Frailty Trajectories and Transitions in Older Adults: A Population-Based Approach Using Linked Administrative Data

Chiara Chiavenna , Bocconi University
Elisabetta Listorti, SDA Bocconi School of Management
Benedetta Pongiglione, UNIVERSITA Di PAVIA

Frailty is a multidimensional geriatric syndrome reflecting vulnerability from accumulated health deficits (Clegg et al., 2013). It is a major determinant of functional ability, service use, and quality of life, and its growing prevalence challenges the sustainability and equity of health and long-term care systems (Hoogendijk et al., 2019). Longitudinal evidence shows that frailty evolves through heterogeneous, non-linear trajectories, with transitions between robustness, vulnerability, and severe frailty (Gill et al., 2006; Bolano et al., 2019). Social disadvantage related to education, income, gender, and migration further amplifies these risks (Welstead et al., 2021). The Lancet Commission on Frailty (Dent et al., 2025) calls for improved measurement, early detection, and policy integration of frailty prevention. This study, part of the DIFF project, uses Lombardy’s integrated administrative data (>10 million residents) to model frailty trajectories and transitions among adults aged 65+. Individual records are linked across hospital, emergency, outpatient, pharmacy, exemption, home-care, and social-care databases through encrypted identifiers. A validated Frailty Index (FI) based on 40 indicators of morbidity, disability, and social vulnerability is computed annually following the cumulative-deficit model (Mitnitski et al., 2001; Rebora et al., 2025). We apply Group-Based Trajectory Modelling to identify subpopulations with distinct FI patterns and Multistate Modelling to estimate transition probabilities and durations between frailty states and death. Preliminary results on 92,000 older adults reveal substantial heterogeneity: most experience gradual FI increases, while ~13% show sharp worsening or temporary recovery. Linked administrative data enable dynamic surveillance and actionable evidence for equitable, preventive, and integrated care.

See extended abstract

 Presented in Session P5. Health, Mortality, and Ageing 1