Identifying High-Need Patient Profiles That Respond to Intensive Care Management: Insights from the Camden Health Care Hotspotting RCT

BackgroundEvaluations of complex care programs for high-need patients have yielded mixed results, and identifying patient subgroups may reveal differential intervention effects. This study aimed to use latent class analysis (LCA) to identify high-need patient subgroups within a randomized trial of the Camden Coalitions Core Model and to examine differences in healthcare utilization and care team engagement. Methods & FindingsWe conducted a post-hoc exploratory analysis of a randomized controlled trial (ClinicalTrials.gov: NCT02090426) involving 780 adults aged 18 [-] 80 years in Camden, New Jersey, who had multiple chronic conditions and frequent hospitalizations. Participants were assigned to receive multidisciplinary care management delivered by nurses, social workers, and community health workers for 3 [-] 4 months following hospital discharge, or to usual care. LCA incorporated medical, behavioral, and social risk factors, as well as prior hospital utilization, to identify patient subgroups. Outcomes included inpatient readmissions and emergency department visits over two consecutive 6-month post-discharge periods, along with service hours delivered to intervention patients. Four patient classes emerged: (1) Behavioral Health & Housing Instability, (2) Multi-system Medical Complexity, (3) Pulmonary Health & Substance Use, and (4) Lower Overall Complexity. In the second 6-month follow-up period, intervention patients had lower readmission rates compared with controls (-6.4 percentage points; 90% CI, -12.2 to -0.5). Subgroup differences included reduced readmissions in Class 4 and fewer emergency department visits in Class 1. Service intensity varied across classes, with Class 1 receiving the highest number of staff hours and Class 2 the lowest. ConclusionPatient segmentation revealed meaningful variation in healthcare utilization outcomes and care team engagement across high-need subgroups, suggesting that tailoring complex care interventions to specific patient profiles may improve program effectiveness and equity.