FEMIS Disability-Identification Algorithm — Targeting Inclusive-Education Grants in Fiji

Fiji's Ministry of Education, with the Nossal Institute for Global Health (University of Melbourne), the Pacific Disability Forum and the Pacific Islands Forum Secretariat (funded by Australia's DFAT), developed a rule-based algorithm combining the UNICEF/Washington Group Child Functioning Module (CFM) with Learning and Support Needs (LSN) data to more accurately flag disability status of students in Fiji's Education Management Information System (FEMIS).

To test whether combining CFM functioning data with LSN data (personal-assistance needs, educational adaptations, assistive-technology use) identifies disability status more accurately than CFM alone, for the purpose of allocating Special and Inclusive Education Grants.

Researchers Beth Sprunt and Manjula Marella surveyed 472 children aged 5-15 across 10 special and 5 mainstream schools in four administrative divisions, with 98 teachers participating, between March and July 2015. The combined data were built into a rule-based (non-machine-learning) algorithm, now embedded in FEMIS as the 'Student Learning Profile', which since a March 2022 national implementation package forms the basis of individualised education plans and grant-eligibility decisions, with defined staff qualification requirements.

The study found a moderate-to-high correlation (Spearman's rho 0.681) between assistance needs and functional difficulty; among children reporting only 'some difficulty' on the CFM, 16.2% still needed substantial assistance — a group the CFM alone would have missed.

The authors note the approach depends on FEMIS's granular data architecture and is not directly transferable to manual, census-based education systems; the speech-impairment subsample was too small for meaningful analysis; and the study covered only five clinical impairment types, with broader accessibility assessed separately.