BACKGROUND: Tuberculosis (TB) is the leading cause of death among HIV-positive patients. We assessed the cost-effectiveness of including lateral-flow urine lipoarabinomannan (LF-LAM) in TB diagnostic algorithms for severely ill or immunosuppressed HIV-positive patients with symptoms of TB in Kenya.
METHODS: From a decision-analysis tree, ten diagnostic algorithms were elaborated and compared. All algorithms included clinical exam. The costs of each algorithm were calculated using a 'micro-costing' method. The efficacy was estimated through a prospective study that included severely ill or immunosuppressed (CD4
RESULTS: The algorithm that added LF-LAM alone to the clinical exam lead to the least average cost per TB case detected (€47) and was the most cost-effective with a cost/DALY averted of €4.6. The algorithms including LF-LAM, microscopy and X-ray, and LF-LAM and Xpert in sputum, detected a high number of TB cases with a cost/DALY averted of €6.1 for each of them. In the comparisons of the algorithms two by two, using LF-LAM instead of microscopy (clinic&LAM vs clinicµscopy) and using LF-LAM along with GeneXpert in sputum instead of GeneXpert in urine along with GeneXpert in sputum, (clinic&LAM&Xpert_sputum vs clinic&Xpert_sputum&Xpert_urine) led to the highest increase in the cost-effectiveness ratios (ICERs): €-7.2 and €-12.6 respectively. In these two comparisons, using LF-LAM increased the number of TB patients detected while reducing costs. Adding LF-LAM to smear microscopy alone or to smear microscopy and Xray led to the highest increase in the additional number of TB cases detected (31 and 25 respectively) with an incremental efficiency estimated at 134 and 344 DALYs respectively. The ICERs were €22.0 and €8.6 respectively.
CONCLUSION: Including LF-LAM in TB diagnostic algorithms is cost-effective for severely ill or immunosuppressed HIV-positive patients.