Objective
We sought to determine demographic and behavioral factors associated with human papillomavirus (HPV) positivity in a community-based HPV self-collection cervical cancer screening pilot project.
Study design
HPV self-collected samples were obtained from 199 women aged 30-69 years in the impoverished urban Ugandan community of Kisenyi, during September through November 2011. Demographic and behavioral information was collected. Descriptive statistics and a logistic regression model were used to analyze factors associated with HPV positivity.
Results
There was overwhelming acceptance of HPV self-collection in this community. High-risk HPV prevalence was found to be 17.6%. Lower levels of formal education (adjusted odds ratio [AOR], 0.40; 95% confidence interval [CI], 0.08–2.03) were associated with higher prevalence of HPV as was use of oral contraception (AOR, 2.01; 95% CI, 0.83–4.90) and human immunodeficiency virus status (AOR, 0.43; 95% CI, 0.14–1.37).
Conclusion
Screening should be targeted and prioritized for women with lower levels of education, oral contraceptive use, and human immunodeficiency virus positivity as they have the highest HPV prevalence in this low-resource population.
Each year over half a million cases of cervical cancer occur worldwide with the greatest burden in the developing world where 85% of deaths occur. Over 3500 women in Uganda are diagnosed with cervical cancer each year and 2500 women will die from their disease. In Uganda, cervical cancer is the most common cancer among women. Seventy percent of cervical cancers are potentially preventable by a vaccine making cervical cancer one of the most preventable and treatable of the gynecological cancers. Despite renewed global calls to address growing rates of cancer in low- and middle-income countries, rates of cervical cancer in Uganda are predicted to rise by 80% by the year 2025. Uganda’s age-adjusted incidence rate of cervical cancer is 47.5 per 100,000, one of the highest in the world. These unacceptably high rates are a result of significant health and human resources challenges that limit the availability of screening programs including lack of trained providers, lack of cytotechnologists, competing health priorities, and lack of political will. Although cytology-based screening has been in place for decades in the developed world, due to the complex infrastructure required for this type of screening, it is not a feasible option in sub-Saharan Africa. High-risk strains of the human papillomavirus (HR HPV) are known to cause >95% of invasive cervical cancer. There are currently no human papillomavirus (HPV) screening programs available in Uganda; cytology and visual inspection with acetic acid–based programs are also very limited. HPV DNA testing as a screening test performs as well as or better than standard cytology-based screening. There is increasing evidence for the accuracy of self-collection of HPV specimens for cervical cancer screening in marginalized and low-resource populations as an option for decreasing barriers of embarrassment and limited provider training in pelvic examination. Although several studies have explored self-collection as part of a clinic-based program, few have evaluated community-based screening programs with women collecting samples in their homes and places of work and none have done so in sub-Saharan Africa. Our hypothesis was that self-collection would be acceptable and feasible and that a substantive proportion of the women would be HR HPV positive. Our objectives for this pilot study were to determine: (1) acceptability and feasibility of an HPV self-testing approach, (2) the prevalence of HR HPV in this population, and (3) factors associated with HR HPV to inform future program development.
Materials and Methods
Study design
Setting
Kisenyi is an impoverished neighborhood in central Kampala, Uganda, with a residential population of approximately 40,000 people, significantly increasing during the day, due to a mobile workforce. Our study was designed to assess the feasibility of implementing home-based HPV self-collection with specific detail given to demographic and behavioral factors that increase a woman’s risk of being HR HPV positive.
Recruitment
A community stakeholder workshop was organized at the Kisenyi Health Unit in September 2011 that engaged community leaders, health care providers, and eligible women. Ethical approval for the study protocol was obtained from the institutional review boards of both Makerere University, Kampala, and the University of British Columbia, Vancouver, Canada, prior to study recruitment. Local community outreach workers, who were hired and trained in survey implementation and cervical cancer education, approached women and offered cervical cancer screening. Women were recruited opportunistically by community outreach workers going into all subvillages; no study participants were recruited at the local health unit. Representative study participants from daytime residents of each of the 4 subvillages in Kisenyi were invited to participate, a method previously effectively demonstrated by our research team. As is common with opportunistic sampling, as people become aware of a study they may be more likely to seek involvement. The number of women approached in each subvillage was recorded to calculate self-testing uptake rates. After enrollment and completing informed consent in the most common local language, Luganda or English, community health workers obtained contact and demographic information and administered a survey of attitudes, knowledge, and beliefs of cervical cancer and HPV.
Survey
The survey tool had previously been tested in Kisenyi and includes demographic data as well as knowledge attitudes and beliefs around HPV, cervical cancer, and screening; it has been described in detail elsewhere. After completing the survey, women were offered the opportunity to provide a self-collected specimen. The community health workers briefly explained how to perform the collection aided by a diagram. Women collected the specimen at the location of recruitment, either in their homes or places of employment.
Follow-up
The specimens were then labeled with a study number and date of birth and transported to Mulago Hospital’s HPV processing laboratory at the end of each day. Women with positive HR HPV results were contacted and invited to attend colposcopy follow-up at Mulago Hospital on specific days, receiving reimbursement for transportation. Treatment was provided at the colposcopy clinic if findings of cervical intraepithelial neoplasia ≥2 were confirmed on biopsy. All study participants were encouraged to attend human immunodeficiency virus (HIV) counseling and testing. All participant information and survey responses were entered into an Access database (Microsoft Office 2010, version 14.0; Microsoft, Redmond, WA) on a password-protected computer; hard copies of the surveys were kept in a locked filing cabinet in the project office.
Participants
Women aged 30-69 years who lived or worked in Kisenyi were eligible for participation. This age range was chosen based on natural history of HPV and cervical dysplasia. Informed consent was required and consent forms were completed in Luganda. If a participant was illiterate, a third party read the form to her and both verbal consent and a thumbprint were obtained. Women who were unable to provide informed consent; were not fluent in Luganda, Swahili, or English; or had undergone a previous hysterectomy were excluded from the study. Our targeted sample size prior to a scaled-up study was based on a 70% uptake rate of HPV self-collection. Approximately 20,000 women between the ages of 30 and 69 years live in Kisenyi; assuming HPV prevalence of 20% from other sub-Saharan populations, a sample size of 200 would give a 95% confidence interval (CI), ±5.5% around the population estimate of HPV prevalence.
Laboratory methods
Qiagen’s care HPV (Hilden, Germany) platform was used to perform the HPV DNA analysis. As a platform designed specifically for HPV DNA testing in low-resource settings, it is simple to run with a relatively fast turnaround time. This platform has the ability to detect 14 types of HR HPV (16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66, and 68), using polymerase chain reaction amplification and reporting results as HR HPV positive or negative. care HPV equipment at Mulago Hospital was in place as part of an ongoing study guided by Makerere University and the Program for Appropriate Technology in Health. The laboratory technicians were employed by Makerere University and received an additional stipend to process the study specimens. Specimens were stored at room temperature in the laboratory until enough were collected to run a batch of 90. If it was not possible to perform the HPV DNA analysis within 7 days due to laboratory workload, the specimens were frozen for DNA processing at a later time. All specimens were processed within 4 weeks of collection, well within acceptable limits for the care HPV platform. Results were provided by the laboratory technicians in written form with the participant’s study number and a positive or negative result then entered into an Excel spreadsheet (Microsoft Office 2010, version 14.0) on a password-protected computer.
Statistical analysis
Women who submitted a swab for HPV evaluation were included in the final analysis (n = 199). Statistical analyses were conducted using software (SPSS, version 14.0; IBM Corp, Armonk, NY). CI were calculated by using the Clopper-Pearson or “exact” method. This method provides more reliable CI with small samples. The exact method was designed to guarantee at least 95% coverage. Contingency tables were generated to summarize observations on categorical variables and for assessing possible association between individual characteristics and HPV-positive infection. Tests of association between HR HPV and individual demographic characteristics were based on the χ 2 of independence. Characteristics significant at the P < .10 level based on a Wald χ 2 statistic were considered in the multiple logistic regression model; also included were characteristics deemed vital to understanding HR HPV risk after a comprehensive review of peer-reviewed literature on the subject, these included marital status, oral contraceptive use, HIV seropositivity, and selected attitudes to screening.
Results
Of the 205 women who were enrolled in the study, 199 provided an HPV self-collected specimen. Reasons for not performing self-collection included concerns regarding menstruation and pregnancy. Women’s median age was 37 (interquartile range [IQR], 32–46) years. Further participant demographics shown in Table 1 are representative of national statistics with a mean age of first coitus of 16 (IQR, 14–16) years and median number of pregnancies of 5 (IQR, 3–6). The majority of the women (48.7%) had completed primary school and may or may not have done some secondary schooling, while 23.2% had no schooling or only some primary schooling. Nearly 60% of women worked outside of the home with >70% renting and not owning their homes. The majority of women, >60%, lived <30 minutes’ walk of the Kisenyi Health Unit. Christianity was reported as the religion for 58.5% of women and 39.1% were Muslim, representing urban distributions in Uganda. The vast majority of women (>97%) had been sexually active with a male partner at some point in their life yet only 8% of women had ever had any type of pelvic examination.
| Demographics | Sample size, n (%) | Prevalence, n (%) | (95% CI) | P value a | Unadjusted OR (95% CI) |
|---|---|---|---|---|---|
| Overall (aged 30-69 y) | 199 | 35 (17.59) | (12.57–23.6) | ||
| Age groups, y | |||||
| 26-30 | 25 | 5 (20) | (6.83–40.7) | 1.2 (0.417–3.45) | |
| 31-35 | 58 | 15 (25.86) | (15.26–39.04) | 2.11 (0.992–4.488) | |
| 36-40 | 36 | 6 (16.67) | (6.37–32.81) | 0.924 (0.352–2.423) | |
| 41-45 | 29 | 3 (10.34) | (2.19–27.35) | .435 | 0.498 (0.142–1.746) |
| 46-50 | 24 | 2 (8.33) | (1.03–27) | 0.391 (0.088–1.747) | |
| 51-55 | 16 | 3 (18.75) | (4.05–45.65) | 1.089 (0.293–4.044) | |
| 56-69 | 11 | 1 (9.09) | (0.23–41.28) | 0.453 (0.056–3.658) | |
| Marital status b | |||||
| Single | 64 | 12 (18.75) | (10.08–30.46) | 1.124 (0.519–2.431) | |
| Common-law union | 84 | 11 (13.1) | (6.72–22.22) | .049 | 0.571 (0.263–1.243) |
| Married | 21 | 2 (9.52) | (1.17–30.38) | 0.463 (0.103–2.084) | |
| Divorced, separated, widowed, no domestic partner | 29 | 10 (34.48) | (17.94–54.33) | 3.053 (1.272–7.326) | |
| Education of participant | |||||
| None or some primary | 64 | 15 (23.44) | (13.75–35.69) | 1.76 (0.833–3.72) | |
| Primary with or without some secondary | 97 | 15 (15.46) | (8.92–24.22) | .312 | 0.75 (0.359–1.566) |
| Secondary or postsecondary | 38 | 5 (13.16) | (4.41–28.09) | 0.622 (0.238–1.836) | |
| Work outside of home | |||||
| No | 84 | 19 (22.62) | (14.2–33.05) | .111 | 1 (Referent) |
| Yes | 115 | 16 (13.91) | (8.17–21.61) | 0.553 (0.265–1.153) | |
| Live in Kisenyi | |||||
| No | 38 | 5 (13.16) | (4.41–28.09) | .425 | 1 (Referent) |
| Yes | 161 | 30 (18.63) | (12.94–25.52) | 1.511 (0.545–4.195) | |
| Religion groups b | |||||
| Christian | 117 | 17 (14.53) | (8.7–22.24) | .288 | 0.604 (0.29–1.258) |
| Muslim | 77 | 17 (22.08) | (13.42–32.98) | 1.637 (0.785–3.414) | |
| Time to walk to nearest health center, min b | |||||
| <30 | 21 | 21 (17.5) | (11.17–25.5) | 0.985 (0.467–2.075) | |
| 30-60 | 8 | 8 (18.18) | (8.19–32.71) | .654 | 1.053 (0.441–2.518) |
| >60 | 3 | 3 (10.71) | (2.27–28.23) | 0.521 (0.148–1.833) | |
| Have had sexual intercourse with male partner b | |||||
| Yes | 193 | 33 (17.1) | (12.07–23.17) | .185 | 0.309 (0.05–1.925) |
| No | 5 | 2 (40) | (5.27–85.34) | 1 (Referent) | |
| Ever had pelvic examination | |||||
| Do not know | 13 | 2 (15.38) | (1.92–45.45) | ||
| No | 170 | 30 (17.65) | (12.23–24.22) | .971 | 1 (Referent) |
| Yes | 16 | 3 (18.75) | (4.05–45.65) | 1.089 (0.293–4.044) | |
| HIV positive | |||||
| No | 180 | 29 (16.11) | (11.06–22.31) | .092 | 0.416 (0.146–1.184) |
| Yes | 19 | 6 (31.58) | (12.58–56.55) | 1 (Referent) | |
| Wealth index (quintiles) | |||||
| Poorest | 23 | 5 (21.74) | (7.46–43.7) | 1.352 (0.466–3.925) | |
| Second | 41 | 10 (24.39) | (12.36–40.3) | 1.716 (0.748–3.94) | |
| Middle | 70 | 14 (20) | (11.39–31.27) | .288 | 1.286 (0.608–2.72) |
| Fourth | 22 | 2 (9.09) | (1.12–29.16) | 0.436 (0.097–1.959) | |
| Wealthiest | 43 | 4 (9.3) | (2.59–22.14) | 0.414 (0.137–1.244) | |
| Ever had sexual intercourse | 199 (97.07) | ||||
| Ever had pelvic examination | 16 (7.80) | ||||
| HIV positive | 19 (9.55) | ||||
| Willingness to collect own sample for HPV testing | 192 (93.66) | ||||
| Age at interview, median (IQR), y | 37 (32–46) | ||||
| Age at first sexual intercourse, media (IQR), y | 16 (14–16) | ||||
| No. of pregnancies, media (IQR) | 5 (3–6) | ||||
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
