The need to address the broader reproductive health issues of people affected by HIV is gaining recognition by international organisations and at the country and programme level.1 One aspect of reproductive health—preventing unintended pregnancies—can have an important effect on the reduction in the number of HIV-positive births. In a previous publication we estimated that current levels of contraceptive use prevent over 173 000 unintended HIV-infected births each year in sub-Saharan Africa, or 474 HIV infected infants per day.2 Moreover, the number of additional HIV-positive births that could be prevented if the unmet contraceptive need of HIV-infected women was fully addressed numbered about 160 000. The desire to prevent unintended pregnancies among women who are HIV positive is particularly great. A recent longitudinal study found that, after learning their status, HIV-positive women are less likely to want children and more likely to use contraceptives.3

To further appreciate the role of contraception in preventing HIV-positive births, we examined the current and potential contribution of family planning programmes in each of the countries receiving funds from the President’s Emergency Plan for AIDS Relief (PEPFAR). Thus, the purpose of this paper is two fold: (1) to estimate the number of HIV-positive births that contraceptive use is already preventing in each of the PEPFAR focus countries; and (2) to estimate the cost savings to each PEPFAR country if unwanted and unintended HIV-positive births were prevented via contraceptive use rather than in antenatal care programmes that provide services, including antiretroviral prophylaxis, to HIV-positive pregnant women. Although the strategy to prevent mother-to-child transmission of HIV (PMTCT) is more comprehensive than antiretroviral prophylaxis for HIV-positive pregnant women attending antenatal care, we use the term “PMTCT services” in this article to refer only to those services in antenatal care.

METHODS

Unintended pregnancies and births averted by current contraceptive use

To estimate the number of HIV-positive births that contraceptive use is currently preventing in PEPFAR countries, we relied on existing data from the 15 countries receiving PEPFAR funds. The main sources included the Demographic and Health Survey (DHS), Joint United Nations Programme on HIV/AIDS (UNAIDS) and the United States Census Bureau International Data Base (see table 1 for sources, inputs and formulae).

Our first step estimated the number of pregnancies currently averted by contraceptive use. This number takes into account each country’s current contraceptive prevalence rate (CPR) and multiplies the CPR by the number of women aged 15–49 years.4^–8 We used the CPR for all women rather than for married women when available. We also used the most recent statistics available regardless of marital status. Finally, we applied a conservative initial annual pregnancy rate of 40%9 and subtracted the number of pregnancies that would have occurred due to contraceptive failure.

The contraceptive failure rate is estimated to be between 3.8% and 19% depending on the contraceptive method mix in each country and the failure rate for typical use of each method.4 10 Failure rates are specific for each country and can be derived from the formulae provided in table 1; however, the average failure rate is 10.2%. Failure rates are higher in countries that rely heavily on traditional methods such as Côte d’Ivoire and Rwanda, and lower in countries that rely on modern methods such as South Africa and Namibia.

The annual number of unintended pregnancies averted to HIV-infected women is the product of the number of unintended pregnancies averted by current contraceptive use and the adult (aged 15–49 years) prevalence of HIV.11^–15 In all countries but Vietnam the adult HIV prevalence is a conservative estimate since the female HIV prevalence is usually higher than the male HIV prevalence.11 12 To estimate the number of HIV-positive births currently averted by contraception per year, we assumed a 30% vertical transmission rate (in the absence of antiretroviral prophylaxis) and applied that proportion to the number of pregnancies averted to HIV-infected women.16

Potential cost savings of increasing contraceptive use

To estimate the cost savings to PEPFAR countries if they prevented HIV-positive births with contraception rather than PMTCT services, we started with estimates of unintended and unwanted births for the population of all women of reproductive age in each country (see table 2 for sources, inputs and formulae). Unintended births are composed of both those that were unwanted (ie, wanted no more children) and those that were mistimed (ie, pregnancies that were wanted later). We make this distinction because some prevented mistimed pregnancies may occur in the future when women no longer want to postpone pregnancy. Mistimed and unwanted births are determined as such by DHS based on respondents’ answers to a series of questions about children born to women in the last 5 years and any current pregnancy. Whether the birth was mistimed or unwanted is based on women’s reports of how they felt at the time they became pregnant.

Estimates of unintended and unwanted births are available from the DHS.4 17 18 Population-based estimates of unintended and unwanted pregnancies for women with HIV are not available; however, studies of infected women suggest that their levels of unintended pregnancies are at least as high as for all women regardless of HIV status (ranging from 51% to 91%).3 19 22

Our next step was to calculate the number of unintended and unwanted births to HIV-infected women. To do this we multiplied the population-based estimates of the unintended or unwanted birth rate by the number of annual births to HIV-infected women.23 To estimate those births that were HIV positive we multiplied the number of unintended and unwanted births to HIV-infected women by the vertical transmission rate of 30% in the absence of antiretroviral prophylaxis.16

To estimate the first year cost savings of preventing HIV-positive births via contraception compared with PMTCT services, we used the cost per HIV-positive birth averted for each intervention that was calculated in a previous study ($663 for the family planning intervention and $857 for PMTCT services).24 This study was conducted from the health system perspective during 1 year with a hypothetical sub-Saharan Africa population. The cost of the family planning programme included the cost of outreach visits to increase use among all women (including those who are HIV-infected and those who are not) and the cost of methods and services to all new users for 1 year. The cost of the PMTCT services included the cost of training providers, HIV counselling and testing for pregnant women, and the completed nevirapine dose for mother and infant. Other studies of PMTCT service costs which offer nevirapine in the Africa region have ranged from $298 to $9258 per HIV-positive birth averted.25^–28^. (Our estimate of $857 is closest to the most recent estimate of $847 per HIV infection averted in the Africa region by Hogan and colleagues in 200525).

The cost of family planning programmes to prevent unintended HIV-positive births is the product of the number of unintended or unwanted HIV-positive births and $663. The cost of PMTCT services is the product of the number of unintended or unwanted HIV-positive births and $857. The cost savings is the difference between the two.

RESULTS

Unintended pregnancies and births averted by current contraceptive use

The annual number of unintended pregnancies currently averted to HIV-infected women ranges from 595 in Guyana to over 400 000 in South Africa, while the number of unintended HIV-positive births currently averted ranges from 178 in Guyana to over 120 000 in South Africa (table 1). The number of births to HIV-infected women and, consequently, the number of HIV-positive births currently averted is positively correlated with the CPR, population size and HIV prevalence. As populations grow and contraceptive prevalence increases, the role that contraception is currently playing to prevent HIV-positive births increases. Despite a decrease in the prevalence of HIV in some countries,11 population growth and CPR increases are the main reasons why our previous estimate of 173 000 HIV-positive births already prevented by contraceptive use annually in sub-Saharan Africa is probably an underestimate.2

Potential cost savings of increasing contraceptive use

Above and beyond those HIV-positive births that are already averted by current contraceptive use, additional infant infections could be prevented if unintended pregnancies were reduced. The estimated annual number of unintended HIV-positive births ranges from 261 in Vietnam to 35 231 in South Africa (table 2). Unwanted HIV-positive births are much lower because they are a subset of unintended pregnancies. We estimate the minimum annual cost savings to prevent unwanted HIV-positive births with contraceptive use compared with PMTCT services to range from $26 000 in Vietnam to over $2.2 million in South Africa. The magnitude of the cost savings is positively associated with the proportion of unwanted pregnancies and with the number of women with HIV.

DISCUSSION

Our analyses show that contraception, despite relatively low CPRs in most PEPFAR countries, is already making a major contribution to preventing unintended HIV-positive births. However, thousands of unintended HIV-positive births still occur each year. Important cost savings in each PEPFAR country could be incurred if women delayed mistimed and prevented unwanted pregnancies with contraception rather than interrupting vertical HIV transmission with PMTCT services and antiretroviral prophylaxis.

Together, strong family planning and PMTCT programmes can maximise the prevention of HIV transmission. Not only does contraception allow women to prevent unwanted pregnancies, it can help women with HIV to delay pregnancies until they are emotionally and physically ready to have more children and until they can access appropriate antenatal care, antiretroviral regimens and safe delivery care. Moreover, knowledge of HIV status is not a prerequisite for contraceptive use, unlike for PMTCT programmes; thus, strengthened family planning programmes will benefit all women including those with HIV who do not know their status.

Because of the way we calculated the cost of the respective family planning and PMTCT services to avert HIV-positive births, family planning programmes will always result in cost savings compared with PMTCT services. A key assumption was that costs do not vary with programme size; in other words, there are constant returns to scale. Although these costs are also estimates and carry their own uncertainty as addressed in a previous study,24 the direction and magnitude are consistent with other studies.28 29 Moreover, these cost savings associated with family planning are minimal estimates since they do not include the costs of antenatal care, of antiretroviral treatment for HIV-positive infants or the social costs of children left as orphans. Also, the cost per HIV-positive birth averted by PMTCT services relies on the single-dose nevirapine regimen which is not the first-line triple antiretroviral therapy approach recommended by WHO.30 Although first-line PMTCT regimens are more efficacious, they will be more costly to implement.31 Finally, we have not quantified the benefits to women, their children and families of preventing unintended pregnancies which are widely known to be substantial.9 32

While family planning programmes can delay pregnancies that are wanted later, some women will eventually get pregnant which will result in PMTCT service costs. For example, one study in Egypt found that 41% of women using methods to space pregnancy gave birth within 3 years compared with 59% of women not using methods who wanted a pregnancy.33 There will be a cost savings in later years, but it will be reduced because we do not take into account later increases in births that occur when women no longer want to postpone childbearing. Thus, the cost savings estimated here are those incurred in the first year rather than the future annualised cost savings.

Not only do women’s desires related to pregnancy and childbearing change over the life course, but they can change within a single pregnancy. Some pregnancies that start out as unintended are reported as wanted births.34 This complex transition affects our ability accurately to measure unintended pregnancy. Moreover, in the context of HIV we do not know how knowledge of one’s HIV status may also influence wantedness over the course of pregnancy and birth.

Our estimates assume that all unintended births can be prevented by family planning programmes, which is, of course, unrealistic just as it is unrealistic to expect PMTCT services to be able to eliminate all vertical HIV transmission. Some women may have unmet need for contraception but may not use methods for a variety of reasons including their husbands’ attitudes, religious prohibitions, lack of knowledge about where to get methods or they have concerns about side effects or costs.35 Not all of these can be overcome and translated into demand.

UNAIDS recently released revised numbers of persons living with HIV. Although we have revised country-specific estimates of the number of births to HIV-infected women in 2006,23 we have not relied on the updated country-specific adult HIV prevalence estimates because this will not be released until approximately July 2008.11 In the new estimates the number of persons living with HIV was reduced by 16% compared with the 2006 estimates; 70% of the overall reduction is due to changes in Angola, India, Kenya, Mozambique, Nigeria and Zimbabwe with the largest contributor to the reduction being India. Since country-specific revisions are not available, some of our estimates of the number of HIV-positive births averted by current contraceptive use specifically in Kenya, Mozambique and Nigeria may change. However, we do not anticipate that these changes will result in any revisions to our conclusions.

Any estimates, including ours, are only as good as the data that we use to make them. The DHS and UNAIDS numbers are the most rigorous population-based estimates available. The biggest limitation to our model comes from the lack of precision. For example, women living with HIV may have different contraceptive prevalence, method mixes and levels of unintended pregnancy than the population of all women from which we draw our data. To the best of our knowledge, the numbers we used approximate those found among HIV-positive women,19^–22 but population-based surveys about the fertility desires and contraceptive use of HIV-positive women are needed. Nonetheless, we have relied on sources that engender confidence in the direction and magnitude of the results.

CONCLUSION

Current attempts to reduce the number of infant HIV infections are already being bolstered by global family planning programmes. These could be substantially strengthened by additional efforts to provide contraception to HIV-infected women who do not wish to become pregnant. Family planning as an HIV prevention approach is cost effective compared with other PMTCT approaches. Achieving the United Nations General Assembly Special Session on HIV/AIDS (UNGASS) goal of reducing infant HIV infections by 50% by 2010 can be better achieved by a comprehensive approach which combines two effective interventions: increasing contraceptive use and targeting antiretroviral programmes for HIV-infected women.36