A quiet prompt in the doctor's computer modestly cut falls among older adults

There was no new drug and no wearable sensor. The eSTEPS trial tested a prompt in the electronic health record that sat on the doctor's screen without interrupting the visit. It made it easier to order strength and balance exercises tailored to the patient, refer them to physical therapy and hand over patient materials.

The analysis included 22 primary care practices in the Mass General Brigham system, assigned in matched pairs, and 11,945 adults aged 65 and older who had screened positive for fall risk. Follow-up ran 12 to 23 months. After statistical adjustment, the rate of patients with at least one recorded fall was 19.3 per 100 patient-years in practices with the prompt, compared with 21.2 in the other practices: about 9% lower. Fall injuries were also slightly lower, and death rates didn't differ.

The odd part is how few people the prompt reached. Only 22.1% of patients in the prompt practices were given an exercise prescription, and nobody knows how many did the exercises.

The counting is imperfect too. Falls were tallied from diagnosis codes, which miss many falls. The prompt practices differed from the others at the start, and before statistical adjustment only self-reported falls showed a significant drop. Everything happened in one health system using one record platform. The open question is what a change this small could do if far more than one in five patients actually got the prescription.