Concordance in the measurement of quality of life and health indicators between two methods of computer-assisted interviews: self-administered and by telephone

The aim of this study was to establish the concordance for quality of life (QOL), disability, and use of health service indicators between two modes of computer-assisted interviews: audio-computer-assisted self-interview (A-CASI) and computer-assisted telephone interview (CATI). High concordance between these modes of data collection would allow comparisons and interchangeable use in cross-sectional or longitudinal assessments. Adult English-speaking women (n = 126) were enrolled from women's health clinics at a public hospital. QOL using the short form 12 version 2, disability (days missed from work, inability to do household activities), and utilization of health services (number of emergency room visits and hospitalizations) were assessed first with A-CASI at the time of enrollment and again (n = 102) with CATI 1 week later. Participants assessed with both modes were 38 years old on average, predominantly African-American, 41% had a high school education or less, and 61% were uninsured. Lin's concordance correlation coefficient or Cohen's kappa was calculated to establish concordance between paired A-CASI and CATI assessments. Concordance between the two interview methods ranged from fair to substantial for the QOL components, (concordance correlation coefficient [CCC] of .76 and .87, respectively), the QOL subscales, and disability indices (CCC range; .53-.91). For health services utilization, there was moderate concordance for emergency room visits (CCC = .70) but only slight concordance for the number of hospitalizations in the past year (CCC = .37). Administering surveys through a telephone or self-administered computer-assisted interview resulted in moderate to substantial agreement for the short form QOL components and fair to substantial for the QOL subscales and disability measures. These findings suggest A-CASI and CATI can be used interchangeably for some QOL scales.