Multimorbidity networks and functional impairment among older adults with and without complex needs: a network-based comparative study

Multimorbidity and resultant functional impairment represent critical global challenges, particularly among populations with complex needs (CN). However, the underlying disease network architecture and the specific pathways linking distinct conditions to functional impairment remain poorly understood. This study aims to disentangle pathological disease patterns from normative aging by comparing age-specific multimorbidity networks and their probabilistic dependencies on functional impairment between populations with CN and non-CN populations. This cross-sectional study utilized data from the Long-Term Care Insurance of Shanghai, China, from January to June 2023. Participants were stratified into CN ( n = 42,499) and non-CN ( n = 135,308) older adults. Statistically significant ( P -values adjusted by Bonferroni correction) and common comorbidity patterns (the pattern with prevalence > 1/10,000 in each dataset) were included, and network analysis was employed to construct undirected multimorbidity networks and compare feature differences in these networks for CN and non-CN older adults, respectively. Additionally, Bayesian network analysis was employed to explore the probabilistic dependencies and pathways between multimorbidity and functional impairment. Finally, age-stratified analyses were conducted to examine these network characteristics and probabilistic relationships across different age subgroups. CN older adults exhibited a multimorbidity network complexity of 2.14, compared with 1.72 in non-CN older adults, with hub-associated connections encompassing the vast majority of chronic conditions. However, the Bayesian network analysis revealed a lower network density in CN older adults (3.89) than in non-CN older adults (4.40). Compared with non-CN older adults, CN older adults exhibited a distinctive profile of hub diseases, characterized by digestive system disorders, Parkinson’s disease, and malignant neoplasms. The most pronounced differences occurred in the 60–74 age group, where cognitive impairment, osteoporosis, and Parkinson’s disease were uniquely observed in CN older adults. The 75–89 age group exhibited greater homogeneity, with thyroid disease, coronary heart disease, and osteoporosis being common to both populations. Furthermore, the primary diseases exhibiting direct probabilistic dependencies with functional impairment were largely consistent across both populations, and neurological diseases played a predominant role. Parkinson’s disease was uniquely linked to functional impairment in non-CN older adults. CN and non-CN older adults exhibit heterogeneity in multimorbidity networks and the probabilistic dependency structures underlying functional impairment. A dissociation exists between the hub diseases within the multimorbidity network and the specific diseases exhibiting direct probabilistic dependencies with functional impairment. Interventions should manage hub diseases while prioritizing the distinctive diseases of CN older adults, particularly among those aged 60–74, that exhibit direct probabilistic associations with functional impairment, thereby minimizing the functional impairment burden in older adults.