Retrospective Validation of a Protocol to Limit Unnecessary Transport of Assisted-living Residents Who Fall

Abstract Objective. Emergency medical services (EMS) often transports patients who suffer simple falls in assisted-living facilities (ALFs). An EMS "falls protocol" could avoid unnecessary transport for many of these patients, while ensuring that patients with time-sensitive conditions are transported. Our objective was to retrospectively validate an EMS protocol to assist decision making regarding the transport of ALF patients with simple falls. Methods. We conducted a retrospective cohort study of patients transported to the emergency department from July 2010 to June 2011 for a chief complaint of "fall" within a subset of ALFs served by a specific primary care group in our urban EMS system (population 900,000). The primary outcome, "time-sensitive intervention" (TSI), was met by patients who had wound repair or fracture, admission to the ICU, OR, or cardiac cath lab, death during hospitalization, or readmission within 48 hours. EMS and primary care physicians developed an EMS protocol, a priori and by consensus, to require transport for patients needing TSI. The protocol utilizes screening criteria, including history and exam findings, to recommend transport versus nontransport with close primary care follow-up. The EMS protocol was retrospectively applied to determine which patients required transport. Protocol performance was estimated using sensitivity, specificity, and negative predictive value (NPV). Results. Of 653 patients transported across 30 facilities, 644 had sufficient data. Of these, 197 (31%) met the primary outcome. Most patients who required TSI had fracture (73) or wound repair (92). The EMS protocol identified 190 patients requiring TSI, for a sensitivity of 96% (95% CI: 93-98%), specificity of 54% (95% CI: 50-59%), and NPV of 97% (95% CI: 94-99%). Of 7 patients with false negatives, 3 were readmitted (and redischarged) after another fall, 3 sustained hip fractures that were surgically repaired, and 1 had a lumbar compression fracture and was discharged. Conclusions. In this cohort, two-thirds of patients with falls in ALFs did not require TSI. An EMS protocol may have sufficient sensitivity to safely allow for nontransport of these patients with falls in ALFs. Prospective validation of the protocol is necessary to test this hypothesis.