The yield, reliability, and validity of a postal survey for screening community-dwelling older people

OBJECTIVES: To assess the yield, reliability, and validity of a postal survey developed to identify older persons in need of outpatient geriatric assessment and follow-up services. DESIGN: A longitudinal cohort study. SETTING: Outpatient primary care clinic at a Department of Veterans Affairs teaching ambulatory care center. PARTICIPANTS: Patients (N = 2,382) aged 65 and older who returned a Geriatric Postal Screening Survey (GPSS) that screened for common geriatric conditions (depression, cognitive impairment, urinary incontinence, falls, and functional status impairment). Validity and reliability testing was performed with subsamples of patients classified as high or lower risk based on responses to the GPSS. MEASUREMENTS: Test-retest reliability was measured by percentage agreement and kappa statistic. The diagnostic validity of the 10-item GPSS was tested by comparing single GPSS items to standardized geriatric assessment instruments for depression, mental status and functional status, as well as direct questions regarding falls, urinary incontinence, and use of medications. Validity was also tested against clinician evaluation of the specific geriatric conditions. Predictive validity was tested by comparing GPSS score with 1-year follow-up data on functional status, survival, and healthcare use. RESULTS: Respondents identified as high risk by the GPSS had scores that indicated significantly greater impairment on structured assessment instruments than those identified as lower risk by GPSS. The overall mean percentage agreement between the test and retest surveys was 88.3%, with a mean weighted kappa of 0.70. In comparison with a structured telephone interview and with a clinical assessment, individual items of the GPSS showed good accuracy (range 0.71-0.78) for identifying symptoms of depression, falls, and urinary incontinence. Over a 1-year follow-up period, the GPSS-identified high-risk group had significantly (P < .05) more hospital admissions, hospital days and nursing home admissions than the lower-risk group. CONCLUSION: A brief postal screening survey can successfully target patients for geriatric assessment services. In screening for symptoms of common geriatric conditions, the GPSS identified a subgroup of older outpatients with multiple geriatric syndromes who were at increased risk for hospital use and nursing home admission and who could potentially benefit from geriatric intervention.