Measurement equivalence using a mixed-mode approach to administer health-related quality of life instruments

To evaluate the effects of mode, order of administration, and the interaction of mode and order on health-related quality of life scales when self-administered by mixed mode (paper-mode and web-mode) for measurement equivalence. Health-related quality of life data was analyzed from the Cancer of the Prostate Strategic Urologic Research Endeavor using the Medical Outcomes Study (MOS) Short Form-36 (SF-36) and the University of California Los Angeles Prostate Cancer Index (UCLA-PCI). A randomized crossover design assigned participants to two groups with a preferred 2-5-day washout period. Cognitive debriefing evaluated participants' mode preference. Of the 245 men enrolled, 85 % completed both modes. The majority were White (97 %), college educated (66 %), reported an annual income >$75,000 (46 %), and a median age of 69 years. Intraclass correlation coefficients were high for each item on both instruments (r = .54-.97). Exact percentage agreement for yes/no items was high (a parts per thousand yen.88). For the SF-36, significant differences were observed for order of administration (physical component and physical function scores) and for the interaction between mode and order (mental component, role emotional, social function, vitality, and mental health scores). For the UCLA-PCI, the largest difference was 12.8 points lower for sexual bother for order of administration by web-mode first (p = .03). Seventy percent preferred the web-mode, 21 % had no preference, and 9 % preferred the paper-mode. Web-mode and paper-mode administrations of the SF-36 and UCLA-PCI are equivalent in men with prostate cancer, implying that mixed-mode survey administration is warranted.