Children with Medical Complexity: Latent Subgroups Across Hospitalizations at Tertiary Children's Hospitals in the US.

Abstract

To identify distinct subgroups of children with medical complexity (CMC) based on patterns of complex chronic conditions (CCCs) using latent class analysis and to examine differences in healthcare utilization across these subgroups. We conducted a retrospective cohort study from January 2022 to October 2025 for hospitalized encounters for CMC, identified by ≥1 CCC, from 49 pediatric tertiary hospitals. Latent class analysis was used to identify distinct patterns of CCCs among hospitalized children. The resulting 6 latent classes were treated as the categorical exposure. Outcomes included intensive care unit admission, invasive mechanical ventilation, medical/surgical complication, length of stay, and in-hospital mortality. Of 1 317 606 hospitalized encounters, latent class proportions were as follows: class 1, 127 069 (9.6%); class 2, 152 680 (11.6%); class 3, 347 525 (26.4%); class 4, 97 602 (7.4%); class 5, 68 644 (5.2%); and class 6, 524 086 (39.8%). Classes varied across type and number of CCCs. For adjusted outcomes, class 1 had the highest probability (34.90% [Q1, Q3, 32.66%, 37.21%]) of intensive care unit admissions. Class 3 had the highest probability of a medical or surgical complication (34.48% [Q1, Q3, 32.32%, 36.72%]). Class 6 had the highest probability of invasive mechanical ventilation (16.31% [Q1, Q3, 15.28%, 17.39%]) and in-hospital mortality (2.87% [Q1, Q3, 2.63%, 3.13%]), as well as the longest lengths of stay (13.60 days [Q1, Q3, 13.08, 14.14 days]). Meaningful subgroups of CMC exist among hospitalized encounters, with differential adverse healthcare utilization. The latent classes identified in this study offer a framework for characterizing heterogeneity among CMC and advancing health services research toward greater specificity. Such frameworks are essential for evaluating the efficacy of tailored interventions.