Validity of a computer touch-screen questionnaire system in back patients.

Abstract

Validation study. A system for patient self-recording on a computer touch-screen was developed. The validity of this method compared with the use of regular paper-and-pen questionnaires was studied. Systematic evaluation of health problems often involves the use of self-reporting questionnaires in order to collect data on different variables. Recording of such data for subsequent analysis requires several steps including filling out of paper forms and secretarial work. Missing values and misrecordings are frequent problems. Seventy-nine patients with back problems visiting our outpatient clinic for surgical evaluation were invited to take part in the study by answering a mailed questionnaire concerning background data, pain, function, quality of life, and depressive symptoms approximately 3 weeks before their visit. At the day of the visit they were asked to again answer the same questions with a replicate response option format displayed on a computer touch-screen. The agreement concerning background history questions, measured by kappa values was generally good (0.71-1.0). For visual analogue scale recordings (0-100) of back and leg pain, the mean of the differences were 1.1 and 2.1, respectively, and the correlations (Pearson) 0.72 and 0.87. The Euro-Qol 5 Dimension Score, the General Function Score, the Zung Depression Scale and the physical dimensions of the Short Form-36 showed a high degree of agreement between paper and screen recordings with difference means close to 0 and reliabilities comparing well to published methodologic errors for paper form versions. There were, however, differences concerning the mental components of the SF-36 with somewhat higher ratings on screen recordings. Missing values were 0% for screen recordings and for paper forms, 2.3% on individual questions and 12% on composite scores. Computerized touch-screen questionnaires virtually eliminates missing values and show good validity and reliability compared to paper forms. The SF-36 tended to produce slightly higher values on touch-screen recordings in questions concerning mental health.