August 25, 2026

@michaelokun

Can we predict the first fall in Parkinson’s before it happens? A first fall can be a sentinel event in Parkinson’s disease, meaning an event that may signal an important change in disease trajectory, independence and future risk. Anis, Alberts and colleagues describe in a new paper in Movement Disorders Clinical Practice how routinely collected cognitive, motor and clinical data might help predict a fall before the first one occurs. Key points: - In 997 folks w/ Parkinson’s, an integrated LASSO model combining clinical, cognitive and motor information predicted future falls w/ an AUC of 0.81. - Among folks who had never fallen, processing speed, walking speed, disease duration and age emerged as important predictors, although prediction was more difficult and the AUC fell to 0.70. - The consequences of that first fall mattered: among 203 first-time fallers, nearly half required an emergency visit, more than half underwent imaging and 12.3% were hospitalized. My take: What resonates w/ me is that as the authors themselves suggest “we need to close the gate before the horse gets out of the barn.” For years, one of our strongest predictors of a future fall has been a previous fall. That is useful, but it is not prevention. This study shifts the question from who will fall again to who might fall for the first time, and importantly reminds us that cognition and processing speed belong in the falls conversation alongside gait and balance. Here are 5 points that resonated w/ me: 1- The first fall in Parkinson’s can change independence, confidence and quality of life, so prediction before that event should be our goal. 2- Falls are not simply a walking problem. Cognition, processing speed, age and disease duration matter too. 3- The Waiting Room of the Future approach is clever because brief assessments can be collected on an iPad before the visit and fed directly into the EHR. 4- An AUC of 0.70 in folks who had never fallen is promising, however it is not yet good enough to determine care by itself and the authors appropriately call for prospective multi-center validation. 5- The real game-changer: prediction triggers intervention.