Field Validation of the Los Angeles Motor Scale as a Tool for Paramedic Assessment of Stroke Severity

Background and Purpose— The Los Angeles Motor Scale (LAMS) is a 3-item, 0- to 10-point motor stroke-deficit scale developed for prehospital use. We assessed the convergent, divergent, and predictive validity of the LAMS when performed by paramedics in the field at multiple sites in a large and diverse geographic region. Methods— We analyzed early assessment and outcome data prospectively gathered in the FAST-MAG trial (Field Administration of Stroke Therapy–Magnesium phase 3) among patients with acute cerebrovascular disease (cerebral ischemia and intracranial hemorrhage) within 2 hours of onset, transported by 315 ambulances to 60 receiving hospitals. Results— Among 1632 acute cerebrovascular disease patients (age 70±13 years, male 57.5%), time from onset to prehospital LAMS was median 30 minutes (interquartile range 20–50), onset to early postarrival (EPA) LAMS was 145 minutes (interquartile range 119–180), and onset to EPA National Institutes of Health Stroke Scale was 150 minutes (interquartile range 120–180). Between the prehospital and EPA assessments, LAMS scores were stable in 40.5%, improved in 37.6%, and worsened in 21.9%. In tests of convergent validity, against the EPA National Institutes of Health Stroke Scale, correlations werer =0.49 for the prehospital LAMS andr =0.89 for the EPA LAMS. Prehospital LAMS scores did diverge from the prehospital Glasgow Coma Scale,r =−0.22. Predictive accuracy (adjusted C statistics) for nondisabled 3-month outcome was as follows: prehospital LAMS, 0.76 (95% confidence interval 0.74–0.78); EPA LAMS, 0.85 (95% confidence interval 0.83–0.87); and EPA National Institutes of Health Stroke Scale, 0.87 (95% confidence interval 0.85–0.88).Conclusions— In this multicenter, prospective, prehospital study, the LAMS showed good to excellent convergent, divergent, and predictive validity, further establishing it as a validated instrument to characterize stroke severity in the field.