Health-related quality of life in children with inflammatory brain disease

Table 3 Impact of baseline patient characteristics and presenting clinical features on trajectories of health-related quality of life (HRQoL) over time

	Child self-reported HRQoL Parameter estimate β (SE^b)	P value	Parent proxy-reported HRQoL Parameter estimate β (SE^b)	P value
Predictors for impaired overall HRQoL trajectories over time
Female Gender (vs. Male)	−7.6 (4.6)	0.10	−3.8 (4.4)	0.39
Seizures	−15.6 (4.6)^a	< 0.01	−11.5 (4.4)	< 0.01
Cognitive Dysfunction	− 15.0 (4.8)^a	< 0.01	−15.2 (4.6)	< 0.01
Hemiparesis	−2.9 (4.6)	0.53	−2.5 (4.4)	0.58
Predictors for impaired physical functioning trajectories over time
Female Gender (vs. Male)	−8.9 (5.6)	0.12	−7.8 (5.9)	0.19
Seizures	−13.4 (5.7)	0.02	−9.9 (6.0)	0.10
Cognitive Dysfunction	−14.4 (5.9)	0.02	−11.4 (6.2)	0.07
Hemiparesis	−7.3 (5.7)	0.20	−1.7 (5.9)	0.77
Predictors for impaired psychosocial functioning trajectories over time
Female Gender (vs. Male)	−7.1 (4.7)	0.13	−1.6 (4.3)	0.71
Seizures	−16.2 (4.7)	< 0.01	−12.4 (4.3)	< 0.01
Cognitive Dysfunction	−15.3 (4.9)^a	< 0.01	−17.7 (4.4)^a	< 0.01
Hemiparesis	−0.1 (4.7)	0.98	−0.1 (4.7)	0.44

Legend: Health-related quality of life (HRQoL) was measured using PedsQL 4.0 Generic Core Scales. Child-reported HRQoL was measured using the child self-reporting PedsQL questionnaire. PedsQL scores are between 0 and 100, with higher values representing better HRQoL. Negative changes in HRQoL associated with a clinical phenotype therefore represent risk factors for impaired HRQoL. 464 child self-reported PedsQL questionnaires were completed and used in this analysis. Each parameter estimate was derived from a separate, univariable linear regression model, adjusted for time, evaluating the impact of the predictor alongside time on HRQoL. Regression models were adjusted for repeated measures with an autoregressive covariance structure to establish whether these relationships changed statistically significantly with time
n = 145 patients
^aasterisk indicates that the HRQoL impairment associated with the presence of the clinical variable at diagnosis is time-dependent, with the gap in HRQoL between patients with the clinical variable at diagnosis and patients who do not have the clinical variable at diagnosis decreasing with increasing time. This was only assessed in clinical parameters with statistically significant parameter estimates
^bSE Standard error

ISSN: 1546-0096