Sensory impairment, loneliness, and the emergence of physical–psychological–cognitive multimorbidity: a cross-national cohort study in China, the United States, and England

The co-occurrence of physical diseases, psychological disorders, and cognitive impairment—here termed physical–psychological–cognitive multimorbidity (PPC-M)—poses growing challenges to healthy aging but its upstream, modifiable determinants remain underexplored. Sensory impairments are common in later life and may be associated with the emergence of PPC-M, partly through psychosocial factors such as loneliness. This study aimed to examine the associations of vision, hearing, and dual sensory impairments with incident PPC-M and to assess whether loneliness partly explains these associations in three national ageing cohorts. We analyzed 15,022 adults aged ≥ 45 years from three nationally representative aging cohorts—CHARLS (China), HRS (United States), and ELSA (England)—followed for 6–8 years. Vision impairment (VI), hearing impairment (HI), and dual sensory impairment (DSI) were self-reported at baseline. Incident PPC-M patterns included: physical–psychological (P1P2), physical–cognitive (P1C), psychological–cognitive (P2C), and triple comorbidity (P1P2C). We fitted Cox proportional hazards models and multi-state survival models, adjusting for demographic factors, educational level, marital status, and health behaviours. Loneliness was assessed using harmonized measures and tested as a mediator within a counterfactual mediation framework. Sensitivity analyses included propensity score matching, competing risk of death, and alternative definitions of exposures and outcomes. At baseline, 52.8% reported no sensory impairment, 10.9% HI only, 17.2% VI only, and 19.1% DSI. A graded association emerged across PPC-M outcomes. Compared with no impairment, DSI showed the strongest associations with P1P2 (HR = 2.18), P2C (HR = 2.36), and P1P2C (HR = 1.56; 95% CI: 1.19–2.03), with VI generally showing stronger associations than HI. Results were directionally consistent across countries and robust in sensitivity analyses. Loneliness statistically accounted for approximately 10–12% of the observed associations across impairment types, suggesting a potential psychosocial explanatory pathway. Multi-state models suggested accelerated transitions from single-domain conditions to combined multimorbidity among those with DSI. Sensory impairments—especially dual impairment—were independently associated with the onset and progression of PPC-M. Loneliness accounted for a meaningful share of these associations, suggesting a potential target for risk stratification and intervention. Integrating accessible and affordable vision/hearing services with loneliness-reduction strategies may help identify older adults at elevated risk of complex multimorbidity and inform equitable healthy-aging strategies across diverse settings.