Healthy Expectancy Through Routine Antenatal STI Screening

Overview

This study will evaluate whether routine screening and treatment for two common sexually transmitted infections, chlamydia and gonorrhoea, during pregnancy can reduce preterm birth and other poor birth outcomes in Botswana, and whether this approach is affordable and cost-effective for the health system. About 2,000 pregnant women attending their first antenatal care visit at up to 10 government clinics in Botswana will be invited to join the study. All women will first receive the usual antenatal care services provided in Botswana, including routine health checks and HIV and syphilis testing. Women who enroll in the study will be randomly assigned to one of two groups: 1. Standard of care group: Women receive routine antenatal care only. 2. Intervention group: In addition to routine antenatal care, women are screened for chlamydia and gonorrhoea using self-collected vaginal swabs at their first antenatal care visit and again in the third trimester. The main outcome of the study is whether screening and treating chlamydia and gonorrhoeae reduces preterm birth (before 37 weeks). Other outcomes include low birth weight, very preterm birth, and maternal health conditions.

Aim 1 Procedures i. Screening, recruitment and enrolment We will identify and enroll 2000 pregnant women presenting for antenatal care (ANC) services at up to ten clinics in Gaborone and surrounding villages. All pregnant women at their first ANC visit will be screened for eligibility after standard HIV and syphilis screening per Botswana National Guidelines. Interested women who meet the age, pregnancy, gestational age, ANC visit, and residence criteria will be read a brief study description. Following confirmation of eligibility by clinic midwife (no cervicovaginitis requiring syndromic treatment), women will be provided with more detailed study information and invited to provide informed consent. Interested women will be read the study consent form, which will include information on Chlamydia trachomatis (CT) and Neisseria gonorrhoea (NG) infections, study risks and benefits, and each potential group assignment. Staff will record reasons for ineligibility or refusal and collect basic de-identified information from clinic logs to use for descriptive analysis. Written informed consent will be obtained from each study participant prior to enrollment. Women who are unable to read and/or write will be required to have an impartial witness present during the above procedures. Participants will be offered copies of the informed consent forms. Forms will be available in English and Setswana. For eligible participants aged 15-17, parental consent will be sought from their parent/guardian (verbal consent by telephone if parent/guardian not present and documented by the study team). All participant recruitment will occur in government clinics in Botswana. The study will be managed by a Study Coordinator and a Nurse Coordinator who will be based at study clinics, supervising a field team consisting of experienced research assistants and a qualified research nurse who are responsible for day-to-day conduct of study procedures including recruitment, enrollment and follow-up. A study clinician will also be on call for clinical questions from the study team and to dispense treatment. The Principal Investigator is responsible for the overall conduct and management of the study. During periods of recruitment, study posters, flyers and other materials may be placed at the clinics alerting potentially eligible pregnant women about the study. All recruitment and materials will be reviewed and approved by the IRB. Study staff will be trained in the study's methods, protocol, and ethical human subjects' research. ii. Randomization procedures Randomization will occur following all routine antenatal procedures and enrollment into the study. Randomization procedures were informed by community engagement. Participants will be randomized in a 1:1 ratio stratified by recruitment site to ensure balanced allocation across study arms within each site. The randomization sequence will be generated and overseen by the study statistician. Randomization and allocation will be implemented through the REDCap randomization module, which will automatically assign participants to study arms after eligibility confirmation, ensuring allocation concealment throughout enrollment. Once assigned, the randomization allocation will be unblinded. To minimize influence of the unblinded nature of the study on outcomes, ongoing data monitoring will not include information about endpoints disaggregated by site or arm. Only the study statistician will review data on endpoints by arm or site. iii. Baseline (first ANC) visit procedures All participants: All participants will receive the standard-of care i.e., routine ANC services conducted by the clinic midwives as per the standard-of-care in Botswana, including a clinical assessment of obstetric history, health status (e.g. weight, blood pressure, anemia, urinalysis), internal pelvic exam, routine labs (e.g. complete blood count, blood type, and Rh factor), provider-initiated counseling and screening for HIV and syphilis, treatment for syphilis if required, tetanus toxoid immunization, and iron and folate supplementation. Women who test positive for HIV antibodies receive CD4 T cell and HIV-1 viral load screening and referral for specialist review and initiation of antiretroviral therapy. These procedures will be done prior to enrollment in the study. Thereafter all participants will be asked to respond to a questionnaire. Questionnaire: An interviewer-administered questionnaire will be administered to all participants in their preferred language (e.g., English or Setswana). The questionnaire will be adapted from measures used in our previous STI screening studies or documented in literature. Data will also be extracted from the participant's obstetric record card and clinic records. The main measures will include 1) demographics and socio-economic status; 2) obstetric, gynecological, prior medical, and sexual health history; 3) sex partner and relationship dynamics, including risk of intimate partner violence; and 4) patient direct and indirect costs associated with transportation, co-pays, potential lost wages due to waiting or scheduling additional visits. Physical, and pelvic examination findings, as well as syphilis and HIV test results from routine antenatal procedures, will be recorded. Intervention arm: In addition to the standard-of-care described above, participants in the intervention arm will receive screening for CT and NG at their first ANC. Specimens collected for screening will include two self-collected vaginal swabs. The first swab will be used for CT/NG screening (with CT/NG assay). The second swab will be stored for future screening of STIs and vaginal infections. Results arising from future testing will not be returned to participants, as they may no longer reflect current clinical status and therefore may not be clinically applicable. CT/NG-positive participants will be treated using US CDC recommended treatments and will receive azithromycin 1g by mouth for CT and ceftriaxone 500 mg by injection for NG. (Note: the Maduo Study achieved a 100% cure rate among participants treated for CT and/or NG infection). Participants who test positive will receive counselling on STIs, the importance of partner notification and treatment, and abstinence from sexual intercourse until seven days after treatment. Participants who test positive will also be scheduled for a test-of-cure approximately four weeks after treatment. All participant-facing study staff will be trained on partner(s) notification counseling. For example, it is important to clarify that the duration of CT infection during pregnancy is unknown, many infections are asymptomatic, and infections do not implicate partner fidelity. Participants will be given several options for partner services: (1) no disclosure, in the case of risk for intimate partner violence; (2) supportive disclosure, where the participant is assisted by trained study staff either in person or over the phone in notifying their partners, and given the option of bringing partners to the study clinics for counselling and treatment; and (3) expedited partner therapy, where the participant can bring treatment to partners prior to partner examination by a healthcare provider. Standard of care arm: At the baseline and third trimester visits, we will ask to collect remnant urine from the routinely collected specimens. The urine specimens will be stored and used to test for CT and NG after delivery to retrospectively assess whether baseline STI prevalence is comparable between study arms and to support secondary analyses of infection dynamics during pregnancy. Because testing of stored specimens will occur retrospectively, the results may not reflect participants' current infection status at the time results become available. Individual results from retrospective testing of stored specimens will not be returned as diagnostic results. However, participants in the SOC group whose stored specimens test positive for CT and/or NG will be notified and offered CT/NG testing, with treatment provided if the repeat test is positive. In addition, all participants in the SOC will be provided with study contact information and informed that they may request STI screening after delivery at any time while the study is ongoing. iv. Ultrasound After enrollment, all participants in both study arms will be booked in for an ultrasound performed by a study practitioner trained in collaboration with the University of Botswana Obstetrics and Gynaecology specialty training programme. After this, participants will be encouraged to attend all regularly scheduled ultrasounds offered through routine care. v. Third trimester visit The third trimester visit will take place between 28 and 32 weeks' gestation. We will try our best to schedule the visit to coincide with a routine ANC visit. Data collection at this visit will include information on pregnancy progress, routine monitoring and clinical examination findings, any incident HIV diagnosis and treatment initiation, healthcare visits or admissions, and medication or supplement use since enrollment. If participants had an early delivery before the visit, this information will be deduced during a routine follow-up phone call to the participant. Research nurses will be adequately trained to collect relevant perinatal history, ascertain overall well-being of the mother and newborn if it was a live birth, and link her to relevant health services if necessary. Participants in the intervention arm will provide two self-collected vaginal swabs. The first swab will be used for CT/NG screening. The second swab will be stored for future screening of STIs and vaginal infections. Results arising from future testing will not be returned to participants, as they may no longer reflect current clinical status and therefore may not be clinically applicable. Results, treatment, partner services and test-of-cure will be delivered as described at baseline. vi. Data collection post-birth We will conduct a full medical record review to collect date of birth, birthweight, delivery method, spontaneous preterm birth, induced preterm labor, and C-section. Our primary outcome, preterm birth, (\<37 weeks gestation), will be calculated based on delivery date, last menstrual period, and ultrasound dating. If last menstrual period and ultrasound are inconsistent, we will re-date based on the ultrasound alone. We will also seek to determine whether the preterm birth was medically indicated (e.g. presence of preeclampsia, placental abruption, intrauterine growth restriction, fetal distress). Secondary outcomes will include infant outcomes: low birth weight (\< 2500g), and a composite measure of preterm or low birth weight; very preterm birth (\<32 weeks gestation); and preterm birth sub-types (spontaneous, medically indicated), which will be abstracted from the infant record; and pregnancy/maternal outcomes: gestational hypertension (one-two elevations of systolic (≥140) or diastolic blood pressure (≥90) after 20 weeks pregnancy), preeclampsia (hypertension with substantial proteinuria after 20 weeks pregnancy),50 and incident HIV infection will be abstracted from the obstetric record. If data are missing, we will check the Integrated Patient Management System (IPMS), a nationwide electronic system that include laboratory records, and thereafter follow-up with participants by phone. v. Retention strategy To ensure high participant retention, we will leverage successful strategies from our prior work in Botswana. These include deploying trained and closely supervised staff, providing reasonable reimbursements to cover time and travel expenses, and collecting multiple contact methods (text, WhatsApp, phone calls) for participants and their trusted associates. For follow-up purposes, study staff will collect participants' locator information, including telephone number, and contact details for two or more trusted persons will be collected to assure follow-up. Participants who do not return for scheduled visits will be actively contacted by both clinic and study staff and encouraged to return for care. vi. Participant reimbursement Participants will be reimbursed 100 Botswana Pula (approximately $7 for each study visit, covering their time and travel costs). vii. Statistical analysis For each outcome and study arm as well as for sub-groups of interest (e.g., parity, age, HIV status), descriptive statistics will be generated using frequencies and proportions for categorical variables and means, medians, and IQRs for continuous variables. Hypotheses testing will be based primarily on generalized linear mixed models (GLMM) with random intercepts for clinics, which will allow us to control for within clinic correlation and account for variability in the outcome due to differences between clinics. Also, the random effect for clinic would account for unbalances in sample size between clinics. Additionally, we will consider including a second random effect for the interaction between Arm and Clinic, as suggested by Feaster et al (2011). The interaction effect will be tested and if significant will be included in the model(s). Else, it will be excluded. This would allow us to make generalizations beyond the clinics included in our study. Furthermore, in all analyses, we will consider the inclusion of other covariates selected based on our conceptual model. Our overall modeling approach will be based on the "purposeful selection of variables" strategy, introduced by Hosmer and Lemeshaw where subject matter significance, relationships among the independent variables (e.g., correlations, confounding, and interactions) and statistical significance will be taken into consideration. In all statistical tests, significance will be determined based on an alpha level of 0.05, except for interactions, where a more liberal alpha level of 0.10 will be used. For our first hypothesis that women in the intervention arm will be less likely to experience preterm birth as compared to women in standard-of-care, we will use a mixed effects log binomial regression with a binary variable indicating whether a participant experienced a preterm birth (yes vs. no) as the outcome variable and study Arm as the primary fixed effect. We examine both the adjusted and unadjusted models. Covariates will be included based on considerations described in the introductory paragraph. Additionally, even though we did not hypothesize any interactions, to ensure the structural integrity of the model, all two-way interactions between the study Arm and covariates as well as among covariates will be tested. Interactions that are not significant will be excluded from the final model, whereas significant interactions will be further evaluated by calculating and testing the simple main effects of the factors involved in the interaction on the outcome. We should note that women who will have experienced loss from miscarriage, stillbirth, twins and missing/LTFU will be excluded from this analysis (i.e., only women who have been given live birth will be included). Throughout the study we will carefully monitor expected events (e.g. miscarriage) and unexpected events to ensure they're not associated with the intervention. For our secondary hypotheses, that women in the intervention will be less likely to experience i) low birth weight (\< 2500g), ii) preterm birth or low birth weight, iii) Very preterm \<32 weeks, or iv) preeclampsia, compared to women in the standard-of-care;, we will use the same method as described above with indicator variables for low birth weight (yes. vs. no), preterm birth or low birth weight (yes vs. no), very preterm birth (yes vs. no), experiencing preeclampsia (yes vs. no) as the outcomes. We will also combine adverse birth outcomes (preterm, low birthweight, miscarriage, or stillbirth), depending on the distribution of the composite measure for the adverse neonatal outcomes, we will be using either linear or log-linear mixed regression with the composite score as the outcome and Intervention Arm as the primary predictor. Finally, we hypothesize that parity will act as a moderator between the intervention and the outcome of experiencing preterm birth, with the intervention expected to be more effective among nulliparous women.. To explore this hypothesis, we will use a mixed effects log binomial regression with a binary variable indicating whether a participant has experienced a preterm birth or not as the outcome and Study Arm, Parity (nulliparous vs. not), and the interaction between the two as primary fixed effects, with the interaction being of main interest. If the interaction is found to be significant, then we will evaluate it by calculating and testing the simple main effects of the intervention among nulliparous and non-nulliparous women, respectively. Only women who have been given live birth will be included in this analysis. We will also explore whether the intervention had a differential effect among women living with HIV (compared to HIV-) and among younger women \<25 years (compared to older, ≥25 years) A similar analytic approach as described for the parity analysis will be used where the interaction between Study Arm and HIV status and age will be of primary interest. If the interaction is found to be significant, then we will evaluate it by calculating and testing the simple main effects of the intervention by HIV status and age group. Only women who will have given live birth will be included in this analysis. Aim 2 Procedures i. Data collection For Aim 2, we will collect factors associated with the costs of care. From the patient perspective, data will be collected through the interviewer-administered questionnaire, and as described above, will include direct costs (e.g. co-pays), transport, and childcare costs, and the opportunity cost of lost wages. From the health care provider perspective, data will be collected through records review and time-and-motion observations. Data will include personnel time and time involved in screening, treatment, counseling, and syndromic assessment; and the costs of supplies and equipment. During the observations, the researchers will have no direct interaction with providers or patients and data will be solely related to their work. No additional consent or visits will be required. ii. Cost analysis From the perspectives of a government payor and the patient, we will assess the financial costs (actual expenditures) and economic costs (value of resources used) of the CT/NG screening and treatment intervention, including planning, start-up and training, counseling, screening, treatment, partner services, tests of cure, and rescreening; among pregnant women asymptomatic for CT/NG at the first ANC visit. Data will be collected prospectively from the planning stage to final outcomes, using in-country data sources and cost data collection spreadsheets and manuals that we previously developed. Cost data will be collected primarily in Botswana Pula and converted to US dollars using market exchange rates. Component costs will be divided into planning, start-up, capital, and recurrent (supplies, personnel, transport). Start-up costs will include materials and personnel time, supplies, and overhead associated with training; Capital costs, which are fixed/one-time expenditures that last longer than one year and cost ≥$100 (e.g. PCR instrument) will be annualized over the expected useful life with a discount rate of 3%. For recurrent costs, we will use a micro-costing, bottom-up and approach. Time-and-motion observations will assess supplies used and personnel type and time involved in each activity. We will conduct approximately 25-100 observations for each activity over the course of study implementation. Overhead and infrastructure will include any additional clinic space necessary to house the PCR. When valuing resources that are paid from the research budget, we will use public sector prices for staff and medication. For example, the costs associated with supply procurement will be obtained from Central Medical Supplies. Personnel salary scales will come from the Botswana Directorate of Public Service Management (DPSM). We will cost the CT/NG test assay cartridges and the PCR instruments at a level commensurate with a potential public sector scale-up. Care will be taken to value volunteer time and to exclude any costs that are incurred only as part of research activities. For the patient perspective, we will collect participant out-of-pocket and opportunity costs associated with any additional (outside of routine care) visits or wait times incurred due to the screening intervention. Wait times will be assessed using REDCap at the conclusion of each visit. Costs associated with any additional visits (e.g. related to receipt of treatment), including participant time, transport, or out-of-pocket expenses will be identified through self-report and record abstraction. To capture the opportunity costs of participant wages/salary earnings foregone, the per capita household income will be computed as total household income divided by total number of household members, with adjustments for children. To value these costs equitably, the mean per capita household income reported at the baseline interview will be used as a proxy of this opportunity cost. We will calculate the unit costs of each component activity (cost per initial screen, treatment, partner treatment, test-of-cure) and the overall cost per person screened, true CT/NG infection treated, and true CT/NG infection cured. An excel-based Budget Impact Analysis (BIA) model will be developed to evaluate the costs for a national point-of-care CT infection screening program for asymptomatic pregnant women. The BIA will follow the guidelines put forth by the International Society for Pharmacoeconomics and Outcomes Research (ISPOR) Task Force on BIA good practices, and the construction and presentation will be informed by the CAB. The BIA model will take the perspective of the government payor and the eligible population will include pregnant women ≥15 years of age in Botswana who are asymptomatic for cervicovaginal infections, and the population size is assumed to remain in a steady state over the time horizon (1-5 years). The parameters associated with uptake of screening, CT/NG prevalence (by sub-groups), receipt of treatment, test-of-cure, and rescreening will be derived from Aim 1. The base model will include the first ANC screen, test-of-cure, and third trimester rescreen using a point-of-care strategy; however, the frequency of screening can be adjusted to assess the change in costs (e.g. removing the third trimester screen). Additionally, the model can allow for adjustments to the costs of the PCR instrument to represent a centralized lab scenario; however, potential decreases in the reach of treatment must be considered. The BIA model will present undiscounted costs, however, the net present value can be calculated if requested by the CAB. To address parameter and structural uncertainty, the model can be used to assess alternative scenarios where structural assumptions and parameter values can be varied. The model and results will be presented and discussed with the CAB to determine face validity. The main outcomes will be the one year and 5-year costs of scale-up to the target population as well as sub-groups. iii. Cost-effectiveness Evaluation: We will develop a decision analytic model to estimate the incremental cost-effectiveness ratio (ICER, difference in costs divided by difference in health outcomes). For the cost, we will consider both healthcare and patient costs. Our primary outcome will be the incremental cost per additional DALY averted compared to the standard-of-care. For example, DALYs associated with preterm birth (primary outcome) and low birth weight (secondary outcome) will be calculated using the Institute for Health Metrics and Evaluation (IHME) Global Burden of Disease. This measure considers 1) the years of life lost due to prematurity or low birth weight-related mortality in the first year of life, and 2) among the proportion of preterm or low birth weight infants surviving past the first year, the average degree of disability and years lived with the disability, using the average lifespan in Botswana and a discount rate of 3%. Recognizing the debate about cost-effectiveness thresholds in LMIC, we will discuss options with CAB policy-makers, including the per-capita gross domestic product (GDP ($7,738; 2022) and the ICERs of currently-implemented HIV- and STI-related programs in Botswana. For example, if the preterm birth rates were 11% in the intervention and 16% in the standard-of-care, resulting in \~50 preterm births/\~238 DALYs averted, the intervention could be considered cost-effective at an incremental cost of $1.8 million, according to the GDP threshold. Further, as participants vary with respect to intervention benefits and/or the cost of delivering it, we will also estimate the cost-effectiveness of implementing the intervention among the predetermined sub-groups (e.g. nulliparous women). iv. Uncertainty analysis: For each analysis, a probabilistic multivariable uncertainty analysis will be undertaken through multiple sampling rounds of both effectiveness and cost parameters. Parameters will be varied according to the confidence intervals generated in Aim 1 or along plausible ranges. The ranges and distributions will be compared with published studies and will be evaluated by experts in Botswana (e.g. CAB members). All cost outcomes will be presented as arithmetic means with 95% confidence intervals. We will calculate the probability that the scenarios are cost-effective by assessing the proportion of ICERs that fall below the willingness to pay threshold. v. Sensitivity Analyses: Sensitivity analyses will be conducted through one-way and multi-way methods to identify key drivers (e.g. test kit cost, utility weights) of the results and assess how changing those parameters impacts the overall costs, per patient cost of each strategy, and the ICER. Parameters to be varied will be informed by the CAB and will likely include the cost of the PCR instrument, the cost of the CT/NG cartridge and collection kit, personnel time and salaries, probabilities of screening and treatment uptake, the prevalence of infection, the measure of effect on preterm birth and other outcomes, and utility weights. Parameters will be varied by their confidence intervals or plausible ranges. For example, the high value for the CT/NG cartridge cost (currently \~$16/cartridge) will be the retail price with tax and customs charges and the low value will be the WHO target price of $1 (excluding the cost of the device). If the intervention is effective, but not cost-effective, we will calculate the necessary cost reductions required to fall below the established thresholds.