Effectiveness of medication withdrawal in older fallers: Results from the improving medication prescribing to reduce risk of falls (IMPROveFALL) trial

Abstract

Objectives: to investigate the effect of withdrawal of fall-risk-increasing-drugs (FRIDs) versus 'care as usual' on reducing falls in community-dwelling older fallers. Design: randomised multicentre trial. Participants: six hundred and twelve older adults who visited an Emergency Department (ED) because of a fall. Interventions: withdrawal of FRIDs. Main Outcomes and Measures: primary outcome was time to the first self-reported fall. Secondary outcomes were time to the second self-reported fall and to falls requiring a general practitioner (GP)-consultation or ED-visit. Intention-to-treat (primary) and a per-protocol (secondary) analysis were conducted. The hazard ratios (HRs) for time-to-fall were calculated using a Cox-regression model. Differences in cumulative incidence of falls were analysed using Poisson regression. Results: during 12 months follow-up, 91 (34%) control and 115 (37%) intervention participants experienced a fall; 35% of all attempted interventions were unsuccessful, either due to recurrence of the initial indication for prescribing, additional medication for newly diagnosed conditions or non-compliance. Compared to baseline, the overall percentage of users of =3 FRIDs at 12 months did not change in either the intervention or the control group. Our intervention did not have a significant effect on time to first fall (HR 1.17; 95% confidence interval 0.89-1.54), time to second fall (1.19; 0.78-1.82), time to first fall-related GP-consultation (0.66; 0.42-1.06) or time to first fall-related ED-visit (0.85; 0.43-1.68). Conclusion: in this population of complex multimorbid patients visiting an ED because of a fall, our single intervention of FRIDs-withdrawal was not effective in reducing falls.