Screening for geriatric problems in the emergency department: Reliability and validity

Objective: To determine the test-retest reliability and concurrent criterion validity of a self-report ED screening questionnaire for adverse outcomes in elders. Methods: A cohort of 1,885 patients aged greater than or equal to 65 years were recruited from the EDs of 4 Montreal hospitals. Patients were excluded if they could not be interviewed because of their clinical status or cognitive impairment and no informant was available. The screening questionnaire, administered in the ED, contained 27 items on social, physical, and mental risk factors, medical history, and use of hospital services, medications, and alcohol. A random sample of 404 patients were invited to participate in a clinical assessment 1-3 weeks after the ED visit, that included re-administration of the screening questionnaire, and standardized instruments to assess disability, social resources, depression, alcohol use and abuse, and current medications. Results: Study data were collected from 221 patients (54.7%), of whom 193 were included in the test-retest reliability analyses and 213 in the analyses of concurrent validity. The concordance correlation coefficient for test-retest reliability of the risk factor score was 0.78 (95% confidence interval: 0.71, 0.83; n = 193). Several screening questions showed moderately good agreement with the appropriate criterion standard, particularly those on visual and hearing impairment, depression, and use of medications. The best subset of 9 screening questions explained approximately half of the variance in the total disability score. Conclusions: The screening questionnaire score has good test-retest reliability, but individual screening questions have, at best, modest concurrent validity. The final set of screening questions should be selected based on their predictive validity.