Evaluation of two health status measures in adults with growth hormone deficiency.

OBJECTIVE To evaluate the psychometric properties of two health status measures for adults with growth hormone deficiency (GHD): Nottingham Health Profile (NHP) and Short-Form Health Survey (SF-36). DESIGN (1) A cross-sectional survey of adults with treated or untreated GHD to assess reliability and validity of the questionnaires. (2) A randomized, placebo-controlled study of 3 months' GH withdrawal from GH-treated adults to assess the sensitivity of the questionnaires to change. PATIENTS (1) A cross-sectional survey of 157 patients with severe GHD (peak GH < 10 mU/l on provocative testing), mean age 48.9 years (range 23-70 years), who had either received GH replacement therapy for at least 6 months immediately prior to the study or had not received GH treatment in the previous 6 months. (2) GH treatment was withdrawn from 12 of 21 GH-treated adults, all with severe GHD (peak GH < 7.7 mU/l on provocative testing), mean age 44.9 years (range 25-68 years). MEASUREMENTS The NHP and SF-36 were used once in the cross-sectional survey, but twice in the GH-withdrawal study, at baseline and end-point (after 3 months). RESULTS (1) Cross-sectional survey. Both questionnaires had high internal consistency reliability with subscale Cronbach's alphas of > 0.73 (NHP) and > 0.78 (SF-36). Calculation of an NHP Total Score, occasionally reported in the literature, was shown to be inadvisable. Overall, patients with GHD were found to have significantly worse perceived functioning than the UK general population in SF-36 subscales of General Health, Bodily Pain, Social Functioning, Physical Functioning, Role-Emotional, Role-Physical, and Vitality. Although neither questionnaire found significant differences between GH-treated and non-GH-treated patients, there were correlations with duration of GH treatment (P < 0.01) for GH-treated patients in SF-36 Mental Health (r = 0.29, N = 87) and SF-36 Vitality (r = 0.33, N = 88), indicating improvement with increasing treatment duration. The SF-36 was also more sensitive than the NHP to sex differences: men had significantly better health status compared with women (P < 0.05) in all SF-36 subscales but Mental Health, but only in one NHP subscale (Physical Mobility). (2) GH-withdrawal study. Significant between-group differences in change were found in SF-36 General Health [t(17) = 2.76, P = 0.013, two-tailed] and SF-36 Mental Health [t(17) = 2.41, P = 0.027, two-tailed]: patients withdrawn from GH reported reduced general health and mental health at end-point. The NHP found no significant change. CONCLUSIONS The SF-36 is a better measure than the NHP of health status of people with GH deficiency because of its greater discriminatory power, with ability to detect lesser degrees of disability. It also has superior sensitivity to some subgroup differences and superior sensitivity to change compared with the NHP. The SF-36 is highly acceptable to respondents, and has very good internal consistency reliability. The SF-36 is recommended to measure the health status of adults with GH deficiency.