Objective: To model the effect of health systems performance on rates of mother-to-child HIV transmission.
Methods: We modeled the effect of variation in performance of the multiple steps of different prevention of mother-to-child transmission (PMTCT) protocols using hypothetical and reported data.
Setting: Data from a PMTCT program in a large province in South Africa was used to compare model predictions with reported outcomes for mother-to-child HIV transmission.
Main Outcome Measure: Perinatal HIV transmission was predicted for infants of 6 weeks of age.
Results: HIV-infected pregnant women who fulfill eligibility criteria are initiated on lifelong antiretroviral treatment, whereas noneligible HIV-infected women and their infants receive single-dose nevirapine in a health system functioning at reported performance levels, and the overall vertical transmission rate would be 19.5%. Adding azidothymidine for women not eligible for lifelong treatment would further decrease the overall transmission rates only marginally to 17%. If the same steps were accomplished at 95% reliability, then the overall transmission rates would be 9.4% and 4.1%, respectively.
Conclusions: Introduction of more effective combination antiretroviral interventions will yield only marginal reductions in childhood HIV infections and mortality unless health systems achieve high levels of performance at each step of the PMTCT pathway. Investment in and support for the mechanisms of delivering and sustaining PMTCT interventions at scale are required if gains in maternal and child survival are to be realized in countries highly affected by HIV.
From the *Institute for Healthcare Improvement, Cambridge, MA; †Department of Pediatrics, University of North Carolina, Chapel Hill, NC; ‡Department of Child Health, University of KwaZulu-Natal, Durban, South Africa; and §Department of Adolescent Health and Development, World Health Organization, Geneva, Switzerland.
Received for publication August 15, 2010; accepted September 21, 2010.
All authors have contributed to the development of the models and writing of the manuscript.
The data for the models were obtained through literature review. Data for the South African health system performance was obtained from the South African District Health Information System.
The model was presented at a WHO internal technical consultation in October 2009.
N. Rollins is a staff member of the World Health Organization.
The authors alone are responsible for the views expressed in this publication and they do not necessarily represent the decisions, policy, or views of the World Health Organization.
The authors have no funding to disclose.
Correspondence to: Nigel Rollins, MB, MD, Department of Child and Adolescent Health and Development, World Health Organization, Avenue Appia 20, 1211 Geneva, Switzerland (e-mail: firstname.lastname@example.org).
Antiretroviral (ARV) drugs used for the prevention of mother-to-child transmission (PMTCT) of HIV can virtually eliminate the risk of childhood HIV infection and improve maternal survival.1 Yet failure of any of the multiple sequential steps of PMTCT care results in cumulative losses of pregnant mothers from PMTCT services, with increased risk of HIV transmission to their infants. Among infants attending immunization clinics in South Africa, 21% of HIV-exposed infants were HIV infected by 6 weeks, despite single-dose nevirapine (sdNVP) being offered through antenatal clinics (ANC).2 The transitioning of PMTCT protocols in high HIV prevalence countries, from sdNVP to combination ARV drug regimens3 and “optional” eligibility criteria for initiating maternal antiretroviral treatment (ART),4 offer the prospect that HIV transmission rates and maternal mortality will decrease. However, these new interventions are likely to have minimal effect unless primary health care systems that deliver PMTCT services in these countries are improved.
The PMTCT “cascade” identifies the sequence of steps needed to deliver ARV interventions to HIV-infected women and their infants: counseling, HIV testing, CD4 testing, dispensing of ARVs at ANC and labor wards, and testing of infants at 6 weeks. However, the output of this sequence, prevention of HIV transmission, is significantly affected by the reliability of the health system to deliver each step in the sequence.5
If the entire multistep PMTCT care pathway is reduced to just 3 steps (ANC access, HIV/CD4 testing, and ARV treatment), the profound effect of even small losses in reliability is evident. If each step is delivered at 95% reliability, then a 5% loss that is compounded at each subsequent step means that only 86% of women will receive the ARV intervention. If the system delivers each step at 80% reliability (20% loss at each step), then only 51% of women will receive the final intervention and at 60% reliability, only 22% will receive ARVs. In real life, the multiple occasions when mothers are “lost” result in failure of interventions that under controlled research conditions, effectively decrease HIV transmission.6,7 It is estimated that in 2008, PMTCT programs in eastern and southern Africa provided ARVs to only 58% of HIV-infected pregnant women (45% in all low- and middle-income countries) who needed them.8
We developed a model that estimated the effect of ARV interventions for PMTCT on population-based vertical transmission rates in infants, taking into account health systems' performance. The proportion of women accessing each point of service delivery could be independently varied. These steps reflect activities that are largely within the control of health systems and are points of potential failure or improvement.
The model was restricted to antenatal and peripartum interventions, and the primary outcome was defined as HIV status of infants born to HIV-infected mothers at 6 weeks of age. It is assumed that the HIV status of all infants is determined at 6 weeks, whether they were part of the PMTCT program or not. The model did not include postnatal transmission due to breast feeding, primary infection of pregnant women, or situations when a pregnant woman might not test positive, such as during the “window period” of a newly acquired HIV infection.
The model assumed vertical transmission rates, measured in infants at 6 weeks, associated with different (or no) ARV interventions as described in published literature9,10 and could adjust transmission rates depending on maternal CD4 count. At each step of the PMTCT process, women who “fell out” of the ideal treatment stream were entered into default streams. For example, if a service offered ART to women meeting either immunological (CD4 count) or clinical criteria, then women who missed this opportunity because of impaired system performance could be designated to receive either nothing or a default intervention such as sdNVP or azidothymidine (AZT)/sdNVP.
We explored 6 scenarios that reflect actual PMTCT practices in different settings:
1. No intervention to HIV-infected women or HIV-exposed infants;
2. HIV-infected women receive sdNVP in labor. HIV-exposed infants receive sdNVP after delivery;
3. HIV-infected women receive AZT any time from 28 weeks of gestation and sdNVP in labor. HIV-exposed infants receive sdNVP and AZT for 1 week after delivery;
4. A 2-tier system in which HIV-infected women who do not fulfill criteria for lifelong ART and their infants receive sdNVP as in setting #2, whereas HIV-infected women who do fulfill criteria are initiated on ART by 34 weeks of gestation. Eligibility for ART was initially set at CD4 count <200 cells per milliliter;
5. A 2-tier system in which HIV-infected women who do not fulfill criteria for lifelong ART and their infants receive AZT and sdNVP as in setting #3, whereas HIV-infected women who do fulfill criteria (CD4 count <200 cells/mL) are initiated on lifelong ART by 34 weeks of gestation;
6. A 2-tier system in which HIV-infected women who do not fulfill criteria for lifelong ART receive triple ARV prophylaxis (eg, AZT/3TC/Kaletra) for the duration of pregnancy, whereas HIV-infected women who do fulfill criteria (CD4 count <200 cells/mL) are initiated on lifelong ART by 34 weeks of gestation.
These intervention scenarios were tested in a model that used reported performance data from South Africa (SA)11,12 for each of the 3 core steps (Fig. 1). The model assumes that CD4 testing is available at all sites. Infants born to HIV-infected women who had not been tested or treated were considered to be at unimpeded risk of transmission. We also ran the model using CD4 count <350 cells per milliliter as the eligibility criteria for initiating ART.
When PMTCT interventions are delivered in a health system functioning at reported SA performance levels for each step, then only one third of mothers and their infants would receive sdNVP or ART interventions (scenario 4) that should reduce HIV vertical transmission rate to about 8% (Fig. 1). The other two thirds of mothers would be lost along the PMTCT pathway, and their infants would be subjected to an unimpeded HIV transmission rate of 25%. When combined, the overall vertical transmission rate would be 19.5% (Surveillance data from SA in 2006 estimated population-based transmission rates to be 20.8%.2]. If the same steps were accomplished at 95% reliability, then the overall transmission rate would be 9.4%. Using the reported SA levels of system efficiency, introduction of AZT to a PMTCT protocol in addition to sdNVP/ART (scenario 5) would further decrease the overall transmission rates only marginally, that is, from 19.5% to 17%. Improving reliability of the system at all steps to 95% efficiency would result in a fall in the transmission rate to 4.1%.
In models of system-wide performance levels that ranged between 60% and 95% (Table 1), a high-performing system (95% reliability at all steps) that provides only sdNVP to all HIV-infected women, irrespective of CD4 count, would achieve population-based transmission rate similar to a low-performing (60%) system that provided ART to all HIV-infected women. In systems that function below 60% efficiency, there is limited impact of any ARV intervention on infant HIV transmission rates when measured at population level.
When performance was varied at different steps of the pathway, failure in the early steps of the PMTCT pathway such as poor access to ANC services or failure of counseling services has a greater impact on overall transmission rates than later losses in efficiency such as at initiating ARVs. In this model, if all steps in the system were delivered with 100% efficiency except HIV testing, which was delivered at 80%, the HIV transmission rate would be 8.2%. By contrast, if all steps were at 100% efficiency except delivery of ARVs (80% efficiency), then the overall transmission rate would be 5.7%. The effect is most significant if women with low CD4 counts are not identified and referred for ART, as infants born to these mothers are at much higher risk of transmission than those with higher (ie, CD4 >200 cells/mL) counts.
Finally, when ART is provided at the reported current performance levels from SA in the 2-tier models, changing the eligibility criteria for ART from CD4 <200 to CD4 <350 has little effect on the number of infants becoming infected (18.2% vs. 15.3%, respectively), although would benefit the additional mothers who would receive ART.
The principal finding is that to reduce the number of infants who become infected with HIV and ensure that mothers receive interventions that will save their lives, each step of the PMTCT pathway needs to be delivered with greater than 90% reliability. Equally important to recognize is that introducing more effective drug interventions, or raising the CD4 threshold at which ART would be offered, will have limited effect unless health system performance is addressed. Nor is it good enough to focus on 1 step of the pathway or only those women who come through the door of the antenatal clinic. Unless the overwhelming majority of pregnant women in the community are reached, the goal of minimal infant transmissions (eg, to less than 5%)13 cannot be reached.
As stated in the Methods section, neither the effects of breastfeeding on later postnatal transmission or acute maternal HIV infection are included in the analyses. The World Health Organization now recommends the provision of ARVs to prevent postnatal transmission, which has the potential to significantly improve HIV-free survival of infants born to HIV-infected mothers.14 However, this opportunity also depends on consistent delivery of specific interventions over an even longer period than antenatal care, that is, approximately 12 months and emphasizes the need for health systems' interventions to improve the reliability and quality of care provided through primary health facilities.
The model helps explain global estimates that only about 45% of HIV-infected women living in low- and middle-income countries ever receive PMTCT interventions.8 With poor reliability for the 3 main steps of the PMTCT pathway, it is hard to envision how new therapies could have any impact on HIV transmission rates or maternal mortality. As PMTCT interventions are directed at HIV-positive women, increasing the number of women who know their status will have the greatest impact on the number of infant infections that can be averted. This will require different strategies as, in some settings, access to any ANC service will be the dominant constraint, while elsewhere, it will be that counseling services are failing to offer testing to all women. Interventions to improve the quality of services within the health system need to be matched by initiatives to increase the demand for services. Simplification of protocols and methods to monitor and improve local performance will help.
Although the magnitude of these tasks seems overwhelming, especially knowing the physical and human resource constraints that prevail in many countries, quality improvement methods15 and other innovations such as task shifting16 can yield significant benefits relatively quickly. However, unless strategies to improve PMTCT delivery are energetically and resolutely pursued, then ARVs will remain in cupboards. Such interventions would not just improve PMTCT services but could strengthen essential maternal and child health services and related health information systems.
The 2008 World Health Report17 calls for reforms so that health systems contribute to health equity, social justice, and the end of exclusion; reforms that acknowledge the complexity of contemporary health systems as they strive to secure healthier communities. This call could not be more timely or appropriate when considering the urgent need to reduce HIV infections and mortality in infants and mothers in southern Africa. Reform means changing the way things are done-not simply exhorting these countries to “try harder” using current strategies. If mothers and infants in the countries most heavily afflicted by the HIV epidemic are to gain the benefit of interventions that have almost eradicated HIV transmission in the North, then immediate and exceptional efforts are needed to harness methods that could improve the quality and reliability of services delivered.
1. Paintsil E, Andiman WA. Update on successes and challenges regarding mother-to-child transmission of HIV. Curr Opin Pediatr
2. Rollins N, Little K, Mzolo S, et al. Surveillance of mother-to-child transmission prevention programmes at immunization clinics: the case for universal screening. AIDS
3. WHO. Towards Universal Access: Scaling up Priority HIV/AIDS Interventions in the Health Sector
. Geneva, Switzerland:/UNAIDS/UNICEF; 2008.
4. WHO. Antiretroviral drugs for treating pregnant women and preventing HIV infection in infants in resource-limited settings. Towards universal access. Recommendations for a public health approach. 2006. Available at: http://www.who.int/hiv/pub/guidelines/en/index.html
. Accessed November 10, 2010.
5. Nkonki LL, Doherty TM, Hill Z, et al. Missed opportunities for participation in prevention of mother to child transmission programmes: Simplicity of nevirapine does not necessarily lead to optimal uptake, a qualitative study. AIDS Res Ther
7. Manzi M, Zachariah R, Teck R, et al. High acceptability of voluntary counselling and HIV-testing but unacceptable loss to follow up in a prevention of mother-to-child HIV transmission programme in rural Malawi: scaling-up requires a different way of acting. Trop Med Int Health
8. WHO. Towards Universal Access. Scaling up Priority HIV/AIDS Interventions in the Health Sector. Progress report
. Geneva, Switzerland: WHO/UNAIDS/UNICEF; 2009.
9. Stover J, Johnson P, Zaba B, et al. The Spectrum projection package: improvements in estimating mortality, ART needs, PMTCT impact and uncertainty bounds. Sex Transm Inf
10. Halima D, Mofenson LM, Ekpini R, et al. International recommendations on antiretroviral drugs for treatment of HIV-infected women and prevention of mother-to-child HIV transmission in resource-limited settings: 2006 update. Am J Obstet Gynecol
. 2007;197(3 suppl): S42-S55.
12. Day C, Barron P, Monticelli F, eds. The District Health Barometer 2007/08
. Durban: Health Systems Trust; 2009.
13. South African National AIDS Council HIV & AIDS and STI National Strategic Plan for South Africa 2007-2011
. Pretoria, South Africa: SANAC; 2007.
15. Barker PM, McCannon CJ, Mehta N, et al. Strategies for the scale-up of antiretroviral therapy in South Africa through health system optimization. J Infect Dis
. 2007;196(Suppl 3):S457-S463.
16. Zachariah R, Ford N, Philips M, et al. Task shifting in HIV/AIDS: opportunities, challenges and proposed actions for sub-Saharan Africa. Trans R Soc Trop Med Hyg