Study design overview

The Enhancing the A in SAFE (ETAS) study is a community randomized study to evaluate cost and community acceptability of enhanced antibiotic trachoma interventions within an endemic county in South Sudan. A population-based prevalence survey will be conducted whereby 30 communities are selected for trachoma assessment. Following the survey, communities will be randomized 1:1 to receive 1 of 2 enhanced antibiotic interventions (Fig. 1). The remaining communities in the county will receive standard-of-care annual MDA. The primary outcomes of the study are the cost of enhanced MDA compared to standard-of-care annual MDA and the community acceptability of enhanced MDA within study communities. A trachoma impact survey will be conducted approximately 12 months following the original trachoma prevalence survey. The total study duration will be approximately 13 months from the start of the community-level census to the time of the 12 month follow up impact survey. This protocol adheres to the Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) guidelines and refers to version 1.2 of the study protocol as revised 5 May 2021.

Fig. 1

ETAS study design summary for communities included in the primary outcome analysis. HH = Household; MDA = Mass Drug Administration

Setting, eligibility, and participants

The study will be conducted within a trachoma endemic county in Eastern Equatoria state, South Sudan (Fig. 2). This area was chosen based on previous survey data demonstrating a high burden of TF among children ages 1 to 9 years ranging from 30 to 48% [7]. The baseline for this study will be timed to coincide with the scheduled programmatic trachoma impact surveys in Eastern Equatoria state. For purposes of this manuscript, the initial survey connected to this study will be referred to as a baseline survey and the survey conducted at completion of the study will be referred to as an impact survey (survey methods described below). As part of the baseline survey to determine TF prevalence in the study county, 30 communities will be randomly selected from a geographically ordered list provided by the South Sudan MOH Trachoma Control Program [8]. All 30 communities will be enrolled in this randomized community study. Communities are eligible to participate if they are in the study county regardless of the individual-community level TF prevalence. There are no exclusion criteria for community-level participation. Within study communities, all households and individuals will be eligible for participation. Children ages 6 months to 1 year will be eligible for azithromycin interventions in this study as they are normally targeted with this medication as part of programmatic interventions and in previous randomized trials, they could potentially serve as an infection reservoir if not treated and are an age group which may benefit from the ancillary effects of azithromycin on mortality [9,10,11,12,13]. Because of the observed mortality benefit to these young children, ongoing studies are now randomizing children in this age group to multiple doses of azithromycin per year at national levels [14].

Fig. 2

ETAS study area in Eastern Equatoria state, South Sudan with 2015 TF prevalence estimates shown

For the 30 communities enrolled in the randomized community study, we expect an average population per community (based on programmatic data) of 649 individuals. This would lead to MDA treatment data on approximately 19,470 individuals of all ages. Since children ages 6 months to 9 years account for approximately 40% of the population, we expect to have MDA treatment data on 260 children per cluster or 7,788 total children in this age range.

Interventions

Antibiotic interventions will take place at the community level. The first intervention strategy will consist of an enhanced antibiotic regimen comprised of a community-wide MDA, followed by 2 additional rounds of targeted treatment to children ages 6 months to 9 years. The targeted treatments will be distributed 2 weeks and 4 weeks after the community-wide treatment, resulting in 3 doses to children 6 months to 9 years. The second strategy will consist of a community-wide biannual MDA approximately 6 to 8 months apart, resulting in 2 doses for the targeted community. Timing of the second dose will be based on access to communities and at least 6 months after the first dose.

Azithromycin, also known in the global trachoma program by the brand name Zithromax®, is a macrolide antibiotic and is used in the treatment of active trachoma. Zithromax® is donated to the South Sudan MOH for MDA by Pfizer Inc., through the International Trachoma Initiative (ITI). The recommended dosage for the treatment of trachoma is a single dose of 20 mg/kg of body weight and is implemented using an “age-height” based dosing strategy for the trachoma MDA campaigns. Azithromycin can be administrated in either tablet form or powder for oral suspension (POS). Dosing and administration for this study will be conducted following procedures outlined in the Zithromax® Management Guide [15]. To reduce the risk of choking, all children participating in the study intervention ages 6 months to 9 years will receive azithromycin by POS only. Populations ineligible for azithromycin based on national guidelines, including infants under 6 months of age, pregnant women in their first trimester, individuals with a previous history of severe adverse events (SAEs), and individuals with severe illness at the time of MDA, will be offered treatment with 1% tetracycline eye ointment (TEO), to be applied twice daily to both eyes for a 6-week period [15]. Trained drug distributors will facilitate drug administration for the appropriate treatment dose by assessing everyone that presents. Those receiving antibiotic will be directly observed by the distribution team.

Throughout this study, existing S, F, and E interventions will continue through standard programming. This includes offering surgery to individuals identified as needing TT surgery, and advocacy work with water, sanitation, and hygiene (WASH) focused non-governmental organizations operating in the country.

Adverse events

Azithromycin is generally well tolerated and adverse events are rare [16,17,18]. The most commonly reported side effects of azithromycin are gastrointestinal effects which include diarrhea, nausea, abdominal pain, and vomiting. As part of MDA training, all drug distributors, community chiefs, and government supervisors are trained on side effects and how to respond to SAEs. All study participants will be informed of potential side effects of azithromycin and screened for any history of SAEs from azithromycin. All drug recipients will be informed to immediately report to the drug distributors or community chief if they experience signs and symptoms of any SAE. All suspected SAEs will be recorded, immediately managed via the nearest health facility, investigated, and reported according to South Sudan MOH guidelines and will be recorded using an SAE reporting form. SAEs will be reported to The Carter Center, Pfizer Inc., and the Institutional Review Boards (IRB) within 24 h. An SAE officer will be stationed within the study area for the duration of the project to monitor reports and assist key government staff in investigating and facilitating rapid responses to suspected SAEs during the intervention period. While this study will include increased intensity of antibiotic pressure, Chlamydia-specific antimicrobial resistance to azithromycin has yet to be demonstrated as part of trachoma MDAs, including those with multiple rounds per year [19].

Treatment masking

Because of the community nature of the MDA treatments, masking will not be possible in this study. However, MDA treatment teams will be different from teams collecting trachoma indicator data. Laboratory and photographic grading team members will be masked to the treatment arm of the study.

Intervention costing

The costing analysis will determine the relative costs of the enhanced MDA strategies, including the triple dose and biannual MDA, which will be compared to each other, and compared to the standard-of-care annual MDA. This study will utilize the Global Health Cost Consortium Reference Case for Estimating the Costs of Global Health Services and Interventions and the Global Health Cost Consortium Principles and Methods Reporting Checklist to ensure the quality and comparability of results [20, 21]. The costing analysis will provide the evidence base to inform resource allocation decisions by the South Sudan MOH and partners within the country.

The study will take a payer perspective and encompass the costs incurred by The Carter Center as the technical and financial supporting implementing partner. This will be used to identify the cost per person treated and total costs associated with each arm of the study and annual MDA. To reduce the potential for bias, cost data will be collected using the same methods from 15 randomly selected communities in the study county receiving the standard-of-care annual MDA. Taking an ingredients-based approach, the study will include a detailed micro-costing database that captures all financial and economic costs incurred. The study will use the community level as the unit of analysis for total costs and calculate the unit cost per person treated using a top-down approach, dividing the total cost by the number of people treated.

The quantity and price of all required resources will be identified to evaluate total intervention costs for each arm of the study as well as the standard-of-care annual MDA (Table 1). To ensure costs such as salaried labor are not double counted, some activities that include staff time, including social mobilization and training, will not be broken down into all specific components. Salaried costs, for example, will be included under central administration and not included in each cost category. Where possible, data will be extracted from The Carter Center and South Sudan MOH databases.

Table 1 Summary of activity costs included in each cost category within the ETAS study

Real world costs will be presented in 2 ways: 1) costs related to the programmatic implementation setting (research activities excluded) and 2) costs related to the study setting (including study and implementation costs) (Supplemental Table S1). Study related costs will be helpful in planning future MDA trials within this setting. Cost categories will include capital costs and recurrent costs. Cost activities will include planning, training, community sensitization, drug transportation, drug administration, supervision, surveillance, reporting, and central administration.

The economic cost of capital items, i.e. those only incurred once and with a value of 100 USD or more, will be annualized based on a 3% discount rate and estimates of the useful life of the item [22]. A “useful life” is defined as the period during which an asset or property is expected to be usable for the purpose it was acquired. For vehicles specifically, we will calculate the daily cost of a vehicle in relation to its useful life and discount rate. The financial costs of capital items will be estimated by dividing by the number of years of useful life without discounting (straight line depreciation). Rental costs will be used to determine the cost of office buildings. If a building is owned, the rental cost will be used as a cost indicator. Recurrent costs will be disaggregated from The Carter Center budget and assigned to the cost data categories for each treatment arm after each intervention phase as well as the standard-of-care annual MDA. Personnel costs will include participation in the intervention (e.g., per diems), as well as the relevant share of that person’s salary and fringe benefits that cover their working time as part of the intervention [23]. The proportion of salaries and benefits that are attributable to MDA activities will be estimated by program staff.

Commodities include the study drug, Zithromax®, which is donated, as well as TEO, which is purchased by The Carter Center. The cost of TEO will be included as a financial and economic cost. The economic costs of donated medicines will also be calculated. List costs will be derived from the International Medical Products Price Guide [24].

To estimate the costs associated with a programmatic setting, study and research costs will be removed to make costs comparable to the control communities. These costs would be those associated with increased supervision and monitoring, increased training for data collection, data collection, and administration of the study, and they will be separated among other costs. Further analysis will use the costs associated with the 2 intervention delivery methods and will estimate the potential costs incurred if each intervention was scaled up region- or country-wide. This will be calculated by multiplying the unit cost per person treated by the number of people requiring antibiotics, as stated in the WHO Alliance for the Global Elimination of Trachoma by the year 2020 (GET2020) database [25].

All costs will be presented in constant 2022 USD. If required, costs will be first converted from the local currency to USD, using the World Bank exchange rate relating to the time period during which the cost data were collected. The values will then be inflated using US inflation rates to the base year of the analysis [26, 27].

Economies of scale

Costing analyses of MDA programs often assume an intervention’s cost per treatment is constant, regardless of the number of individuals treated. This is a misleading way to consider the delivery costs of MDA, due to the occurrence of economies/diseconomies of scale [28]. This analysis will explore the effect of economies of scale by presenting the cost per person treated at 3 different coverage levels (60%, 80%, 90%).

Sensitivity analysis

Univariate and multivariate sensitivity analysis will be conducted to identify main cost drivers and identify a range of potential program costs. The sensitivity analysis will explore changes to the cost of drugs, treatment delivery, security, and per diems, and assumptions around the useful life of capital items. Costs of personal protective equipment associated with COVID-19 prevention will be removed in the sensitivity analysis. The sensitivity of the discount rate will also be varied from 0–6%.

Community acceptabilityAntibiotic coverage

Prior to the MDA interventions, a household census will be conducted in all 30 communities whereby all members of all households will be enumerated and the name, sex, and age of everyone will be recorded via electronic census forms. A household member is defined as any person who has lived for at least one month in the household and eats and sleeps in the household. Global positioning system (GPS) coordinates will be captured for each household. The study community census will be used to establish the population to be targeted for study interventions and will provide the denominator for MDA coverage calculations. New residents migrating into study areas after the census but prior to MDA treatment will be added to the census as they are detected. Further, residents unknown to the study team during the MDA treatment will also be added to the census and subsequently treated, if eligible.

During the MDA, each administered dose will be recorded into the electronic record created during the community census by a trained data collector who travels with the distribution team. This record will be used for both study arms to determine the antibiotic coverage during each of the subsequent treatment rounds, allowing for the calculation of the number of doses each participant received. The definition of coverage for this study is: the number of individuals ages 6 months to 9 years who received the drug in study communities in each treatment round divided by the number of children ages 6 months to 9 years enumerated in the baseline census. A secondary analysis will calculate the antibiotic coverage in each round for the age groups 10 to 15 years and ≥ 15 years. Further analyses on compliance will examine the extent and predictors of missed rounds among the 2 study arms.

The dataset on treatment coverage will be compiled quickly by the study coordinator and will be used to monitor the success of the intervention. If the coverage is lower than expected, < 85% among children ages 6 months to 9 years, the treatment team will be redeployed to conduct a community mop-up to improve treatment coverage.

While MDA treatment coverage will be monitored using the census and treatment records, treatment validation will also be conducted. Validation will be based on the WHO endorsed Supervisor’s Coverage Tool [29]. After each round of treatment (either community-wide or targeted) a random selection of 5 study communities will be chosen for treatment validation. A new team will return to a random sample of 20 households in each selected community to assess self-reported treatment within 1 week of treatment. Self-reported coverage from the modified Supervisor’s Coverage Tool will then be compared to the coverage reported in the treatment register.

Community awareness and acceptability

A mixed methods approach, including focus group discussions (FGD) and community awareness surveys, will be used to elucidate insights from several groups of key community members on the acceptability and feasibility of enhanced MDA strategies and of targeting children for interventions (Fig. 3).

Fig. 3

Timing of mixed methods study collection as part of the ETAS study. DD = Drug Distributor; FGD = Focus Group Discussion; MDA = Mass Drug Administration

Three thematic areas will be the focus of the FGDs: 1) Pre-MDA FGDs with drug distributors and MDA field supervisors; 2) Post-MDA FGDs with participating drug distributors; and 3) Post-MDA FGDs with participating community members. The community awareness surveys will be conducted among community members to determine their level of awareness of the MDA campaign following mobilization activities in targeted communities (Table 2).

Table 2 Qualitative sample size and participant eligibility for the ETAS study

Mothers, caretakers, and chiefs from communities receiving enhanced interventions will be recruited for FGDs using purposive sampling. This approach ensures that participants are from different communities and have participated in the relevant activities that will be discussed during their FGD. One FGD will only include chiefs, because they have significant influence within the communities and other participants are unlikely to disagree with them publicly. Drug distributors from the county, recruited for the intervention, will also participate in the FGDs. Each FGD will be conducted for approximately 1.5 h by qualified research assistants who have received a 3-day training on research techniques. One moderator and 1 assistant will be assigned per FGD and will follow a comprehensive interview guide on key themes about the research question. All FGDs with community members and chiefs will be conducted in the local Toposa language. Assistants will audio record the discussions, verbally translate and record the Toposa audio into English, and then transcribe the English audio version. FGDs with MDA field supervisors will be conducted in English. FGD assistants will closely observe participants, their significant interactions, and the dynamics within the group, taking detailed notes on collective themes.

A community survey will be completed following MDA mobilization activities to gauge the community’s knowledge and awareness of the upcoming MDA campaign. The survey will be conducted at 3 time points. The first will be conducted before the first round of MDA in each study arm. Two communities will be randomly selected from each study arm for the survey. The census data from that community (previously collected) will be used to randomly select 20 households to survey. One person from each household will be interviewed for the survey by trained interviewers, with preference given for the female caregiver. The survey will be conducted a second time prior to the child-only treatment rounds in the thrice-treatment study arm, and then the survey will be conducted a third time in the bi-annual study arm prior to the second round of MDA.

Surveys and trachoma monitoring

Household data collection and trachoma monitoring will take place through population-based surveys at baseline (prior to interventions) and approximately 12 months later. Sample size calculations will assume a TF prevalence of 4% among children ages 1 to 9 years with a precision of 2%, an assumed design effect of 3.0, and a 95% confidence level, for a total sample size of 1,104 children. Adjusting for an assumed non-response rate of 20% yields a population size of 1,325 children targeted for surveying. A multi-level cluster random sampling procedure will be used whereby 30 communities will be selected via simple random sample from a list provided by the South Sudan MOH Trachoma Control Program. We will employ the random sort technique to ensure an equal probability of selection across communities in the sampling frame. An independent sampling statistician will generate random values in SAS 9.4 (SAS Institute Inc.) with the ranuni function and a seed to allow replication of the sample should there be a need to augment the cluster list with additional clusters resulting from shortages in availability due to unforeseen problems in access or eligibility. In the second stage 6 segments of 5 households will be randomly selected through community mapping and segmentation for a total of 30 households in each of the 30 communities. Given an estimated 3.7 people per household, of which 40% are children, we would expect to survey on average 45 children ages 6 months to 9 years present per community. Therefore, for trachoma outcomes, we anticipate examining a total of 1,350 children ages 6 months to 9 years. A new random sample of 30 households from each of the same 30 study communities will be selected for the 12-month impact survey, plus 4 additional communities will be randomly selected from the county for inclusion to monitor trends in trachoma outcomes under standard-of-care, annual MDA, and to inform sample size calculations for a potential future efficacy trial. At each survey, a household questionnaire will collect demographic, socioeconomic, and water, sanitation, and hygiene (WASH) data using standard questions found in trachoma impact surveys [7].

Trained and certified trachoma graders will examine all individuals according to the WHO simplified grading system for trachoma using a 2.5 × magnification loupe and adequate light [30]. Tarsal conjunctivae photographs will be taken of both eyes of each child ages 6 months to 9 years using a cellscope device, a smartphone attachment that illuminates and magnifies the conjunctivae [31]. Photographs will be graded for the signs of trachoma. Conjunctival swabs to detect Chlamydia trachomatis (Ct) infection will be collected from children ages 6 months to 9 years whereby a gloved grader will firmly draw a swab 3 times over the left conjunctivae, rotating at each pass. Conjunctival swabs will be tested for infection and infectious load using polymerase chain reaction [32]. With the availability of additional funding, a sample of positive conjunctival swabs from each timepoint will be tested for antimicrobial resistance markers within the conjunctival space using genomic sequencing techniques.

Data collection and management

A custom-built Android-based electronic data capture system (Conexus, Los Gatos, CA, USA) will be used for the census, data collection, and treatment phases of the study. This data is stored in a NoSQL database in Salesforce (Salesforce, San Francisco, CA, USA). Study participants will be assigned unique identification numbers to be used to link the census and treatment data and will not contain personal identifying information. Survey data will be collected using Open Data Kit (ODK) Collect to submit XML forms to NEMO, a form design and aggregation software built from the ODK software development kit. Qualitative data will be audio recorded, and some written notes may be recorded by facilitators. At the end of each day, or as frequently as possible, data collection devices will be connected to the internet so that data can be uploaded to the study server. All computers and other electronic devices will be password protected. An independent data monitoring committee will not be formed for this study, however, data from all phases of the study will be reviewed by data management staff to ensure quality. Upon completion of the study, de-identified data will be made available upon request of the South Sudan MOH.

Ethics approval and consent to participate

This study has been approved by the South Sudan MOH Research Ethics Review Board at the Directorate of Policy, Planning, Budgeting, and Research; and the IRB of Emory University (00002467). Because of the low level of literacy within the population, verbal informed consent and assent will be sought from the community leader and each individual prior to distribution of study medications. Before the census or surveys are conducted in a community, oral informed consent will be obtained, first from the community leader, then from the adult household representative of those houses included in the study, and then from each participating individual. Within a household, field staff will fully explain the study purpose and what is included in the study. Field staff will explain that data collection tools will not seek sensitive or private information and that respondents will be free to terminate the interview at any time without need of explanation. Declining to participate in the survey or questionnaire will not influence their ability to receive any treatment indicated in the survey. The oral informed consent form script will be read to the participants prior to data collection. Oral consent will be recorded electronically. Major changes to the protocol will be reported to the Review Boards at Emory University and in South Sudan.

Risks and benefits of the study will be communicated to study staff and communities through community meetings. The medications to be administered are safe and used routinely by the South Sudan MOH and have been found to be safe. The trachoma grading is non-invasive, requiring only a visual inspection. Collection of ocular swabs for detection of trachoma infection is associated with minor discomfort due to the swab abrading the conjunctiva. Qualitative data collection will not ask questions that cover invasive or sensitive issues. There are unavoidable opportunity costs, though screening all the members of a household and conducting the study data collection should take less than 30 min. The physical, social, or psychological risk associated with participation in this survey is no more than the risk associated with daily life. There are no economic risks associated with this study. All individuals with signs of trachoma will be provided TEO, and individuals identified with TT, the blinding stage of trachoma, will be registered and informed when free eyelid surgery will be provided in their county. There will be additional indirect community benefits in that the data from the study may inform decisions about future form and delivery of the trachoma control programs. Compensation will be provided to FGD participants to reflect the extended time devoted to participating in the discussion.

Dissemination plan

Reports detailing study findings will be shared firstly with the South Sudan MOH and local health departments. Information gained from the study will also be shared with communities during MDA trainings and post-MDA review meetings. The results will further be shared with local partner organizations, as well as international organizations through program reviews, meetings, scientific conferences, and peer-reviewed journals. Study data will also inform future trachoma-specific intervention effectiveness trials in South Sudan and other countries seeking to accelerate progress towards eliminating trachoma as a public health problem. De-identified trial data will be made publicly available upon request of the South Sudan MOH after publication of primary outcomes.