To quantify patient preferences when making decisions as to whether to accept latent tuberculosis infection (LTBI) preventive treatment, using a discrete choice experiment (DCE).
A DCE survey was developed and administered to LTBI patients. Each patient was given 10 random choices along with two fixed choices to check consistency. Two hypothetical treatment options and one opt-out option were presented in each choice task. Latent class analysis was conducted to estimate preferences for six key treatment attributes.
Among the 214 respondents, 194 (90.7%) who provided valid DCE responses and complete sociodemographic information were included. Results consistently suggested that respondents were averse to higher risk of active tuberculosis and side effects and longer treatment. A three-latent-class model with five covariates was chosen. Forty-seven percent of the respondents were assigned to class 1, 32% to class 2, and 21% to class 3. Although all six attributes were shown to significantly influence the respondents' treatment decision, the risk of active tuberculosis, chance of liver damage, and frequency of clinic visits were the most important ones. Significant preference heterogeneity was observed in two attributes: frequency of clinic visits (P 0.01). Class 1 individuals were most likely to have children. Class 2 had the highest employment rate. Class 3 respondents tended to choose the opt-out option on DCE tasks and were more likely to be born outside Canada, have higher education, and be unemployed.
Respondents consistently preferred preventive treatment with higher effectiveness, fewer side effects, and shorter length. Substantial preference heterogeneity existed among respondents.