. 2019 Dec 17;14(12):e0226405. doi: 10.1371/journal.pone.0226405

Table 4. Behavioral outcomes adjusting for severe malaria complications and treatment.

	Children ≥6 years
	N (obs.), N	Coefficient (95% CI)	P
Adjusting for disease severity
Socio-emotional function^a
Internalizing behavior	357, 149	0.01 (-0.21, 0.24)	0.919
Externalizing behavior	357, 149	0.55 (0.22, 0.87)	0.001
Executive function^b
Global Executive Composite	251, 137	0.48 (0.14, 0.82)	0.006
Emergent Metacognition Index	251, 137	0.27 (-0.09, 0.62)	0.139
Behavior Regulation Index	251, 137	0.67 (0.32, 1.01)	0.0002
Adjusting for acidosis
Socio-emotional function^a
Internalizing behavior	334, 139	-0.06 (-0.28, 0.16)	0.605
Externalizing behavior	334, 139	0.51 (0.18, 0.85)	0.003
Executive function^b
Global Executive Composite	228, 127	0.45 (0.09, 0.81)	0.016
Emergent Metacognition Index	228, 127	0.23 (-0.15, 0.62)	0.225
Behavior Regulation Index	228, 127	0.64 (0.27, 1.01)	0.0008
Adjusting for jaundice
Socio-emotional function^a
Internalizing behavior	357, 149	0.01 (-0.22, 0.23)	0.957
Externalizing behavior	357, 149	0.51 (0.18, 0.83)	0.003
Executive function^b
Global Executive Composite	251, 137	0.45 (0.12, 0.79)	0.009
Emergent Metacognition Index	251, 137	0.23 (-0.12, 0.59)	0.190
Behavior Regulation Index	251, 137	0.64 (0.30, 0.99)	0.0003
Adjusting for fluid resuscitation and transfusion
Socio-emotional function^a
Internalizing behavior	357, 149	-0.01 (-0.23, 0.22)	0.957
Externalizing behavior	357, 149	0.54 (0.22, 0.87)	0.001
Executive function^b
Global Executive Composite	251, 137	0.48 (0.14, 0.82)	0.006
Emergent Metacognition Index	251, 137	0.26 (-0.10, 0.62)	0.151
Behavior Regulation Index	251, 137	0.68 (0.34, 1.02)	0.0001

Abbreviations: CI, confidence interval; P, P-value; N (obs.), number of observations in the model; N, the number of children in the analysis.

^aAssessed using the Child Behavior Checklist (CBCL)

^bAssessed using the Behavior Rating Inventory of Executive Function (BRIEF)

All linear mixed models were fitted with a subject specific random intercept and visit as a categorical variable (baseline, 6 months, 12 months, 24 months).

Adjusted models included age, sex, height-for-age, weight-for-age, socioeconomic status, home environment, maternal education, preschool exposure, number of seizures during hospitalization (in cerebral malaria), parenteral antimalarial treatment (quinine vs. artemisinin), year of enrollment and test administrator as fixed effects. As coma is part of LODS, the presence of coma was not included in the model with LODS.