. Once patients returned home, they worked with a team of nurses, social workers, community health workers, and health coaches in the months after discharge. The team conducted home visits, scheduled and accompanied patients to initial primary and specialty care visits, coordinated follow-up care and medication management, measured blood pressure and blood glucose levels, coached patients in disease-specific self-care, and helped patients apply for social services and appropriate behavioral health programs.

To test the program's effect on readmissions, researchers randomly assigned 800 hospitalized patients with medically and socially complex conditions and at least one additional hospitalization in the prior six months to receive either the program (n=399) or usual care (n=401). They collected hospital discharge data through March 31, 2018, from the four Camden, N.J., hospital systems. The primary outcome was hospital readmission within 180 days after discharge. Results were published Jan. 9 by the New England Journal of Medicine.

Overall, 782 (98%) patients had complete outcomes data and were included in the analysis sample. The trial population was 50% male, with about 17% who were younger than age 44 years, 55% who were between ages 45 and 64 years, and 28% who were age 65 years or older. About 55% were non-Hispanic black, 30% were Hispanic, and 15% were non-Hispanic white. Nearly all (94%) were not employed, and 44% received a diagnosis of substance use disorder during the index admission. Medicare and Medicaid were the primary payers for 48% and 45% of participants, respectively.

The 180-day readmission rate was 62.3% in the intervention group and 61.7% in the control group, for an adjusted between-group difference of 0.82 percentage point (95% CI, −5.97 to 7.61 percentage points; P=0.81). The intervention also had no significant effect on any of the secondary outcomes (number of readmissions, proportion of patients with at least two readmissions, hospital days, hospital charges, hospital payments received, mortality) or within any of the prespecified subgroups (number of admissions in the previous year, two or at least three; preferred language, English or other). In contrast with these results, a comparison of admission rates for the intervention group in the six months before and after enrollment misleadingly suggested a substantial decline in admissions in response to the intervention because it did not account for a similar decline in the control group.

The authors noted that the trial was not powered to detect smaller differences that could be clinically meaningful or to analyze effects within specific subgroups. In addition, usual care evolved during the trial period, as the Coalition was leading a citywide campaign to connect Medicaid patients with primary care within seven days after discharge, they said.

Approaches to care management that are designed to connect patients with existing resources may be insufficient for those with complex needs, the authors said. “The results suggest both the challenges of reducing readmissions in a medically and socially complex superutilizer population and the importance of conducting randomized evaluation of interventions such as this one, which, because they target high-cost patients, are likely to show substantial regression to the mean in observational studies,” they concluded.