We propose to assess the impact of mode of administration (self-administered versus interviewer administered) and language of administration (English versus Spanish) on response to four leading generic health-related quality of life (HRQOL) measures (SF-36, EQ-5D, Quality of Well-Being scale, Health Utilities Index).
The specific aims of this project are to: 1) evaluate the psychometric equivalence by mode and language of administration for the 4 generic HRQOL measures in terms of missing data, central tendency, variance, reliability, item difficulty, and item discrimination (slopes); 2) examine correlations between self and interviewer-administered responses to the same HRQOL measure, and between the same constructs measured by different measures; 3) predict self-administered HRQOL responses from interviewer-administered responses (and vice versa); 4) estimate the respondent burden of self-administration versus interviewer-administration for the different measures; and 5) estimate the English and Spanish language readability of the 4 generic HRQOL measures. This 48-month effort will involve two primary data collection activities. First, we propose to supplement the baseline data collection in the three centers (UCSD, UCLA, University of Wisconsin) of Project 2 (n =200 arthritis--hip replacement, n = 200 vision impairment--cataract extraction with lens replacement, n =200 heart failure--medication treatment) by conducting 600 telephone interviews. Second, we will collect data from 400 adult Hispanic patients (200 English-speaking and 200 Spanish-speaking) from two sites in Los Angeles: AltaMed Senior BuenaCare and Roybal Comprehensive Health Center. This project will provide essential information about the equivalence of SF-36, QWB, HUI, and EQ-5D data collected in different modes and language of administration. Results will enable researchers to decide whether it is appropriate to pool data by mode and language. In addition, the study will yield information that will allow estimates of responses in one mode of administration (e.g., self-administered) from the other (e.g., interviewer-administered). If the results show that the different modes of assessment are equivalent or at least predictable from one another, then this will allow greater flexibility in collecting data using different modes and mixed mode data collection. In addition, the study will provide important information about the readability of widely used generic HRQOL measures that can be useful in deciding the appropriateness of their use with lower-educated and less literate subgroups.
Showing the most recent 10 out of 52 publications
