Reproducibility package for Child-Direct Disability Screening in Tanzanian Primary Schools: Estimating Administrative Undercounts and Implications for Disability Prevalence in Low- and Middle-Income Countries

Policy for children in low- and middle-income countries depends on knowing who they are and what they need. For children with disabilities, that information is limited, costly to collect, and often understated in existing data systems. Administrative school records in Tanzania report disability prevalence of 0.56%—an order of magnitude below caregiver reported household surveys across 51 LMICs, which suggest rates near 14%. Caregiver surveys provide a broader picture but are expensive to implement at scale, limiting their use for policy.

We evaluate an alternative: administering the Child Functioning Module (CFM) by directly asking primary school children in classrooms across nine regions of Tanzania. We benchmark these child-direct reports (3,246 children) against caregiver reports from UNICEF MICS6 (483,164 children, 51 countries), Tanzania’s Annual School Census, and optometrist-led vision screening. Overall prevalence and severity are similar to caregiver reports, but child-direct data identify substantially higher rates of difficulties in vision, hearing, communication, and social interaction. For vision and hearing, available medical screening evidence from Tanzania and other settings likewise indicates higher rates than caregiver reports. Child-direct estimates exceed administrative records by a factor of 13.2 (95% CI: 10.3–16.2). We therefore estimate a lower-bound prevalence of 14.4% (95% CI: 14.1–14.7) across 51 LMICs.

These findings suggest that school-based, child-direct screening is a feasible way to improve identification, supporting more effective use of development budgets for classroom accommodations, teacher deployment, and examination arrangements.