Context: the flow of USG global health funding

As illustrated in figure 1, the USG is the largest global health donor worldwide, funding several key health domains, including HIV/AIDS; tuberculosis (TB); malaria; maternal and child health (MCH); FP and RH; nutrition; global health security; neglected tropical diseases; and water, sanitation and hygiene (WASH). In 2018, USG global health assistance totalled $10.8 billion.3

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Government contributions to global health assistance for 2017,in millions of US$.45

Of this, up to $9 billion was potentially subject to PLGHA, to the extent that such funding was ultimately provided to foreign NGOs, directly or indirectly.3 The USG does not reveal how much funding is granted (directly or through subgrants) to foreign NGOs; it is unclear if they track this information or not. A recent Kaiser Family Foundation report analysing USG global health funding obligated by the US Agency for International Development (USAID) for fiscal years 2013–2015, showed that, had PLGHA been in place during that time, at least 1275 foreign NGOs (half as prime and half as subrecipients) and $2.2 billion would have faced restrictions. These 1275 foreign NGOs carried out global health activities spanning major health domains in at least 91 countries.10

As shown in figure 2, the bilateral funds impacted by PLGHA are given to US or foreign prime recipients and subrecipients who may channel them to actors throughout the health system. Though MOHs and public clinics are exempt from PLGHA, funding to these entities—as well as to foreign NGOs—could be terminated if a foreign NGO upstream decides not to certify PLGHA and has consequently reduced their direct funding or technical support of the public sector.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

USG Global health assistance bound by PLGHA.

US Secretary of State Mike Pompeo announced an expansion of PLGHA on 26 March 2019. As represented in figure 3, this expanded interpretation stipulates that foreign NGOs that receive USG global health assistance as a prime awardee or subawardee are prohibited from providing any (including non-USG) financial support to any foreign NGO that engages in activities prohibited by the policy. It extends the policy beyond the organisations receiving USG money to subgrantees of separate donor projects. Organisations must ensure that their foreign NGO subgrantees comply with the PLGHA policy, even if these subgrantees do not receive any USG global health assistance, from any source. This expansion essentially gags organisations that work with funds provided by other donors.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

The 2019 PLGHA Expansion.46 Note: PLGHA is commonly referred to as the “Global Gag Rule” or “GGR”. The above figure uses “GGR” to refer to PLGHA.

In a late May 2019 communication, USAID informed its implementing partners that no changes will be made to the language of the ‘financial support’ prohibition contained in the standard provisions included in their grants and cooperative agreements as recipients of USG global health assistance. As a result, the expanded interpretation of the ‘financial support’ requirement announced by Secretary Pompeo in March 2019 is in effect, and NGOs must apply the requirement to any new financial agreements with foreign NGO partners, as well as existing agreements. Implementing partners are responsible for determining how to conduct the due diligence that may be necessary to ensure compliance with the ‘financial support’ requirement, along with all of the other requirements contained in the standard provisions in their assistance agreements with USG.

Impact of previous iterations of the Mexico City policy

There is some peer-reviewed research and grey literature on the impact of previous iterations of the Mexico City policy on access to key sexual and reproductive health (SRH) services and on overall health outcomes. Bendavid et al examined the association between exposure to the Mexico City policy during the George W Bush presidency and the likelihood of induced abortion among women of reproductive age across 20 countries in Sub-Saharan Africa.11 They found that women living in countries defined as heavily exposed to the Mexico City policy had 2.55 times the odds of self-reported abortion compared with women living in less exposed countries. The authors posited that, in the countries with high exposure to the policy, foreign NGOs that did not certify compliance and thus lost USG funding had been significant providers of FP and RH services.11 These services were not ‘replaced’ by other USG grantees willing to certify the policy. Thus, the policy led to an increase in the number of unintended pregnancies, and subsequently, the number of women seeking voluntary terminations.

Rodgers expanded on this study using Demographic and Health Survey (DHS) data for 51 countries in four regions (Latin America and the Caribbean, South/South-East Asia, Sub-Saharan Africa, and Eastern Europe and the Middle East) from 1994 to 2008.12 A separate regression analysis for each region showed that in Latin America and the Caribbean and in Sub-Saharan Africa, the odds of having an induced abortion in a high exposure country after the Mexico City policy (during the George W Bush presidency) went into effect increased by factors of approximately 3 and 2, respectively. However, these findings did not hold in Eastern Europe and the Middle East or South/South-East Asia. Rodgers posited the same causal pathway as Bendavid et al in Sub-Saharan Africa and Latin America and the Caribbean. She stated that the null finding in other regions was explained by replacement funding and other contextual factors.

Two studies by Jones used DHS data from Ghana to estimate the impact of the Mexico City policy on pregnancy, abortion and child health outcomes. One study found that the 1984 and 2001 versions of the policy were associated with increases in pregnancies, particularly in rural areas; and that approximately 20% of the additional rural pregnancies that occurred when the policy was in place ended via induced abortion.13 As with the abovementioned studies, Jones attributed this finding to the constraints on FP services resulting from the policy. The second study is the only peer-reviewed study where Mexico City policy related health outcomes outside of SRH were quantitatively examined. This study found that the weight-for-age and height-for-age of children in Ghana were negatively associated with the 2001 reinstatement of the Mexico City policy.14 Jones hypothesised that, due to restricted access to voluntary safe abortion, a greater proportion of babies born may have experienced reduced nutrition and healthcare ‘as a result of their unwantedness’.14

In June 2019, Brooks, Bendavid and Miller published a study on changes in modern contraceptive use, pregnancies and abortion among women in 26 Sub-Saharan African countries between periods when the Mexico City policy was rescinded (1995–2001 and 2009–2014) and when it was in effect (2001–2008). They found that women living in countries defined as heavily exposed to the Mexico City policy had 12% more pregnancies, 13.5% less use of modern contraception and 40% more abortions when the policy was in effect compared with when it was rescinded. Because it compared periods of exposure to the Mexico City policy to periods of non-exposure to the policy, this is the first study to demonstrate that changes in these SRH outcomes reflect the status of the policy’s implementation in Sub-Saharan Africa. The authors joined Bendavid et al, Rodgers and Jones in suggesting that the observed changes are the result of disruptions to contraceptive service delivery when the Mexico City policy is active.15

We searched ‘Mexico City Policy AND impact’, ‘Mexico City Policy AND evaluate’, ‘Global Gag Rule AND impact’ and ‘Global Gag Rule AND evaluate’ on PubMed and did not identify any additional peer-reviewed studies.

In addition to these peer-reviewed studies, grey literature describes the impact of prior iterations of the Mexico City policy on foreign and US NGOs, access to contraception and RH care services, and on health systems more broadly. Anecdotal data from these sources indicate weakened HIV prevention efforts, contraceptive stock-outs, shuttered clinics, obstructed access to safe abortion in countries where it is legal, the disruption of mobile and community-based healthcare serving youth and rural populations, and pervasive fear of advocating for and sharing information about legal abortion among NGO staff and health workers. The reports suggest that these impacts were attributable to several factors, including foreign NGOs deciding not to certify and thus losing funding; foreign NGOs changing their programming as a result of certifying; foreign NGOs certifying and overinterpreting the Mexico City policy; and a general chilling about abortion policy, advocacy and practice at the country level.16–20

Overinterpretation, often referred to as the ‘chilling effect’, can stem from incomplete or incorrect information received from USAID missions and/or from fear of jeopardising an important source of funding. In some instances, certifying NGOs self-censored or limited activities far beyond what the Mexico City policy required. Research in Uganda, Ethiopia, Kenya, Peru and Nepal found that while the Mexico City policy was in place during the George W Bush presidency, discussions about national abortion law reform included fewer stakeholders, and lawmakers reported lack of access to critical information.21 22 Service delivery can become chilled as well; several foreign NGOs that complied with the Mexico City policy stopped providing RH services that were in accordance with national law and permitted under the policy.18

The pattern of reinstating and rescinding the Mexico City policy can deepen the chilling effect. Research conducted in Ethiopia soon after the Obama administration rescinded the Mexico City policy found that the flip-flop nature of the policy led some NGOs to decide against taking on the risks of working on abortion altogether, irrespective of the policy’s recent reversal. Further, organisations reported confusion about what activities were and were not allowed during periods when the policy was rescinded, and recipients of FP assistance reported being burdened by the costs associated with changing their practices in accordance with the policy’s re-instatement.19

Our hypothesis: the potential health systems impact of disruptions to USG global health assistance due to PLGHA

We expect that reduced access to FP will recur in the case of PLGHA. The funding and service delivery landscapes have changed significantly since prior iterations of the Mexico City policy were in place, so we expect changes in domains beyond FP. Moreover, PLGHA applies to a much greater amount of funding, and thus to many more organisations and services. There is no entity that systematically tracks whether foreign NGOs decline to certify PLGHA and thus lose expected funds, forego seeking USG global health assistance due to PLGHA, misinterpret PLGHA or change their activities so that they can certify PLGHA and receive USG global health assistance. Taking into account the current global health funding and health systems contexts and our review of research on prior iterations, we describe how we think these decisions made by USG global health assistance grantees, or potential grantees, might affect health systems in low and middle income countries.

In the last decade, global health guidelines and funding have supported health service integration, as opposed to vertical funding of disease-specific programmes.23 24 Because of this shift, patients seeking RH care are more likely to also access additional health services (and vice versa) than they would have been under the Mexico City policy. Referral systems and integrated services connect patients across multiple domains of the health system including HIV/AIDS, TB control, FP and MCH. This integration can be more efficient, increase service uptake, improve patient satisfaction and improve health outcomes.25–30

As illustrated in figure 4, in this context of service integration, funding disruptions could have a ripple effect. For example, an HIV service organisation that certifies the policy may stop referring their clients with unintended pregnancies to the local RH clinic, as they overinterpret the policy and believe they cannot refer patients to facilities that provide abortion, PAC or emergency contraception. Alternatively, an HIV service organisation that declines to certify the policy and loses funding as a result may be forced to close clinics, reduce staff and/or reduce services, which in turn affects the referral network(s) that they participated in. Importantly, these disrupted referral networks affect populations served by certifying NGOs, those served by non-certifying NGOs and those served by organisations to whom the policy does not apply (eg, US NGOs).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Example* Referral System Disruptions Due to the Expanded PLGHA, for Certifying and Non-Certifying Foreign NGOs.

Further, while USG global health assistance may be directed to a specific domain or population, the money is also used for overheads and expenses shared across the programmes. This can include salaries, infrastructure and commodities. Thus, loss of income resulting from a foreign NGO’s decision not to certify PLGHA can result in funding as well as referral disruptions beyond the funding stream/activities subject to the policy.

Funding cuts in one area could also affect services that seemingly have little to do with abortion. For example, Amin et al describe a collaboration between a local NGO, funded by USAID, and local government to provide expanded immunisation programme and microcredit assistance, along with FP services, in rural Bangladesh.31 Microcredit loans are a source of financial empowerment for poor women, and are often the first step in breaking down informational and cultural barriers to health services, particularly FP.32–35 If the foreign NGO described by Amin et al declined to certify PLGHA, or misinterpreted the restrictions, resulting impacts could include fewer referrals for rural women to antenatal and delivery care, disruption of childhood immunisations, and decreased access to microcredit assistance for women.

The disruptions to funding streams (figure 2), referral systems (figure 4) and service delivery (figure 5) can contribute to siloing within each of these areas. Referrals and integrated service delivery will likely be weakened or disrupted entirely by PLGHA. Creating new referral relationships to replace or improve those impacted by PLGHA requires additional time and resources—those who do not certify the PLGHA must search for and establish alternative partnerships, and those who do may need to secure additional funding to allow them to fill gaps. Moreover, collaboration at the country and global levels among foreign NGOs who do and do not certify can decrease. Just as certifying NGOs overinterpret the policy to mean that they cannot refer clients to non-certifying NGOs, so too may they understand the policy to imply that they cannot coordinate or collaborate in national or global level discussions about RH or other areas.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5

Service provision disruptions for PLGHA certifying and non-certifying foreign NGOs.

As outlined in table 1, research assessing our hypothesis about health systems impacts of PLGHA should explore the impacts of the policy on funding, client referral, coordination, policy discussions and health service provision among certifying and non-certifying NGOs, the public sector, and the private sector. Research should be especially attentive to equity, as rural and other populations may disproportionately lose access to services. While FP services may be the most affected, in many cases, WASH, MCH, HIV, and other areas will also be affected. Thus, health systems research should include NGO and public sector actors in these domains. The extent and mechanisms of overinterpretation should be explored. What kinds of overinterpretation occur? When it does, on what basis? What reasons do the relevant decision-makers give for their interpretation? A comprehensive understanding of overinterpretation can inform efforts to mitigate the harm associated with PLGHA. Moreover, some research should assess the governance of abortion and FP at the global level. Deliberation may be ‘gagged’ at fora for shared strategy development and agenda setting.

A pre-PLGHA and post-PLGHA comparison will help to surface some of most important changes, though there are important challenges to note. The dates the policy comes into effect may vary, as PLGHA is applied when grants are renewed or newly negotiated. Moreover, it may be difficult to get accurate information about likely impacts, as many actors within health facilities and subrecipient NGOs may not know the cause of changes, or are unable to attribute changes to PLGHA as opposed to other phenomena, such as funding cuts to UNFPA or domestic political changes. Many stakeholders may be reluctant to discuss overinterpretation and the chilling effect, as well as PLGHA and abortion more generally, as they fear alienating the USG or see little to be gained in discussing the politically and culturally sensitive issue of abortion. Also, pre-PLGHA and post-PLGHA is an imperfect comparison, as it will not necessarily surface NGO decisions to forego funds or activities.

As noted, there is scant peer-reviewed research on the impact of changes in overseas development assistance policy. Research should amply describe methods, as well as how triangulation and other strategies were used to draw conclusions about attribution. There are co-occurring phenomena that could both exacerbate the impact of PLGHA and make establishing causality challenging, including closing civil society space for women’s rights in many countries,36 President Trump’s decision to stop funding UNFPA,37 and disruptions in the production of critical FP commodities.38

Ripple effects of funding disruptions: 1 year into the expanded policy

The impact of PLGHA will emerge over time, as research is ongoing. However, some early data are available; much of it supports our hypothesis presented above. amfAR and Johns Hopkins University conducted an electronic survey of current prime PEPFAR implementing partners (n=286) from May to September 2018, and found that PLGHA impacted 33% of them across 31 countries. Some of the impacts reported by the PEPFAR prime partners included a reduction in the provision of non-abortion related services such as HIV, contraception, cervical cancer screening and adolescent health counselling.39 These reductions may be due to overinterpretation and/or the disruption of referral networks.

Multiple organisations have collected and shared anecdotal evidence since the advent of PLGHA. This evidence suggests that PLGHA has affected multiple health domains and populations within the first year of implementation, including programmes related to HIV, WASH and Zika.40 Some of these impacts are summarised in table 2. Among other populations, these programmes serve people living in rural areas, adolescent girls and young women, people living with disabilities, and refugee populations.40 We speculate that PLGHA has inequitable effects because outreach activities and activities in isolated rural areas are the most expensive per client served, and thus may be the first to be cut (S Beare, S Medina, personal communication, 11 June 2019).

Additional research conducted by the members of our GGR research working group (see acknowledgements) will yield more in-depth, longitudinal quantitative and qualitative data, including on the impact on services provided by foreign NGOs that opt to comply.

Approaches to mitigate negative impacts on health systems

Given that the Mexico City policy has been reinstated and rescinded several times, the RH funding and advocacy community was better positioned to anticipate some of the policy’s harms in 2017 compared with 2001. One day after PLGHA was announced, the Government of the Netherlands launched the ‘She Decides’ initiative. The purpose of She Decides is to provide a political platform to ‘support the fundamental rights of girls and women to decide freely and for themselves about their sexual lives’.41 In March 2017, more than 50 governments attended a conference organised by the government of Belgium and co-sponsored by the governments of Denmark, Sweden and the Netherlands. The conference raised $190 million in pledges to support SRH and rights globally.42 As of March 2018, the movement had mobilised $450 million.43 However, the magnitude of the funds affected by PLGHA exceeds any amount of replacement funding that can be expected to be mobilised, even if She Decides combined with support from other sources. Moreover, it is likely that pledges to the movement do not represent new funding to global health and development, such that there are opportunity costs for donor commitments elsewhere.

One donor has tried to mitigate anticipated harm by introducing their own conditions. The Swedish International Development Agency issued guidance in response to PLGHA stating that they would review partnerships with foreign NGOs who may not be able to fulfil the scope of their agreements because they opt to comply with PLGHA.44 Some report that new conditionalities also mean that foreign NGOs feel less able to chart their own course. In other words, they are caught between the policy preferences of bilateral donors.40