Assessing the impact of patient self‐selection on the costs to treat latent tuberculosis infection ( LTBI ) with isoniazid and transitional rifampin

Rationale, aims and objectives This article examines costs to treat latent tuberculosis infection ( LTBI ) in an urban clinic population and highlights the potential effectiveness of an alternative transitional treatment regimen. Methods Patients who experienced side effects on the gold standard of 9 months of isoniazid (9 INH ) were either continued on isoniazid or transitioned to 4 months of rifampin (4 RIF ). I use multilevel T obit models with selection to analyse whether transitioning to 4 RIF is less costly than remaining on 9 INH among patients experiencing side effects. Results Results reveal that self‐selection is present in this clinic data. Using an ordered probit parametric selection rule to account for selection, I find transitioning patients to 4 RIF costs significantly less than continuing on 9 INH . This result is especially sensitive to selection: not controlling for selection demonstrates that patients transitioning to 4 RIF actually cost significantly more to treat. Post hoc analysis revealed that switching to 4 RIF significantly reduced the dropout probability among these patients. Conclusion Future work should more carefully assess the clinical attributes, including effectiveness, of treatment with 9 INH and transitional 4 RIF as alternative treatment for LTBI .