Cost-Effectiveness Analysis of Prevention of Mother-to-Child Transmission of HIV | Infectious diseases of poverty

study design

Since the perspective of the study was society, all PMTCT costs were included, regardless of who paid for them. Based on the results of the pilot survey, HIV PMTCT funding channels included Chinese central government, Sichuan provincial government, Liangshan prefectural government, county governments, MCHs, organizations non-governmental organizations (NGOs) and individuals. Expenditures were made at MCHs, CDCs, general hospitals, and per person. All costs were collected from the above institutions and individuals through two different types of questionnaires.

To collect HIV PMTCT costs and effectiveness data in 2017, the study was conducted from December 2018 to January 2019. Based on high, medium and low HIV prevalence rates, Annually reported prevalence of HIV-positive pregnant women in the 17 counties of Liangshan Prefecture was 0.5–3.2%, 0.1–0.4%, and 0.0–0.1%, respectively . For each level of prevalence, two counties were sampled based on purposeful sampling for the collection of cost data from MCHs, CDCs, and general hospitals.

To collect individual costs, HIV-positive pregnant women registered in the national HIV, syphilis and HBV PMTCT information system were recruited at the local clinic for a questionnaire survey. Individual costs included meals due to hospital visits, transportation due to hospital visits, infant formula, hospitalization and medical expenses during pregnancy and labour. The loss of working day income due to hospital visits was not included, given the low level of daily income of Liangshan residents. Total individual costs in Liangshan were calculated by multiplying the average individual survey cost by the sampling weight.

To estimate the sample size, the individual indirect costs were set as standard deviation (σ) at $74.1, the tolerated error (δ) was set at $14.8, and ({mu }_{alpha })((alpha =0.05)) was 1.96. The above parameter for sample size calculation was estimated based on the local consumption level. The sample size calculation formula was as follows:

$$n={left(frac{{u}_{alpha }sigma }{delta }right)}^{2},$$

the estimated minimum sample size of HIV-positive pregnant women (not) was obtained as 96.

And another method of sample size estimation has been cited, the sample size criterion for pharmacoeconomic evaluation. The criterion suggested that the minimum sample size of each group should not be less than 100 cases for a high quality pharmacoeconomic evaluation study [15].


The costs were divided into three parts according to their traceability: direct medical costs, direct non-medical costs and indirect costs. Direct medical costs covered the cost of drugs for ART and prophylaxis, HIV EID, HIV preliminary screening test reagents, HIV re-inspection test reagents, laboratory consumables, hospitalization and laboratory equipment. Direct non-medical costs covered the salaries of those providing intervention services; individual transport, individual meals, infant formula; and financial assistance for pregnant women. Indirect costs included staff training expenses, propaganda expenses, office expenses and office equipment expenses.

According to Chinese Government Accounting Standard No. 3 – Fixed Assets, the useful life has been set at 15 years for office furniture, 10 years for laboratory equipment, 6 years for office equipment and 5 years each for the micropipettes. The scrap rate was set at 0% for the micropipette and 5% for the other fixed assets [16]. The annual depreciation of fixed assets was determined using the straight-line method:

$$Annual, Depreciation = Cost , of , Asset* , left( {{1 }{-}Scrap , Value , Rate} right)/Useful , Lifetime.$$

Cost-effectiveness analysis

Efficiency analysis

In this study, the number of pregnant women tested for HIV (NOT1) and the number of HIV-positive pregnant women (NOT2) were analyzed. To calculate the number of pediatric infections averted by the HIV infection of their mothers (NOT3), NOT2 was multiplied by the difference between the rate (34.8%) without any intervention in China and the current rate (9.0%). The formula to calculate NOT3 was the following:

$$N_{3} = N_{2} *(34.8 – 9.0% ).,$$

Health utility analysis

The years of life (AV) gained by NOT3 were calculated as a measure of health utility using a mortality table. Here, 76.9 years – the life expectancy of people in Sichuan province in 2017 and the age-specific death rate of the sixth national census of China have been referenced [17]. The cost utility ratio (CUR) was used to measure the cost per LY gained, as follows:

where VS was the cost of PMTCT, and you was the number of AL gained by NOT3.

Cost-benefit analysis

The benefits of PMTCT were of two types: direct benefits and indirect benefits. The direct benefit was defined as the saving in the cost of ART for NOT3, and it was calculated based on a Markov model with the R-4.1.1 statistical software heemod package (WN Venables, DM Smith and R Core Team). To create the Markov model, it was assumed that the disease process of HIV/AIDS exhibits the following states of health: HIV state, AIDS state, and absorbing state of death. It was also assumed that the life expectancy of children infected with HIV from the mother was 25 years according to the literature. [18], and that the cycle time of the Markov model was one year. Costs after 2017 were discounted to 2017 at the rate of 3% per year (Fig. 1).

Fig. 1

Markov model of HIV/AIDS for ART. AIDS acquired immunodeficiency syndrome, ART antiretroviral therapy, HIV human immunodeficiency virus

The indirect benefit, defined as the economic value created by pediatric infections averted, was calculated by multiplying the estimated LYs by the age-weighted productivity of GDP per capita based on human capital theory. The age-weighted productivity of people aged 0-14, 15-44, 45-59, and over 60 is 0.15, 0.75, 0.8, and 0.1 , respectively [27].

The benefit-cost ratio (BCR), defined as the proportion of net output to inputs in economic terms, was taken as the total net benefit per cost of PMTCT. It was calculated as follows:

$$BCR=frac{sum b}{sum c}=frac{sum left(Direct Benefit+Non-Direct Benefitright)-PMTCT Cost}{PMTCT Cost}.$$

Sensitivity analysis for BCR

A one-sided sensitivity analysis was performed for the key variables to which the BCR was expected to be sensitive. It was assumed that PMTCT costs could vary by ± 25%. The range of transition probabilities and costs of ART in the Markov model was between the lower and upper levels in Table 1, while the life expectancy of pediatric infections averted ranged from 15 to 35 years and the rate of The discount varied between 0 and 10%. Life expectancy in Liangshan ranged from 60 to 85 years and its GDP per capita ranged from US$1,295.8 to US$8,393.9 (Table 2).

Table 1 Parameters of the HIV/AIDS Markov model for ART
Table 2 Cost-effectiveness of PMTCT of HIV in Liangshan Prefecture, Sichuan Province, China

Sara H. Byrd