Psychometric properties of two measures of psychological well-being in adult growth hormone deficiency

Abstract

Background: Psychometric properties of two measures of psychological well-being were evaluated for adults with growth hormone deficiency (GHD): the General Well-being Index, (GWBI) - British version of the Psychological General Well-being Index, and the 12-item Well-being Questionnaire (W-BQ12). Methods: Reliability, structure and other aspects of validity were investigated in a cross-sectional study of 157 adults with treated or untreated GHD, and sensitivity to change in a randomised placebo-controlled study of three months' growth hormone (GH) withdrawal from 12 of 21 GH-treated adults. Results: Very high completion rates were evidence that both questionnaires were acceptable to respondents. Factor analyses did not indicate the existence of useful GWBI subscales, but confirmed the validity of calculating a GWBI Total score. However, very high internal consistency reliability (Cronbach's alpha = 0.96, N = 152), probably indicated some item redundancy in the 22-item GWBI. On the other hand, factor analyses confirmed the validity of the three W-BQ12 subscales of Negative Well-being, Energy, and Positive Well-being, each having excellent internal reliability (alphas of 0.86, 0.86 and 0.88, respectively, N from 152 to 154). There was no sign of item redundancy in the highly acceptable Cronbach's alpha of 0.93 (N = 148) for the whole W-BQ12 scale. Whilst neither questionnaire found significant differences between GH-treated and non-GH-treated patients, there were correlations (for GH-treated patients) with duration of GH treatment for GWBI Total (r = -0.36, p = 0.001, N = 85), W-BQ12 Total (r = 0.35, p = 0.001, N = 88) and for all W-BQ12 subscales: thus the longer the duration of GH treatment (ranging from 0.5 to 10 years), the better the well-being. Both questionnaires found that men had significantly better overall well-being than women. The W-BQ12 was more sensitive to change than the GWBI in the GH-Withdrawal study. A significant between-group difference in change in W-BQ12 Energy scores was found [t(18) = 3.25, p = 0.004, 2-tailed]: patients withdrawn from GH had reduced energy at end-point. The GWBI found no significant change. Conclusion: The W-BQ12 is recommended in preference to the GWBI to measure well-being in adult GHD: it is considerably shorter, has three useful subscales, and has greater sensitivity to change.