Table 3

Examples of CDATs (children <5 years) used in LMIC countries sourced from published literature

Tool and country	Type of tool and item selection (translation±adaptation)	Reliability	Validity	Comments
	Method of assessment	Internal consistency	Criterion
	Age	Interobserver	Discriminant
		Test-retest	Construct
Kilifi Developmental Checklist (KDC) Kenya28	Mixture of items from government local screening test, Western tools and LMIC tools developed elsewhere	Excellent*	Items range from poor to good correlation with parental report†	Aimed to be used by low-skilled workers. Identified that secondary school level education was minimal for an assessor.
	Direct and Report	Excellent‡
	11–109 months	Excellent‡	Done
Kilifi Developmental Inventory (KDI) Kenya30	Assessment based on KDC with additional items from western tools	Excellent*	Good correlation with maternal reports†	Designed to monitor changes over time. Involved parental focus groups so procedures acceptable to community.
	Direct observation	Excellent‡	Sensitive to neurodevelopmental disorders and underweight children.
	6–35 months	Excellent‡
Developmental Milestone Checklist Kenya45	Screening Items from Western tools	Good*	Very good correlation with KDI†.30	Structured interview. Future plans to develop cut-offs using clinical samples. Can be administered by those with little training.
	Report		Stunted children had lower scores.
	3–24 months	Good‡	Done
Guide to monitoring child development. Turkey46	Screening Core ideas were adapted from three seminal western models of child development.	Excellent*	Excellent agreement with comprehensive evaluation§ Sensitivity 0.88; 95% CI 0.69 to 0.96 Specificity 0.93; 95% CI 0.83 to 0.97	Brief, open-ended precoded interview with parent. No specific cognitive domain questions. May not be appropriate for monitoring ‘at-risk’ populations.
	Report	Good§
	0–24 months
Rapid neurodevelopmental assessment RNA 0–2 years Bangladesh47	Screening Mixture of neurological examination and neurodevelopment from western tools		Significant differences in scores in children with neurodevelopment impairment	Used in a small high-risk population and identified infants requiring intervention for retinopathy of prematurity. RNA identified additional problems such as impairments in vision and hearing.
	Direct	Excellent§	Detected difference in scores between urban rural children in older age groups
	0–24 months
Rapid neurodevelopmental assessment. RNA 2–5 years Bangladesh48	Screening Mixture of neurological exam and development assessment from western tools		Sensitivity 80–90% Specificity 60–78%	Comparison with ‘gold-standard’ tests already developed for Bangladesh.
	Direct	Excellent§	Stunted children had lower scores.
	24–60 months
Malawi Developmental Assessment Tool. Malawi32	Screening Western developmental tool and new items locally devised	Excellent49		Good strategy to devise socioemotional items.
	Direct and report	60–80% of items at least fair§	Sensitivity 97% Specificity 82% Sensitive to malnutrition except social developmental domain
	0–72 months	99–100% scoring at least fair§	Good predictive validity50 51
INTERGROWTH-21st Neurodevelopment Assessment29	Screening Items from selection of Western and LMIC tools			Project specific criteria to create a tool for use in children from middle-class and upper class families across low-income, middle-income and high-income settings, carried out by non-specialists. Also evaluates vision, hearing and sleep-wake cycle.
		Direct and report	Good
		24 months	Excellent
Systematic approach to assess nutritional influences. Indonesia17	Assessment Items from selection of Western tools	Items at least fair*		Clear hypothesis driven tool creation. May not be suitable as a general developmental assessment.
	Direct	At least fair agreement¶	Expected relationship with maternal depression and maternal education
	22–55 months	Items at least fair†
Interpretation of statistical tests are listed below.16 52 These need to be interpreted in context of application and population.17 53
Levels of κ Levels§, R†, intraclass‡ correlation and α coefficients*.		Level of proposed agreement¶ (%)		Levels of clinical or practical significance
<0.4		<70%		Poor
0.4–0.59		70–79		Fair
0.60–0.74		80–89		Good
0/75–1.00		90–100		Excellent

CDAT, child development assessment tool; LMICs, low-income and middle-income countries.