Table 2.

Quality assessment by outcome

	Quality assessment					Summary of findings
				Directness		No of events*
No of studies	Design	Limitations	Consistency	Generalizability to population of interest	Generalizability to intervention of interest	Intervention	Control	RR / SMD (95% CI)
Neonatal mortality: moderate outcome specific quality of evidence
Three	RCT	No significant heterogeneity, fixed effect model used	One of the three studies suggest benefit	All studies from the developing countries	Two of the studies used sunflower while one used coconut oil as emollient	04	11	RR: 0.73 [0.56, 0.94]
Hospital acquired infection: moderate outcome specific quality of evidence
Three	RCT	No significant heterogeneity, fixed effect model used	All studies suggest benefit	All studies from the developing countries	Studies used sunflower and coconut oil	19	33	RR: 0.50 [0.36, 0.71]
Weight: moderate outcome specific quality of evidence
Four studies (five data sets)	RCT	No significant heterogeneity, fixed effect model used	Three studies suggest benefit	All studies from the developing countries	Studies used coconut, sunflower and soybean oil	149	151	SMD: 98.04 [42.64, 153.45]
Weight gain (g/kg/day): moderate outcome specific quality of evidence
Two studies (three data sets)	RCT	Significant heterogeneity so a random effect model used	One study suggested benefit	All studies from developing countries	Studies used coconut and sunflower oil	95	97	SMD: 1.57 [0.79, 2.36]
Length: moderate outcome specific quality of evidence
Two studies (three data sets)	RCT	Significant heterogeneity so a random effect model used	Two studies suggest benefit	All studies from developing countries	Studies used coconut and sunflower oil	124	128	SMD: 0.33 [-0.15, 0.81]
Head circumference: moderate outcome specific quality of evidence
Two studies (three data sets)	RCT	No significant heterogeneity, fixed effect model used	None of the studies suggest benefit	All studies from developing countries	Studies used coconut and sunflower oil	124	128	SMD: 0.05 [-0.30, 0.41]