The STAMP trial showed urine-based tuberculosis (TB) screening in unselected people with HIV (PWH) hospitalized in Malawi and South Africa (SA) reduced 2m all-cause mortality by 2.8% and increased TB diagnoses by 7.3%. We examined the cost-effectiveness of this screening strategy, projecting outcomes at longer time horizons.
We used the CEPAC-International model to project clinical and economic outcomes of 2 TB screening strategies among hospitalized PWH: (1) Intervention: testing sputum with Xpert MTB/RIF and urine with Xpert and Determine TB-LAM; (2) Standard of Care: sputum Xpert alone. The modeled cohort matched the trial cohort (median CD4 219/µL [Malawi], 236/µL [SA]). Costs of Xpert/LAM were US$26/$3 in Malawi and $15/$3 in SA. We calibrated model output at 2m to STAMP trial outcomes and then projected longer-term outcomes including life expectancy (LE), costs, and incremental cost-effectiveness ratios (ICERs), discounted at 3%/y. We considered the Intervention cost-effective if its lifetime ICER was less than that of second-line antiretroviral therapy in each country: $740/year of life saved (YLS) in Malawi and $950/YLS in SA. Informed by the trial, in the base case true TB prevalence in Malawi/SA was 18%/28%, proportion of patients providing sputum was 39%/75%, and probability of empiric TB treatment was 4%/10%, all of which were varied in sensitivity analysis. We estimated the 5y clinical and budget impact of implementing the Intervention countrywide in Malawi and SA.
Model-generated absolute reductions in mortality by the Intervention in Malawi/SA were 3.5%/2.2% at 2m, and LE increased by ~0.5y (undiscounted) in both settings. The Intervention’s lifetime ICER was $490/YLS in Malawi and $850/YLS in SA (Table). The Intervention’s ICER was lower (more attractive) at higher TB prevalence, lower proportion of patients providing sputum, and lower empiric treatment rate. When we modeled a modified intervention, testing urine with only LAM was more cost-effective and possibly cost-saving. Implementing the Intervention countrywide over 5y among hospitalized PWH was associated with ~35,000 and ~171,000 YLS in Malawi and SA, with budget impact of $30 million in Malawi and $228 million in SA (6.6% and 2.3% increases in total health expenditures in these populations).
Urine-based TB screening for hospitalized PWH would increase life expectancy and is cost-effective in resource-limited settings.