Table 1.
Key model parameters to assess the cost-effectiveness of different strategies for the prevention, diagnosis, and treatment of AHD among people living with HIV in Malawi
| Base case value | |||
|---|---|---|---|
| Sex at birth4, 17 | |||
| Female | 63% | ||
| Male | 37% | ||
| Age, years4, 17 | 33·2 (9·3) | ||
| Initial CD4 count, cells per μL4* | 330 (110) | ||
| CD4 count <200 cells per μL with or without WHO stage 3 or 4 disease | 90 (50) | ||
| CD4 count ≥200 cells per μL with WHO stage 3 or 4 disease | 250 (50) | ||
| No AHD (CD4 count ≥200 cells per μL and no WHO stage 3 or 4 disease) | 400 (130) | ||
| Cohort distribution | |||
| CD4 count <200 cells per μL with WHO stage 3 or 4 disease4, 18 | 12·4% | ||
| CD4 count <200 cells per μL without WHO stage 3 or 4 disease4, 18 | 8·6% | ||
| WHO stage 3 or 4 disease with CD4 count ≥200 cells per μL | 4·0% | ||
| No AHD (CD4 count ≥200 cells per μL and no WHO stage 3 or 4 disease) | 75·0% | ||
| HIV care continuum† | |||
| Virological suppression at 6 months from antiretroviral therapy initiation | |||
| Integrase strand transfer inhibitor-based regimen19, 20 | 92% | ||
| Protease inhibitor-based regimen21 | 73% | ||
| Loss to follow-up over 12 months22 | 8·5% | ||
| Return to care after 12 months of being lost to follow-up, monthly23 | 1·3% | ||
| Return to care upon developing symptoms of new opportunistic infection‡ | 50% | ||
| Tuberculosis† | |||
| CD4-stratified tuberculosis prevalence§ | |||
| Active tuberculosis disease24 | 10–37 | ||
| Latent tuberculosis disease24 | 20–47 | ||
| Tuberculosis symptoms25, 26§¶ | 51–87% | ||
| Monthly active tuberculosis-related mortality among those untreated7 | 7% | ||
| CD4-stratified test diagnostic yield among outpatients with symptoms27§‖ | |||
| Xpert | 68–70% | ||
| Xpert plus LAM | 72–85% | ||
| Test specificity27, 28 | |||
| Xpert | 98% | ||
| Xpert plus LAM | 95% | ||
| Probability of receiving empiric tuberculosis treatment24§ | 7–30% | ||
| RHZE efficacy for drug-susceptible-tuberculosis treatment29 | 98% | ||
| Tuberculosis preventive therapy efficacy for preventing tuberculosis disease30, 31** | 43% | ||
| Cryptococcal infection† | |||
| Cryptococcal disease prevalence32 | |||
| CD4 count <100 cells per μL | 7% | ||
| CD4 count 100–200 cells per μL | 2% | ||
| Monthly mortality from untreated cryptococcal meningitis33 | 78% | ||
| Cryptococcal antigen test sensitivity34 | 98% | ||
| Cryptococcal antigen test specificity34 | 98% | ||
| Fluconazole efficacy for preventing cryptococcal meningitis35, 36 | 72% | ||
| Other opportunistic infections† | |||
| Opportunistic infection incidence (stratified by CD4 count and antiretroviral therapy status), monthly§ | |||
| Severe malaria37 | 0·02% | ||
| Serious bacterial infections7, 38, 39, 40 | 0·04–3·68% | ||
| Other WHO stage 3 or 4 disease38 | 0·25–4·59% | ||
| Opportunistic infection mortality | |||
| Severe malaria41 | 28·1% | ||
| Serious bacterial infections7 | 30·0% | ||
| Other WHO stage 3 or 4 disease7 | 18·7% | ||
| Co-trimoxazole efficacy in preventing incident opportunistic infections | |||
| Severe malaria42, 43 | 88·4% | ||
| Serious bacterial infections42, 43 | 49·8% | ||
| Other WHO stage 3 or 4 disease44, 45 | 15·0% | ||
| Quality of life, utility† | |||
| Age-stratified and sex-stratified46§ | 0·860–0·910 | ||
| Tuberculosis47 | 0·620 | ||
| Acute opportunistic infection, 1 month | |||
| Severe malaria48 | 0·52 | ||
| Serious bacterial infections48 | 0·54 | ||
| Other WHO stage 3 or 4 disease48 | 0·50 | ||
| Cryptococcal meningitis, 1 month48 | 0·48 | ||
| Major drug toxicity, 1 month48 | 0·75 | ||
| AHD care continuum | |||
| Percentage of eligible people who have the test performed | |||
| Sputum Xpert†† | 79% | ||
| Urine LAM49 | 91% | ||
| Tuberculosis preventive therapy | NA | ||
| Cryptococcal antigen‡ | 75% | ||
| Co-trimoxazole | NA | ||
| Percentage of people with positive test results who initiate treatment | |||
| Sputum Xpert50 | 91% | ||
| Urine LAM50 | 91% | ||
| Tuberculosis preventive therapy†† | 79% | ||
| Cryptococcal antigen‡ | 90% | ||
| Co-trimoxazole‡ | 90% | ||
| Costs (2023 USD)† | |||
| HIV care | |||
| TDF–3TC + DTG, monthly51 | $4 | ||
| AZT–3TC + LPV–r, monthly51 | $19 | ||
| AHD care | |||
| CD4 count, per test52 | $6 | ||
| Sputum Xpert, per test53 | $16 | ||
| Urine LAM, per test54 | $6 | ||
| RHZE treatment, monthly54 | $12 | ||
| Tuberculosis preventive therapy, monthly54 | $1 | ||
| Cryptococcal antigen screening54 | $4 | ||
| Fluconazole pre-emptive therapy, monthly54 | $5 | ||
| Co-trimoxazole prophylaxis, monthly54 | $1 | ||
Data are % or mean (SD) unless otherwise stated. AHD=advanced HIV disease. AZT–3TC + LPV–r=zidovudine and lamivudine with lopinavir–ritonavir. LAM=lateral flow lipoarabinomannan. NA=not applicable. RHZE=rifampicin, isoniazid, pyrazinamide, ethambutol. TDF–3TC + DTG=tenofovir disoproxil fumarate and lamivudine with dolutegravir.
Values of initial CD4 count are square root transformed (appendix p 4).
See appendix for additional details (pp 6–9).
Assumption (ie, when no data are available to inform a parameter estimate, then an assumption is made and the estimate in sensitivity analysis is varied.)
Range shows input parameters that are stratified by CD4 count, age, and/or sex (appendix pp 18, 20, 23, 25 for additional details).
Based on the WHO-recommended four-symptom screen, comprising current cough, fever, night sweats, and weight loss.
Diagnostic yield, defined as the total proportion of tuberculosis cases identified by the tests; calculated by multiplying the sensitivity of the tests by the proportion of people living with HIV who could provide the diagnostic sample; data from Broger and colleagues.27
Tuberculosis preventive therapy prevents initial tuberculosis infection and the progression of latent tuberculosis infection to active tuberculosis disease with an efficacy of 43% over a period of 30 months; this effect lasts for 24 months after completing tuberculosis preventive therapy.
Maphosa T, unpublished data from the Evaluation of Advanced HIV Disease Differentiated Care Model in Malawi study.