Acceptability and feasibility of tests for infection, serological testing and photography to define need for interventions against trachoma

(PLoS NTD word limit = 300, current word count =300) 22

Trachoma causes blindness due to repeated conjunctival infection by Chlamydia 23 trachomatis (Ct). Transmission intensity is estimated, for programmatic decision-making, by 24 prevalence of the clinical sign trachomatous inflammation—follicular (TF) in children aged 1–9 25 years. Research into complementary indicators to field-graded TF includes work on conjunctival 26 photography, tests for ocular Ct infection, and serology. The perceived acceptability and 27 feasibility of these indicators among a variety of stakeholders is unknown. 28 Methodology: Focus group discussions (FGDs) with community members and in-depth 29 interviews (IDIs) with public health practitioners in Tanzania were conducted. FGDs explored 30 themes including participants' experience with, and thoughts about, different diagnostic 31 approaches. The framework method for content analysis was used. IDIs yielded lists of perceived 32 strengths of, and barriers to, implementation for programmatic use of each indicator. These were 33 used to form an online quantitative survey on complementary indicators distributed to global 34 stakeholders via meetings, mailing lists, and social media posts. 35

Sixteen FGDs and 11 IDIs were conducted in October–November 2022. In general, all 36 proposed sample methods were deemed acceptable by community members. Common themes 37 included not wanting undue discomfort and a preference for tests perceived as accurate. Health 38 workers noted the importance of community education for some sample types. The online survey 39 was conducted in April–May 2023 with 98 starting the questionnaire and 81 completing it. 40 Regarding barriers to implementing diagnostics, the highest agreement items related to 41 feasibility, rather than acceptability. No evidence of significant differences was found in 42 responses pertaining to community acceptability based on participant characteristics. 43

All of the indicators included were generally deemed acceptable by all 44 stakeholders in Tanzania, although community education around the benefits and risks of 45 different sample types, as well as addressing issues around feasibility, will be key to successful, 46 sustainable integration of these indicators into trachoma programs. 47 48 49 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint 3 Author Summary (limit = 200, current =200) 50 Trachoma is a disease that causes blindness through conjunctival infection with the bacterium 51 Chlamydia trachomatis. Trachoma is targeted for global elimination by 2030. To know whether 52 population-level interventions are required, we must know how intensely conjunctival C. 53 trachomatis is being transmitted in a population. The current proxy recommended by the World 54 Health Organization is prevalence of a clinical sign of active (inflammatory) trachoma: 55 trachomatous inflammation—follicular. However, this indicator has several drawbacks. Policy-56 makers are considering the utility of a number of complementary indicators, including 57 conjunctival photography and tests for infection and serology. 58 We sought the opinions of different stakeholders to determine the acceptability and feasibility of 59 complementary indicators for use in trachoma programs. In Tanzania, we undertook focus group 60 discussions with community members and in-depth interviews with public health practitioners. 61 We also conducted an online survey of global stakeholders. We found that all the proposed test 62 types were acceptable to stakeholders in Tanzania; common themes included not wanting undue 63 discomfort and a preference for test types perceived to be accurate. Community education and 64 building trust was deemed critical. From the online survey, the most agreed-upon barriers to 65 implementation of each method were related to concerns about feasibility, rather than 66 acceptability. 67 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint 4

68 Trachoma is a neglected tropical disease (NTD) that causes blindness as a result of 69 repeated conjunctival infection by the bacterium Chlamydia trachomatis (Ct).1 Transmission 70 intensity is estimated, for programmatic decision-making, via field grading of the clinical sign 71 trachomatous inflammation—follicular (TF). A major goal of the World Health Organization 72 (WHO) Alliance for the Global Elimination of Trachoma is the reduction of TF prevalence in 1–73 9-year-olds (TF1–9) to <5% in all formerly endemic districts worldwide, contributing to the 74 achievement of global elimination of trachoma as a public health problem.2 A key component of 75 the WHO-recommended SAFE (surgery, antibiotics, promotion of facial cleanliness and 76 environmental improvement) strategy that is designed to be used to achieve this target is mass 77 drug administration (MDA) of antibiotics to clear infection. The number of MDA rounds is 78 dependent on the category of TF1–9 within an evaluation unit (generally equivalent to a district) 79 as measured by population-based prevalence surveys3 on the assumption that TF1–9 is well-80 correlated with prevalence of Ct infection. 81 However, TF1–9 has several limitations as the sole indicator for trachoma. First, the 82 relationship between TF1–9 and infection is not predictable,4 especially post-MDA,5 and the 83 reduction in TF1–9 often lags behind the reduction in Ct infection.6 Second, conjunctival follicles 84 (the features that lead to a diagnosis of TF) are not necessarily specific to trachoma, sometimes 85 occurring due to other causes.7 Furthermore, in areas with low prevalence of trachoma, it 86 becomes both costlier to train field graders8 and more difficult for graders to pass the inter-grader 87 agreement test.9-11 88 There has been ongoing research into indicators that could be alternative or 89 complementary to TF, including conjunctival photography,12 tests for ocular Ct infection,5 and 90 serology to assess an individuals’ previous exposure to Ct infection by detecting antibodies to the 91 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint 5 pathogen.13 It is important to note that each of these indicators measures a different signal: 92 clinical grading (both field-based and through photography) measures signs of inflammation, 93 infection tests are a direct measure of current infection with Ct, and serology measures a history 94 of previous exposure to the pathogen. Therefore, each of these indicators could provide different 95 measures of the public-health threat caused by trachoma; i.e., indicators should be considered 96 "complementary" rather than "alternative" to each other. 97 Following a WHO meeting in 2016 reviewing existing data, it was concluded that there 98 was insufficient evidence to support routine use of Ct infection or serology tests to inform 99 trachoma elimination programs.14 However, a WHO informal workshop in December 2021 100 supported the use of age-stratified data on Ct infection and serological data on the presence of 101 anti-Ct antibodies for decision-making in districts where TF1–9 either remains at or above the 5% 102 elimination threshold at the second impact survey (districts with "persistent" TF) following 103 MDA interventions, or falls <5% but subsequently returns to ≥ 5% during the surveillance period 104 after cessation of MDA (districts with "recrudescent" TF).15 105 WHO is now producing guidelines on the use of serology in trachoma elimination 106 programs. As part of the guideline development process, the Guideline Development Group must 107 consider both the acceptability of the proposed intervention and the values and preferences of the 108 people affected by the recommendations, which will partially determine the strength of any 109 recommendations.16 Research to date has indicated the value complementary indicators could 110 have for trachoma surveillance purposes. A literature review assessing post-elimination 111 surveillance systems for multiple diseases, including trachoma, suggested that opportunities for 112 inclusion of infection testing and serological markers in trachoma elimination settings continue 113 to be explored.17 A 2019 survey of stakeholders found that the top-cited barrier to trachoma 114 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint 6 eradication is inadequate surveillance tools and systems to monitor for recrudescence.18 While 115 eradication is a different goal than elimination as a public health problem, the need for practical 116 strategies to measure trachoma after the cessation of MDA remains. 117 With 2030 as the target for global elimination of trachoma as a public health problem,19 a 118 better understanding of the acceptability and feasibility of complementary indicators for 119 trachoma detection by all stakeholders is urgently needed. In this study, we aimed to determine 120 the opinion of a wide variety of stakeholders through a mixed-methods approach: a qualitative 121 study of community members and public health practitioners in Tanzania, followed by a 122 quantitative online global survey of stakeholders. 123

124 Qualitative study in Tanzania 125 Ethics 126 The qualitative study was approved by the London School of Hygiene & Tropical Medicine 127 Research Ethics Committee (Ref: 28028), the Task Force for Global Health ethical review body, 128 and Tanzania's National Institute of Medical Research (Ref: NIMR/HQ/R.8a/Vol.IX/4123). 129 Work was conducted in collaboration with Tanzania's health ministry through the National NTD 130 Control Programme, region- and district-level health departments, and village leaders. Written, 131 informed consent was obtained from all participants prior to their inclusion by trained research 132 assistants fluent in the local languages (Swahili and/or Maa) with an impartial witness present for 133 illiterate participants. Permission to record the focus group discussions (FGDs) and in-depth 134 interviews (IDIs) and publish quotes was obtained at enrollment. 135 Setting/Sampling 136 Tanzania was chosen based on the support of its health ministry, history and strength of its 137 trachoma program, presence of districts with persistent and recrudescent TF as well as districts 138 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint 7 having met the elimination threshold, and access to the "Maasai corridor", an area spanning the 139 Kenya-Tanzania border with high levels of persistent and recrudescent TF among pastoralists 140 including the Maasai ethnic group. FGDs with community members were conducted in four 141 districts in Tanzania, consisting of two districts with TF1–9 <5% and two with TF1–9 ≥ 5% at most 142 recent survey, chosen to be representative of pre- and post-elimination settings. Districts were 143 chosen in three selected regions (Dodoma, Singida, and Arusha) to ensure a diversity of 144 trachoma program experience, geographic location, and ethnic group. The current TF1–9 category 145 in Tanzania and the four selected districts are shown in Figure 1. Within each chosen district, 146 two villages were randomly selected for the FGDs by assembling a list of all villages in each 147 district, assigning each village a random number, and selecting the two villages with the lowest 148 randomly generated numbers. 149 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint 8 Figure 1: Map of the four districts selected for the study and their current trachomatous 150 inflammation—follicular (TF) prevalence category in children 1–9-years-old, Tanzania. The 151 boundaries and names shown and the designations used on this map do not imply the expression 152 of any opinion whatsoever on the part of the authors, or the institutions with which they are 153 affiliated, concerning the legal status of any country, territory, city or area or of its authorities, 154 or concerning the delimitation of its frontiers or boundaries. Prevalence map source: Trachoma 155 Atlas.20 156 157 158 For the FGDs, community members aged 18–60 years who had experience caring for children 159 were eligible to participate. Participants were identified using a snowball sampling approach,21 160 with index participants chosen by the village leader. Snowball sampling continued until the 161 target group size (n=10) for each FGD was reached. Separate FGDs were held with men and 162 women. 163 IDIs were conducted with public health practitioners in Tanzania. We sought interviews with a 164 wide variety of public health practitioners, including laboratory personnel, NTD program 165 officers, and non-governmental organization (NGO) program staff at all levels. Participants were 166 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint 9 selected based on experience working in NTD programs and their availability, with a target of 2–167 3 respondents for each of the three regions selected for FGDs, plus practitioners at the national 168 level who were based in Dodoma. 169 Data collection 170 FGDs and IDIs were conducted by trained research assistants in the participants' preferred 171 language using pre-designed topic guides (Supplemental File 1), exploring themes such as 172 previous experience with each of the test types: TF status via field grading and photography, 173 infection testing, and serology, as well as (for community members) thoughts about the different 174 sample collection methods: field grading, conjunctival photography, eye swabs, and dried blood 175 spots, and (for public health practitioners) thoughts about the feasibility of each test type. In the 176 FGDs, visual aids were used to help participants understand the different sample collection 177 methods. Sessions were recorded using an electronic audio recorder, transcribed by a native 178 speaker and then translated by them into English. Translations were checked for accuracy by the 179 researcher who carried out data collection. 180 Data analysis 181 Transcripts were entered into the qualitative data management software Nvivo22 and coded based 182 on both an a priori list of codes pertaining to each test type (Supplemental File 2) as well as 183 broader emergent themes. The framework method for content analysis23 was used to chart data 184 into a framework matrix. For the FGDs, comparative analysis was done to compare responses 185 between communities based on gender and setting (TF1–9 ≥ 5% or TF1–9 <5%). The coding 186 framework and analysis plan were reviewed by a second researcher prior to finalization of 187 analysis. For the IDIs, responses pertaining to strengths and barriers for each of the sample types 188 were used to formulate the stakeholder survey questionnaire. 189 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint 10 Quantitative survey of stakeholders 190 Ethics 191 The quantitative survey was approved by the London School of Hygiene & Tropical Medicine 192 Research Ethics Committee (Ref: 28589) and the Task Force for Global Health ethical review 193 body. Informed consent was obtained electronically in the preferred language of the participant, 194 from the four languages available: English, Spanish, French, and Portuguese. All questions 195 (other than consent) were optional, and the survey responses were anonymous. At the end of the 196 survey window, survey data were downloaded from SurveyMonkey24 prior to deletion of the 197 survey and data from the platform. 198 Setting/sampling 199 Stakeholders working in trachoma, including national program coordinators or other national 200 program staff, members of NGOs, donor organizations, and academic organizations were asked 201 to participate in an online survey to assess their agreement with the strengths and barriers to each 202 test type identified by the IDIs with Tanzanian public health practitioners. Stakeholders were 203 required to be over the age of 18 years. A list of email addresses of stakeholders meeting these 204 inclusion criteria was assembled using publicly available sources, such as organizational 205 websites, publications, and meeting reports with participant lists. Snowball sampling was 206 employed by asking key members of partner organizations to recommend additional 207 stakeholders. We solicited responses from as many people as possible who met our criteria. 208 Data collection 209 The online survey began during the annual meeting of the WHO Alliance for the Global 210 Elimination of Trachoma by 2020 (GET2020) held in Istanbul, Turkïye, in April 2023. The 211 GET2020 meeting convenes a variety of stakeholders; representatives from all trachoma-212 endemic, formerly-endemic, and suspected-endemic countries are invited, as well as members of 213 related NGOs and academic or research organizations. The survey was hosted on 214 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint 11 SurveyMonkey.24 A link to the survey was available via email sent to the previously assembled 215 list of email addresses, through advertisements at the GET2020 meeting, and through a post on 216 the social media platform X (formerly known as Twitter) from EMHE. Prior to the start of the 217 survey window, the survey was piloted with colleagues with expertise in trachoma and online 218 survey forms to gain feedback on the questionnaire and the ease of using the survey platform. 219 The questionnaire (Supplemental File 3) included questions on the participant's age group, 220 gender identity, organizational role, primary country/countries of trachoma work, level of 221 agreement with the importance of various perceived barriers to implementation at the program 222 level and strengths for different complementary indicators on a 5-point Likert scale (with the 223 options of: Strongly Agree, Agree, Neither Agree nor Disagree, Disagree, Strongly Disagree), 224 and identification of further knowledge gaps remaining, entered as free text, that may prevent 225 implementation of these indicators for trachoma surveillance. The strengths and barriers 226 presented in the questionnaire were obtained from the responses of IDI participants. 227 Data analysis 228 Descriptive analysis was performed using R.25 Since the total target population is unknown (due 229 to the survey being publicly available via an online link) no response rate could be calculated. 230 Frequency tables were created to determine the percent of respondents that strongly agreed or 231 agreed with, neither agreed nor disagreed with, and disagreed or strongly disagreed with each 232 proposed strength and barrier. The "strongly agree" and "agree" responses were combined and 233 the "strongly disagree" and "disagree" responses were combined to determine the percent of 234 respondents who agreed or disagreed with a statement, respectively. In order to determine if 235 responses pertaining to community acceptability differed between national program staff and 236 other stakeholders, responses pertaining to these barriers were analyzed on the basis of the 237 participant's organizational and personal role (country staff, as defined as a respondent whose 238 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint 12 organizational role was "Government" and personal role was "Programmatic/Implementation", 239 compared to all other stakeholders) by calculating the median and interquartile range (IQR) of 240 responses to these questions. The 2-tailed Wilcoxon rank-sum test was used to calculate p-values 241 comparing the responses of the two groups. Key knowledge gaps entered by respondents as free 242 text responses were cleaned and aggregated. 243

244 Qualitative study in Tanzania 245 Data collection took place in Tanzania from October to November 2022. A total of 160 246 community members participated in 16 FGDs, with ten people per FGD. The reported population 247 of selected villages ranged from roughly 1,100 to 5,400. Equal numbers of men and women 248 participated. Fourteen FGDs were conducted in Swahili and two were conducted in Maa. A total 249 of 11 public health practitioners participated in IDIs. The roles of participants included current 250 and former country-level staff, NGO staff, lab technicians, regional NTD coordinators, and one 251 healthcare worker. The framework matrix used to assess community responses for each test type 252 by gender and setting is provided as Supplemental File 4. 253 TF field grading 254 According to community members in Tanzania, key advantages of field grading of TF were the 255 familiarity of this method (particularly noted in FGDs with women), the lack of pain or adverse 256 consequences from this method, and the ability to receive results immediately. A perceived 257 disadvantage of field grading of TF, mentioned in half of the FGDs, was the possibility of short-258 term discomfort or adverse consequences that may be caused by the strong light of torches used 259 as part of the field grading process: 260 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint 13 "I think that [the] examination method of being graded with torches causes effects. For example, 261 eyes turn red and keeps on shedding a lot of tears after a certain period of time. I recommend that 262 they should innovate a new method for trachoma examination." (31– 35-year-old woman, TF1–9 263 <5% district) 264 In the IDIs, most of the public health practitioners who provided advantages of field grading of 265 TF mentioned that this method is both familiar and acceptable to communities, which agrees 266 with community feedback from the FGDs. Very few IDIs mentioned any disadvantages with 267 field grading of TF, and most disadvantages related to general concerns with needing to train 268 staff to conduct the examination, which is not necessarily specific to this method. None of the 269 IDIs mentioned the possibility that communities may dislike this method due to fear of adverse 270 consequences from the torch used as part of the examination. 271 Conjunctival photography 272 A majority of FGDs mentioned that a key advantage of conjunctival photography is a perceived 273 increase in the accuracy of grading via photo versus field grading due to the ability for graders to 274 have ample time: 275 “It is possible that [TF field grading] won’t be conducted as it is required because you are in [a] 276 hurry but when you are taking a photo it is a good thing because a person who will examine that 277 photo will have plenty of time to do that.” (26–30-year-old man, TF1–9 <5% district) 278 Other advantages of conjunctival photography mentioned by FGD participants were the 279 familiarity with the concept of imaging for diagnostics, and the lack of pain and adverse 280 consequences from this method. Each of these advantages was particularly often noted in the 281 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint 14 FGDs with men. A majority of FGDs (particularly with men) mentioned that a key disadvantage 282 is a fear of long-term adverse consequences caused by the photo flash: 283 “Some people are worried that when you are photographed you will go blind or your eye will be 284 destroyed.” (46–50-year-old man, TF1–9 <5% district) 285 Several FGDs (particularly those with men) also indicated a belief that conjunctival photography 286 is not as accurate as other methods: 287 “A photo is not enough; we would like to get another examination method because the camera 288 won’t show the disease that I have in my eyes or in my eyelid clearly.” (31–35-year-old woman, 289 TF1–9 <5% district) 290 None of the FGD participants mentioned that they had concerns related to confidentiality, usage 291 of the photos, or the potential identifiability of photos taken. 292 In the IDIs, none of the respondents mentioned a high level of community acceptability of 293 conjunctival photography as a specific strength of this method. Several IDI participants 294 mentioned that community fears related to flash photography causing adverse consequences and 295 mistrust that the entire face would be photographed (leading to concerns by community members 296 around the confidentiality of their and their children’s images) may be potential barriers to this 297 method. Several mentioned that education would be needed in order for communities to 298 participate: 299 "The most important thing is to give education and sensitization. If you will just tell a parent that 300 you want to take a photo, he/she might think you have [a] hidden agenda with the child." 301 (Program coordinator) 302 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint 15 In terms of feasibility, while there was an acknowledgement from a few IDI participants that the 303 fragility of camera equipment may be a concern in the field, most indicated that this would not 304 be a true barrier to implementation, especially given the benefits of conjunctival photography. 305 Eye swabs 306 A majority of FGDs mentioned increased perceived test accuracy (especially compared to field 307 or photo grading) as a key advantage of tests for ocular Ct infection. Half of FGDs (particularly 308 those with men) indicated that an advantage of the test for ocular Ct infection is that this method 309 is perceived as modern, high-tech, "professional", and/or legitimate (as it is approved by an 310 authorizing body prior to use). 311 Most FGDs (particularly those with men) mentioned that a key disadvantage of eye swabs is a 312 fear of long-term adverse consequences from the test procedure: 313 "I won’t agree with the use of eye swab method, I will even tell my family members not to 314 participate in that method because it has effects, it may leave you with a wound in the eye." (31–315 35-year-old-man, TF1–9 ≥ 5% district) 316 In addition, half of FGDs (particularly those with men) mentioned a fear of pain from the test 317 procedure, especially with regard to children. Several FGDs indicated that the complicated 318 nature of the test (with the need for sample collection, storage, and testing) may cause test errors 319 or failures. 320 From the IDIs, many respondents said that eye swabs would be acceptable to communities only 321 with proper education as to the risks and benefits of participation. Some IDI respondents said 322 that training would be needed in order for the sample collector to not harm the eyes of 323 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint 16 participants, with the importance of this being two-fold: to minimize harm to participants and 324 increase participation rates: 325 "…if the trained personnel will not be careful when taking [the] sample, she/he can harm a child 326 and these children can inform each other that it hurts when they are taking [the] sample and this 327 will make all children not participate in [the] examination." (Lab technician) 328 In terms of feasibility, several IDI respondents said that a desire for samples to be processed at 329 the local level (in order to increase turn-around time) is likely unfeasible due to a lack of current 330 lab capacity at the district level. This was mentioned by both public health practitioners at the 331 national and lower levels of administration. 332 Blood spots 333 Key advantages to blood spots mentioned in a majority of FGDs were the familiarity with blood 334 testing and (particularly among FGDs with women) the lack of pain or adverse consequences 335 from the testing procedure. Half of FGDs (particularly those with men) indicated that the high 336 perceived accuracy of the results is an advantage to this method: 337 “There are many diseases that cannot be seen easily but when you take a blood sample to the lab 338 it will be a perfect solution.” (51–55-year-old man, TF1–9 <5% district) 339 Several FGDs mentioned that a disadvantage to blood spots is a lack of trust by participants that 340 the sample will be used only to test trachoma; with fears especially around testing for HIV or 341 other sexually transmitted infections without participants' consent. Many FGDs with participants 342 who mentioned this fear indicated that it could be mitigated with community education, as 343 community members had simply not been made aware that the presence of antibodies to 344 trachoma could be tested for via blood spots prior to participating in the FGDs. 345 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint 17 One FGD among the Maasai ethnic group mentioned a general dislike of blood testing in 346 particular: 347 “We, as Maasai people don’t like blood testing most of the time. For example, HIV and TB are 348 prevalent in Maasai society and when people are asked to test using blood samples, they run 349 away.” (46–50-year-old woman, TF1–9 ≥ 5% district) 350 When IDI participants mentioned advantages of blood spots, community acceptability was not 351 often specifically mentioned as a key strength of this method. When community acceptability 352 was mentioned, the need for education was underscored. One IDI participant mentioned the 353 potential for distrust related to sample use (specifically related to fears around HIV testing) 354 among community members. Some IDI participants also mentioned the potential for community 355 acceptability issues if community participants expect on-the-spot results (as they would receive 356 for other types of blood tests they undergo as part of routine health care, such as malaria testing), 357 and again underscored the need for community education as to the purpose of the test in the 358 context of trachoma programming. 359 In terms of feasibility, key strengths mentioned by IDI participants included the familiarity of 360 blood spot collection and testing by health systems and the ability for high throughput of these 361 sample types, although several participants mentioned a lack of lab capacity to store and test 362 samples at a local level as a potential issue. While familiarity of blood testing by health systems 363 was often mentioned, IDI respondents did not mention the potential for integration of trachoma 364 monitoring into existing health systems via the use of blood spots as a specific advantage of this 365 method. 366 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint 18 When asked, the majority of FGDs had at least one participant who would hypothetically 367 participate in each sample collection method. No qualitative difference in responses between 368 districts with a TF1–9 ≥ 5% vs. <5% was detected (15 FGDs in districts with a TF1–9 ≥ 5% vs. 15 369 FGDs in districts with a TF1–9 <5% with at least one respondent who said they would 370 hypothetically participate in each method). 371 FGDs with men generally indicated more willingness to hypothetically participate in all of the 372 testing methods compared to FGDs with women (16 FGDs with men vs. 14 FGDs with women 373 with at least one respondent who said they would hypothetically participate in each method). 374 While responses from FGDs with men and women were largely quite similar, FGDs with men 375 often mentioned the importance of perceived accuracy, a lack of adverse consequences from the 376 testing procedure, and the benefit of tests perceived as modern or high tech as important to 377 determining the acceptability of the different methods. FGDs with women mentioned the 378 importance of test familiarity and a lack of undue discomfort or harm from participating in the 379 test procedure as important to determining the acceptability of the different methods. 380 Quantitative survey of stakeholders 381 The online survey was live from April 24 to May 31, 2023. Responses from 98 people who 382 consented to participate were received. Respondents reported living and working in 42 different 383 countries and represented each trachoma-endemic WHO Region (African Region, Eastern 384 Mediterranean Region, Region of the Americas, South-East Asia Region, and the Western 385 Pacific Region). A majority (60%) of respondents were men. Most respondents were between the 386 ages of 35– 49 years, worked for an NGO, and worked in a programmatic/implementation role 387 (Table 1). 388 Table 1: Self-reported demographic information of respondents to stakeholder survey 389 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint 19 Characteristic n % Gender Man 55 60% Woman 37 40% Age group 18–34 years 5 5% 35–49 years 45 48% 50–64 years 32 34% 65 + 11 12% Organizational Role Government 20 22% Non-governmental organization 46 50% University or research 22 24% Other 4 4% Personal Role Academic/Research 24 26% Programmatic/Implementation 57 61% Other 12 13% 390 Eighty-one respondents answered questions beyond providing demographic information. Table 2 391 includes results for questions related to respondents' agreement with strengths and barriers to 392 each indicator. Almost all (79, 98%) of respondents agreed that TF grading being WHO-393 recommended for diagnosing trachoma is a strength of field grading of TF. Besides this, the 394 strengths of field grading of TF respondents most agreed with were that TF grading is familiar 395 (71, 88%), acceptable to communities (69, 85%), harmless/does not cause side effects (66, 81%), 396 and that participants are provided results on-the-spot (66, 81%). The most agreed-upon barrier 397 indicated for field grading of TF was that it is a subjective measure that is prone to human error 398 (65, 80%). 399 Table 2: Stakeholder Survey – Frequency of responses to strength/barrier survey items taken 400 from in-depth interview responses from public health practitioners in Tanzania 401 Survey ite m 3-w a . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint 20 Total respon- dents Strongly agree or agree n (%) Field grading of TF Strength TF grading is recommended by the World Health Organization for diagnosing trachoma in communities. 81 79 (98 ) TF grading is a familiar method. 81 71 (88 ) TF grading is acceptable to communities. 81 69 (85 ) TF grading is harmless/does not cause side effects. 81 66 (81 ) TF grading provides results on the spot. 81 66 (81 ) TF grading is accurate. 79 42 (53 ) Barr ier TF grading is subjective and prone to human error. 81 65 (80 ) TF grading is not as sensitive or specific as other methods. 81 50 (62 ) Everting the eyes of children can be difficult. 80 49 (61 ) Conjunctival p hotography Strength Multiple graders can grade the same photo, which will increase accuracy and help determine the status of "edge cases." 73 61 (84 ) Increased costs of conjunctival photography will be less than those for eye swabs or blood spots. 73 41 (56 ) The potential to use Artificial Intelligence to grade photos will provide more accurate

than field grading of TF. 71 34 (48 ) The photo can be used to determine the presence or absence of other eye diseases. 73 30 (41 ) Conjunctival photography is easy to implement. 72 29 (40 ) Grading the photo away from the field will reduce field time compared to TF grading. 73 24 (33 ) Conjunctival photography is more accurate than field grading of TF. 73 21 (29 ) Conjunctival photography is more sensitive than field grading of TF. 72 18 (25 ) Barr ier Increased costs compared to TF grading due to need for camera equipment. 72 43 (60 ) Increased costs compared to TF grading due to increased personnel training/time needed. 72 41 (57 ) Community participation may be low if participants do not receive individual results. 71 39 (55 ) Community participation may be low due to fear/mistrust of having photo taken. 72 38 (53 ) Conjunctival photography grading is subjective and prone to human error. 72 36 (50 ) Community participation may be low due to fear of side effects due to photo flash. 72 27 (38 ) Conjunctival photography is perceived as less accurate compared to other methods. 73 18 (25 ) Infection testing Strength Infection testing can detect pre-symptomatic trachoma. 68 55 (81 ) Eye swabs can be tested for multiple pathogens/diseases. 70 55 (79 ) Infection testing is more accurate than field grading of TF at identifying trachoma. 70 54 (77 ) Infection testing is more accurate than other methods at identifying trachoma. 70 52 (74 ) Eye swabs/infection testing are acceptable to communities as long as education is provided. 69 51 (74 ) Eye swabs have a low chance of causing side effects. 70 51 (73 ) . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint 21 Barr ier Extra work needed to plan for logistics of field collection and sample processing. 71 68 (96 ) Desire for samples to be tested by nearest lab may be a challenge. 71 63 (89 ) Increased costs compared to TF field grading due to storage of samples. 71 62 (87 ) Increased costs compared to TF field grading due to equipment needed. 71 62 (87 ) Requirement for cold chain a challenge due to need for stable electricity. 70 61 (87 ) Increased costs compared to TF field grading due to increased personnel training/time needed. 70 48 (69 ) Poorly trained personnel might cause damage to eyes of participants. 71 40 (56 ) Community participation may be low if participants do not receive individual results. 70 36 (51 ) Community participation may be low due to fear of sample collection process or side effects. 70 34 (49 ) Infection testing may be less accurate than TF field grading due to time between collection and testing. 71 18 (25 ) Serology Strength Blood spots (for other diseases) are familiar to health systems. 66 51 (77 ) Blood spots are easy to collect. 66 47 (71 ) Blood spots are easier to transport and store than eye swabs. 64 42 (66 ) Blood spots (for other diseases) are familiar and acceptable to community members. 66 39 (59 ) Serological testing is more accurate than TF grading at identifying trachoma. 66 29 (44 ) Serological testing is fast. 65 26 (40 ) Serological testing can detect pre-symptomatic trachoma. 66 21 (32 ) B arr ier Extra work needed to plan for logistics of field collection and sample processing. 66 58 (88 ) Increased costs compared to TF field grading due to equipment needed. 66 55 (83 ) Lack of current lab capacity. 65 52 (80 ) Increased costs compared to TF field grading due to increased personnel training/time needed. 66 45 (68 ) Increased costs due to storage of samples. 65 41 (63 ) Increased danger of working with blood/risk of medical errors, compared to other diagnostics. 66 38 (58 ) Field conditions challenging, samples may be destroyed in the field. 66 35 (53 ) Community participation may be low due to misunderstanding or mistrust of test purpose. 66 35 (53 ) Community participation may be low if participants do not receive individual results. 64 31 (48 ) Community participation may be low due to fear of finger pricking. 66 28 (42 ) TF = trachomatous inflammation—follicular 402 The most agreed-upon strength for conjunctival photography was that multiple graders can grade 403 the same photo, increasing accuracy and helping to determine the status of borderline cases (61, 404 84%). In contrast, a third of respondents (24, 33%) said that a strength of conjunctival 405 photography is that the ability to grade the photo away from the field would reduce field time. 406 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint 22 The most agreed-upon barrier to implementation of conjunctival photography is that it would 407 increase the costs compared to TF field grading due to the need for camera equipment (43, 60%), 408 followed closely by increased costs due to the need for increased personnel training and time (41, 409 57%). 410 For infection testing, the most agreed-upon strength was this indicator's ability to detect infection 411 in the absence of clinical signs (55, 81%). The most agreed-upon barriers for this indicator all 412 pertain to challenges around the feasibility of implementation, including the extra work required 413 to plan for logistics of field collection and sample processing (68, 96%), the challenge posed by a 414 desire for samples to be tested by the nearest lab (63, 89%), increased costs compared to field 415 grading of TF posed by the need to store samples (62, 87%) and purchase needed equipment (62, 416 87%), as well as the challenge posed by the requirement for maintaining a cold chain (61, 87%). 417 In contrast to items about feasibility, there were lower levels of agreement with barriers related 418 to low community participation, due either to a lack of receiving individual results (36, 51%), or 419 fear of the sample collection process or side effects (34, 49%). 420 The most agreed-with strengths of serology include the familiarity of health systems with blood 421 spots (for other diseases) (51, 77%), as well as an acknowledgement that blood spots are both 422 easy to collect (47, 71%), and easier to transport and store compared to eye swabs (42, 66%). 423 The most agreed-with barriers to the implementation of serology included the extra work needed 424 to plan for the logistics of field collection and sample processing (58, 88%), the increased costs 425 compared to TF field grading due to equipment needed (55, 83%), and the lack of current lab 426 capacity (52, 80%). Less than half of respondents agreed that community participation may be 427 low if participants do not receive individual test results (31, 48%) or due to a fear of finger 428 pricking (28, 42%). 429 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint 23 The results of the comparison of responses to barriers related to community acceptability by 430 country staff vs. other stakeholders are presented in Table 3. (In this analysis, a higher score 431 indicates a stronger level of agreement that the statement is a barrier to implementation of the 432 indicator.) Country staff had a higher agreement compared to other stakeholders that barriers 433 around community acceptability of conjunctival photography would pose a challenge to 434 implementation of this indicator, while other stakeholders had higher agreement that barriers 435 related to community acceptability of serology would pose a challenge to implementation of this 436 indicator. However, none of the comparisons were significant at an alpha of 0.05. 437 Table 3: Stakeholder agreement with proposed barriers of complementary indicators to 438 trachomatous inflammation—follicular detection by affiliation (country staff vs. other 439 stakeholders) 440 Barrier Country staff a (n=14) Other stakeholders (n=59) b median (IQR) median (IQR) p value c Conjunctival p hotography Community participation may be low due to fear/mistrust of having photo taken. 4.0 (3.6–4.4) 3.0 (2.0–4.0) 0.13 Community participation may be low due to fear of side effects due to photo flash. 3.5 (2.5–4.5) 3.0 (2.0–4.0) 0.46 Community participation may be low if participants do not receive individual results. 4.0 (3.5–4.5) 4.0 (3.5–4.5) 0.37 Infection testing Community participation may be low due to fear of sample collection process or side effects. 3.0 (2.0–4.0) 3.5 (2.5–4.5) 0.70 Community participation may be low if participants do not receive individual results. 4.0 (3.4–4.6) 3.5 (3.0–4.0) 0.67 Serology Community participation may be low due to fear of finger pricking. 2.0 (1.1–2.9) 3.0 (2.0–4.0) 0.08 Community participation may be low due to misunderstanding or mistrust of test purpose. 3.0 (2.1–3.9) 4.0 (3.5–4.5) 0.38 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint 24 Community participation may be low if participants do not receive individual results. 2.5 (1.6–3.4) 4.0 (3.0–5.0) 0.14 IQR = Interquartile range, a: organizational role = "Government" and personal role = "Programmatic/Implementation", b: not all respondents answered every question, c: as measured by a 2- tailed Wilcoxon-rank-sum test. Note: A higher score indicates a stronger level of agreement that the statement is a barrier to implementation of the indicator. 441 Key knowledge gaps entered by respondents are presented as Supplemental File 5. Categories of 442 knowledge gaps identified include: setting and interpreting meaningful thresholds using 443 complementary indicators, understanding test characteristics (e.g. sensitivity and specificity) of 444 each indicator, refining survey design to allow the inclusion of complementary indicators, 445 interpreting serological data, understanding clinical signs and the progression of disease, 446 integrating complementary indicators for trachoma into existing systems, understanding the 447 acceptability and feasibility of complementary indicators, and identifying and mitigating 448 potential issues with training and lab quality assurance. 449

450 From the FGDs, we found that each diagnostic method explored was generally acceptable 451 to community members in Tanzania, with key themes such as not wanting to suffer harm from 452 the diagnostic process and the importance of (perceived) test accuracy apparent. From the IDIs 453 with public health practitioners in Tanzania, we found that these health workers also considered 454 each diagnostic method to be acceptable to community members, as long as appropriate 455 education on the risks and benefits of participation is provided. 456 When global stakeholders were asked to indicate their agreement that given barriers 457 would hinder implementation of complementary indicators, the highest agreement items tended 458 to be those around feasibility, rather than acceptability. The highest consensus of any strength or 459 barrier listed was agreement that the WHO recommendation for field grading of TF is a benefit 460 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint 25 of this method, with 98% of respondents in agreement. When results around community 461 acceptability were stratified based on country staff vs. other stakeholder, no statistically 462 significant differences were found, indicating broad agreement among respondents in the 463 stakeholder survey that community acceptability of each method is unlikely to hinder 464 implementation. This was concordant with the findings from community members in Tanzania, 465 who rated each test type generally acceptable. 466 A study in the Bijagós Islands, Guinea Bissau on different diagnostic tests and sample 467 types for trachoma surveillance (including clinical exams, eye swabs, and finger-prick blood 468 samples) reported that although all the studied test types were generally acceptable, there was a 469 preference among community members for laboratory-based testing; these results were 470 considered more accurate than clinical examination.26 Additionally, sample types that did not 471 require close proximity to the eye (i.e., finger-prick blood samples) were preferred.26 Our study 472 had similar results, with participants noting a perception of improved accuracy from laboratory-473 based methods, although expressing some unease with the eye swab and blood spot procedures, 474 especially with regards to children. This similarity in findings is reassuring given the different 475 locations and population types, indicating some generalizability of the results. Considered in 476 parallel, these results suggest that acceptability would not be a widespread barrier to 477 implementation. 478 Different studies have shown varying levels of acceptability of blood collection. Research 479 in sub-Saharan Africa has shown that collecting blood or other bodily samples from individuals 480 can be challenging, with fears related to blood-stealing and intentional spreading of disease 481 concerning public health interventions documented since colonial times.27 Individual reticence to 482 participation in a trial in The Gambia involving blood collection via finger stick for malaria 483 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint 26 screening was linked to fear of blood-taking due to depletion of life force and fear of 484 exploitation.28 In contrast, a study on the feasibility of bloodspot collection for the surveillance 485 of human African trypanosomiasis in the Democratic Republic of the Congo found that refusal to 486 participate in bloodspot collection was rare, although reasons for non-participation were not 487 analysed.29 Finger-prick blood samples are routinely taken for malaria testing,30 which is 488 commonplace in most places that are also trachoma-endemic.31 Our community member 489 participants expressed broad acceptance of blood spots, with two notable exceptions. 490 First, among the Maasai ethnic group one FGD (out of four) noted a general cultural 491 dislike of blood-giving. This may be related to a noted lack of trust of non-Maasai visitors, an 492 incomplete understanding of the purpose of interventions, and a perception of misalignment 493 between community and government priorities, which has eroded trust in this institution.32 A 494 study conducted in the Sinya Ward (a Maasai community in Longido District) found that, while 495 this community does perceive trachoma as a problem,33 competing priorities can make it difficult 496 to participate in trachoma-related interventions, such as MDA.32 A lack of sensitization around 497 MDA has previously been associated with poor uptake of azithromycin in this area,34 498 underscoring the need for effective messaging on health programs in these and other under-499 served communities, which, as we near trachoma elimination, may represent the last foci of TF. 500 Second, concerns explicitly about the purpose of blood tests were raised (among a variety 501 of ethnic groups), with specific fears mentioned around testing for HIV without the participants' 502 consent. These fears were relayed both by community participants themselves and recognized as 503 a potential reason for low community acceptance by public health practitioners. In this regard, 504 the familiarity and commonplace nature of blood testing is a disadvantage, as community 505 members are aware that their blood samples could be tested for many other diseases, some of 506 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint 27 which (such as HIV) still carry a level of stigma.35 However, familiarity with blood spots was 507 also perceived as an advantage of this method, both by community members, who appreciated 508 the lack of pain or harm caused by the sample procedure, and by public health practitioners, who 509 noted experience with sample collection, processing, and testing by an array of health systems in 510 Tanzania. 511 While public health practitioners in Tanzania did often recognize the familiarity of blood 512 testing (for other diseases) by health systems as an advantage of this method, our study found a 513 lack of acknowledgement of the potential utility of serology for the integration of trachoma 514 monitoring with other disease programs. This group also recognized that a lack of local lab 515 capacity would make the quick processing of blood samples a challenge, which was corroborated 516 as a barrier to implementation by global stakeholders. The Global NTD road map 2021– 2030 517 calls for concerted action to achieve integration across NTD programs and for increased 518 investment in developing lab capacity in order to meet NTD targets.19 The long-term benefit of 519 investments in increasing and maintaining lab capacity of health systems for infection testing and 520 serology should be underscored to stakeholders at all levels. A coordinated approach is needed to 521 leverage the investments and achievements made to date across different disease programs, 522 enabling the building, strengthening and maintenance of laboratory capacity and networks.36 523 Community members found conjunctival photography to be an acceptable method for 524 trachoma diagnosis, noting a general familiarity with the concept of medical imaging and 525 potential benefits of this method compared to field grading of TF. Another study on the 526 acceptability of conjunctival photography among community members in Tanzania found similar 527 sentiment around the utility of photography for grading borderline cases (contributing to greater 528 overall accuracy), but noted an element of concern among community members regarding the 529 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint 28 exposure of children to a photo flash.37 However, this fear pertained only to traditional digital 530 single-lens reflex (DSLR) cameras, compared to photography via smartphone.37 In addition, our 531 study did not find that issues around photo use and confidentiality were a concern among 532 community members. However, these issues were recognized as potential concerns by public 533 health practitioners in our study, as well as both community members and Tropical Data trainers 534 in the previous Tanzanian qualitative study.37 Consent materials for activities incorporating 535 conjunctival photography for research or programmatic use should be very clear as to the risks 536 and benefits of participating. 537 We found that generally, FGDs with men were more favorable to each of the test types 538 compared to FGDs with women, with FGDs with men more likely to have at least one 539 participant indicate that they would hypothetically allow the children under their care to 540 participate in each test type compared to FGDs with women. A previous systematic review and 541 meta-analysis on the acceptance of the COVID-19 vaccine, which included studies from each 542 WHO region, found that male gender was associated with an increased likelihood of vaccine 543 acceptance.38 A systematic review on this topic in sub-Saharan Africa found similar results.39 In 544 terms of care-giving behavior, a multi-country questionnaire similarly found that fathers are 545 more willing than mothers to vaccinate their children against COVID-19, possibly due to 546 differences in risk-taking behaviors among fathers vs. mothers.40 It may be advisable to target 547 messaging around the safety of and risk of discomfort or adverse consequences from each 548 diagnostic method specifically to women caregivers, especially as they are often the decision-549 makers in regards to children’s healthcare.41 550 The high level of stakeholder agreement that a WHO recommendation of field grading of 551 TF is beneficial indicates that additional WHO guidance around complementary indicators may 552 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint 29 be advantageous. In informal workshops, tests for infection and serology have already been 553 advocated for by WHO,15 and the forthcoming guidelines on their use are eagerly awaited. 554 Seventy unique questions regarding knowledge gaps preventing the interpretation and 555 implementation of complementary indicators were identified by global stakeholders. The 556 collation and categorization of these questions will hopefully serve as a useful resource for 557 helping to define future research priorities as we approach the trachoma elimination endgame. 558 Our study had several limitations. Messaging that not all tests would yield individual 559

was included in the FGD topic guide and stated by the research assistant while conducting 560 the FGD. However, it is clear from the transcripts that participants did not adequately understand 561 that they would not always receive individual results for each sample type, making conclusions 562 difficult to apply to surveillance settings, where participants could not expect individual results 563 and/or treatment. Furthermore, we note that asking community members if they would 564 hypothetically participate in each sample collection type may yield different results than actually 565 presenting participants with the opportunity to be tested. In addition, community FGDs and IDIs 566 only took place in a selection of sites within one country. There are likely to be differing views 567 across the trachoma-endemic world. 568 In attempting to evaluate the opinion of a variety of trachoma stakeholders, the 569 stakeholder survey was disseminated to as wide an audience as possible, resulting in an 570 opportunistic sample. While an opportunistic sample has implications for generalizability, the 571 multiple avenues of survey dissemination, the offering of multiple languages of the survey, and 572 the diverse geographic representation of respondents indicate that the results are likely fairly 573 representative of trachoma stakeholders as a whole. 574 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint 30 Strengths and barriers from the IDIs were used to formulate the questions for the 575 stakeholder survey. In order to preserve the intent of these responses, the items in the stakeholder 576 survey could have been perceived as overly positive (for the strengths) and negative (for the 577 barriers), which may have had an effect in biasing stakeholders' level of agreement with each 578 item. In addition, the factual accuracy of the statements offered by the IDIs (for example, that 579 one test may be more "accurate" than another) is debatable, which may have hindered 580 interpretation of participants' agreement with these items. There may have also been potential 581 strengths and barriers to each method not identified by IDI respondents (such as the potential 582 benefit of being able to test blood spots for multiple infections). In each case, to avoid biasing 583 the selection of strengths and barriers presented in the questionnaire, care was taken to preserve 584 the original content and meaning from the IDIs. However, future surveys may wish to provide 585 more comprehensive lists of strengths and barriers and more neutral question wording (for 586 example, by allowing respondents to rate the acceptability of different methods, rather than their 587 agreement with subjective statements). 588 Future work would be needed in order to study the opinions of communities in other 589 regions to have more confidence in the generalizability of our results. A deeper exploration into 590 the attitudes and perceptions of blood testing among the Maasai ethnic group, along with the 591 effectiveness of potential education or sensitization strategies, is warranted in order to ensure this 592 group is equitably represented in future trachoma programming. In addition, a study of non-593 response rates in studies where additional indicators are collected would yield information about 594 the acceptability of different sample collection methods for trachoma detection during real-world 595 program activities. 596

597 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint 31 We found that conjunctival photography, infection testing, and serology were all 598 generally acceptable to community members in Tanzania. Critical themes included participants 599 not wanting to suffer harm from the diagnostic process and the importance of (perceived) test 600 accuracy. Many of the perceived disadvantages mentioned by community members in Tanzania, 601 such as fear of undue discomfort or adverse consequences of participating in the diagnostic 602 process or mistrust of the test purpose (for serology) are potential focus areas for outreach by 603 healthcare workers and other partners prior to implementation. This need for community 604 education on the risks and benefits of participation of each of the methods was underscored in 605 interviews with public health practitioners in Tanzania. According to both Tanzanian public 606 health practitioners and global stakeholders, questions of feasibility remain, especially for 607 infection testing and serology. The fact that in general, all methods were acceptable to 608 community members is positive, as it may be that each indicator has a role to play for trachoma 609 elimination surveillance purposes. 610 611 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint 32 Funding 612 Study activities were funded by the International Trachoma Initiative. AWS is a staff member of 613 the World Health Organization. RBs salary was funded by the Wellcome Trust (206275/Z/17/Z). 614 Competing Interests 615 I have read the journal's policy and the authors of this manuscript have the following competing 616 interests: KKR, PJH, and PME are employees of the International Trachoma Initiative (ITI), a 617 program of The Task Force for Global Health, which receives an operating budget and research 618 funds from Pfizer Inc., the manufacturers of Zithromax (azithromycin). EMHE receives salary 619 support from ITI, and PM and JMM were contractors of ITI for the duration of field activities in 620 Tanzania. Pfizer, Inc. had no role in in the study conception or design; data collection, analysis, 621 or interpretation; writing of the report; or the decision to publish this work. 622 Authors' Contributions 623 KKR, EMHE, and AWS conceived of the study; KKR, RB, TDH, GK, TBM, and EMHE 624 designed the study protocols; PME and PJH reviewed the Tanzania study protocol; GK, JMM, 625 PM, SGM, and KKR carried out field coordination; KKR, JMM, and PM carried out data 626 collection; KKR conducted data analysis; KKR, TBM, JMM, PM, and EMHE conducted data 627 interpretation; KKR prepared the original draft of the manuscript; all authors critically reviewed 628 the manuscript and approved the final manuscript for publication. 629

630 The authors which to thank the participants of the focus group discussions, in-depth interviews, 631 and the survey respondents for their involvement in this project. 632 Data availability statement 633 Study data will be made available upon reasonable request to the corresponding author. 634 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint 33 Disclaimer 635 The authors alone are responsible for the views expressed in this article and they do not 636 necessarily represent the views, decisions or policies of the institutions with which they are 637 affiliated. 638 639 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint 34

640 1. Gambhir M, Grassly NC, Burton MJ, Solomon AW, Taylor HR, Mabey DC, et al. Estimating the 641 Future Impact of a Multi-Pronged Intervention Strategy on Ocular Disease Sequelae Caused by 642 Trachoma: A Modeling Study. Ophthalmic Epidemiol. 2015;22(6):394-402. 643 2. World Health Organization. Validation of elimination of trachoma as a public health problem. 644 Geneva, Switzerland: World Health Organization; 2016. 645 3. World Health Organization. Report of the third Global Scientific Meeting on Trachoma 646 Elimination. Baltimore, MD: Johns Hopkins University; 2010. 647 4. Wright HR, Taylor HR. Clinical examination and laboratory tests for estimation of trachoma 648 prevalence in a remote setting: what are they really telling us? Lancet Infect Dis. 2005;5(5):313-20. 649 5. Ramadhani AM, Derrick T, Macleod D, Holland MJ, Burton MJ. The Relationship between Active 650 Trachoma and Ocular Chlamydia trachomatis Infection before and after Mass Antibiotic Treatment. PLoS 651 Negl Trop Dis. 2016;10(10):e0005080. 652 6. Solomon AW, Harding-Esch E, Alexander ND, Aguirre A, Holland MJ, Bailey RL, et al. Two doses 653 of azithromycin to eliminate trachoma in a Tanzanian community. N Engl J Med. 2008;358(17):1870-1. 654 7. Butcher R, Sokana O, Jack K, Sui L, Russell C, Last A, et al. Clinical signs of trachoma are prevalent 655 among Solomon Islanders who have no persistent markers of prior infection with Chlamydia 656 trachomatis. Wellcome Open Res. 2018;3:14. 657 8. Trotignon G, Jones E, Engels T, Schmidt E, McFarland DA, Macleod CK, et al. The cost of mapping 658 trachoma: Data from the Global Trachoma Mapping Project. PLoS Negl Trop Dis. 2017;11(10):e0006023-659 e. 660 9. Harding-Esch EM, Burgert-Brucker CR, Jimenez C, Bakhtiari A, Willis R, Bejiga MD, et al. Tropical 661 Data: Approach and Methodology as Applied to Trachoma Prevalence Surveys. Ophthalmic 662 Epidemiology. 2023;30(6):544-60. 663 10. Jimenez C, Gower E, Harding-Esch E, West SK. Harnessing photography and image recognition 664 technology to aid in the elimination of trachoma. Community Eye Health. 2022;35(114):24. 665 11. Solomon AW, Pavluck AL, Courtright P, Aboe A, Adamu L, Alemayehu W, et al. The Global 666 Trachoma Mapping Project: Methodology of a 34-Country Population-Based Study. Ophthalmic 667 Epidemiology. 2015;22(3):214-25. 668 12. Harding-Esch EM, Naufal F, Saboya M, Jimenez C, West SK. Use of photography for support of 669 trachoma grading: progress report. 2021. 670 13. Martin DL, Saboyà-Díaz MI, Abashawl A, Alemayeh W, Gwyn S, Hooper PJ, et al. The use of 671 serology for trachoma surveillance: Current status and priorities for future investigation. PLoS Negl Trop 672 Dis. 2020;14(9):e0008316-e. 673 14. World Health Organization. Trachoma Alternative Indicators Study Data review. Geneva, 674 Switzerland: World Health Organization; 2016. 675 15. World Health Organization. Informal consultation on end-game challenges for trachoma 676 elimination. Task Force for Global Health, Decatur, United States of America. Geneva: World Health 677 Organization; 2022. 7–9 December 2021. 678 16. World Health Organization. WHO handbook for guideline development - 2nd ed. Geneva, 679 Switzerland: World Health Organization; 2014. 680 17. Senyonjo L, Downs P, Schmidt E, Bailey R, Blanchet K. Lessons learned for surveillance strategies 681 for trachoma elimination as a public health problem, from the evaluation of approaches utilised by 682 Guinea worm and onchocerciasis programmes: A literature review. PLoS Negl Trop Dis. 683 2021;15(1):e0009082-e. 684 18. Oldenburg CE, Aragie S, Amza A, Solomon AW, Brogdon J, Arnold BF, et al. Can we eradicate 685 trachoma? A survey of stakeholders. Br J Ophthalmol. 2021;105(8):1059-62. 686 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint 35 19. World Health Organization. Ending the neglect to attain the Sustainable Development Goals – A 687 road map for neglected tropical diseases 2021–2030. Geneva , Switzerland World Health Organization; 688 2020. 689 20. International Trachoma Initiative. Global Atlas of Trachoma Decatur, Georgia 2020 [Available 690 from: trachomaatlas.org]. 691 21. Magnani R, Sabin, Keith, Saidel, Tobi, Heckathorn, Douglas. Review of sampling hard-to-reach 692 and hidden populations for HIV surveillance. AIDS. 2005;19:S67-S72. 693 22. Lumivero. NVivo 13 (2020, R1). 694 23. Gale NK, Heath G, Cameron E, Rashid S, Redwood S. Using the framework method for the 695 analysis of qualitative data in multi-disciplinary health research. BMC Medical Research Methodology. 696 2013;13(1):117. 697 24. SurveyMonkey Inc. San Mateo, California, USA [Available from: www.surveymonkey.com ]. 698 25. R Core Team. R: A language and environment for statistical computing. Vienna, Austria: R 699 Foundation for Statistical Computing; 2021. 700 26. Sahota RS, Sanha S, Last A, Cassama E, Goncalves A, Kelly AH, et al. Acceptability and perceived 701 utility of different diagnostic tests and sample types for trachoma surveillance in the Bijagos Islands, 702 Guinea Bissau. Trans R Soc Trop Med Hyg. 2021;115(8):847-53. 703 27. Geissler PW, Pool R. Editorial: Popular concerns about medical research projects in sub-Saharan 704 Africa – a critical voice in debates about medical research ethics. Tropical Medicine & International 705 Health. 2006;11(7):975-82. 706 28. O'Neill S, Dierickx S, Okebe J, Dabira E, Gryseels C, d'Alessandro U, et al. The Importance of 707 Blood Is Infinite: Conceptions of Blood as Life Force, Rumours and Fear of Trial Participation in a Fulani 708 Village in Rural Gambia. PLoS One. 2016;11(8):e0160464. 709 29. Inocencio da Luz R, Phanzu DM, Kiabanzawoko ON, Miaka E, Verlé P, De Weggheleire A, et al. 710 Feasibility of a dried blood spot strategy for serological screening and surveillance to monitor 711 elimination of Human African Trypanosomiasis in the Democratic Republic of the Congo. PLoS Negl Trop 712 Dis. 2021;15(6):e0009407. 713 30. World Health Organization. Malaria surveillance, monitoring and evaluation: a reference 714 manual. Geneva, Switzerland World Health Organization; 2018. 715 31. World Health Organization. World malaria report 2022. Geneva; 2022. 716 32. Mtuy TB, Bardosh K, Ngondi J, Mwingira U, Seeley J, Burton M, et al. Understanding hard-to-717 reach communities: local perspectives and experiences of trachoma control among the pastoralist 718 Maasai in northern Tanzania. J Biosoc Sci. 2021;53(6):819-38. 719 33. Mtuy TB, Burton MJ, Mwingira U, Ngondi JM, Seeley J, Lees S. Knowledge, perceptions and 720 experiences of trachoma among Maasai in Tanzania: Implications for prevention and control. PLoS Negl 721 Trop Dis. 2019;13(6):e0007508-e. 722 34. Desmond N, Solomon AW, Massae PA, Lema N, Anemona A, Foster A, et al. Acceptability of 723 azithromycin for the control of trachoma in Northern Tanzania. Trans R Soc Trop Med Hyg. 724 2005;99(9):656-63. 725 35. Gooden TE, Mkhoi ML, Mdoe M, Mwalukunga LJ, Senkoro E, Kibusi SM, et al. Barriers and 726 facilitators of people living with HIV receiving optimal care for hypertension and diabetes in Tanzania: a 727 qualitative study with healthcare professionals and people living with HIV. BMC Public Health. 728 2023;23(1):2235. 729 36. Mfuh KO, Abanda NN, Titanji BK. Strengthening diagnostic capacity in Africa as a key pillar of 730 public health and pandemic preparedness. PLOS Glob Public Health. 2023;3(6):e0001998. 731 37. Bisanzio D, Butcher R, Turbé V, Matsumoto K, Dinesh C, Massae P, et al. Accuracy, acceptability 732 and feasibility of photography for use in trachoma surveys: a mixed methods study in Tanzania. Int 733 Health. 2023. 734 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint 36 38. Wang Q, Yang L, Jin H, Lin L. Vaccination against COVID-19: A systematic review and meta-735 analysis of acceptability and its predictors. Prev Med. 2021;150:106694. 736 39. Azanaw J, Endalew M, Zenbaba D, Abera E, Chattu VK. COVID-19 vaccine acceptance and 737 associated factors in 13 African countries: A systematic review and meta-analysis. Front Public Health. 738 2022;10:1001423. 739 40. Goldman RD, Yan TD, Seiler M, Parra Cotanda C, Brown JC, Klein EJ, et al. Caregiver willingness 740 to vaccinate their children against COVID-19: Cross sectional survey. Vaccine. 2020;38(48):7668-73. 741 41. McCauley AP, West S, Lynch M. Household decisions among the Gogo people of Tanzania: 742 determining the roles of men, women and the community in implementing a trachoma prevention 743 program. Soc Sci Med. 1992;34(7):817-24. 744 745 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 1, 2024. ; https://doi.org/10.1101/2024.01.29.24301764doi: medRxiv preprint