Fall history is an independent predictor of adverse health outcomes and utilization in the elderly

To determine whether a history of falls predicts functional decline, adverse health events, and hospitalization.

Twelve-month prospective cohort study.

Participants were members of a Medicare managed care program. Outcomes were assessed each quarter and included functional status, healthcare utilization (hospitalization), and adverse events (hospitalizations, nursing home placement, or death). Subject healthcare utilization diaries were corroborated with health system data files.

At baseline, 70% reported no falls (NF), 18% had 1 fall (F), and 12% reported 2 or more falls (RF). Fall status predicted functional decline; new ADL deficits were seen in 18% of NF, 28% of F, and 55% of RF (P < or = .0001). Following adjustment for baseline function, this association remained predictive (adjusted odds ratio [OR] for new ADL deficits: 3.5, P = .007; and for new ADL and IADL deficits: 12.0, P= .0001). Fall frequency was a univariate predictor of adverse events (hospitalizations, nursing home placement, or death) and of hospital utilization alone. One or more adverse event(s) occurred in 18% of NF, 22% of F, and 38% of RF (P = .049). Hospitalization occurred in 16% of NF, 22% in F, and 35% of RF (P = .03). Following adjustment for likelihood of future hospitalization (P(ra)), these associations remained predictive for RF (adjusted OR for one or more adverse event[s]: 2.4, P = .05; OR for hospitalization 2.4, P = .06).

Fall history predicts decline in function, hospitalization, and adverse events among a Medicare managed care population and remains independently predictive of poor outcomes after controlling for baseline function and likelihood of future hospitalization.