1.1.1. Human papillomavirus and preventable deaths from cervical cancer.

In May 2018, the WHO Director-General declared a Joint Global Commitment to Cervical Cancer Elimination. In the same year, approximately 311,000 women died from cervical cancer worldwide and an estimated 570,000 were newly diagnosed with the condition [1]. Around 90% of the deaths occurred in low- and middle-income countries, and cervical cancer is consistently the first or second most common cause of female cancer-related death in sub-Saharan Africa [1]. Based on current trends, the number of deaths from cervical cancer in Africa are expected to double by 2040, while modelling suggests that Australia, and certain other high-income countries, are on target to reach elimination goals over the same timeframe. Thus, cervical cancer is a condition of profound and increasing inequality.

HPV vaccines will be critical to reversing these trends and ultimately to reaching elimination targets in sub-Saharan Africa. There are currently six licensed HPV vaccines. Cervarix^®, Cecolin^® and Walrinvax^® are bivalent vaccines targeting high-risk HPV types 16 and 18. Gardasil^® and Cervavac^® is are quadrivalent vaccines, additionally incorporating HPV types 6 and 11 which cause anogenital warts, while Gardasil9^® is a nonavalent vaccine including a further five high-risk HPV-types (31, 33, 45, 52 and 58) [2,3]. In clinical trials undertaken in high-income countries, these vaccines demonstrated high efficacy against vaccine-type cervical cancer in HPV-naïve 16–26-year-olds [4]. Varying degrees of cross-protection against vaccine-related HPV types have also been reported. Immunobridging studies, based on the comparison of HPV-neutralizing antibody titres, have subsequently been used to support the use in 9–14-year-old females which the WHO recommend be rolled out globally [3,4]. These vaccines were first licensed as a three-dose schedule and later, a two-dose schedule was approved based on immunogenicity and effectiveness data [3]. Currently, a single-dose HPV vaccine schedule has been allowed for by the WHO for most of the HPV vaccines, based on efficacy, effectiveness, and immunogenicity data from several observational studies and clinical trials [3,5–12]. Monitoring the population-level impact of the vaccines in the context of expanding programmatic use in West- and elsewhere in sub-Saharan Africa is now critical. There are no efficacy data from this setting. Furthermore, geographical variation in the prevalence of certain HPV types, as well as distinct, but poorly characterized, behavioural, biological, and socio-demographic exposures in sub-Saharan Africa are likely to influence vaccine impact and make extrapolating data from other settings unreliable. Any impact of pathogen and host-genetic diversity on transmission and persistence is similarly unknown. Meta-analyses of data generated in a limited number of high-income countries indicate that, while the prevalence of vaccine-type HPV has been reduced following vaccine introduction, cross-protection against non-vaccine-type HPV is dependent on achieving high coverage [13]. Similarly, a meta-analysis examining the impact of vaccine introduction on the prevalence of non-vaccine type HPV reported cross-protection against HPV 31, but little or no reduction in HPV 33 and HPV 45, as well as a trend towards an increase in certain high-risk, non-vaccine HPV types [14]. Also, a recent systematic review, reported an increase in HPV vaccine effectiveness, according to the number of vaccine doses received, which would not be predicted based on immunogenicity data from clinical trials alone [15]. Findings from two large reviews of single-dose HPV vaccination schedules indicated that a single dose was highly effective against HPV 16/18, and although the antibody titre levels were stable over the 10-year follow-up periods they were significantly lower than those generated by two- or three-dose HPV vaccine schedule [5,10]. While a single-dose HPV vaccine will significantly simplify logistics, reduce cost, and ease supply constraints, it will be crucial to assess longer-term immunity, potential waning effects, and the context-specific population-level impact of the vaccines following introduction. The reports emphasize the limitations of the data available, even from high-income countries, and the significant risks of bias in current analyses due to the lack of contemporaneously collected data on other exposures. They emphasize the importance of impact studies thoroughly characterising the baseline population prior to vaccine introduction to allow future estimates to be appropriately adjusted [13,15]. In addition, modelling studies have been used extensively to predict the population-level impact of HPV vaccine introduction and have been highly influential in guiding policy [16]. Factors demonstrated to substantially influence estimates of impact include baseline type-specific HPV prevalence and age distribution; individual level vaccine efficacy and age-stratified population coverage; sexual activity patterns including mixing between age groups and sexual activity risk groups; and co-morbidities including malnutrition and other STI. These data further reinforce the fact that impact data from high income countries cannot be assumed to translate to low-income settings in West-Africa, given that differences in many of these factors are likely, but unquantified.

1.1.1.1. Rationale and impact: There are no systematically collected data on HPV prevalence in The Gambia. However, in a small study in 20–49-year-old females attending a sexual health clinic in Banjul, the urban capital of the country, HPV 52, which is not included in Gardasil, was the most common high-risk type detected [17]. Only one vaccine-type HPV (HPV 16) was detected amongst the 18 high-risk HPV types reported. In a second study, conducted in a rural setting in The Gambia in 1999, HPV 16 and 18 accounted for 30 of the 93 (32%) of the high-risk HPV types detected in 15–54-year-olds [18]. Establishing sufficiently precise, baseline population estimates of HPV prevalence, and characterizing associated exposures, will allow vaccine impact to be prospectively monitored and confounding minimized. Such data will also enhance predictions from current models, through their context-specific calibration. The dynamic nature of current HPV vaccine policy, as emphasised by the latest WHO Strategic Advisory Group of Experts (SAGE) working group on HPV immunization [19], underscores the critical importance of timely, context-specific data in shaping vaccine policy in West Africa. SAGE has recommended a single-dose schedule in younger women, reflecting a considerable interest in this approach, as well as a two-dose schedule with a 6-month interval in women older than 21 years [19,20]. As of the end of 2024, 20 African countries (excluding The Gambia) have adopted the WHO’s single-dose regimen in their national immunization program [21]. While this schedule offer advantages, in terms of facilitating coverage, and addressing supply and cost constraints, robust data to monitor its impact is essential. Furthermore, with new HPV vaccine manufacturers emerging from developing countries, and their vaccines pre-qualified by WHO, the availability of HPV vaccines is expanding for supply to Gavi-supported countries. While these vaccines also have the potential to overcome supply problems and to reduce cost, ensuring they offer comparable longer-term protection, particularly cross- and herd-protection, given, for example, the use of different adjuvants, will be essential. Modelling studies, to examine the value of higher valency vaccines, and of catch-up vaccinations across wider age cohorts also need context-specific data on HPV types, age distributions, and other exposures from West Africa, for their recommendation to be applied here.

The Gambia introduced a two-dose HPV vaccination schedule to be administered to 9–14-year-old girls in 2019 as a national school-based campaign program [21]. But this program has not been sustained, was largely interrupted by the COVID-19 pandemic and the coverage has been very low, 23% for the first dose in 2024 [21]. This presents a key opportunity to generate these vital baseline data. Fig 1 below illustrates the age of the future HPV-vaccinated female cohort year-on-year.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. HPV vaccine introduction and future age coverage.

https://doi.org/10.1371/journal.pone.0332842.g001

Additionally, measuring the prevalence of other STIs – NG, CT, TV, MG, syphilis, as well as HIV, hepatitis B and C, will strengthen data quality, enable timely surveillance updates, and improve disaggregated reporting to inform STI and hepatitis programme delivery. These efforts are significant in Western and Central Africa, where service coverage and funding remain limited according to the WHO [22].

1.1.2.. Antimicrobial resistance.

The WHO warns that NG may soon become untreatable due to the global spread of resistance to ceftriaxone and azithromycin, alongside pre-existing resistance to other agents [23]. However, sub-Saharan Africa is underrepresented in WHO-GASP (WHO Global Gonococcal Antimicrobial Surveillance Program), with only 2 of 47 countries submitting AMR data. This gap is critical in West Africa, where syndromic STI management, unregulated antibiotic access, and limited contact tracing likely accelerate resistance. Whole-genome sequencing now reliably predicts AMR in NG and allows mapping of resistant strain spread [24]. The approach also allows high-resolution molecular typing to be undertaken and thus local, regional, and international spread of resistant strains to be mapped [24]. The proposed study will provide the first data of this type from sub-Saharan Africa.

For MG, resistance to azithromycin and emerging resistance to moxifloxacin are growing concerns. Yet, only 72 of 8,966 MG samples included in a recent global AMR meta-analysis were from sub-Saharan Africa—none from West Africa [25].

1.1.2.1.Rationale and impact: Findings will directly inform antibiotic recommendations for syndromic STI treatment in West Africa, addressing a key driver of AMR. Moreover, they will support the WHO Global STI Vaccine Roadmap, which calls for better data on NG burden and resistance to model vaccine impact and accelerate development [26].

2.2.1. Primary objectives.

• To establish the baseline, population prevalence of high-risk and low-risk, vaccine-type, and non-vaccine type HPV infection through the measurement of urinary HPV-DNA.
• To quantify the effects of the following exposures on HPV prevalence:
• Age at sexual debut; number and age(s) of life-time sexual partner(s); sexual mixing patterns including within and outside polygamous marriages; use of barrier and other contraceptives.
• The presence of other STIs.
• Age at menarche, parity, age at first and subsequent pregnancies, breast feeding.
• Cervical cancer screening and/or treatment.
• Other socio-demographic characteristics (education level, employment, access to water and sanitation facilities, smoking, nutritional status).

2.2.2. Secondary objectives.

• To establish the population prevalence of STIs – Neisseria gonorrhoea (NG), Chlamydia trachomatis (CT), Trichomonas vaginalis (TV), Mycoplasma genitalium (MG), syphilis, human immunodeficiency virus (HIV), hepatitis B and hepatitis C, and their relationship to common symptoms of STIs and other adverse health outcomes related to STIs (e.g., stillbirths, infertility) [29–33].
• To determine the frequency of antimicrobial resistance alleles in NG and MG and associated strain variation

2.2.3. Exploratory objectives.

Exploratory objectives which may be explored within the survey include:

• To determine the seroprevalence of HPV types and other relevant infections (e.g., herpes simplex virus) as a marker of past infection and/or vaccination.
• To establish pathogen and host genetic diversity and their impact on infection prevalence.
• To measure the clustering of HPV types and other STIs, within polygamous marriages and other relevant groups, as a marker of transmissibility.

In examining the clustering of infections within polygamous marriages and other relevant groups, a limited number of males will also be recruited.

2.3.1. Primary.

• HPV types 16, 18, 31, 33, 45, 52, 58, 26, 35, 39, 51, 53, 56, 59, 66, 68, 69, 73, 82 (high-risk types), 6, 11, 40, 42, 43, 44, 54, 61, and 70 (low-risk types) urinary prevalence rates. (Gardasil® has been used in The Gambia and contains HPV 6, 11, 16 and 18; Gardasil 9® includes HPV 6, 11, 16, 18, 31, 33, 45, 52, 58; Cervarix® contains HPV 16 and 18; Cecolin® included HPV 16 and 18.)

2.3.2. Secondary.

• NG, CT, TV, MG urinary prevalence rates.
• Syphilis, HIV, hepatitis B and hepatitis C seroprevalence rates.
• Frequency of known AMR alleles in NG and MG.

2.3.3. Exploratory.

• HPV type-specific and additional infection (e.g., herpes simplex virus) seroprevalence rates.
• Pathogen and host genetic diversity as measured by targeted (e.g., human leucocyte antigen typing) or whole genome sequencing (e.g., NG).

2.4.1. Overall design summary.

We will conduct a probability-based, multistage cluster survey of females aged 15–49 years using the Basse Health and Demographic Surveillance System as the sampling frame, accounting for clustering at district, village, compound, and household levels. Villages will be selected with probability proportional to size. Following informed consent, data on sociodemographic, obstetrics, sexual mixing, and exposures associated with HPV, other STIs, HIV, and hepatitis B/C will be collected via face-to-face interviews and ACASI in local languages for sensitive topics. First-void urine (Colli-Pee™) will be tested for type-specific HPV, Neisseria gonorrhoeae, Chlamydia trachomatis, Trichomonas vaginalis, and Mycoplasma genitalium; HIV, syphilis, and hepatitis B/C will be diagnosed using WHO-prequalified rapid tests. Whole blood will be stored for plasma and DNA. A sub-study of polygamous marriages and other sexual mixing groups will include both sexes to compare HPV/STI transmissibility, with additional targeted sampling as needed to refine estimates.

2.5.1. Eligibility.

Eligible participants are self-identified females aged 15–49 years, resident in a randomly selected household within the Basse HDSS at the survey visit (defined as having slept there the previous night), able to comply with study procedures, and providing written informed consent/assent. Age will be self-reported or, when available, verified with official documents. This definition includes transient populations. For the sub-study, males in polygamous marriages and other specified groups meeting the same criteria will also be enrolled.

2.5.2. Sampling strategy– population prevalence survey.

The primary aim of the survey is to generate overall estimates of the population prevalence of individual HPV types as well as high- and low-risk, vaccine- and non-vaccine HPV types in 15- to 49-year-olds in the Basse HDSS region. Estimates will be generated for the other STIs, HIV, Hep B and C using the same sampling methodology.

To minimize bias, a probability-based, multistage, random sampling process will be employed. The Basse HDSS will be the sampling frame for the survey. As the sample will be large (>10,000), no age stratification will be necessary, and it is expected that the age distribution will be similar to values in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Age and sex distribution of residents in the Basse HDSS.

https://doi.org/10.1371/journal.pone.0332842.t001

2.5.3. Participant recruitment – polygamous marriage and other targeted recruitment.

Participants not selected through the random selection process above and who will not therefore be part of the primary prevalence survey may be approached to take part in the sub-study which aims to better understand the dynamics of transmission in defined sexual mixing groups – predominately polygamous marriages. This will include residents in households, compounds or villages not selected for the prevalence survey. In this case the aim will be to recruit the male (generally the husband) and co-wives. At least two co-wives will be recruited for each male. Sensitisation, consent, as well as data and sample collection and follow-up, will continue in the same way as for the main survey. Given that these individuals are not selected at random in the same way, they will not contribute to the overall survey estimates.

2.6.1. Sensitization.

Selected villages will be visited for community sensitization before the survey, during which a central location (e.g., health post) will be identified for survey activities. Selected households will be approached for individual sensitization with all eligible females (15–49 years), usually in consultation with household heads. Consenting participants will be invited to the central location for data and sample collection.

2.6.2. Informed consent and assent procedures.

Written or thumb-printed informed consent will be obtained from all participants; for those <18 years, parental or guardian consent and participant assent will be required. Consent will be conducted in the participant’s preferred language, verbally translated by trained staff from an approved English script. A literate witness fluent in English will attest to translations. If assent is withheld, minors will not be enrolled regardless of parental views.

Note: Only one parent or guardian will be required to provide written consent but in general both parents should agree to participation. The Gambian local languages are not written thus written translation has been proven to be ineffective.

2.6.3. HIV and other pre-test and post-test counselling.

During informed consent, participants will be advised of planned testing. All will receive pre-test counselling from trained staff for HIV and hepatitis B/C in accordance with national and international guidelines. Ongoing verbal consent post-counselling will be documented; those declining testing will be referred to government counselling/testing services. Post-test counselling for HIV and hepatitis B/C will be provided based on point-of-care results, with onward referral as appropriate.

2.7.1. Questionnaire.

Sociodemographic and obstetric data as well as data on family structure and common symptoms of STIs [29–33] will be collected through face-to-face interviews (S1 Appendix) conducted using an online or offline custom-designed REDCap database.

2.7.2. Audio computer-assisted self-interview (ACASI).

Data on sexual history, sexual health, and other topics prone to social desirability bias will be self-administered via ACASI (developed by the survey team on a commercially available software platform provided by SurveyCTO) in Mandinka, Fula, Wolof, and Serahule—spoken but unwritten languages—precluding written English use. Visual prompts of familiar local objects on a tablet touchscreen will guide responses, limited to categorical (including ordered) options due to the inability to enter free text or numerical values. Despite these constraints, ACASI is considered essential for minimizing social desirability bias and enhancing data validity.

2.8.1. First void urine collection.

Participants will self-collect first-void urine using a ColliPee™ device, which captures a defined 20 mL volume into a nucleotide-preserving medium after at least two hours without urination. Residual urine is discarded to the toilet. Collection will follow standardized verbal and pictorial instructions. This method is established for HPV surveys and for detecting NG, CT, MG and TV, though not yet for guiding interventions.

2.8.2. Blood sample collection.

3.0mL blood samples will be collected into a clotted tube from all participants by peripheral venepuncture using standard aseptic technique. A small volume of the sample will be retained in the sampling syringe to be used for the designated point of care testing Fig 2.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Collipee Device.

https://doi.org/10.1371/journal.pone.0332842.g002

2.8.3. HPV genital swab collection.

In males recruited to the sub-study examining HPV and STI transmission within polygamous marriages, a saline wetted nylon flocked swab will be collected sampling the penile shaft, glans, foreskin (if applicable) and scrotum [34–36].

2.8.4. Point of care testing.

2.8.4.1. Pregnancy: A highly sensitive (detects hCG at a concentration of ≥ 25mIU/mL) urinary pregnancy test, certified against an appropriate regulatory standard, will be undertaken of urine samples obtained fresh in the clinic immediately prior to testing. The manufacturer’s instructions for the performance of the test will be followed.

2.8.4.2. HIV, hepatitis B and C, syphilis: WHO-prequalified point-of-care tests (POCTs) will be used to detect HIV-1/2, syphilis, and hepatitis B/C, with sensitivity and specificity >99% and minimal inter-reader variability. Given their diagnostic reliability and WHO-recommended use, no additional laboratory confirmation will be performed; however, plasma samples will be stored for potential later verification.

2.9.1. Urine nucleic acid extraction.

Nucleic acids will be extracted from the first void urine samples using a process optimized to maximise nucleic acid recovery and hence the diagnostic sensitivity of the downstream assays [37–40]. Genomic DNA of HPV, C. trachomatis, N. gonorrhoeae, M. vaginalis and (T. vaginalis), will be extracted from 2 mL of first void urine using the QIAamp DSP mini kit (CAT 61304). Extracts showing high RT-PCR cycle threshold (Ct>=35) will be repeated with twice the volume of urine (4 mL of first void urine).

2.9.2. Urinary HPV laboratory testing.

The Allplex^TM HPV28 Detection Assay (cat. no. HP10372X; Seegene) will be used to detect HPV DNA. This real-time PCR assay identifies 28 HPV types – 19 high-risk types (16, 18, 31, 33, 45, 52, 58, 26, 35, 39, 51, 53, 56, 59, 66, 68, 69, 73, and 82) and 9 low-risk types (6, 11, 40, 42, 43, 44, 54, 61, and 70). It covers HPV types included in the vaccines – Gardasil™ (the vaccine introduced in The Gambia [HPV types 6, 11, 16, 18]), Gardasil9™(6, 11, 16, 18, 31, 33, 45, 52, 58), and Cervarix™ (16, 18).

Although the assay is validated for cervical and self-collected vaginal specimens (but not urine), it is currently adopted by the International Agency for Research on Cancer (IARC) for use in HPV prevalence surveys [41]. Its adoption in the PHASE survey ensures comparability with other international surveys. This is further supported by our successful participation in the Global HPV Genotyping Proficiency Study, in which our assay was deemed proficient for detecting HPV types 6, 11, 16, 18, 31, 33, 45, 51, 52, 56, 58, 59, 68a and 68b (S4 Appendix).

Given variable sensitivities and applicability to clinical care, additional assays may also be undertaken for HPV genotyping on samples collected in the survey based on initial results and/or resource availability [42].

2.9.3. Urinary STI testing.

The commercial Allplex™ STI Essential Assay real time PCR (RT-PCR) assay will be used for the detection of CT, NG, MG and TV DNA (SD9801X; https://www.seegene.com/assays/allplex_sti_essential_assay). The assay is suitable for screening for NG, CT, MG and TV on first void urine samples. There are currently no POCT for these four infections which comply with the WHO ASSURED criteria – meeting all requirements of the given target product profiles for these assays [43].

In addition to the 4 four pathogens of interest, STI Essential Assay can detect three additional bacteria DNA: M. hominis, Ureoplasma parvum and U. urealyticum. RT-PCR results will not be reported for clinical care as routine testing and treatment for these pathogens is not recommended in symptomatic or asymptomatic individual by the European STI Guidelines Editorial Board.

2.9.4. Antimicrobial resistance testing.

Alleles known to confer antimicrobial resistance in NG and MG will be identified by RT-PCR based on established protocols [25,44]. Macrolide resistance-associated mutations (MRAMs), predominantly associated with 23S rRNA gene mutations at A2058 and A2059, and quinolone resistance-associated mutations (QRAMs), predominantly associated with parC gene mutations (ParC S83 and D87 codons) will be measured [44–46]. Commercially available RT-PCR kits (Seegene Allplex™ MG & MoxiR Assay [SD10233], Allplex™ MG & AziR Assay [SD10170] and Allplex™ NG &DR Assay [SD10368]) will be used for AMR profiling.

2.9.5. Additional exploratory outcome data.

2.9.5.1. Whole genome sequencing of NG: Whole-genome sequencing of NG (~100 randomly selected NG-positive samples) will be performed using a newly developed method generating complete genomes from first void urine. Genomic DNA will be extracted from first void urine samples using a customised protocol for host DNA depletion and/or NG DNA enhancement by species-specific DNA baits specifically designed against the 2016 WHO NG reference strains. Genomic DNA will be sequenced using Oxford Nanopore Technology (ONT) for long read sequencing. When necessary, non-NG sequence reads will be filtered out using the NG reference sequence. Dedicated in-house bioinformatics pipelines will be used for AMR determinants detection ([Comprehensive Antibiotic Resistance Database – CARD https://card.mcmaster.ca]; Antimicrobial Resistance [NG-STAR]),), for typing (NG-MAST [www.ng-mast.net], MLST [https://pubmlst.org/], and phylogenetic analyses.

2.10.1. Sampling frame.

The Basse HDSS (Dec 2023) census was used to define the sampling frame, identifying 59,950 eligible females aged 15–49 across 205 villages (population >30). Data were available at village, compound, and household levels. We assumed sampling coverage of 20%, 40%, and 50% at each respective level, inflating coverage for estimated loss to follow-up of 25%, 30%, and 30%. All eligible female subjects in a household were selected.

To improve geographic balance, the region was divided into four longitudinal tracts:

• West to East: −14.55 to −14.35 (9% of population),
• −14.35 to −14.1 (51%),
• −14.1 to −13.9 (27%),
• • −13.9 (13%).

Villages were selected using weighted simple random sampling (by population size × longitudinal weight), due to poor distribution. A 10,000-iteration simulation was run to minimize imbalance in regional sampling proportions. From the 10 best-balanced outcomes, one was randomly chosen. This process was repeated 1,000 times, yielding 55 final villages. Inclusion probabilities were calculated based on village selection frequency across simulations.

Within selected villages, compounds and households were selected by simple random sampling due to limited population accuracy at finer levels. This process generated a final sampling frame of 21,342 subjects.

2.10.2. Precision simulation.

To assess precision, we simulated 1,000 surveys assuming:

• HPV prevalence of 25%,
• Household-level intra-class correlation coefficient (ICC) of 0.1,
• Coefficient of variation: 20% (between villages), 10% (between compounds), 10% (between households).

We modelled 11 levels of missingness from worst-case (sample size = 5,686) to full retention (sample size = 21,342), reducing missingness from 100% to 0% in 10% steps. The worst-case scenario assumed 25% of villages, 30% of compounds/households, and 50% of participants were missing. Fig 3 displays 95% confidence intervals for HPV prevalence estimates across these scenarios. Results show precision is stable for 25% prevalence but would be more sensitive for rarer outcomes Fig 4.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Precision of HPV estimate versus dropouts.

https://doi.org/10.1371/journal.pone.0332842.g003

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Distribution of the chosen villages in the study area (south Bank of upper river region).

https://doi.org/10.1371/journal.pone.0332842.g004

2.10.3. Variance estimation and weighting.

Standard errors were estimated using Taylor series linearization to account for the three levels of clustering. Since compounds and households were selected by simple random sampling, their inclusion probabilities were based on the proportion selected at each stage. Village-level inclusion probabilities were derived from the proportion of simulations in which each village was selected.

To address potential bias from design-based weights, which could overemphasize compounds with extreme prevalence and low inclusion probability, a uniform compound-level weight was applied, assuming equal probability for all compounds within a village. This reduced the influence of outliers and improved estimate stability.

Final sampling weights were the inverse of the product of inclusion probabilities at village, compound, and household levels. A finite population correction was applied due to the large sample size.

2.10.4. Participant selection.

Participant selection (sampling frame) will be undertaken by the survey statistician, independent of the field team. The target sample size is a minimum of 10,000 participants, selected based on practical considerations and diminishing returns in precision beyond this threshold, as illustrated in Fig 3. Selected compounds in each village will be surveyed sequentially. To preserve feasibility and avoid over-sampling, no more than 75% of the eligible female population in any village will be surveyed. This will be monitored in real time by the survey team, and recruitment in a village will cease once this threshold is reached.

If 15- to 49-year-old females, who are not listed in the Basse HDSS database, are found to be resident in a household (slept in the household the previous night), they will be included in the survey. This approach helps capture transient populations who may have distinct infection risk profiles.

Villages, compounds, households, and individuals who decline participation will not be replaced. Similarly, missing individuals (non-existent participants due to HDSS inaccuracy, migrated participants, deaths etc) will not be replaced as this has been accounted for in the sampling frame. This pattern of missingness is expected to be consistent across villages. If the target sample size is not met, additional villages will be selected using the same procedure, excluding those initially sampled.

2.10.5. Sub-study of polygamous marriages.

This parallel-group study compares infection clustering within polygamous marriages to higher-level clusters (households, compounds, villages) and between infections, as a proxy for transmissibility/persistence. One arm comprises wives from polygamous marriages; the other, single wives randomly recruited into null clusters. Clusters are limited to sizes 2, 3, or 4 (weighted 10%, 40%, 50% respectively) to reflect population structure. For each arm, intracluster correlation coefficients (ICCs) will be estimated via ANOVA using Smith’s method [47,48], and differences assessed by combining variance estimates to derive 95% CIs [49]. A significant difference is defined as a lower 95% CI > 0. Power was assessed via simulation (≥1,000 iterations), generating designs with specified prevalence, ICCs, and effect sizes; power is the proportion of ICC differences where the lower CI exceeds 0. For example, with a prevalence of 15%, ICCs of 0.02 (compound) and 0.12 (marriage), and 500 clusters per arm, power exceeds 80%.

3.1.1. Staff and training.

The survey will be conducted by personnel with appropriate qualification, training, and experience to successfully undertake a survey of this nature in The Gambia. Training on all procedures will be undertaken and documented before any survey activities commence.

3.1.2. Standard operating procedures and study specific procedures.

All activities in the survey will be undertaken according to written survey specific procedures (SSP) and/or MRCG standard operating procedures (SOP) as appropriate. SSP outline in detail all aspects of survey conduct, including field, laboratory and data management activities and is written based on the approved protocol. They aim to ensure consistent, high standards of practice across the entire team and serve as reference documents for staff undertaking given procedures. MRCG SOP provide guidance on local ethical and regulatory requirements.

3.1.3. Laboratories and testing.

Point-of-care tests (HIV, hepatitis B and C and syphilis) that have been assessed and prequalified by the WHO will be used throughout. Additional testing (HPV and other STIs) will be carried out using commercially available validated assays established with appropriate controls at MRCG. We successfully participated in the Global HPV Genotyping Proficiency Study, in which our assay was deemed proficient for detecting HPV types (S4 Appendix).

3.2.1. Data queries.

The eCRF will incorporate on-entry error checks to flag missing, inconsistent, or implausible data in real time for immediate correction. Batch validation checks will be run regularly to ensure cross-field consistency, with queries resolved by field and laboratory teams. Cross-checks will verify alignment between REDCap and ACASI datasets and between recorded and stored biological samples.

5.6.1. Sexually transmitted infections (NG, CT, MG, TV, Syphilis).

Appropriate first-line antibiotic therapy will be provided for participants diagnosed with these STIs. Antibiotic choices will be based on international best practice, adjusted according to local guidelines and/or knowledge of antibiotic resistance patterns when appropriate.

5.6.2. HIV.

Following post-test counselling, participants will be referred on to their preferred HIV treatment centre (which may not be their nearest centre for reasons of confidentiality). When possible and requested, the first clinic attendance will be facilitated by the survey team.

5.6.3. Hepatitis B and C.

Participants testing positive for hepatitis B or C will be reviewed by the MRCG hepatitis clinic to assess disease severity and treatment eligibility. Periodic clinics within the study area, coordinated by the PHASE survey team, will facilitate these assessments. The MRCG team will provide treatment and follow-up as indicated. Pregnant women testing positive for hepatitis B will receive an alert card to support timely newborn hepatitis B vaccination, in line with prevention guidelines. Additional antiviral prophylaxis for preventing mother-to-child transmission is not routinely available in The Gambia.

5.6.4. HPV.

Although HPV testing is established in cervical cancer screening internationally, the validity of first-void urine testing for this purpose remains under evaluation [50–53]. WHO recommends cervical cancer screening from age 30 years, or age 25 years in women living with HIV, due to high spontaneous HPV clearance in younger women [51]. Furthermore, the use of HPV testing as a screening test for cervical disease in women under 30 years in any setting is not recommended based on the high rates of spontaneous viral clearance. Even in those over 30 years of age, HPV screening opportunistically and without rigorous pilot testing of the programme as a whole is not recommended [54]. In addition, Urinary HPV testing, while useful for population estimates (pooled sensitivity 87% [95% CI 78–92%], specificity 94% [95% CI 82–98%]), is not currently recommended for screening purposes [38,55,56].

Health promotion messages during the survey will emphasise accessing available screening services, particularly for participants testing HIV-positive and/or high-risk HPV (HR-HPV).

Following consultation with local experts and the International Agency for Research on Cancer (IARC), all participants aged ≥25 years with HR-HPV will be eligible for screening. This lower-than-WHO age threshold reflects local reports of rising cervical cancer incidence in younger women and was considered to outweigh the potential risk of overtreatment. Scientists at IARC agreed with the decision to lower the screening age in the context of this population-based survey after some deliberation.

Eligibility for cervical cancer screening will be based on the IARC classification of HPV types [57,58]. Specifically, only individuals infected with Group 1 types (HPV 16, 18, 31, 35, 39, 45, 51, 52, 56, 58, 59, and 66) and Group 2A type (HPV 68) will be considered to have HR-HPV types of interest. All other detected HPV types will be considered low-risk and reported as such.

Screening will be performed using visual inspection with acetic acid (VIA) by trained physicians, midwives, and nurses from the SOS Gambia Clinic, who will also provide treatment and follow-up. The PHASE survey team will coordinate screening in batches at accessible locations within the study area. Separate informed consent will be obtained for the use of screening data in further research.

6.2.1. Data collection in the field.

Questionnaire data will be captured directly into a REDCap™ eCRF, designed with the survey team and data manager, in accordance with GCP and regulatory requirements. REDCap’s offline capability is critical given intermittent internet in the field. ACASI and REDCap data will be integrated for analysis, with ongoing validation to ensure completeness, participant ID alignment, and consistency. Full procedures are outlined in the data management plan.

6.2.2. Laboratory data.

The collection of laboratory samples in the field and the arrival and initial processing and storage of samples in the laboratory will be captured in the REDCap database to allow their easy reconciliation. All samples will be barcoded for storage and the barcodes will also be captured in the database.

Laboratory results on primary and secondary outcomes (HPV types, NG, CT, MG, TV) will be integrated into the REDCap database in blocks as they are generated.