The 2024–2025 EDHS included the UNICEF/Washington Group Child Functioning Module for ages 5–17 for the first time, providing national estimates. Investigators reported an overall design-weighted prevalence of any functional difficulty of 9.5% (95% CI 8.1–11.1). Daily depressive symptoms were present in 6.8% and daily anxiety in 6.6% of children; affective problems therefore dominated the measured burden. Seventy point three percent of affected children had only affective difficulty. By contrast, difficulty in at least one non-affective domain affected 2.8% (95% CI 2.2–3.7), and each individual non-affective domain fell in the 0.5–1.4% range.
The study team defined functional difficulty as a response of “a lot of difficulty” or “cannot do at all” in any of 11 CFM domains, or reports of daily anxiety or depression. The source did not report additional clinical validation or objective testing; the analysis relied on caregiver-reported module responses captured during the EDHS interview.
Prevalence varied by region. The lowest regional prevalence was 1.5% in Southwest Regional State, while Oromia Regional State had the highest observed prevalence at 16.5%. The data indicated strong within-family clustering. The cohort showed a sibling correlation of 67.7%, signalling that functional difficulty aggregates within households. Cluster-level intraclass correlation (ICC) was 20.9%, indicating community-level clustering beyond family effects. These clustering metrics reflect the multilevel structure the investigators modelled — children nested in mothers and mothers nested in survey clusters.
Geographic heterogeneity and high family-level correlation suggest local social, environmental, or reporting factors play a large role. The preprint does not provide causal explanations for regional differences; the dataset and code are available per the authors for further exploration.
In adjusted multilevel logistic regression models, the authors found non-attendance at school was associated with substantially higher odds of non-affective functional difficulty (adjusted odds ratio [AOR] 2.82, 95% CI 2.01–3.96). Sex differences were observed: girls had lower odds of non-affective difficulty compared with boys (AOR 0.63, 95% CI 0.48–0.84). The models accounted for the hierarchical survey design with children nested within mothers and within clusters.
The investigators separated outcomes into any difficulty, non-affective difficulty, and affective difficulty for modelling. The paper reports that these associations were robust to the sensitivity analysis that excluded interviewers with statistically outlying response patterns (see next section). The source did not report additional mediation analyses or longitudinal follow-up because the EDHS is cross-sectional.
The study team identified large interviewer effects concentrated among a small subset of interviewers. Six interviewers — representing 7.9% of the weighted sample — contributed 58.2% of weighted cases of daily anxiety or depression. Within the same clusters, those interviewers recorded daily affective symptoms for 56.4% of children, whereas other interviewers in the same clusters recorded daily symptoms for 2.1% of children. These discrepancies suggest differential reporting or interviewer-driven measurement artefact.
To examine the influence of these outliers, the authors reran analyses excluding 11 interviewers flagged statistically as outliers. After exclusion, estimated prevalence of any functional difficulty fell to 4.6% (95% CI 3.6–5.8) and prevalence of daily anxiety or depression dropped to 3.3%. Despite the reduced point estimates, the associations between school non-attendance and non-affective difficulty, and between sex and non-affective difficulty, remained materially unchanged. The source did not provide qualitative investigation of interviewer training records or field monitoring reports; such operational details were not reported in the preprint.
Investigators used cross-sectional EDHS data linked between children and mothers; weighted sample counts cited were 11,122 children with a weighted n = 10,726 and 4,739 mothers. The analysis code, including the record-linkage procedure, is available in Additional file 2 and the DHS Program data can be requested on registration. Multilevel logistic regression models included random intercepts for mothers and clusters; interviewer effects were examined with random-intercept models and within-cluster comparisons.
Key limitations arise from the survey and analytic design. The EDHS is cross-sectional; the preprint therefore cannot establish causation between school non-attendance and functional difficulty. The functional measures are caregiver-reported responses to the CFM; the authors did not report clinical validation against objective assessments. Large interviewer-related measurement variability substantially affected prevalence estimates, indicating that field implementation and interviewer behaviour materially influence results. The study team noted that excluding outlier interviewers changed prevalence estimates but did not change primary associations; the preprint does not report whether further field investigation or re-interviews were conducted.
Unanswered questions left in the paper include the drivers of regional variation, the clinical severity or service needs of children with affective versus non-affective difficulty, and the operational steps required to reduce interviewer-related bias. The authors recommend consideration of appropriate interventions and resource allocation, but the preprint does not specify programmatic approaches or thresholds for service planning.