Skip to main content
. 2021 Oct 28;11(10):e046829. doi: 10.1136/bmjopen-2020-046829

Table 4.

Multivariable linear regression demonstrating the adjusted association of intimate partner violence (IPV) and composite scores for developmental domains in complete case analysis sample

Composite cognitive (n=549) Composite language (n=598) Composite motor
(n=521)
Adjusted* Adjusted† Adjusted‡ Adjusted‡
Coefficient (95% CI) Coefficient (95% CI) Coefficient (95% CI) Coefficient (95% CI)
Emotional IPV score −0.32 (−0.60 to 0.04)* −0.35 (−0.69 to 0.01)* −0.58 (−0.95 to 0.20)*
Physical IPV score −0.42 (−0.75 to 0.09)*

Adjusted models were run only where IPV subtype was associated with outcome explored (p<0.05) in bivariate analyses. For cognitive and language scores, only emotional IPV was associated in bivariate analyses and therefore run in multivariable models. Due to collinearity, emotional and physical IPV were run in separate multivariable models for motor development. IPV subtypes, depression and alcohol dependence were included as continuous variables in all models.

*Adjusted for child sex, HIV exposure, length-for-age z-scores, weight-for-age z-scores and weight-for-length z-scores at 24 months.

†Adjusted for maternal education, recruitment site, weight-for-age z-score at birth, child sex and HIV exposure.

‡Adjusted for child age at assessment and weight-for-length z-scores at 24 months.