Objective: To develop and retrospectively analyze an instrument that rapidly characterizes pretreatment stroke severity for use in prehospital acute stroke clinical trials.
Methods: The Los Angeles Motor Scale (LAMS) was constructed by assigning point values to Los Angeles Prehospital Stroke Screen (LAPSS) items of facial weakness, arm strength, and grip to yield a total 0-5 scale. The concurrent, discriminant, and predictive validities of the LAMS were probed using data from 90 patients enrolled in acute stroke clinical trials. Predictive performance of the LAMS was compared with that of the initial full National Institutes of Health Stroke Scale (NIHSS) and the five-item shortened NIHSS (sNIHSS) in projecting long-term outcomes on standard functional end points.
Results: LAMS score at entry averaged mean 2.6, median 2. Entry LAMS scores correlated closely with entry NIHSS scores (r=0.75). LAMS score correlations with three-month functional outcome measures were robust. Receiver operator curve analyses (c statistic) for performance in predicting three-month outcomes were: three-month modified Rankin-LAMS 0.75, sNIHSS 0.69, NIHSS 0.74; three-month Barthel Index-LAMS 0.77, sNIHSS 0.76, NIHSS 0.82; three-month NIHSS-LAMS 0.76, sNIHSS 0.62, NIHSS 0.70; and three-month GOS-LAMS 0.55, sNIHSS 0.67, NIHSS 0.76. Considering dichotomized three-month measures, entry LAMS scores were markedly lower in patients destined for excellent outcome, e.g., three-month modified Rankin score<or=1, mean entry LAMS score 2.2 versus 3.7, p=.0002.
Conclusions: A motor score derived from the LAPSS rapidly quantifies stroke severity in the field and predicts functional outcomes with accuracy comparable to that of the full NIHSS and the sNIHSS.