Impact of the 2013 WHO guidelines for screening and treatment of cervical precancerous lesions on women’s screening rates, by HIV-status, in East and Southern Africa: A regression discontinuity design analysis

Introduction

Cervical cancer is the leading cause of death among women (1,2). According to recent global estimates, East and Southern African regions had the highest incidence of new cervical cancer cases across all regions (40 and 36 cases per 100,000 women years, respectively), and Africa had the highest mortality rates globally in 2020 (2). In 2013, the World Health Organization (WHO) issued comprehensive guidelines since 2010. It recommended cervical pre-cancer screening among women aged 30__49, and sexually active girls and women as soon as they test HIV-positive to prevent invasive cancer among women living with HIV (WLHIV). WLHIV are at higher risk of developing cervical cancer than women with an HIV-negative status (3). The guidelines indicated that the evidence for screening women with an HIV positive status to prevent cervical cancer was of lower quality than that for women with a negative or unknown HIV status (3). Many African countries (38 of 54) formulated cervical cancer prevention and treatment strategies based on the WHO guidelines (4). The WHO developed the guidelines from randomised controlled trials, systematic reviews, and expert guidance, among others (1,5). However, the impact of the 2013 WHO cervical pre-cancer screening and treatment guidelines and the age to start screening at the population level has not been studied.

Studies have examined the policy-level impacts of the WHO guidelines to expand the HIV treatment eligibility criteria and the “Treat all” guidelines to provide immediate anti-retroviral therapy (ART) to people living with HIV (PLHIV) regardless of their CD4 count (6,7,8). These studies found that the guidelines expanded the enrolment and retention of PLHIV in care (6,7,8). Research is required to identify whether the 2013 WHO cervical pre-cancer guidelines changed dynamics around screening for vulnerable women in east and southern Africa. The guidelines may have helped decision makers choose and implement the optimal screening and treatment strategy applicable in their context. Conversely, the new guidelines might have led decision makers to defer screening, for adolescent girls and young women living with HIV (AGYWLHIV) to age 30 and over. Examining whether trends in screening changed after the 2013 WHO guidelines may inform screening efforts for 2020–2030.

The WHO developed the 2013 guidelines to guide decision makers about which screening and treatment strategy to use in their settings. It recommended an overall screen-and-treat strategy with the treatment for precancerous lesions provided soon after a positive screening test result instead of after a sequence of tests and examinations. Besides nine recommendations, the guidelines contain a decision-making flow chart for choosing the best screen-and-treat strategy for a particular setting. It includes flow charts for women with a negative, unknown, or positive HIV status. It did not specify the age to start screening women with an HIV-positive status (3). The WHO later recommended starting screening WLHIV starting at age 25 (5). Stratifying analyses by HIV status and age groups, we examined the impact of the age, 25, to start screening on screening rates in Ethiopia, Malawi, Rwanda, Tanzania, Zambia, and Zimbabwe. We then used a regression discontinuity design (RDD) analysis to examine the impact of the WHO guidelines on women’s screening rates. The RDD methodology allows us to estimate whether the guidelines impacted screening (10).

Methods

Results

Table 1 shows the sample, comprising 73179 women, 6680 (9.1%) with a positive, 4328 (5.9%) with an unknown and 62171 (85.0%) with a negative HIV status. The samples of women per country ranged from 10157 in Malawi to 15455 in Tanzania. WLHIV were generally older than women with an unknown or a negative HIV status. Self-reported screening rates were higher in all surveyed countries among WHLHIV (17.7%, n = 1179) than among women with an unknown (12.1%, n = 525) or negative HIV status (6.5%, n = 4022) (p<0.001). The screening rates ranged from 10.2% (n = 50) and 6.6% (n = 930) among WLHIV or women with a negative HIV status in Rwanda to 22.5% (n = 342), 14.9% (n = 155) and 9.6% (n = 816) among WLHIV, women with unknown, or negative HIV status in Zambia.

Over two-thirds of women were ever married in all surveyed countries. More than half resided in rural areas except in Ethiopia, where the survey was conducted in urban areas. Less than half of women had primary level education in Ethiopia, Zambia, and Zimbabwe, whereas more than half had primary level of education in Malawi, Rwanda, and Tanzania. More women were from households in the wealth quintile 5, except for Tanzania (Table 1).

Adolescent girls and young women living with HIV (AGYWLHIV) aged 15LJ24 reported screening rates that were at least two times lower than their peers with a negative or unknown HIV status in all countries. In contrast, adolescent girls and young women (AGYW) aged 15LJ24 with a negative or unknown HIV status reported higher screening rates than women aged 25LJ34 or 35LJ49 with a negative or unknown HIV status (Figure 1). The differences were significant (Supplementary Table 1).

• Download figure
• Open in new tab

Figure 1:

Self-reported cervical pre-cancer screening rates among women by country, HIV status, and age group, presented as percentages (PHIA 2015L18).

Figure 2 shows the number of women screened for cervical pre-cancer in all six countries by HIV status, and year last screened. The period of women’s screening ranged from 2008 to 2018. Beside Rwanda, all countries surveyed had data before and after 2014. Rwanda had data for women with unknown HIV status after 2014, which were excluded from the RDD estimation.

• Download figure
• Open in new tab

Figure 2:

Number of self-reported screened for cervical pre-cancer among women stratified by country, HIV status and year screened (PHIA 2015L18).

Table 2 shows no evidence the guidelines increased or reduced the screening rates in the analyses overall stratified by the women’s HIV status, or the country-level (Table 2), or by age group (results not presented). The ATE was small and less than 0.001% in the pooled analysis. In the country-level analysis, the ATE was small (<0.001%) for all women, and for women with a negative or a positive HIV status in Malawi, Rwanda, and Zambia (Table 2). The RDD was valid. The assumptions held for implementing the RDD method, for women with unknown or HIV-positive status for pooled analysis; not for all women or women with an HIV-negative status. Although the related change in ATE for WLHIV in Ethiopia and Tanzania were significant, the RDD was not valid (Supplementary Table 2).

Discussion

This study examined the impact of the 2013 WHO guidelines for screening and treating cervical precancerous lesions for cervical cancer prevention on women’s screening rates in six East and Southern African countries. We found no evidence that the guidelines negatively affected women’s screening rates. However, AGWLHIV reported screening less often than their peers with a negative, or unknown HIV status, or older women. The age to start screening WLHIV should be lowered from 25 to 15 to screen more AGYWLHIV aged 15_24. Countries should adopt, implement, and monitor the impact of the guidelines on screening.

Our main finding was that the 2013 WHO cervical pre-cancer guidelines did not reduce or increase women’s screening rates. This result applied to women with unknown or HIV- positive status in pooled analyses. Our result supports the evidence that shows that the recommendations often have heterogeneous impacts on objectives. The guidelines sometimes have unintended consequences (6,7,8,16,17,18). In the United States of America (USA), state insurance mandates to cover costs for Pap smear tests, increased Pap testing rates by 1.3 percentage points among women aged 19_64. The impacts were more pronounced in black and Latino women than in white women (18). In sub–Saharan Africa, the recommendations to expand ART eligibility increased ART initiation and retention in care with greater impacts on PLHIV who met or nearly met the eligibility criteria to access ART (6,7). The recommendation did not prevent people with advanced HIV disease from initiating ART (6,7). However, one study found that the “Treat all” guidelines may have reduced pre-ART CD4 monitoring, weakening advanced HIV disease management (8).

One reason for our findings is that several countries began their cervical cancer prevention programmes around 2008, with government support, followed by a rapid scale-up. Adopting and implementing the new guidelines might have not imposed significant resource reallocations in countries’ cervical pre-cancer screening programmes. The 2013 guidelines simplified the screening and treatment process, encouraging decision-makers to use the screening and treatment strategy suitable to their context. Further comprehensive cervical cancer control guidelines were launched in December 2014, providing additional support and guidance to decision-makers and stakeholders in accelerating cervical cancer screening and treatment of pre-cancerous lesions (19). Altogether, the WHO established the 2013 cervical pre-cancer screening guidelines accompanied by deliberate efforts, including training, and media outreach, to support decision-makers in countries in adopting and implementing the guidelines (3). This result suggests that the guidelines and the efforts surrounding their dissemination are crucial in ensuring decision-makers effectively manage screening programmes and prevent unintended consequences, such as, reducing screening for vulnerable women.

We found that AGYWLHIV reported a screening rate that was at least two times lower than that of their peers with a negative or unknown HIV test result, or older women. Their peers with a negative or unknown HIV status reported higher screening rates than older women with a negative or unknown HIV status. This result reveals a systematic exclusion of AGYWLHIV from screening, an intended result of implementing the 2013 WHO guidelines. Although the guidelines recommended screening sexually active girls and women immediately after they tested HIV-positive, decision-makers might not have prioritized screening AGYWLHIV. They might have only screened them if AGYWLHIV or their providers requested the screening (11,15). AGYWLHIV might have not been aware of their HIV-positive status and less likely to access ART, and associated services including cervical cancer screening than older WLHIV (20,21). Further, AGYWLHIV might have not accessed ART and screening because of stigma and discrimination associated with HIV. In this study, we found higher screening rates among older WLHIV than any group of women. Policy makers might have prioritised screening older WLHIV than AGYWLHIV or other sub-groups of women, showing that the countries followed the WHO recommendations.

However, the potential risk to cervical cancer for AGYWLHIV in East and Southern Africa is worth re-examining. AGYW aged 15_24 are vulnerable. They accounted for 81% of the 250,000 new HIV infections among young people in 2021 (22). AGWYLHIV may be at higher risk of human papillomavirus (HPV) infection, and cervical cancer than their peers without HIV (23). HPV takes longer to clear in AGYWLHIV than among their peers without HIV (24). More AGYWLHIV in East and Southern Africa could be screened by lowering the age to start screening WLHIV. In the USA, stake holders proposed lowering the age to start screening from 25 to 21 (25). One reason for this proposal is that cervical cancer is more prevalent among women aged 25_29 in the USA. Screening women younger than 25 would reduce the risk of cervical cancer in women aged 25_29 (25). A further reason is the lack of national cancer screening programmes to screen and detect cervical pre-cancer lesions among young women (25). The countries in our study do not have large scale comprehensive national screening programmes that reach a diverse group of women at risk of cervical cancer (4,11). The WHO included guidance for screening WLHIV aged 25_49 years in the 2021 guidelines (1,5). The 2021 guidelines excluded AGYWLHIV from screening, including those born with HIV, who may have an elevated risk of cervical cancer. It argued that the evidence on the utility of screening AGYWLHIV was limited and that WLHIV face more screening over their life with increased potential harm from multiple screenings (23). Lowering the age to start screening WLHIV to include 15_24 years-olds, and prioritizing AGWYLHIV in screening, may lead to screening younger women who do not need screening (25). The potential for over- screening and its associated risks may be justified in East and Southern Africa where screening rates are low and the risk to cervical cancer among AGWLHIV is high.

Our study has several limitations and strengths. Women may have recalled the responses inaccurately because they are self-reported. Countries are flexible in adapting the guidelines to their context (19). However, they may align their screening programmes to the WHO screening guidance. The results are valid for the respondents in the local areas of the regression, typical of RDD analyses. Although the RDD assumptions for this study were valid for women with unknown or positive HIV status in pooled stratified analyses, the sample size was small for several variables reducing the variability of the observations for the estimations. A strength of this study is that we chose the optimal bandwidths apriori using data-driven bandwidth selection procedure, preventing us from selecting biased bandwidths. Another strength is that we used quality population-based data with high response rates. The assumptions of the RDD were valid for women with unknown or HIV positive status. Therefore, we could make causal inference of changes in women’s screening rates based on the 2013 WHO guidelines for these group of women.

Conclusion

We found no evidence that the 2013 WHO guidelines for screening and treating pre- cancerous lesions for the prevention and treatment of cervical cancer reduced or increased women’s screening rates. However, AGYWLHIV reported screening less often than their peers or older women. Policy makers may need to lower the age to start screening women living with HIV from 25 to 15 years to screen more AGWYLHIV. Studies are required to examine the impact of the guidelines including the age to start screening women living with HIV on cervical cancer screening to screen more vulnerable AGYWLHIV.

Authors’ contributions

Conceptualization, DC. Methodology, DC. Validation, OK, JE, SA, AN, CO, JB, AS and OK. Formal analysis, DC, JE, SA, CO, JB, AN, AS and OK; data curation, DC writing—original draft preparation, DC.; writing—review and editing, DC, OK, JE, SA, AN, CO, JB, AS and OK; visualization, DC. All authors have read and agreed to the published version of the manuscript.

Conflict of Interest Declaration

The authors declare no conflict of interest.

Ethics declaration

This study did not require ethical review because the personal details in the data are de- identified, and the data is publicly available. It can be accessed with a request at https://phia-data.icap.columbia.edu