Bayesian multilevel modelling of the association between socio-economic status and stunting among under-five-year children in Tanzania

Table 5 Multivariable analysis of the association between SES indices and CIAF

		Survey year
Model/Factors	2004/5	2010	TDHS 2015/16
Model 1: MLE-MGLMM	AOR (95% CI)	AOR (95% CI)	AOR (95% CI)
A: Fixed factors
MPI	0.99 [0.95, 1.03]		1.09 [1.04, 1.13]***
WAMI		0.37 [0.25, 0.54]***
B: Random factors
Variance (region)	0.064 [0.026, 0.155]	0.035 [0.014, 0.083]	0.0246 [0.0102, 0.0595]
Variance (enumeration area)
Model 2: BE-MGLMM	BPOR (95% CCI)	BPOR (95% CCI)	BPOR (95% CCI)
A: Fixed factors
MPI	0.99 [0.9, 1]		1.1 [1.1, 1.2]
WAMI		0.36 [0.3, 0.4]
B: Random factors
Variance (region)	0.07 [0.02, 0.14]	0.03 [0.01, 0.067]	0.03 [0.008, 0.075]
Variance (region > enumeration)	0.007 [0.003, 0.013]	0.007 [0.002, 0.013]	0.012 [0.003, 0.026]

Adjusted for
Child characteristics: age, sex, size at birth, diarrhoea status, birth type, birth orders, duration breastfed, type of delivery, fever, place of delivery. Maternal characteristics: age, age at first birth, marital status, BMI, media access, decision making, ANC attendant and ANC visits. Household characteristics: age of household head, sex of the household head, food security, total no. of under-fives, household size and residence area
BPOR means Bayesian posterior odds ratio, CCI means credible confidence intervals, AOR means adjusted odds ratio, CI means credible intervals
***represents statistical significance at < 1%

ISSN: 2072-1315