Objectives: The purpose of this study was to examine response rate, mode effects, and reliability of the Stroke Impact Scale (SIS) in a veteran stroke population using mail and telephone modes of administration.
Methods: Patients who had suffered a stroke were identified using national VA administrative data and International Classification of Diseases, 9th Revision codes in 13 participating Veterans Affairs hospital. Stroke was confirmed by reviewing electronic medical records. Patients were randomized to SIS administration by mail or telephone at 12-weeks after their stroke. Comparison of response rate, nonresponse bias, domain scores, administration costs, and instrument reliability were performed.
Results: Four hundred fifty-eight patients with stroke were identified, validated, and randomly assigned into 2 administration groups. No significant cluster effect was observed. Response rates for mail and telephone were 45% and 69%, respectively. Mail nonresponders were more likely to have had severe stokes, cognitive deficits, and be unmarried. No difference was observed between telephone responders and nonresponders. Responders in mail and telephone modes were not different, and the SIS score distribution did not indicate the presence of mode effects. Test-retest reliability was good to excellent in the mail group (0.77-0.99) except social participation (0.62). Test retest reliability was excellent in the telephone mode (0.90-0.99) except emotion (0.68).
Conclusions: Telephone mode of survey administration yielded a higher response rate, less bias in responder selection, and higher test-retest reliability. The cost of telephone administration was 2 times the cost of mail. Mode effects in SIS score distribution were not observed in this study but additional research with larger sample sizes is needed to provide more definitive evidence.