Measuring patient-reported outcomes in tuberculosis

• Main Author: Guo, Na
• Language: English
• Published: University of British Columbia 2010
• Online Access: http://hdl.handle.net/2429/28757

Objectives: The primary objectives of this thesis were to: (1) measure health-related quality of life (HRQL) and health state utility values (HSUVs) among patients with active tuberculosis (TB) disease and latent tuberculosis infection (LTBI); (2) investigate the relationship betwee HRQL and adverse drug reactions (ADR)among active TB patients; (3) quantify patients' preferences for LTBI preventive treatment. Methods: Two groups of patients were administered questionnaires: (1) Short-Form 36 (SF-36), Health Utility Index (HUI) and a Visual Analog Scale (VAS) were administered to 119 LTBI and 114 active TB patients at baseline and 3 months of their treatment. (2) A discrete choice experiment (DCE) survey was developed and administered among 214 LTBI patients. Conditional logit and latent class analysis were conducted to quantify respondents' preferences toward six treatment attributes (i.e. treatment length, clinic visit frequency, and risk of developing active TB, liver damage, skin rash and fatigue). Results: The baseline SF-36, HUI-2, HUI-3, Short-Form 6D (SF-6D) and VAS scores from active TB patients were significantly lower than those from LTBI patients. Major ADRs were shown to have significant impacts on active TB patients' HRQL and patients with lower baseline SF-36 scores were more likely to develop ADRs during the treatment. The three health utility instruments (HUI-2, HUI-3, and SF-6D) displayed acceptable construct validity when applying among TB population. However, they did not generate identical HSUV scores for the same individual. The DCE study results showed that all six attributes significantly influenced respondents' treatment decision and preference estimates were reasonable and consistent with our hypotheses. Substantial preference heterogeneity was observed among respondents. Latent class analysis assigned respondents into three groups and five socio-demographic factors significantly predicted the class assignment (i.e. origin of birth, education, employment, had children or not, and use of over-the counter medications). Conclusions: Active TB disease and the treatment associated ADR have substantial impacts on patients' HRQL. HRQL measurements might have the potential to predict patients' treatment outcomes. The DCE technique provides a useful tool of understanding patients' preferences surrounding health care products. this work demonstrates the value and importance of incorporating patient-reported outcome measurements into clinical research and practice.