Skip to main content
. 2020 Mar 12;16(4):e12991. doi: 10.1111/mcn.12991

Table 2.

Multivariable zero‐inflated negative binomial regression model comprising nine risk factors for child food contamination, out of characteristics of 58 caregivers and their households in Maputo, Mozambique

Logistic portion
Risk factor Coefficienta z 95% CI p value
Soap available—Yes (ref: No) 20 20 [18, 22] <.01
HWWS after cleaning nappies—Yes (ref: No) 21 16 [18, 24] <.01
Type of food—Salad (ref: Not salad) −21 −15 [−24, −18] <.01
Food stored after cooking (ref: Cooked and not stored; not calculated: Uncooked food) −22 −20 [−24, −19] <.01
Food stored in covered container (ref: Stored uncovered; not calculated: Not stored after cooking) 20 20 [18, 23] <.01
Constant −20 −22 [−22, −18] <.01
Negative binomial portion
Risk factors Coefficientb z 95% CI p value
Type of food—Porridge (ref: Not porridge) −3.2 −1.7 [−6.9, 0.59] .10
Type of food—Cooked vegetables (ref: Not cooked vegetables) 1.1 1.5 [−0.36, 2.5] .14
Food refrigerated (ref: Stored at ambient temperature;not calculated: Not stored) −25 −3.0 [−42, −8.8] <.01
Wealth quintile—2nd (ref: 1st quintile; not calculated: 3rd, 4th, 5th quintile) −2.7 −1.1 [−7.6, 2.1] .27
Constant 6.4 12 [5.3, 7.4] <.01

Abbreviations: CI, confidence interval; HWWS, handwashing with soap.

a

Coefficient relates to the increase in log odds of excess zeros.

b

Coefficient relates to the increase in log count.