PURPOSE:Trachoma surveys are used to estimate the prevalence of trachomatous inflammation-follicular (TF) to guide mass antibiotic distribution. These surveys currently rely on human graders, introducing a significant resource burden and potential for human error. This study describes the development and evaluation of machine learning models intended to reduce cost and improve reliability of these surveys. METHODS:Fifty-six thousand seven hundred twenty-five everted eyelid photographs were obtained from 11,358 children of age 0 to 9 years in a single trachoma-endemic region of Ethiopia over a 3-year period. Expert graders reviewed all images from each examination to determine the estimated number of tarsal conjunctival follicles and the degree of trachomatous inflammation-intense. The median estimate of the 3 grader groups was used as the ground truth to train a MobileNetV3 large deep convolutional neural network to detect cases with TF. RESULTS:The classification model predicted a TF prevalence of 32%, which was not significantly different from the human consensus estimate (30%; 95% confidence interval of difference, -2 to +4%). The model had an area under the receiver operating characteristic curve of 0.943, F1 score of 0.923, 88% accuracy, 83% sensitivity, and 91% specificity. The area under the receiver operating characteristic curve increased to 0.995 when interpreting nonborderline cases of TF. CONCLUSIONS:Deep convolutional neural network models performed well at classifying TF and detecting the number of follicles evident in conjunctival photographs. Implementation of similar models may enable accurate, efficient, large-scale trachoma screening. Further validation in diverse populations with varying TF prevalence is needed before implementation at scale.

BACKGROUND:Promotion of facial cleanliness is recommended for the elimination of blinding trachoma, largely because of observational studies that have found an association between various measures of facial uncleanliness and trachoma. However, when a field grader assesses both facial cleanliness and trachoma, associations may be biased. Assessment of photographs of the face and conjunctiva by masked graders may provide a less biased estimate of the relationship between facial cleanliness and trachoma. METHODS:Face photographs, conjunctival photographs, and conjunctival swabs were obtained on a random sample of 0-9-year-old children from each of 40 communities in Amhara region, Ethiopia. Face photographs were assessed for the presence of seven measures of an unclean face (i.e., wet nasal discharge, dry nasal discharge, wet ocular discharge, dry ocular discharge, food, dust/dirt, and flies) by three independent masked photo-graders. Conjunctival photographs were similarly graded in a masked fashion for signs of clinically active trachoma. Conjunctival swabs were processed for Chlamydia trachomatis DNA. RESULTS:Of 2073 children with complete data, 808 (39%) had evidence of clinically active trachoma, 150 (7%) had evidence of ocular chlamydia infection, and 2524 (91%) had at least one measure of an unclean face. Dry ocular discharge had the strongest association with clinically active trachoma (age- and sex-adjusted prevalence ratio [PR] 1.4, 95% CI 1.2-1.6) and ocular chlamydia infection (PR 1.9, 95%CI 1.3-2.9), although significant associations were observed between each of the measures of facial uncleanliness and trachoma. CONCLUSIONS:Masked assessment of face and conjunctival photographs confirmed prior observational studies that have noted associations between various measures of facial uncleanliness and trachoma. The causal relationship between facial uncleanliness and trachoma is unclear since many features used to measure facial cleanliness (e.g., ocular discharge, nasal discharge, and flies) could be consequences of antecedent ocular chlamydia infection. TRIAL REGISTRATION:NCT02754583, clinicaltrials.gov.

OBJECTIVES:A 6-week course of tetracycline eye ointment is an alternative to single -dose oral azithromycin in annual mass drug administration for trachoma control. Compliance with the recommended tetracycline eye ointment regimen has not been well characterised when administered as part of a trachoma control program. METHODS:A routine mass drug administration for trachoma was carried out in 40 communities in the Amhara region of Ethiopia. Two tubes of tetracycline eye ointment, to be administered twice daily for 6 weeks, was offered to all children under 6 months of age, to pregnant women who declined to take azithromycin, and to all individuals with a macrolide allergy. Seven weeks following the mass drug administration, a treatment compliance survey was performed for all community members documented to have received tetracycline eye ointment during the mass drug administration. RESULTS:Of the 491 individuals documented as having received tetracycline eye ointment from the treatment records, 367 completed the survey, of which 214 recalled being offered tetracycline eye ointment. A total of 105 (49%) respondents reported taking ≥1 daily dose of tetracycline eye ointment on most days of the week for at least the first week. Only 20 (9%) respondents reported taking at least 1 tetracycline eye ointment dose per week for 6 weeks. The most common reasons for low compliance included 'saving it for a future infection' and 'stopped because I (or my child) seemed healthy'. The odds of low compliance were greater for those who reported not having adequate counselling (e.g., odds ratio [OR] 5.3, 95% CI 2.5-28.9 when low compliance was defined as not taking a tetracycline eye ointment dose for most days of at least the first week). CONCLUSIONS:Compliance with tetracycline eye ointment was low when administered by a trachoma program during a routine mass drug administration, especially for those reporting inadequate counselling. Further research with a larger sample size and varied settings is warranted to better understand and improve compliance.

The purpose of this study was to investigate the correlation between bacterial load of Chlamydia trachomatis as measured from quantitative polymerase chain reaction (qPCR) and the relative clinical severity of trachomatous inflammation. Individuals with trachoma from rural communities in Ethiopia had photographs taken as well as swabs obtained of the upper tarsal conjunctivas. Conjunctival swabs were processed with PCR assay, which provided quantitative results of ocular chlamydial load. A series of 125 conjunctival photographs were ranked from least to most severe according to clinical severity for follicular and papillary conjunctivitis. Higher intensity rankings of trachomatous inflammation were associated with higher chlamydial load for both follicular inflammation (Spearman's ρ = 0.43; P <0.001) and papillary inflammation (Spearman's ρ = 0.50; P <0.001). Rankings of trachomatous inflammation may be a clinically meaningful indicator of trachoma.

Background WHO promotes the SAFE strategy for the elimination of trachoma as a public health programme, which promotes surgery for trichiasis (ie, the S component), antibiotics to clear the ocular strains of chlamydia that cause trachoma (the A component), facial cleanliness to prevent transmission of secretions (the F component), and environmental improvements to provide water for washing and sanitation facilities (the E component). However, little evidence is available from randomised trials to support the efficacy of interventions targeting the F and E components of the strategy. We aimed to determine whether an integrated water, sanitation, and hygiene (WASH) intervention prevents the transmission of trachoma. Methods The WASH Upgrades for Health in Amhara (WUHA) was a two-arm, parallel-group, cluster-randomised trial in 40 rural communities in Wag Hemra Zone (Amhara Region, Ethiopia) that had been treated with 7 years of annual mass azithromycin distributions. The randomisation unit was the school catchment area. All households within a 1.5 km radius of a potential water point within the catchment area (as determined by the investigators) were eligible for inclusion. Clusters were randomly assigned (at a 1:1 ratio) to receive a WASH intervention either immediately (intervention) or delayed until the conclusion of the trial (control), in the absence of concurrent antibiotic distributions. Given the nature of the intervention, participants and field workers could not be masked, but laboratory personnel were masked to treatment allocation. The WASH intervention consisted of both hygiene infrastructure improvements (namely, construction of a community water point) and hygiene promotion by government, school, and community leaders, which were implemented at the household, school, and community levels. Hygiene promotion focused on two simple messages: to use soap and water to wash your or your child's face, and to always use a latrine for defecation. The primary outcome was the cluster-level prevalence of ocular chlamydia, measured annually using conjunctival swabs in a random sample of children aged 0-5 years from each cluster at 12, 24, and 36 month timepoints. Analyses were done in an intention-to-treat manner. This trial is ongoing and is registered at ClinicalTrials.gov, NCT02754583. Findings Between Nov 9, 2015, and March 5, 2019, 40 of 44 dusters assessed for eligibility were enrolled and randomly allocated to the trial groups (20 dusters each, with 7636 people from 1751 households in the intervention group and 9821 people from 2211 households in the control group at baseline). At baseline, ocular chlamydia prevalence among children aged 0-5 years was 11% (95% CI 6 to 16) in the WASH group and 11% (5 to 18) in the control group. At month 36, ocular chlamydia prevalence had increased in both groups, to 32% (24 to 41) in the WASH group and 31% (21 to 41) in the control group (risk difference across three annual monitoring visits, after adjustment for prevalence at baseline: 3.7 percentage points; 95% CI -4 .9 to 12.4; p=0.40). No adverse events were reported in either group. Interpretation An integrated WASH intervention addressing the F and E components of the SAFE strategy did not prevent an increase in prevalence of ocular chlamydia following cessation of antibiotics in an area with hyperendemic trachoma. The impact of WASH in the presence of annual mass azithromycin distributions is currently being studied in a follow-up trial of the 40 study clusters. Continued antibiotic distributions will probably be important in areas with persistent trachoma. Copyright (C) 2021 The Author(s). Published by Elsevier Ltd.

Multiplex bead assays (MBAs) for serologic testing have become more prevalent in public health surveys, but few studies have assessed their test performance. As part of a trachoma study conducted in a rural part of Ethiopia in 2016, dried blood spots (DBS) were collected from a random sample of 393 children aged 0 to 9 years, with at least two separate 6-mm DBS collected on a filter card. Samples eluted from DBS were processed using an MBA on the Luminex platform for antibodies against 13 antigens of nine infectious organisms: Chlamydia trachomatis, Vibrio cholera, enterotoxigenic Escherichia coli, Cryptosporidium parvum, Entamoeba histolytica, Camplyobacter jejuni, Salmonella typhimurium Group B, Salmonella enteritidis Group D, and Giardia lamblia. Two separate DBS from each child were processed. The first DBS was run a single time, with the MBA set to read 100 beads per well. The second DBS was run twice, first at 100 beads per well and then at 50 beads per well. Results were expressed as the median fluorescence intensity minus background (MFI-BG), and classified as seropositive or seronegative according to external standards. Agreement between the three runs was high, with intraclass correlation coefficients of > 0.85 for the two Salmonella antibody responses and > 0.95 for the other 11 antibody responses. Agreement was also high for the dichotomous seropositivity indicators, with Cohen's kappa statistics exceeding 0.87 for each antibody assay. These results suggest that serologic testing on the Luminex platform had strong test performance characteristics for analyzing antibodies using DBS.

Background. Current guidelines recommend community-wide mass azithromycin for trachoma, but a targeted treatment strategy could reduce the volume of antibiotics required. Methods. In total, 48 Ethiopian communities were randomized to mass, targeted, or delayed azithromycin distributions. In the targeted arm, only children aged 6 months to 5 years with evidence of ocular chlamydia received azithromycin, distributed thrice over the following year. The primary outcome was ocular chlamydia at months 12 and 24, comparing the targeted and delayed arms (0-5 year-olds, superiority analysis) and the targeted and mass azithromycin arms (8-12 year-olds, noninferiority analysis, 10% noninferiority margin). Results. At baseline, the mean prevalence of ocular chlamydia in the 3 arms ranged from 7% to 9% among 0-5 year-olds and from 3% to 9% among 8-12 year-olds. Averaged across months 12-24, the mean prevalence of ocular chlamydia among 0-5 year-olds was 16.7% (95% confidence interval [CI]: 9.0%-24.4%) in the targeted arm and 22.3% (95% CI: 11.1%-33.6%) in the delayed arm (P =.61). The final mean prevalence of ocular chlamydia among 8-12 year-olds was 13.5% (95% CI: 7.9%-19.1%) in the targeted arm and 5.5% (95% CI: 0.3%-10.7%) in the mass treatment arm (adjusted risk difference 8.5 percentage points [pp] higher in the targeted arm, 95% CI: 0.9 pp-16.1 pp higher). Conclusions. Antibiotic treatments targeted to infected preschool children did not result in significantly less ocular chlamydia infections compared with untreated communities and did not meet noninferiority criteria relative to mass azithromycin distributions. Targeted approaches may require treatment of a broader segment of the population in areas with hyperendemic trachoma.

Abstract Objective To determine whether a water, sanitation and hygiene intervention could change hygiene behaviours thought to be important for trachoma control. Methods We conducted a cluster-randomized trial in rural Ethiopia from 9 November 2015 to 5 March 2019. We randomized 20 clusters to an intervention consisting of water and sanitation infrastructure and hygiene promotion and 20 clusters to no intervention. All intervention clusters received a primary-school hygiene curriculum, community water point, household wash station, household soap and home visits from hygiene promotion workers. We assessed intervention fidelity through annual household surveys. Findings Over the 3 years, more wash stations, soap and latrines were seen at households in the intervention clusters than the control clusters: risk difference 47 percentage points (95% confidence interval, CI: 41–53) for wash stations, 18 percentage points (95% CI: 12–24) for soap and 12 percentage points (95% CI: 5–19) for latrines. A greater proportion of people in intervention clusters reported washing their faces with soap (e.g. risk difference 21 percentage points; 95% CI: 15–27 for 0–5 year-old children) and using a latrine (e.g. risk difference 9 percentage points; 95% CI: 2–15 for 6–9 year-old children). Differences between the intervention and control arms were not statistically significant for many indicators until the programme had been implemented for at least a year; they did not decline during later study visits. Conclusion The community- and school-based intervention was associated with improved hygiene access and behaviours, although changes in behaviour were slow and required several years of the intervention.

Diarrhea is a leading cause of death among children aged less than five years globally. Most studies of pediatric diarrhea rely on caregiver-reported stool consistency and frequency to define the disease. Research on the validity of caregiver-reported diarrhea is sparse. We collected stool samples from 2,398 children participating in two clinical trials in the Amhara region of Ethiopia. The consistency of each stool sample was graded by the child's caregiver and two trained laboratory technicians according to an illustrated stool consistency scale. We assessed the reliability of graded stool consistency among the technicians, and then compared the caregiver's grade with the technician's grade. We also tested if the illustrated stool consistency scale could improve the validity of caregiver's report. The weighted kappa measuring the agreement between the two laboratory technicians reached 0.90 after 500 stool samples were graded. The sensitivity of caregiver-reported loose or watery stool was 15.5% (95% confidence interval [CI]: 9.7, 24.2) and the specificity was 98.4% (95% CI 97.1, 99.1). With the illustrated scale, the sensitivity was 68.5% (95% CI: 58.5, 77.1) and the specificity was 86.1% (95% CI: 79.3, 90.9). The results indicate that caregiver-reported stool consistency using the terms "loose or watery" does not accurately describe stool consistency as graded by trained laboratory technicians. Given the predominance of using caregiver-reported stool consistency to define diarrheal disease, the low sensitivity identified in this study suggests that the burden of diarrheal disease may be underestimated and intervention effects could be biased. The illustrated scale is a potential low-lost tool to improve the validity of caregiver-reported stool consistency.