Evaluation of the self-sampling for cervical cancer screening in Bolivia

The development of cervical cancer depends on high-risk human papillomavirus (HR-HPV) persistent infection in the uterine cervix [1]. The 2–10 year transformation process leading to invasive cancer provides ample opportunity to detect, prevent and cure true precursor lesions [1]. Although cervical cancer is widely preventable, it is the fourth most common cancer among women throughout the world, being a real public health issue, especially in developing countries, as 85% deaths occur in low and middle income countries [2]. The situation in Bolivia is particularly alarming as the incidence, being 47 per 100,000 women, is estimated to be the highest in Latin America with a mortality rate of 21 per 100,000 women¹ (standardized incidence and mortality rates by age) [2].

Cervical cancer-related deaths have nevertheless declined significantly in developed countries because of extensive cytology screening. However, similar initiatives in developing countries have not been equally successful because of the complexity of the required elements, such as laboratory expertise, depending on high operating costs [3, 4]. It is now well known that the HR-HPV screening test has marked advantages over cytology screening test, with sensitivity of about 90% for detecting high-grade intraepithelial neoplasms and high negative predictive value [5, 6].

Prevention of cervical cancer in Bolivia is primarily based on Papanicolaou smear cytology test (Pap) and more recently on visual inspection under acetic acid (VIA) [7, 8]. Screening, offered by the first level of care, responsible for prevention, is available free of charge, for sexually active women until 64 years old [7]. Nevertheless, Pap smear coverage, from 2005 to 2016, do not exceed 16.6% and coverage of VIA in 2015 and 2016 does not exceed 19% [7, 8].

The health personnel estimated that 50 to 80% of Pap screened women were lost to follow-up, mainly because of delays in result delivery [3].

Bolivia’s Ministry of health has introduced the HPV vaccine in 2016 as an alternative to reduce the incidence of cervical cancer. In 2017, 80% of the target population (girls between 10 and 12 year old) were vaccinated with the first dose [9].

Bolivia is characterized by significant disparities between rural and urban areas, including education and access to basic health services as some of the variables contributing to this difference [10, 11]. Although these differences have been reduced in recent years, the inequalities still persist among the poorest population [12, 13]. Poor coverage of the Pap smear, Pap poor quality, low follow-up of Pap positive cases, lack of information on cervical cancer prevention, poor human resources in health and low credibility in the health system, besides economic, cultural, and social barriers have been identified as the main factors involved in the high cervical cancer incidence in Bolivia [3, 14]. Large coverage disparity has been observed when stratified by area of residence, being much lower in rural areas than in urban areas and increasing the risk of dying from cervical cancer up to three fold in dispersed rural areas. This is linked to various factors, including low educational, cultural and economic level and limited access to health services, diagnostic tests and treatment [15].

The screening coverage rate is a key component in cervical cancer prevention programs. Vaginal self-testing of HR-HPV could reduce screening barriers for those women and therefore increase their screening coverage [16–18]. Indeed, self-collection is easy to perform, provides privacy, is less embarrassing and more comfortable to patients compared to samples collected by health personnel [19, 20]. Furthermore, self-collected samples have been shown to have sensitivity similar to that of samples collected by physicians [21]. We previously reported that a simple, cheap and transport safe method based on the smearing of vaginal and cervical cells on a glass slide (dry samples) using cotton swab gave satisfactory results to detect HR-HPV DNA, providing similar HR-HPV detection results as the physician collected samples [22].

The objectives of this study were to obtain information among the Bolivian women about their level of knowledge on HPV and cervical cancer and to evaluate the degree of the vaginal self-sampling test acceptability and confidence in comparison to gynecologist sampling at urban, peri-urban and rural areas, in order to evaluate the need for an adapted strategy within each three geographical areas, potentially corresponding to a characteristic population. Furthermore, this study assessed the impact of the HR-HPV self-sampling test on screening coverage in a restricted Bolivian peri-urban area.

This cross sectional study is based on the analysis of two surveys with a non-probabilistic stratified sampling by geographical areas (including disparities in access to education, health system and financial income). The evaluation of coverage impact through self-sampling was performed in a peri-urban area, with a mixed population (in regards to the disparities already mentioned above). The Bio-ethical committee of “Universidad Mayor de San Simon” approved the study protocol (October 30th, 2014). The signed informed consent was an indispensable requirement for the inclusion of patients. Pregnant women over 20 weeks, calculated on the basis of last menstrual period, and women with hysterectomy were excluded from this study. This study was divided in three phases. For surveys, a person speaking Quechua and Spanish was available to read and help to fill the survey files in case of analphabetism.

In order to reduce the cervical cancer incidence and mortality in the department of Cochabamba in Bolivia, we aimed to develop and assess a low cost HPV self-sampling screening. We believe that this strategy could improve the poor screening programs developed in our country. Not only the HPV DNA detection rate in self-sampling has been reported to be in good agreement with physician-sampling, but also the self-sampling method seems better accepted among women [24, 25]. It could therefore enhance cervical cancer screening program coverage, a primary prevention strategy in countries with low resources [26].

In our study, in the Survey 1, we confirmed that the level of knowledge about cervical cancer and HPV, in the three geographical areas, was low, diminishing in function of the distance between the area of residence and the capital city (urban area), and of the education level (the null hypothesis is rejected). A study in Cameroon mentioned that education about HPV and cervical cancer is associated to a greater acceptability towards the HPV self-sampling test. The provided information had a positive impact on the patient’s concerns, about her ability to perform the test [27]. Therefore, not only reduced access but also lack of knowledge about these issues could potentially participate to the low cervical cancer screening coverage observed in the Bolivian remote areas. This concern should therefore always be addressed in future screening campaign in Bolivia.

According to the results obtained in the survey 2, most of the women found, in the three geographic areas, self-sampling easy, comfortable and without pain. The level of the self-sampling acceptance was thus good, allowing to consider it as a promising screening tool to break sociocultural barriers. A similar result was observed in a study performed on three different populations (Uganda, Nicaragua and India) as 75% women found that self-sampling was easy and 90% found it very comfortable [28]. However, we observed in survey 2 that there were no clear preference to use it as a form of cervical sampling for the detection of cervical cancer in the three areas, despite a tendency in the rural area to recommend it. This result denoted a distrust in the self-collected vaginal sample. Other studies concluded similarly and found that, despite great acceptability for self-sampling, many women were concerned about test accuracy to detect the lesions and even more concerned by the risk of not having taken correctly the sample [28, 29]. This result was unexpected as in our previous study in Bolivia, in a population of women having previous experience with Pap test at 76%, self-sampling was clearly preferred. This striking difference between this previous study and the present study might be due to the fact that in the previous study women were not attending health center and could have therefore more hesitation to visit a gynecologist [22], while in our study (Phase 2) women were already attending health center. The context and the population in which both studies were performed were thus clearly different and it could have had some influence on the collected responses.

A meta-analysis in which six studies were reviewed about self-sampling acceptability revealed that the majority of women interviewed preferred the self-sampling test which did not require a gynecological examination [25]. The vaginal self-sampling could therefore potentially increase cervical cancer screening coverage [29].

In our third study, assessing self-sampling on cervical cancer screening coverage in a peri-urban area (9th District), we observed up to a 4.7 fold higher screening coverage compared to the previous year Pap coverage. Similarly, another study in South America mentioned a 4 fold increase in screening coverage using self-sampling [30] and similar results were recently obtained in other countries of Latino America [31, 32]. Although, in our third phase study, we did not register the number of women rejecting self-sampling, we observed during the campaigns that most women were very likely to use it after a detailed explanation about the HPV infection and its relationship with cervical cancer. In agreement with our results obtained in the Survey 1 and 2, we believe that intervention with this strategy (primary information and awareness transfer before offering self-sampling) will provide more all women confidence to perform self-sampling, a strategy that could counteract their lack of self-assurance and even their lack of knowledge. Many women, not screened in one of the 8 health centers assigned to the 9th district, were screened by the vaginal self-sampling test. Although awareness campaigns were associated to the self-sampling offer and present in the health centers, these campaigns were not similar. It is therefore possible that this co-intervention was necessary to achieve a successful coverage. Campaigns performed through visits at home or in populated areas to offer only self-sampling were transitorily done by unsubsidized volunteers, while campaigns for Pap and HPV detection performed in health centers were permanently present under the form of posters and by health personnel sensitization. It is therefore clear, that further studies will be necessary to aim at maintaining similar good results with same limited resources. The follow-up of the HR-HPV positive patients was however low and unsatisfactory. This low level of follow-up was not due to a lack of health personnel, but mainly due to a lack of women concern (as the number of women going to health centers for follow-up was very low). Further investment in women encouragement should be provided in screening campaigns to improve the follow-up rate and keep the benefit of a successful coverage rate. It will be interesting to evaluate which barriers and factors are associated to the low observed follow-up rate. Although self-sampling screening could be beneficial to reach a higher screening coverture rate, the follow-up strategy should take into account the difficulty to motivate women to visit a health center. A screen-and-treat strategy (a screening based on HPV detection or VIA, followed by cryotherapy or LEEP, if necessary) could therefore be even more beneficial for this population. However, the cost, sensitivity and specificity of the strategy should be carefully evaluated.

If self-sampling would be routinely used in Bolivia, additional infrastructures and guidelines will be needed, especially to provide clear information, among others to prevent insecurity feelings during samplings, to increase appropriated patient engagement, to improve the follow-up rate, but also to reduce the risk of patient anxiety after a positive result diagnosis transfer [33].

In conclusion, the self-sampling is well accepted by the Bolivian surveyed women, regardless of demographic differences. This suggested that in Bolivia also, cervical cancer detection coverage can be expanded using this form of cervical sampling (self-sampling). Knowledge about cervical cancer and HPV should be optimized to improve cervical cancer prevention programs. This aspect is fundamental to improve confidence and motivation of women performing self-sampling Studies to assess follow-up efficacy in regards to screening strategy should be further performed.