On March 29, 2016, the WHO Director-General terminated the Public Health Emergency of International Concern for the Ebola virus disease (EVD), an outbreak of hemorrhagic fever in West Africa. By then, an estimated 30,000 people had been infected with EVD, and over 10,000 people had died as a direct consequence, the majority in Guinea, Liberia, or Sierra Leone (1).

In December 2013, the index Ebola case in the West Africa outbreak was infected, and the disease was identified as Ebola in March 2014 (2). During the next year, the number of confirmed Ebola cases and deaths in the three most affected countries increased weekly, until joint national and international efforts to stop the epidemic finally led to its gradual decline by the second half of 2015. Before the outbreak, health indicators in West Africa were among the worst in the world, particularly in terms of maternal and child health (3). In this fragile context, the outbreak was a significant shock to the health systems and led to fear and mistrust of the health services, the closure of health facilities, and the deaths of health service staff (4). During the outbreak, health workers were 21–32 times more likely to be infected with Ebola than the general adult population (5). While the international response was considerable – hundreds of Ebola Treatment Units were built – it came late and focused mainly on isolating EVD cases, thereby leaving the treatment and provision of services for the regular burden of disease and routine health care largely unattended (6). For example, previous studies completed by the authors who assessed nationwide routine health service delivery data during the EVD epidemic in Sierra Leone found that hospital functioning was significantly reduced during the outbreak. This information suggests that the indirect effects of EVD on health system functioning might have adversely affected more people than the virus itself (7, 8).

Studies on the indirect health effects of the EVD outbreak are now starting to emerge. While the studies all add important pieces of information, few of them were done at a nationwide level and even fewer covered the whole affected area. Therefore, a systematic review is important at this early stage in order to create a comprehensive picture. By outlining the evidence for what affects the outbreak had on health, it may be possible to help guide policy on health system strengthening and develop resilience for withstanding future shocks. The aim of this study is to assess the magnitude of indirect health effects of the West Africa EVD outbreak, based on existing literature.

The search strategy for the review was developed in collaboration with librarians at the Karolinska Institutet university library. After several pilot searches, the main search was completed in March 2016 and included the following databases: Medline (Ovid), Embase, Web of Science, Global Health (Ovid), POPline, and PubMed. Keywords used in the search included variations in terms for Ebola, health, health systems, health facility, and illness/disease. Relevant gray literature was identified from the various actors on-site in the affected countries during the outbreak. WHOLIS (WHO plus iris), WorldCat Libraries, and Base Bielefeld databases were also searched for gray literature. Full documentation of the search strategy can be found in Supplement 1. The study was registered on PROSPERO under registration number CRD42016036318 (9).

The focus of the review is the West Africa Ebola outbreak, due to its unprecedented magnitude. The affected population was defined as persons residing in Guinea, Liberia, or Sierra Leone, or in the affected areas of Nigeria, Senegal, and Mali. Studies from countries with three or less cases of EVD were not included, since the indirect impact of a small number of cases is likely to be negligible. Direct health effects were defined as morbidity and mortality due to infection with the Ebola virus. Indirect effects refers to changes in health outcomes caused by the effect of the West Africa EVD outbreak on the health systems, including but not limited to increased maternal morbidity and mortality, a reduction in HIV-infected patients receiving antiretroviral treatment (ARTs), an increase in malaria cases due to termination of intermittent preventative treatment (IPT) programs, fewer children being treated for diarrhea and acute respiratory infections (ARI) and hospital in-patient admittance and essential surgery.

Because of the scarcity of literature currently available on the indirect effects of the Ebola outbreak, the eligibility criteria were deliberately broad. Studies were included in the analysis if they contained information on indirect health effects of EVD and had either an abstract or full-text available in the English language. Studies were excluded if they reported information only on the direct effects of EVD, if mental health was the outcome under study, or if the publication was a commentary article. Gray literature and several reports were included in the review; all reports were assessed for potential bias regarding the goals of the author organization. An inclusive approach was taken, and several modeling studies are therefore included in the analysis.

The searches produced 3354 articles after deduplication. A total of 2788 articles were excluded during title screening and 442 were excluded after abstract screening. About 117 articles were read in full, and 22 studies were included in the final review. Other relevant studies were identified for inclusion through the reference lists of included studies. Figure 1 is showing a Prisma diagram of the search and screening procedure. Two of the authors performed the screening process independently, and in case of uncertainty about inclusion, a consensus was reached among the authors after discussion. All articles were organized in an EndNote X7 library.

Information on location of the study, the objective, the study design, the outcome(s) under study, and results was extracted into a spreadsheet file for analysis. After extraction, information was grouped by outcome for the analysis, resulting in three broad categories: reproductive, maternal, and child health services (RMNCH); HIV/AIDS, tuberculosis or malaria; and other outcomes, such as cause of death. The evidence was rated for quality, using the GRADE approach as a guide, and divided into one of the four quality levels: high, moderate, low, and very low (11). In addition to study quality, the relevance of each study to the aim of the review was assessed.

The review included 22 studies, of which 19 were peer-reviewed scientific papers or policy briefs. The remaining three were reports by UNICEF, Doctors of the World, and Voluntary Services Overseas. Eight of the studies focused on Sierra Leone, six on Guinea, and three on Liberia. The remaining five studies included all three countries. Key health indicators for each country can be found in Table 1. A summary of the studies can be found in Table 2.

Eleven studies investigated how EVD impacted RMNCH, 10 focused on HIV/AIDS, tuberculosis, or malaria, and 4 reported other outcomes of changes in patient admissions and surgery and causes of death. There was some overlap within the different health outcome categories in the studies. The results of each category are described below. The main results of this review are visualized in Figure 2, which aims to provide an overview of the indirect effects the EVD outbreak had on health and health-care utilization.

This review indicates that the magnitude of indirect health effects due to the West Africa Ebola outbreak is substantial, but it is also clear that further studies needs to be done post-disaster to better understand the overall health effects of a crisis of this scale. All studies included in this systematic review show a decrease in health service provision or utilization. Estimates from the various studies suggest both morbidity and mortality to have increased after the onset of EVD in Sierra Leone, Guinea, and Liberia. RMNCH was especially affected and the decrease in facility deliveries, cesarean sections, and volume of ANC and PNC visits is likely to have a significant impact on maternal and newborn health. Emergency obstetric care such as cesarean sections is the most effective mean to lower maternal mortality (33) and the unmet obstetric need in these three countries was very high even before the EVD outbreak (34). A decrease in the already low number of women accessing emergency obstetric care in Sierra Leone, Liberia, and Guinea are likely to cause additional deaths among mothers and newborns.

Even if the adult HIV prevalence is estimated at below 2% in all three countries, with around 20% of infected individuals being on ART, HIV/AIDS remains one of the top 10 causes of mortality. Several studies saw an increase in both ART defaulters and admission of new patients. Service provision during this time remained fairly stable, indicating that the decrease in treatment was due to the reluctance among patients to visit health facilities during the EVD outbreak. The authors of one of the studies conclude that during a crisis such as the Ebola outbreak, HIV care is challenging but not impossible (21). However, the EVD outbreak is likely to have a long-term impact on the burden of HIV/AIDS in these countries, due to disrupted treatment and a potential increase of infectiousness as well as a delayed onset of treatment for newly infected people.

Malaria was a substantial burden in Sierra Leone, Guinea, and Liberia before the EVD outbreak. The studies included in this review clearly show that a disruption in both preventative and curative measures for malaria would increase morbidity and mortality. For instance, IPT in pregnancy has repeatedly been shown to decrease cases of malaria amongst pregnant women, which in turn prevents the adverse effects of malaria in pregnancy, including low birth weight and neonatal deaths (35). Several studies in this review showed a substantial decrease in cases of malaria seen at health facilities after the onset of the EVD outbreak, increasing the risk of severe and cerebral malaria, especially in children. However, malaria care was maintained during the outbreak at a greater extent than expected in some of the models.

This review shows a decrease in vaccination coverage for pentavalent vaccines in Guinea and Sierra Leone, and the potential effect of a disruption in measles vaccination programs. For countries with low vaccination coverage before the outbreak, the effects could be more severe. The suspension of vaccine programs during the outbreak will have long-standing adverse consequences, including the potential for a higher rate of transmission of vaccine-preventable diseases and a lower proportion of fully vaccinated children. Making large-scale nationwide vaccination campaigns a main priority in the recovery phase for these three countries could mitigate adverse outcomes.

By May 2015, 1.45% of Guinea’s doctors, nurses, and midwives had died of Ebola; in Liberia, 8.07% died and 6.85% died in Sierra Leone. In countries where the number of health workers was already low, the deaths are sure to have a lasting impact on population health, as is highlighted in a recent policy paper by Evans et al. (17, 32). Here, the authors estimate that approximately 25,000 additional deaths per year could occur due to the death of health-care workers.

Interestingly, a study on obstetric health care in Sierra Leone showed that those facilities that remained open during the outbreak continued to perform the same proportion of both deliveries and cesarean sections as they did before the Ebola outbreak. This suggests that the nationwide decrease in deliveries and cesarean sections is due to the closure of key facilities, or in some cases, the conversion of hospitals into Ebola Treatment Units. The same study also found that government hospitals were able to maintain their health service provision to a much larger extent than both non-profit and for-profit private facilities (8). It is clear that many health-care workers kept working despite the harsh and frightening conditions at the time of the outbreak. Importantly, this review shows that to a large extent, it was not health service provision that failed, but rather the uptake of health services by the population that decreased after the onset of the Ebola outbreak. Not all studies suggested reasons for this, but references have been made to the fear of nosocomial transmission of the disease and mistrust of authorities. Health centers and hospitals became centers of disease transmission and by ignoring cultural traditions and beliefs, mistrust increased. Furthermore, service uptake may have decreased as a result of misinformation on whether facilities were open and providing services during the outbreak. However, it is not possible to identify a cause for of the decreases from the results of this review, and it is likely that disruption of services also played a role in the decline in service delivery. Although no conclusions can be drawn on equity in access to health care during the EVD outbreak, it is plausible that both access and uptake of health services was differentially affected depending on where in the country people reside.

Understanding the damaging effect the West Africa Ebola outbreak had on health and health services in the affected countries can help guide policy development on health system strengthening and hopefully increase their resilience for withstanding future shocks. At this early point after the end of the EVD epidemic, the full magnitude of the indirect effects is yet to be understood and documented. However, it can be seen already that the adverse impact of Ebola stretches far beyond Ebola cases and deaths. This review indicates that indirect effects can be long term and highlights the importance of international support to routine health service delivery and the maintenance of vaccination programs and preventative and curative malaria programs, both in general but especially in times of emergencies. The most important response to an outbreak of this magnitude comes now, before the next outbreak – when we can learn from past mistakes and build back stronger and more resilient health systems.

Conception or design of the work: KR and DS. Data collection: KR and DS. Data analysis and interpretation: KR and DS. Drafting the article: KR. Critical revision of the article: KR, DS, AE, and JS. Final approval of the version to be published: KR, DS, AE, and JS.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.