Study design, period and setting
A community-based cross-sectional study design was used. Fieldwork was undertaken from March 11, 2017 to April 26, 2107. The estimated population of Wadla district was 128,170 with 64,574 males and 63,596 females. There were 28,414 households in this district with an average of 4.5 persons per house. The district had 1 general hospital, 7 health centers, and 20 health posts.
Population
The sampling frame was children aged 1–5 years in 150 rural villages of Wadla district. The study units were rural households that had preschool children.
Sample size determination
We estimated the required sample size using the single population proportion formula.We assumed, based on previous surveys, an observed prevalence of active trachoma (12.1%) (14), which we wished to estimate with 95% confidence within ±5%. We used a design effect of 1.5, and allowed for 10% non- response rate. Through multiplying the sample size by the design effect of 1.5 and incorporating a 10% non-response rate, we estimated 273 children needed to be fram ed in selected households.
Sampling technique
A multistage cluster sampling technique was applied. Wadla district had 20 kebeles (sub-districts) with 247 villages. Twelve of the kebeles were rural, whereas eight of the kebeles were urban. Regarding the villages, 150 of the 247 villages were rural. We used simple random sampling to select 30 of the 150 rural villages. There were 967 households in the selected 30 villages, but only 499 of those households had preschool children. Thus, those 499 households were visited. Heads of households were interviewed for socio-demographic and economic information, plus housing and environmental conditions, and all children aged between 1 and 5 years who had been resident in the district for at least six months were invited to be examined. Graders used the WHO simplified trachoma grading scheme to grade signs of trachoma (15) (Figure1).
The sample size calculated was 273 using single population proportion formula, but as the sampling procedure was cluster sampling, the numbers of screened children were 596 from all villages.
Definitions
Clean face: A face of child that was free of eye discharges, nose discharges and flies at the time of examination
Preschool: Children aged greater than and equal to 1 year and less than or equal to five years
Village: A grouping of homes that contained at least 30 households organized as one peasant association
Fly in home: When there are countable flies in the house during data collection by observation
Active trachoma: The presence of at least one of the two signs of active trachoma according to the WHO simplified trachoma grading scheme (TF or TI) in at least one eye (16).
Trachomatous inflammation—follicular (TF): the presence of five or more follicles each having a diameter of at least 0.5mm in the central part of the upper tarsal conjunctiva (16).
Trachomatous inflammation—intense (TI): a pronounced inflammatory thickening of the upper tarsal conjunctiva that obscures more than half of the normal deep tarsal blood vessels (16).
Trachomatous scarring (TS): the presence of easily visible scarring in the upper tarsal conjunctiva (16)
Trachomatous trichiasis (TT): the presence of at least one eyelash rubs on the eyeball or evidence of removal of in-turned eyelashes in the two weeks before examination (16).
Corneal opacity (CO): the presence of easily visible corneal opacity over the pupil (16)
Exclusion and inclusion criteria
All children in the appropriate age range who had lived in the district for at least 6 months, who were resident in selected villages and available at the time of study were invited to be included. Children who were seriously ill or for whom informed consent was not given by parents or guardians were excluded.
Measurements
The outcome variable was active trachoma and measured by physical examination. However, a number of dependent variables were considered that includes socio-demographic, enviromental, hygiene and sanitation, and children’s demography.
Data collection tools and procedures
Face to face interviews, observation using a checklist and clinical eye examination were used. Experienced health informatics professionals using structured interview questions that prepared from literatures (17, 18) collected data on socio-demographic status, environmental, and housing conditions. All questionnaires on socio-demographic status, housing, and environmental condition, observation checklist, and eye examination tools were pretested and validated before data collection in Kosomender, Meket district, a district bordering Wadla to the north. A household wealth index was developed using composite indicators for rural residents’assets: livestock ownership, size of agricultural land and quantity of crop production.
Two integrated eye care workers performed the eye examination. Those integrated eye care workers were ophthalmic nurses who had been previously trained for a total duration of one month for the purposes of contributing to the 2013–2014 national trachoma survey. The Carter Center delivered that previous training using both pictures and live patients as media of instruction. However, for the purpose of this survey, the trachoma graders undertook refreshment training for 5 days. This involved examination of 58 live patients and 100 pictures of different trachoma signs. Trainers, whose grades were used as the gold-standard assessment assessed graders. The training was also delivered for interviewers. Interviewers assisted graders by recording clinical grades, and data related to each household's socio-demographic status and environmental situation. The training was emphasized on the objectives and procedures of data collection, and mode of communication between graders and interviewers.
When undertaking the actual fieldwork for the study, graders initially observed the eyelashes and cornea of study subjects, looking for TT and CO, then everted the upper lid and inspected the upper tarsal conjunctiva for TF, TI, and TS. Binocular lenses (×2.5) and penlight torches were used (4) to magnify the examined eye.
Data analysis and presentation
The data were checked for completeness, coded, and entered into Epi-info version 7, and then transferred to SPSS version 23 for analysis. The data were checked for normality using Hosmer-Lemeshow-goodness-of-fit. A univariable analysis model were carried out, and variables that had p-value of <0.25 in a binary logistic regression model were included to the multivariable logistic regression analysis. Potential co-linearity was also considered and tested using multi co-linearity model in considering tolerance and variance inflection factor (VIF). Variables with a p-value of < 0.05 in the multivariable logistic regression analysis were considered as statistically significant. A principal component analysis was performed to categorize households’ wealth into poorer, poorest, middle, richer, and richest. However, for the presentation of the variables, the wealth index grouped into three; lowest, middle, and highest. The procedure of eye examination and results reporting presented in figure2. Both active trachoma and cicatricial trachoma were modelled as outcome variables. Thus, children were screened for both Active and cicatricial trachoma (Figure2).
Data quality assurance
The questionnaire was prepared in English and translated to Amharic, then re-translated to English (to check for accuracy) by individuals fluent in both English and Amharic. Both graders and one of the researchers, principal investigator had participated previously in a community-based trachoma survey. The interviewers also had previous experience in community-based data collection.