[Evidence-based medicine and the validation of quality-of-life questionnaires: the Spanish version of the MOS-HIV questionnaire for the evaluation of the quality of life in patients infected by HIV]

Abstract

Background: To validate the Spanish version of the MOS-HIV questionnaire for measuring quality of life in HIV-infected patients.

Methods: 275 HIV-infected patients distributed in three diagnostic categories (asymptomatic, symptomatic and AIDS) and 40 blood donors as healthy controls were recruited from 23 Spanish hospitals and answered the MOS-HIV. Feasibility was assessed by analysing the mean time of administration and the percentage of patients with missing response. Internal consistency was assessed using Cronbach's alpha. A subsample of 48 stable patients completed the questionnaire on two occasions to assess test-retest reliability using the Intraclass Correlation Coefficient. Validity was assessed by calculating the area under the Receiver's Operating Characteristics (ROC) curve between patients and controls, and by correlating MOS-HIV scores with the VAS of Euroqol, the number of symptoms, the CD4+ count and viral load.

Results: The mean time of administration was 16 minutes and 18.9% of patients did answer the MOS-HIV with some missing response. All the scales showed high internal consistency (Cronbach's alpha = 0.78-0.89). Test-retest reliability of Physical Health (PHS) and Mental Health (MHS) Summary Scores was 0.58 y 0.85, respectively. The MOS-HIV discriminated between patients and controls (ROC = 0.77, CI = 0.69-0.85 and ROC = 0.80, CI = 0.72-0.88, for the PHS and MHS, respectively). The correlation of the scales of the MOS-HIV with the CD4+ count and viral load was low (r = -0.26 to 0.3), and with the number of symptoms and the VAS of Euroqol were moderate (r = -0.28 to -0.49 and r = 0.4 to 0.66, respectively).

Conclusions: Considering evidence-based medicine principles, the Spanish version of the MOS-HIV has shown adequate feasibility, reliability and validity for its use in clinical research, but their use in clinical practice is limited because of their time of administration, the low reproducibility of some scales and the features of their scoring system.