Need for and use of contraception by women before and during COVID-19 in four sub-Saharan African geographies: results from population-based national or regional cohort surveys

Introduction

Sexual and reproductive health experts have predicted devastating consequences of the COVID-19 pandemic on the health of women.^{1, 2} Gender disparities are expected to be exacerbated by COVID-19^{3, 4} due to restricted access to health services, increased risk of gender-based violence, and severe economic effects.^{4, 5, 6} These forecasts draw heavily on data from previous epidemics, most notably, the 2016 Ebola outbreak in west Africa, which resulted in a substantial decline in family planning services (65% in Liberia; 23% in Sierra Leone).⁷ Although the effect of COVID-19 on sexual and reproductive health is largely unknown, global projections suggest the 2020 expected satisfied need for modern contraception could drop from 77% to 71% because of 60 million fewer contraceptive users.⁸

Adverse effects of the COVID-19 pandemic on the sexual and reproductive health of women are expected to be most prominent in low-income and middle-income countries (LMICs), which already face a multitude of challenges related to sexual and reproductive health service delivery.⁹ One modelling exercise projects that 15 million additional unintended pregnancies, leading to 28 000 maternal deaths, would occur over the course of one year if COVID-19-related health-care service disruptions affected 10% of women in need of sexual and reproductive health services in LMICs.¹⁰ Given mobility restrictions,¹¹ shortages of contraceptive commodities,^{6, 12} and fear of exposure to COVID-19 when seeking services,¹⁰ these projections might be an underestimate.

Concerns surrounding the impact of COVID-19 on the sexual and reproductive health of women are particularly acute in sub-Saharan Africa. Although the region has achieved success in controlling SARS-CoV-2 spread via early lockdowns and restrictions on movement, these preventive measures took a toll on the African economy and health-care systems, potentially reversing recent progress in the area of sexual and reproductive health.¹¹ Notably, before the onset of COVID-19, sub-Saharan Africa had an overall increase in modern contraceptive use;¹³ however, challenges in achieving universal access to sexual and reproductive health services, including family planning, remain substantial. In 2019, only 55% of sub-Saharan African women's need for contraception was satisfied and the unintended pregnancy rate (91 per 1000 women per year) was the highest globally.¹⁴

Early projections have been essential in raising awareness about the possible effects of COVID-19 on the sexual and reproductive health of women in sub-Saharan Africa and highlight the urgent need for evidence-based policies that balance prevention of COVID-19 transmission with access to sexual and reproductive health services. Few studies provide such timely information, and many rely on health systems data,^{15, 16} which do not necessarily reflect population-level changes in contraceptive use. Individual-level data from Burkina Faso and Kenya highlight little change in contraceptive use as a result of COVID-19;¹⁷ however, population-level data are needed to capture overall trends.

Across four sub-Saharan African geographies at various stages of the demographic transition and with diverse sociopolitical contexts, we aimed to examine population-level changes in the need for and use of contraception in women during the COVID-19 pandemic, determine if these changes differed by sociodemographic characteristics, and compare observed changes during COVID-19 with trends in the 2 years preceding the pandemic. We hypothesised that women's need for contraception would increase during the pandemic in relation to economic uncertainties. We anticipated that contraceptive use among women in need would decrease due to difficulties accessing services.

Methods

Results

Our analyses were restricted to 7245 women in union who were interviewed at baseline and COVID-19 follow-up. Characteristics of the analytic sample are presented in table 2. The mean age of women was 31·8 years (SD 8·3) in Burkina Faso, 34·6 years (7·4) in Kinshasa, 32·6 years (7·7) in Kenya, and 35·9 years (6·4) in Lagos. Women's educational attainment differed substantially across geographies. Most women reported a partial loss of household income during the COVID-19 pandemic; complete loss of income was most common in Kinshasa. Need for contraception and contraceptive use by those in need varied across the geographies we studied.

Between baseline and COVID-19 follow-up surveys, the proportion of women in need of contraception increased in all areas studied, although changes were only significant in Lagos (table 3; appendix 2 pp 2–3). Contraceptive use among women in need increased significantly in the two rural geographies, with a 17·37 percentage point increase in rural Burkina Faso (30·7% to 48·1%) and a 7·35 percentage point increase in rural Kenya (71·6% to 78·9%). Increases in contraceptive use in Kenya were similar between women in urban and rural areas (5·24 and 7·35 percentage points, respectively). No differences in contraceptive use were noted in other urban settings.

These overall trends mask several distinct patterns by sociodemographic groups (table 3). Specifically, there was an increase in the need for contraception among nulliparous women in all geographies studied. Nulliparous women in rural Burkina Faso and urban and rural Kenya had the greatest increases in need (38·88, 35·27, and 35·07 percentage points, respectively). Need for contraception also increased for women with higher education (6·17 percentage points) and those who experienced partial income loss due to COVID-19 (10·48 percentage points) in Lagos.

Contraceptive use among women in need rose for specific subpopulations during the pandemic (table 3). Much greater proportions of nulliparous women in rural Kenya and urban Burkina Faso were using contraception during the COVID-19 pandemic. Significant increases in contraceptive use were also observed for almost every sociodemographic subgroup among rural women in Kenya and Burkina Faso. Although similar increases were noted among some subgroups in urban Kenya, such as older women and those with more children, broader changes by sociodemographic characteristics were not consistently observed in other urban settings. We found an increase in contraceptive use among women who reported partial income loss due to the COVID-19 pandemic in rural Burkina Faso. In contrast to these patterns, we noted non-significant drops in contraceptive coverage among young women in Lagos and among nulliparous women in Kinshasa.

Non-significant increases in the need for contraception during the COVID-19 pandemic (except in Lagos, where the increase was significant; figure 1) are consistent with trends observed in the year before the pandemic, except in urban Burkina Faso and Kinshasa, where such changes marked a departure from the previously observed decline.

The rise in contraceptive use during the COVID-19 pandemic followed a period of stagnation and decline in Kenya and rural Burkina Faso (figure 2). Conversely, in Kinshasa and Lagos, contraceptive use stabilised during the COVID-19 pandemic after years of substantial increases.

Discussion

This analysis across four diverse geographies in sub-Saharan Africa does not support the anticipated deleterious effect of COVID-19 on women's access to and use of contraception in the earliest stages of the pandemic. Instead, we found that the COVID-19 pandemic was associated with marginal increases in women's need for contraception and larger increases in women's contraceptive use, although these trends vary by setting, women's characteristics, and reproductive life stage. Our results provide early evidence that the pandemic has so far, not altered the recent growth in contraceptive use, noted in several countries.¹³ Our results align with individual-level data indicating minimal shifts in contraceptive status in Kenya and Burkina Faso.¹⁷

Stratified analyses by residence revealed discernable differences in key contraceptive indicators. In contrast to our initial hypothesis, we found an increase in contraceptive use in the COVID-19 period in Kenya and rural Burkina Faso, but similar rises were not observed in Kinshasa or Lagos, where previous progress in meeting women's demand for contraception seemed to have stalled. Increases in use of contraception occurred in settings where women in union were more likely to rely on longer-acting contraceptive methods before the pandemic, including in rural Burkina Faso and Kenya, where 55% and 48% of contraceptive users, respectively, were using a long-acting method pre-pandemic. However, no increases in contraceptive use were noted in settings with a lower reliance on these methods (Kinshasa and Lagos). These findings could reflect an important benefit of long-acting methods in reducing unmet need for contraception by sustaining use,²¹ even in the face of service disruptions and potential commodity shortage. However, the substantial increase in contraceptive prevalence in rural areas also suggests fewer disruptions to sexual and reproductive health services in these settings than have been noted in facility-based studies in Ethiopia and South Africa.^{15, 16} The variability in contraceptive trends across geographies at the early stages of the COVID-19 pandemic in sub-Saharan Africa might also reflect differences in health system preparedness, as some countries, such as Burkina Faso, acted to prevent shortages by accruing large stocks of contraceptives.²² A thorough understanding of site-specific pandemic responses is crucial for monitoring sexual and reproductive health indicators in the region, as the effect of COVID-19 restrictions on women's access to contraception might be delayed, with areas requiring urgent interventions not yet apparent.

Furthermore, some differences in sexual and reproductive health indicators were observed by COVID-19-related economic losses, particularly in rural areas. Changes in contraceptive use did not accompany COVID-19-related income loss in Kinshasa, where 67% of women reported complete income loss. These descriptive results confirm the emerging concern with the impact of the COVID-19 pandemic on economic insecurity in sub-Saharan Africa.²³ Economic concerns have been linked to changes in fertility intentions, as evidenced by sanitary crises²⁴ and economic downturns.²⁵ Similarly, we found increases in use of contraception among women who had partial economic losses in Kenya and rural Burkina Faso. Such changes might reflect stronger intentions to postpone or limit childbearing²⁶ in the COVID-19 context among women who are able to access sexual and reproductive health care, whereas those unable to afford such services might benefit from targeted outreach programmes and free contraceptive provision.

Our results show further differences in women's sexual and reproductive health behaviours during the COVID-19 pandemic, according to their reproductive life stage and socioeconomic circumstances. An increase in the need for contraception was found among nulliparous women across all geographies studied. This finding might signal a desire to delay a first birth in the early stages of the pandemic, although the levels of need for contraception remained low among nulliparous women in all geographies studued. In urban Burkina Faso and rural Kenya, more women were able to act on their fertility preferences to prevent or delay pregnancy by using contraception. Importantly, this increase, albeit modest, occurred despite substantial normative barriers that discourage contraceptive use for women who still want children.^{27, 28} However, our results also indicate that these trends in contraceptive use are context-dependent, given that young women in Lagos and nulliparous women in Kinshasa had a drop in contraceptive coverage (albeit non-significant given large CIs). Exploration of motivation for this decline is warranted, as young women who are more likely to rely on coital-dependent methods might have less need for contraception if sexual activity patterns have changed, or might face greater access barriers due to the closures of government youth-friendly services.

This study is not without limitations. Foremost, telephone-based data collection was implemented given COVID-19 safety concerns, resulting in substantial sample restrictions (36·7–42·3% of the initial cohort excluded). Although post-stratification weights limit sample distortion, it is possible that sample selection could bias our COVID-19 survey estimates. Specifically, previous data collected in Burkina Faso indicate substantial differences in contraceptive use estimates between telephone and face-to-face interviews.²⁹ Our results might similarly face such biases because of survey modality, as evidenced by large increases in contraceptive use. Furthermore, to minimise survey length, information about sexual activity was unavailable in the COVID-19 follow-up survey, which led to the exclusion of non-cohabiting partnered women from this analysis. The lack of data on women's sexual activity might have also impacted our estimates of women's need for contraception, as we assumed that women who were in union at baseline were and remained sexually active between surveys. Additionally, these analyses focus solely on women's contraceptive use, although we recognise the importance of couple reproductive decision making. Lastly, analyses were limited by small subgroup sample sizes, rendering large CIs, and preventing analysis for some subgroups (ie, nulliparous and young women).

Although these data were collected in the first few months of the COVID-19 pandemic, the combination of trends in need for and use of contraception could help policy makers understand potential trends in unintended pregnancy that might arise. Overall rates of unintended pregnancy might decline in rural settings of Burkina Faso and in Kenya, given substantial uptake of contraception. Our findings also identify subgroups of women who might have an increased risk of unintended pregnancy due to potential decreases in contraceptive use, specifically young women in Lagos and nulliparous women in Kinshasa.

In conclusion, our results provide early insights into the need for and use of contraception after restrictive COVID-19 lockdown measures across four settings in sub-Saharan Africa. Data suggest a minimal population-level effect on contraceptive coverage despite sustained demand for contraception, although trends vary by geography. Greater contraceptive coverage among women in need was observed in Kenya and rural settings in Burkina Faso, whereas younger women in Lagos and nulliparous women in Kinshasa might face heightened risks of unintended pregnancy related to declines in contraceptive coverage (albeit non-significant). Sustained use of contraception in most settings might reflect the fact that there were no contraceptive shortages (as commodity stocks had not run out yet) or that the effect of shortages was mitigated because of a high proportion of users relying on methods with prolonged duration of use (eg, injectables, implants, intrauterine devices, or multi-pack contraceptive pills). Although results from Kenya and Burkina Faso are encouraging, we warn against assumptions that women are protected from the future risk of unintended pregnancy, as our reported trends might not be sustainable throughout prolonged economic hardship and service disruption. As the health crisis turns into an economic crisis, we anticipate both the demand for contraception and supply of contraceptives will shift with decreasing levels of satisfied demand for contraception, resulting in increased unintended pregnancies.

Data sharing

Data are available by request at pmadata.org.

Declaration of interests

All authors report grants from the Bill & Melinda Gates Foundation during the conduct of the study.

Supplementary Materials