FULL-LENGTH ORIGINAL RESEARCH

Correlates of health-related quality of life in children with

drug resistant epilepsy
*†Lauryn Conway, *†Mary Lou Smith, ‡Mark A. Ferro, §Kathy N. Speechley, ¶Mary B. Connoly,
#O. Carter Snead, #**Elysa Widjaja, and the PEPSQOL Study Team1

Epilepsia, 57(8):1256–1264, 2016

doi: 10.1111/epi.13441

SUMMARY
Objective: Health-related quality of life (HRQL) is compromised in children with epi-
lepsy. The current study aimed to identify correlates of HRQL in children with drug
resistant epilepsy.
Methods: Data came from 115 children enrolled in the Impact of Pediatric Epilepsy
Surgery on Health-Related Quality of Life Study (PEPSQOL), a multicenter prospec-
tive cohort study. Individual, clinical, and family factors were evaluated. HRQL was
measured using the Quality of Life in Childhood Epilepsy Questionnaire (QOLCE), a
parent-rated epilepsy-specific instrument, with composite scores ranging from 0 to
100. A series of univariable linear regression analyses were conducted to identify signif-
icant associations with HRQL, followed by a multivariable regression analysis.
Results: Children had a mean age of 11.85  3.81 years and 65 (56.5%) were male. The
mean composite QOLCE score was 60.18  16.69. Child age, sex, age at seizure onset,
duration of epilepsy, caregiver age, caregiver education, and income were not signifi-
cantly associated with HRQL. Univariable regression analyses revealed that a higher
number of anti-seizure medications (p = 0.020), lower IQ (p = 0.002), greater seizure
Lauryn Conway is a frequency (p = 0.048), caregiver unemployment (p = 0.010), higher caregiver depres-
doctoral student in the sive and anxiety symptoms (p < 0.001 for both), poorer family adaptation, fewer family
Department of resources, and a greater number of family demands (p < 0.001 for all) were associated
Psychology at the with lower HRQL. Multivariable regression analysis showed that lower child IQ
University of Toronto. (b = 0.20, p = 0.004), fewer family resources (b = 0.43, p = 0.012), and caregiver unem-
ployment (b = 6.53, p = 0.018) were associated with diminished HRQL in children.
Significance: The results emphasize the importance of child cognition and family vari-
ables in the HRQL of children with drug-resistant epilepsy. The findings speak to the
importance of offering comprehensive care to children and their families to address
the nonmedical features that impact on HRQL.
KEY WORDS: Pediatric epilepsy, Health-related quality of life, Family.

Epilepsy in children can often have catastrophic conse- including not only the disease state but also the person’s
quences on multiple domains of health-related quality of life physical, psychological, and social well-being. This term
(HRQL). HRQL is defined by the World Health Organiza- refers to the impact, both subjective and objective, of dys-
tion (WHO) as a broad, multidimensional construct function associated with illness or injury, and treatment.1 A
Accepted May 19, 2016; Early View publication June 28, 2016.
*Department of Psychology, Hospital for Sick Children, Toronto, Ontario, Canada; †Department of Psychology, University of Toronto, Toronto,
Ontario, Canada; ‡Departments of Psychiatry and Behavioural Neurosciences and Pediatrics, McMaster University, Hamilton, Ontario,
Canada; §Departments of Paediatrics, Epidemiology & Biostatistics, University of Western Ontario, London, Ontario, Canada; ¶Division of Neurology,
Department of Pediatrics, BC Children’s Hospital, Vancouver, British Columbia, Canada; Divisions of #Neurology; and **Division of Diagnostic
Imaging, Hospital for Sick Children, Toronto, Ontario, Canada
Address correspondence to Elysa Widjaja, Hospital for Sick Children, 555 University Avenue, Toronto, ON M5G 1X8, Canada. E-mail: elysa.
widjaja@sickkids.ca
1
PEPSQOL Study Team members available in Appendix.
Wiley Periodicals, Inc.
© 2016 International League Against Epilepsy

1256

Key Points
• This is the first study to comprehensively examine a
wide range of predictors of HRQL in children with
drug-resistant epilepsy
• Lower child IQ, fewer number of resources available
to families, and caregiver unemployment were
uniquely associated with reduced HRQL
substantial body of literature suggests that HRQL is
impaired in children with epilepsy,2 who have higher rates
of behavioral, cognitive, emotional, social, and academic
problems compared to healthy children or children with
other chronic health conditions.3–10
In light of these findings, a predominant objective of the
management of pediatric epilepsy is to both improve and
prevent potential declines in patient HRQL.11 Correspond-
ingly, a comprehensive understanding of the factors that
may predict and explain variations in children’s HRQL war-
rants rigorous investigation, given the potential implications
for both prognosis and intervention. A recent meta-analysis
identified several clinical and psychosocial risk factors
associated with HRQL, including duration of epilepsy, sei-
zure type, frequency, and severity, number and side effects
of anti-seizure medications, presence of a comorbidity, par-
ental anxiety, and family socioeconomic status.12 There is
also some suggestion that both cognitive and behavior prob-
lems correlate with children’s HRQL.4,13,14
Noting an underrepresentation of studies examining fam-
ily factors in his meta-analysis, Ferro12 underscored the
need for more methodologically sound research aimed at
identifying meaningful family factors that impact HRQL in
children with epilepsy. Family factors include caregiver
psychopathology, such as depression and anxiety, and the
family environment more broadly, such as satisfaction with
family relationships, resources available to aid families’
adaptation, accumulation of stressful events, and so on.
Depression and anxiety are present in up to 56% of parents
of children with epilepsy15,16 and have been shown to corre-
late with children’s HRQL.12,13,15,17 Families of children
with epilepsy experience significantly more stress, anxiety,
and restrictions in family life,18 higher levels of dissatisfac-
tion with their social situation,16 poorer parent–child rela-
tionships,19 and more problems with family functioning and
adaptation.19 A few studies have documented associations
between family factors and HRQL in children with new-
onset epilepsy11,17,20,21 as well as a mixed sample of chil-
dren with active and medication-managed epilepsy.14 At the
more severe end of the spectrum, drug-resistant epilepsy is
defined as failure to achieve sustained seizure freedom with
adequate trials of two or more tolerated and appropriately
chosen antiseizure medications.22 It is unknown if family
1257
HRQL in Children with Epilepsy
factors will also influence HRQL in children with drug
resistant epilepsy, or whether medical factors including the
severity of seizures and number of antiseizure medications
may have a dominant effect on HRQL in these children.
In line with the aforementioned dearth in the literature,
the current study aimed to investigate potential correlates of
HRQL in a multicenter sample of children with drug resis-
tant epilepsy. Potential correlates included child age, sex,
age at seizure onset, duration of epilepsy, number of anti-
seizure medications, seizure frequency, intelligence quo-
tient (IQ), caregiver education, caregiver age, caregiver
work status, socioeconomic status (income), caregiver psy-
chopathology (depression and anxiety), and family environ-
mental factors (including family adaptation, resources and
demands).
Methods
Participants
Data were collected as part of the Impact of Pediatric Epi-
lepsy Surgery on Health-Related Quality of Life (PEPS-
QOL) study, a multicenter prospective cohort study.
Participants were recruited during the period of evaluation
of candidacy for epilepsy surgery from eight centers across
Canada that treat children with drug resistant epilepsy,
including the Hospital for Sick Children in Toronto, BC
Children’s Hospital in Vancouver, Alberta Children’s
Hospital in Calgary, Ste. Justine Children’s Hospital in
Montreal, IWK Health Center in Halifax, The Children’s
Hospital in Winnipeg, Royal University Hospital in Saska-
toon, and McMaster Children’s Hospital in Hamilton.
Patients with drug-resistant localization-related epilepsy
(assessed by clinical semiology and/or electroencephalogra-
phy) age 4–18 years were eligible for inclusion in the study.
Exclusion criteria included prior or planned resective or
nonresective (e.g., corpus callosotomy) epilepsy surgery, or
vagus nerve stimulator placement; neurometabolic disor-
ders, neurodegenerative disorders, and genetic epilepsy syn-
dromes such as genetic generalized epilepsy and epileptic
encephalopathies. Furthermore, children and families were
excluded if they were unable to complete the questionnaires.
The overall response rate was 76.35%, ranging from
50.00% to 100.00% across centers (see Table 1).
The current study utilized baseline data from the longitu-
dinal PEPSQOL study. One-third of participants underwent
or are scheduled to undergo surgery; however, the period of
evaluation is still underway for some of the sample.
Measures
Children’s HRQL was assessed using the original (Aus-
tralian) version of the Quality of Life in Childhood Epilepsy
Questionnaire (QOLCE), a 76-item parent-rated epilepsy-
specific instrument.23,24 The five main QOLCE domains
include physical activity (physical restrictions and energy/
fatigue), cognition (attention/concentration, memory,
Epilepsia, 57(8):1256–1264, 2016
doi: 10.1111/epi.13441
1258
L. Conway et al.

events and changes in life events experienced by a family

Table 1. Initial number of participants approached,
during the previous year. All instruments have been shown
subsequent numbers recruited, and response rates by
center to have acceptable reliability and validity.26,28,30–32 Addi-
tional details on the aforementioned measures are available
Participants Participants Total Response in a Supporting Information.
Centre approached who refused enrolled rate, %a
Clinical data including age at seizure onset, duration of
The Hospital for 106 25 81 76 epilepsy, number of antiseizure medications, and seizure
Sick Children
BC Children’s 16 4 12 75
frequency were collected from medical records. Sociode-
Hospital mographic information including child age, sex, caregiver
Alberta Children’s 4 0 4 100 education, caregiver age, caregiver work status, and socioe-
Hospital conomic status (income) were obtained through a parent
Ste. Justine 10 3 7 70 questionnaire.11,20 Finally, the intelligence quotient (IQ) of
Children’s
Hospital
all patients was assessed using the Wechsler Abbreviated
IWK Health 1 0 1 100 Scale of Intelligence-II (for ages 6–18 years)33 or the
Center Wechsler Preschool and Primary Scale of Intelligence-IV
The Children’s 4 2 2 50 (for ages 4–5 years).34
Hospital at The study protocol received approval from the ethics
Winnipeg
Royal University 3 1 2 67
boards of all participating hospitals and consent was
Hospital obtained from each participant.
McMaster 8 2 6 75
Children’s Statistical analyses
Hospital The statistical software SPSS 21.0 (IBM Corporation,
Overall 152 37 115 76
Armonk, NY, U.S.A.) was used for data analyses. Descrip-
a
Response rate calculation: [(total enrolled/participants tive statistics used to describe the sample included mean
approached) 9 100].
and standard deviation for continuous measures, and fre-
quency and percentages for categorical variables. A series
of univariable analyses were conducted using linear regres-
language, and other cognition), well-being (depression, anx- sion models to assess associations between potential corre-
iety, control/helplessness, and self-esteem), social activity lates (child age, sex, age at seizure onset, duration of
(social interactions, social activities, and stigma), and epilepsy, number of anti-seizure medications, seizure fre-
behavior.23 The composite HRQL score is the unweighted quency, IQ, caregiver education, caregiver age, caregiver
average of 16 QOLCE subscales, ranging from 0 to 100. work status, socioeconomic status [income], caregiver psy-
Higher scores indicate better HRQL. The QOLCE has good chopathology [depressive and anxiety symptoms], and fam-
validity and reliability.23,24 In the current sample, Cron- ily environment [adaptation, resources and demands]) and
bach’s a = 0.93. The “not applicable” response option was HRQL. Variables that showed a significance of p < 0.05
excluded from QOLCE items due to Research Electronic were included in the subsequent multivariable linear regres-
Data Capture (REDCap) system constraints, which would sion model. To assess whether patterns of associations var-
have impacted the calculation of both sub-scale and overall ied by QOLCE domain, we fit the resulting multivariable
totals. REDCap is a secure web application for managing linear regression model for the outcomes of physical activ-
online databases.25 ity, cognition, well-being, social activity, and behavior.
Caregiver’s mood (i.e., depression and anxiety) was Finally, we used the Bootstrap approach of random sam-
assessed with the Quick Inventory of Depressive Symp- pling with replacement with 1,000 iterations to validate the
tomatology (QIDS)26 and the Generalized Anxiety Disor- multivariable regression for the composite QOLCE scores.
der (GAD) – 7.27,28 In terms of the family environment, The assumptions of residual normality, linearity, and
family adaptation was measured with the Family Adaptabil- homoscedasticity were evaluated in the multivariable
ity, Partnership, Growth, Affection, and Resolve (Family regression model, and no violations were detected. There
APGAR),29,30 which assessed satisfaction with family was also no evidence of multicollinearity among the factors
relationships. Family resources were quantified using two as assessed by variance inflation factor values (VIF < 2 for
subscales (family mastery and health, extended family all variables). Before proceeding with data analysis, all vari-
social support)11 from the Family Inventory of Resources ables were screened for missing data. Three variables
for Management (FIRM),31 which assessed resources avail- (FIRM, FILE, and QIDS) had < 1% missing data, whereas
able to aid families’ adaptation to stressful events. Finally, IQ and income had 7.8% and 8.7% missing data, respec-
family demands were measured using the Family Inventory tively. There were no missing data in the remaining vari-
of Life Events and Changes (FILE),32 which quantified the ables. Little’s MCAR test indicated that data were missing
accumulation of simultaneous normal and non-normal life completely at random, v² (42) = 48.81, p = 0.190.
Epilepsia, 57(8):1256–1264, 2016
doi: 10.1111/epi.13441
 1259
HRQL in Children with Epilepsy

Correspondingly, the fully conditional specification multi-

Table 2. Characteristics of children with epilepsy,
ple imputation method (5 imputations) was implemented
caregivers, and families
using the Multiple Imputation procedure in IBM SPSS ver-
sion 21.0 (IBM Corporation). Pooled estimates of the five Sample (n = 115)
imputations are reported. The level of statistical signifi- Child characteristics
cance was set at 0.05. Age, years (SD) 11.85 (3.81)
Sex, n (%)
Male 65 (56.5)
Results Female 50 (43.5)
Age (years) at seizure onset (SD) 6.29 (3.86)
Sample (N = 115) characteristics are presented in Duration (years) of epilepsy (SD) 5.53 (3.90)
Table 2. Children had a mean age of 11.85  3.81 years, Seizure frequency, n (%)
High (daily or weekly) 63 (54.8)
an average IQ of 85.03  18.80 (range: 40–125), and 65
Low (monthly or less) 52 (45.2)
(56.5%) were male. The mean age at seizure onset was Number of antiseizure medications, n (%) 1.84 (0.85)
6.29  3.86 years and the average duration of epilepsy was IQ (SD) 85.03 (18.80)
5.53  3.90 years. Fifty-five percent of the sample experi- Quality of life, QOLCE (SD) 60.18 (16.69)
enced seizures daily or weekly. Caregivera characteristics
Age, n (%)
Most of the caregivers completing the questionnaires
20–24 1 (0.9)
were between 40 and 49 years of age (52.2%), were female 25–29 3 (2.6)
(83.5%), and were employed (68.7%). Mean scores on care- 30–39 32 (27.8)
giver’s depression (QIDS) and anxiety (GAD) were 40–49 60 (52.2)
5.82  3.90 (range 0–17) and 4.90  4.80 (range 0–20), 50–59 18 (15.7)
60–69 1 (0.9)
respectively. Family environment measures had the follow-
Sex, n (%)
ing mean scores: Family APGAR 7.17  2.35 (range 0– Male 19 (16.5)
10), FIRM 49.83  10.98 (range 24–72), and FILE Female 96 (83.5)
8.79  6.41 (0–30). Employment status, n (%)
The mean composite QOLCE score was 60.18  16.69 Employed 79 (68.7)
Not employed 36 (31.3)
(range 18.56–92.44). Box-and-whisker plots of QOLCE
Education, n (%)
scores by the five main QOLCE domains are presented in College/university education 75 (65.2)
Figure 1. The mean scores for each domain were as follows: Less than college/university education 40 (34.8)
physical activity 51.68  15.94, cognition 54.37  23.46, Depressive symptoms, QIDS (SD) 5.82 (3.90)
well-being 68.49  15.21, social activity 65.89  26.58, Anxiety symptoms, GAD (SD) 4.90 (4.80)
Family characteristics
and behavior 64.39  15.15.
Adaptation, Family APGAR (SD) 7.17 (2.35)
Resources, FIRM (SD) 49.83 (10.98)
Univariable analyses Demands, FILE (SD) 8.79 (6.41)
Results of the univariable linear regression models fit for Annual household income, n (%)
the outcome of HRQL (composite QOLCE scores) are pre- Prefer not to say 10 (8.7)
<$10,000 2 (1.7)
sented in Table 3. Child age, sex, age at seizure onset, and
$10,000–$19,999 3 (2.6)
duration of epilepsy were not significantly associated with $20,000–$29,999 5 (4.3)
HRQL. Furthermore, caregiver and family sociodemo- $30,000–$39,999 3 (2.6)
graphic factors (i.e., caregiver age, caregiver education, and $40,000–$49,999 3 (2.6)
socioeconomic status [income]) were also nonsignificant. $50,000–$59,999 7 (6.1)
$60,000–$69,999 8 (7)
In children, greater number of antiseizure medica-
$70,000–$79,999 13 (11.3)
tions (b = 4.17, SE = 1.80, p = 0.020), greater seizure $80,000–$89,999 4 (3.5)
frequency (b = 6.11, SE = 3.09, p = 0.048), and lower $90,000–$99,999 15 (13)
IQ (b = 0.25, SE = 0.08, p = 0.002) were associated ≥$100,000 42 (36.5)
with lower HRQL. Furthermore, unemployment (b = 8.43, a
Caregiver refers to the person completing the questionnaires.
SE = 3.28, p = 0.010), higher depressive symptoms
(b = 1.75, SE = 0.37, p < 0.001), and higher anxiety
symptoms (b = 1.36, SE = 0.30, p < 0.001) in caregivers
were associated with lower HRQL in children. Family char- events (b = 0.77, SE = 0.12, p < 0.001) were associated
acteristics (adaptation, Family APGAR; resources, FIRM; with higher HRQL. Finally, higher family demands
and demands, FILE) were also significantly associated with (b = 1.15, SE = 0.22, p < 0.001) were associated with
QOLCE scores. Specifically, higher satisfaction with family lower HQRL.
relationships (b = 2.21, SE = 0.63, p < 0.001) and greater Given that one domain of the QOLCE captures cognition
resources available to aid families’ adaptation to stressful and that one of our correlates was IQ, we conducted an
Epilepsia, 57(8):1256–1264, 2016
doi: 10.1111/epi.13441
1260
L. Conway et al.

Table 4. Multivariable linear regression model fit for the

outcome of HRQL (composite QOLCE scores)
Parameter b Coefficient 95% CI
IQ 0.20* 0.07, 0.34
Number of antiseizure medications 2.87 6.00, 0.23
Seizure frequency 0.50 5.73, 4.75
Caregiver’s work status 6.53* 1.11, 11.96
Caregiver’s depressive symptoms, 0.54 1.42, 0.36
QIDS
Caregiver’s anxiety symptoms, GAD 0.26 1.02, 0.49
Family adaptation, APGAR 0.08 1.14, 1.23
Family resources, FIRM 0.43* 0.10, 0.80
Family demands, FILE 0.27 0.80, 0.22
Figure 1.
Box-and-whisker plots of QOLCE scores by five main domains. *p < 0.05.
The box-and-whisker plots display the minimum, lower quartile
(25th percentile), median, upper quartile (75th percentile), and
maximum value for each QOLCE domain).
Epilepsia ILAE The set of factors (number of antiseizure medications, sei-
zure frequency, IQ, caregiver work status, caregiver QIDS,
caregiver GAD, Family APGAR, FIRM, and FILE) were
Table 3. Simple linear regression models fit for the significantly associated with QOLCE scores: R = 0.652,
outcome of HRQL (composite QOLCE scores) F9,112 = 8.50, p < 0.001. The coefficient of determination
suggests that 43% of variance in QOLCE scores can be
b Coefficient
Parameter (SE) 95% CI p-Value accounted for by the set of factors. Child IQ, caregiver work
status, and family resources were significantly associated
Child characteristics
Sex 2.92 (3.14) 9.08, 3.24 0.353 with QOLCE scores. Specifically, higher IQ (b = 0.20,
Age 0.37 (0.41) 1.18, 0.44 0.367 SE = 0.07, p = 0.004) and greater resources available to
Age at seizure onset 0.12 (0.41) 0.92, 0.68 0.768 aid families’ adaptation to stressful events (b = 0.43,
Duration of epilepsy 0.25 (0.40) 0.1.03, 0.54 0.542 SE = 0.17, p = 0.012) were associated with higher HRQL.
Seizure frequency 6.11 (3.09) 12.16, 0.06 0.048*
Alternatively, caregiver unemployment (b = 6.53,
Number of antiseizure 4.17 (1.80) 7.69, 0.65 0.020*
medications SE = 2.79, p = 0.018) was associated with lower child
IQ 0.25 (0.08) 0.10, 0.41 0.002* HRQL.
Caregiver characteristics To assess whether patterns of associations varied by
Age 1.50 (1.98) 2.38, 5.39 0.448 QOLCE domain, we fit the multivariable linear regression
Education 1.69 (3.27) 4.74, 8.11 0.607
model for the outcomes of physical activity, cognition,
Work status 8.43 (3.28) 2, 14.85 0.010*
Depressive symptoms, 1.75 (0.37) 2.47, 1.02 0.000** well-being, social activity, and behavior (see Table 5). The
QIDS set of factors (number of antiseizure medications, seizure
Anxiety symptoms, GAD 1.36 (0.30) 1.95, 0.77 0.000** frequency, IQ, caregiver work status, caregiver QIDS,
Family characteristics caregiver GAD, Family APGAR, FIRM, and FILE) were
Income 0.89 (0.57) 0.23, 2.02 0.118
significantly associated with the following: social activity,
Family adaptation, APGAR 2.21 (0.63) 0.97, 3.46 0.000**
Family resources, FIRM 0.77 (0.12) 0.53, 1.01 0.000** R = 0.304, F9,112 = 5.01, p < 0.001: cognition, R = 0.401,
Family demands, FILE 1.15 (0.22) 1.58, 0.72 0.000** F9,112 = 7.67, p < 0.001; well-being, R = 0.329, F9,112 =
5.58, p < 0.00; social activity, R = 0.327, F9,112 = 5.52,
*p < 0.05, **p < 0.001.
p < 0.001; and behavior, R = 0.362, F9,112 = 6.44,
p < 0.001. As evidenced in Table 5, patterns of associations
additional univariable linear regression model fit for com- varied by QOLCE domain. IQ was significantly associated
posite QOLCE scores with cognition subscales excluded, to with all QOLCE domains (b = 0.15–0.46, p = 0.000–
assess whether the association between IQ and HRQL 0.043) with the exception of well-being. Family resources
remained significant. Results indicated that IQ remained (FIRM) was significantly associated with cognition
significantly associated with HRQL even when cognition (b = 0.73, p = 0.003), well-being (b = 0.44, p = 0.010),
subscales were excluded: R = 0.216, F1,114 = 5.60, and behavior (b = 0.50, p = 0.002). Caregiver work status
p = 0.022. Higher IQ was associated with higher HRQL. was significantly associated with physical activity
(b = 0.6.56, p = 0.024) and social activity (b = 0.15.87,
Multivariable analysis p = 0.001). Finally, number of antiseizure medications was
Results of the multivariable linear regression model fit significantly associated with social activity (b = 6.50,
for the composite QOLCE scores are presented in Table 4. p = 0.017).
Epilepsia, 57(8):1256–1264, 2016
doi: 10.1111/epi.13441
 1261
HRQL in Children with Epilepsy

Table 5. Multivariable linear regression models fit for Table 5. Continued.

the outcomes of physical activity, cognition, well-being,
b (SE) 95% CI p Value
social activity, and behavior
Behavior
b (SE) 95% CI p Value IQ 0.23 (0.07) 0.09, 0.36 0.001*
Physical activity Number of antiseizure 0.51 (1.52) 2.48, 3.50 0.739
IQ 0.15 (0.08) 0.01, 0.30 0.040* medications
Number of antiseizure 2.87 (1.65) 6.11, 0.36 0.082 Seizure frequency 2.18 (2.57) 2.86, 7.22 0.396
medications Caregiver’s work status 2.00 (2.66) 3.26, 7.17 0.462
Seizure frequency 0.54 (2.80) 4.96, 6.03 0.849 Caregiver’s depressive 0.22 (0.44) 1.08, 0.65 0.622
Caregiver’s work status 6.56 (2.90) 0.88, 12.24 0.024* symptoms, QIDS
Caregiver’s depressive 0.29 (0.48) 1.23, 0.65 0.545 Caregiver’s anxiety 0.17 (0.37) 0.90, 0.55 0.642
symptoms, QIDS symptoms, GAD
Caregiver’s anxiety 0.42 (0.41) 1.22, 0.37 0.299 Family adaptation, 0.241 (0.59) 1.41, 0.92 0.686
symptoms, GAD APGAR
Family adaptation, 0.43 (0.65) 0.85, 1.70 0.514 Family resources, FIRM 0.50 (0.17) 0.18, 0.83 0.002*
APGAR Family demands, FILE 0.31 (0.25) 0.79, 0.18 0.214
Family resources, FIRM 0.21 (0.18) 0.14, 0.56 0.422
*p < 0.05, **p < 0.001.
Family demands, FILE 0.29 (0.27) 0.82, 0.24 0.278
Cognition
IQ 0.46 (0.10) 0.25, 0.66 0.000**
Number of antiseizure 2.21 (2.31) 6.73, 2.32 0.339
medications
To validate the multivariable regression for the compos-
Seizure frequency 0.54 (3.90) 7.10, 8.18 0.889 ite QOLCE scores, we used the Bootstrap approach of ran-
Caregiver’s work status 3.30 (4.02) 4.58, 11.18 0.412 dom sampling with replacement with 1,000 iterations. The
Caregiver’s depressive 0.55 (0.66) 1.85, 0.75 0.405 overall Bootstrap model was significant: R = 0.637,
symptoms, QIDS F9,94 = 7.13, p < 0.001. IQ (b = 0.20, p = 0.007), family
Caregiver’s anxiety 0.46 (0.56) 1.54, 0.63 0.413
symptoms, GAD
resources (b = 0.44, p = 0.010), and caregiver work status
Family adaptation, 0.55 (0.89) 2.30, 1.20 0.536 (b = 6.66, p = 0.022) were uniquely associated with child
APGAR HRQL. These findings suggest that results of our multivari-
Family resources, FIRM 0.73 (0.25) 0.25, 1.21 0.003* able regression model are robust.
Family demands, FILE 0.15 (0.37) 0.88, 0.57 0.678
Well-being
IQ 0.02 (0.07) 0.15, 0.12 0.815 Discussion
Number of antiseizure 2.23 (1.54) 5.25, 0.79 0.147
medications HRQL can be regarded as the most important health out-
Seizure frequency 0.26 (2.61) 5.35, 4.86 0.920 come in any chronic health condition.35 Previous research
Caregiver’s work status 1.58 (2.71) 3.71, 6.87 0.561 has shown that children with epilepsy have compromised
Caregiver’s depressive 0.55 (0.45) 1.43, 0.34 0.224
symptoms, QIDS
HRQL2 compared to both healthy children and children
Caregiver’s anxiety 0.13 (0.38) 0.58, 0.83 0.739 with other chronic health conditions.3–5 The empirical
symptoms, GAD investigation of risk factors associated with HRQL consti-
Family adaptation, 0.40 (0.61) 0.80, 1.60 0.511 tutes an important line of inquiry that has the potential to
APGAR inform both prognostic information and targets for interven-
Family resources, FIRM 0.44 (0.17) 0.11, 0.77 0.010*
Family demands, FILE 0.43 (0.25) 0.93, 0.07 0.092
tion. Correspondingly, the current study examined a wide
Social activity range of possible correlates of HRQL in children with drug-
IQ 0.25 (0.12) 0.01, 0.49 0.043* resistant epilepsy. Results of the multivariable model sug-
Number of antiseizure 6.50 (2.71) 11.50, 1.18 0.017* gested that lower child IQ, fewer number of resources avail-
medications able to aid families, and caregiver unemployment were
Seizure frequency 0.71 (4.60) 9.72, 8.29 0.877
Caregiver’s work status 15.87 (4.76) 6.54, 25.18 0.001*
uniquely associated with diminished HRQL in children with
Caregiver’s depressive 0.60 (0.79) 2.15, 0.95 0.449 drug-resistant epilepsy. Conversely, there was no evidence
symptoms, QIDS to suggest that child age, sex, age at seizure onset, duration
Caregiver’s anxiety 0.32 (0.67) 1.62, 0.99 0.637 of epilepsy, caregiver education, caregiver age or family
symptoms, GAD socio-economic status influenced child HRQL.
Family adaptation, 0.79 (1.07) 2.90, 1.31 0.459
APGAR
Our results concur with previous work indicating that
Family resources, FIRM 0.46 (0.30) 0.12, 1.04 0.123 diminished HRQL in children with epilepsy is associated
Family demands, FILE 0.52 (0.45) 1.39, 0.36 0.247 with lower cognitive functioning4,11,13,17 and fewer family
Continued
resources.11,17,20,21 We also found that caregiver work status
was independently associated with HRQL. The influence of

Epilepsia, 57(8):1256–1264, 2016

doi: 10.1111/epi.13441
1262
L. Conway et al.
caregiver employment has received considerably less atten-
tion in the literature as compared to the other risk factors
examined. Yong et al.13 reported a significant association
between paternal career and HRQL in children with epi-
lepsy, whereas Ferro et al.17 did not identify caregiver
employment as a significant predictor of HRQL in a sample
of children with new-onset epilepsy. More broadly, com-
pared to healthy controls, epidemiologic data suggest that
parents of children with chronic illness report lower
employment rates and devote less time to leisure activi-
ties.36 Given the documented associations between holding
a satisfying job and reduced parenting stress in caregivers of
children with disabilities,37 future research should continue
to investigate caregiver work status as a potential correlate
of child HRQL in order to clarify the current discordant
findings. It may be that some parents of children with drug-
resistant epilepsy have to give up their employment in order
to deal with the demands of managing their children’s
epilepsy.
In response to the call for more methodologically sound
research aimed at identifying meaningful family factors
that impact child HRQL,12 the current study extended
recent work in children with new-onset epilepsy and a
mixed sample of children with active and medication-
managed epilepsy to a sample of children with drug resis-
tant epilepsy. In the univariate analyses, six of the nine
significant correlates were family factors, and two of these
remained independently associated in the multivariate
analyses, highlighting the important contextual effects of
caregiver psychopathology and the family environment on
child HRQL. Furthermore, of the family factors examined,
the amount of resources available to aid families’ adapta-
tion to stressful events was uniquely associated with child
HRQL. Through the lens of stress and coping theory,38 a
dearth of caregiver coping resources may increase parent-
ing stress and potentially lead to ineffective parenting
behavior.19 The family unit serves as a primary system for
the mediation of challenging life events39 and, in children
with chronic health conditions, may act as a buffer against
the harmful effects of illness-related stressors on the
child’s adaptation.20,32,39 Factors associated with the fam-
ily environment are amenable to change and, as such, our
findings indicate potential targets for future interventions.
Within the pediatric epilepsy literature, several interven-
tions incorporating both children and their families have
been reported.40 These interventions aimed to increase
medical knowledge about epilepsy, as well as educate
families about the management of their children’s symp-
toms.40 Three studies reported a postintervention increase
in parent and child knowledge41–43; two resulted in
improved family functioning42 or parental anxiety44; two
reported no impact on level of caregiver depression or
anxiety43,45; and one was not evaluated.46 Beyond educat-
ing parents about the management of their children’s epi-
lepsy, our findings suggest that the implementation and
Epilepsia, 57(8):1256–1264, 2016
doi: 10.1111/epi.13441
evaluation of family centered care practices that address
caregiver well-being and provide linkages to supportive
community resources may result in more promising HRQL
for children and families.
Cognitive functioning has been consistently identified as
a risk factor for reduced HRQL in children with epi-
lepsy.4,11,13,17 However, it is important to note that one
domain of the QOLCE encapsulates cognition through
attention/concentration, memory, language, and other cog-
nition subscales. As such, we conducted a sensitivity analy-
sis to determine whether the association between IQ and
HRQL remained significant after excluding these subscales
from the total composite QOLCE score. We found that IQ
remained significantly associated with HRQL; however, the
magnitude of the association was diminished. Although this
finding should not detract from previous and future research
aimed at understanding the role of cognitive problems in
optimizing HRQL for children with epilepsy and their fami-
lies,17 the strength of the associations between risk factors
(e.g., low IQ) that are already captured in the QOLCE (e.g.,
cognition) and child HRQL should be interpreted with
caution.
Important caveats of this study methodology must be
acknowledged. Given the cross-sectional design, causal
relationships between identified risk factors and child
HRQL cannot be inferred. The results apply to children with
largely focal onset and drug resistant epilepsy. Furthermore,
the current study employed a parent-report measure for the
outcome of child HRQL as well as explanatory variables.
Despite its widespread use, the QOLCE has important limi-
tations,47,48 including increased respondent burden due to
the length of the questionnaire and low internal consistency
reliability for some of the smaller subscales (e.g., stigma).
In addition, the possibility of caregiver reporting bias could
affect study validity.17 Investigations of informant discrep-
ancy, “the proxy problem”, within this domain have yielded
equivocal findings. A few studies have reported concor-
dance between parent and child reports of child HRQL,49,50
whereas others have documented discordance.51–53 Further-
more, a recent study by Ferro (2014) identified informant as
a moderator of associations between risk factors and child
HRQL.12 Correspondingly, to improve our understanding
of robust risk factors of HRQL, future researchers should
remain cognizant of the proxy problem and overcome its
limitations through the implementation of multi-informant
methodologic approaches. We note that excluding the “not
applicable” response option from QOLCE items is not ideal,
as it limits respondents’ options. Finally, it is important to
note that caregiver refers to the person who completed the
questionnaire in the current study, of which 83.5% were
female. Moving forward, examinations of child behavior,
mental health, convulsive status epilepticus, and use of
emergency medication in children with drug-resistant epi-
lepsy are warranted, as these factors have been associated
with child HRQL.47,54–56

Conclusion
This is the first study to comprehensively examine a wide
range of correlates of HRQL in a multicenter sample of chil-
dren with drug resistant epilepsy. Lower child IQ, fewer
resources available to aid families, and caregiver unemploy-
ment were independently associated with diminished child
HRQL. Our findings underscore the importance of consider-
ing the family context when evaluating HRQL in children
with epilepsy. Given that family factors are amenable to
change, they can serve as potential targets for future inter-
ventions aimed at improving HRQL in children with drug-
resistant epilepsy.
Acknowledgments
PEPSQOL was funded by a grant from the Canadian Institutes for
Health Research (MOP-133708) to Dr. Widjaja and Dr. Smith. Ms. Conway
is supported by a doctoral fellowship from the Social Sciences and Humani-
ties Research Council.
Disclosure
None of the authors has any conflicts of interest to disclose. We
confirm that we have read the Journal’s position on issues involved in
ethical publication and affirm that this report is consistent with those
guidelines.
Appendix: PEPSQOL Study Team
Cristina Go (Hospital for Sick Children, Toronto), Rajesh Ramachan-
drannair (McMaster Children’s Hospital, Hamilton), Lionel Carmant (Ste.
Justine Children’s Hospital, Montreal), Jeffrey Buchhalter (Alberta Chil-
dren’s Hospital, Calgary), Paula Brna (IWK Health Center, Halifax), Fran
Booth (The Children’s Hospital, Winnipeg), Salah Almubarak (Royal
University Hospital, Saskatoon), and Simon Levin (Children’s Hospital at
London Health Sciences Centre, London).
References
1. Spieth LE, Harris CV. Assessment of health-related quality of life in
children and adolescents: an integrative review. J Pediatr Psychol
1996;21:175–193.
2. Wu DY, Ding D, Wang Y, et al. Quality of life and related factors in
Chinese adolescents with active epilepsy. Epilepsy Res 2010;90:16–
20.
3. Austin JK. Comparison of child adaptation to epilepsy and asthma. J
Child Adolesc Psychiatr Ment Health Nurs 1989;2:139–144.
4. Williams J, Steel C, Sharp GB, et al. Anxiety in children with epilepsy.
Epilepsy Behav 2003;4:729–732.
5. Rodenburg R, Stams GJ, Meijer AM, et al. Psychopathology in chil-
dren with epilepsy: a meta-analysis. J Pediatr Psychol 2005;30:453–
468.
6. Austin JK, Risinger MW, Beckett LA. Correlates of behavior problems
in children with epilepsy. Epilepsia 1992;33:1115–1122.
7. Austin JK, Huster GA, Dunn DW, et al. Adolescents with active or
inactive epilepsy or asthma: a comparison of quality of life. Epilepsia
1996;37:1228–1238.
8. Dunn DW, Austin JK, Huster GA. Symptoms of depression in adoles-
cents with epilepsy. J Am Acad Child Adolesc Psychiatry
1999;38:1132–1138.
9. Berg AT, Langfitt JT, Testa FM, et al. Global cognitive function in
children with epilepsy: a community-based study. Epilepsia 2008;49:
608–614.
1263
HRQL in Children with Epilepsy
10. Reilly C, Atkinson P, Das KB, et al. Academic achievement in school-
aged children with active epilepsy: a population-based study. Epilepsia
2014;55:1910–1917.
11. Speechley KN, Ferro MA, Camfield CS, et al. Quality of life in chil-
dren with new-onset epilepsy: a 2-year prospective cohort study. Neu-
rology 2012;79:1548–1555.
12. Ferro MA. Risk factors for health-related quality of life in children with
epilepsy: a meta-analysis. Epilepsia 2014;55:1722–1731.
13. Yong L, Chengye J, Jiong Q. Factors affecting the quality of life in
childhood epilepsy in China. Acta Neurol Scand 2006;113:167–173.
14. Fayed N, Davis AM, Streiner DL, et al. Children’s perspective of qual-
ity of life in epilepsy. Neurology 2015;84:1830–1837.
15. Li Y, Ji CY, Qin J, et al. Parental anxiety and quality of life of epileptic
children. Biomed Environ Sci 2008;21:228–232.
16. Thompson PJ, Upton D. The impact of chronic epilepsy on the family.
Seizure 1992;1:43–48.
17. Ferro MA, Camfield CS, Levin SD, et al. Trajectories of health-related
quality of life in children with epilepsy: a cohort study. Epilepsia
2013;54:1889–1897.
18. Hoare P, Kerley S. Psychosocial adjustment of children with chronic
epilepsy and their families. Dev Med Child Neurol 1991;33:201–215.
19. Rodenburg R, Meijer AM, Dekovic M, et al. Family factors and psy-
chopathology in children with epilepsy: a literature review. Epilepsy
Behav 2005;6:488–503.
20. Ferro MA, Avison WR, Campbell MK, et al. The impact of maternal
depressive symptoms on health-related quality of life in children with
epilepsy: a prospective study of family environment as mediators and
moderators. Epilepsia 2011;52:316–325.
21. Austin JK, Perkins SM, Johnson CS, et al. Self-esteem and symptoms
of depression in children with seizures: relationships with neuropsy-
chological functioning and family variables over time. Epilepsia
2010;51:2074–2083.
22. Kwan P, Arzimanoglou A, Berg AT, et al. Definition of drug resistant
epilepsy: consensus proposal by the ad hoc Task Force of the ILAE
Commission on Therapeutic Strategies. Epilepsia 2010;51:1069–1077.
23. Sabaz M, Cairns DR, Lawson JA, et al. Validation of a new quality of
life measure for children with epilepsy. Epilepsia 2000;41:765–774.
24. Sabaz M, Lawson JA, Cairns DR, et al. Validation of the quality of life
in childhood epilepsy questionnaire in American epilepsy patients.
Epilepsy Behav 2003;4:680–691.
25. Harris PA, Taylor R, Thielke R, et al. Research electronic data capture
(REDCap) – a metadata-driven methodology and workflow process for
providing translational research informatics support. J Biomed Inform
2009;42:377–381.
26. Bernstein IH, Rush AJ, Trivedi MH, et al. Psychometric properties of
the Quick Inventory of Depressive Symptomatology in adolescents. Int
J Methods Psychiatr Res 2010;19:185–194.
27. Spitzer RL, Kroenke K, Williams JB, et al. A brief measure for assess-
ing generalized anxiety disorder: the GAD-7. Arch Intern Med
2006;166:1092–1097.
28. L€owe B, Decker O, M€uller S, et al. Validation and standardization of
the Generalized Anxiety Disorder Screener (GAD-7) in the general
population. Med Care 2008;46:266–274.
29. Smilkstein G. The family APGAR: a proposal for family function test
and its use by physicians. J Fam Pract 1978;6:1231–1239.
30. Smilkstein G, Ashworth C, Montano D. Validity and reliability of the
family APGAR as a test of family function. J Fam Pract 1982;15:303–
311.
31. McCubbin HI, Thompson AI, McCubbin MA. FIRM: Family Inven-
tory of Resources for Management. Family assessment: resiliency,
coping and adaptation. Inventories for research and practice. Madi-
son, WI: University of Wisconsin Publishers; 1996.
32. McCubbin HI, Thompson AI, McCubbin MA. FILE: Family Inventory
of Life Events and Changes. Family assessment: resiliency, coping and
adaptation. Inventories for research and practice. Madison, WI:
University of Wisconsin Publishers; 1996.
33. Wechsler D. Wechsler Abbreviated Scale of Intelligence (WASI) man-
ual. San Antonio, TX: The Psychological Corporation; 2011.
34. Wechsler D. Technical and interpretive manual for the wechsler pre-
school and primary scale of intelligence. San Antonio, TX: The Psy-
chological Corporation; 2012.
35. Thurman DJ, Beghi E, Begley CE, et al. Standards for epidemiologic
studies and surveillance of epilepsy. Epilepsia 2011;52:2–26.
Epilepsia, 57(8):1256–1264, 2016
doi: 10.1111/epi.13441
1264
L. Conway et al.
36. Kuhlthau KA, Perrin JM. Child health status and parental employment.
Arch Pediatr Adolesc Med 2001;155:1346–1350.
37. Warfield ME. Employment, parenting, and well-being among mothers
of children with disabilities. Ment Retard 2001;39:297–309.
38. Lazarus RS, Folkman S. Stress, appraisal, and coping. New York:
Springer Pub. Co; 1984.
39. Chiou H-H, Hsieh L-P. Parenting stress in parents of children with epi-
lepsy and asthma. J Child Neurol 2008;23:301–306.
40. Duffy LV. Parental coping and childhood epilepsy: the need for future
research. J Neurosci Nurs 2011;43:29–35.
41. Tieffenberg J, Wood E, Alonso A, et al. A randomized field trial
of ACINDES: a child-centered training model for children with
chronic illnesses (asthma and epilepsy). J Urban Health 2000;77:
280–297.
42. Austin JK, McNelis AM, Shore CP, et al. A feasibility study of a Fam-
ily Seizure Management Program: ‘Be Seizure Smart’. J Neurosci Nurs
2002;34:30–37.
43. Shore CP, Perkins SM, Austin JK. The Seizures and Epilepsy Educa-
tion (SEE) program for families of children with epilepsy: a prelimi-
nary study. Epilepsy Behav 2008;12:157–164.
44. Jantzen S, M€ uller-Godeffroy E, Hallfahrt-Krisl T, et al. FLIP&FLAP –
a training programme for children and adolescents with epilepsy, and
their parents. Seizure 2009;18:478–486.
45. Snead K, Ackerson J, Bailey K, et al. Taking charge of epilepsy: the
development of a structured psychoeducational group intervention for
adolescents with epilepsy and their parents. Epilepsy Behav
2004;5:547–556.
46. Wohlrab GC, Rinnert S, Bettendorf U, et al. famoses: a modular edu-
cational program for children with epilepsy and their parents. Epilepsy
Behav 2007;10:44–48.
47. Reilly C, Atkinson P, Das KB, et al. Factors associated with quality of
life in active childhood epilepsy: a population-based study. Eur J Pae-
diatr Neurol 2015;19:308–313.
48. Ferro MA, Goodwin SW, Sabaz M, et al. Measurement equivalence of
the newly developed Quality of Life in Childhood Epilepsy Question-
naire (QOLCE-55). Epilepsia 2016;57:247–435.
Epilepsia, 57(8):1256–1264, 2016
doi: 10.1111/epi.13441
49. Ferro MA, Avison WR, Campbell MK, et al. Do depressive symptoms
affect mothers’ reports of child outcomes in children with new-onset
epilepsy? Qual Life Res 2010;19:955–964.
50. Vetter TR, Bridgewater CL, McGwin G. An observational study of
patient versus parental perceptions of health-related quality of life in
children and adolescents with a chronic pain condition: who should the
clinician believe? Health Qual Life Outcomes 2012;10:1.
51. Baca CB, Vickrey BG, Hays RD, et al. Differences in child versus
parent reports of the child’s health-related quality of life in chil-
dren with epilepsy and healthy siblings. Value Health 2010;13:
778–786.
52. Upton P, Lawford J, Eiser C. Parent–child agreement across child
health-related quality of life instruments: a review of the literature.
Qual Life Res 2008;17:895–913.
53. Sattoe JN, van Staa A, Moll HA. The proxy problem anatomized:
child-parent disagreement in health related quality of life reports of
chronically ill adolescents. Health Qual Life Outcomes 2012;10:1.
54. Colonnelli MC, Cross JH, Davies S, et al. Psychopathology in children
before and after surgery for extratemporal lobe epilepsy. Dev Med
Child Neurol 2012;54:521–526.
55. Ferro MA, Chin RF, Camfield CS, et al. Convulsive status epilepticus
and health-related quality of life in children with epilepsy. Neurology
2014;83:752–757.
56. Baca CB, Vickrey BG, Caplan R, et al. Psychiatric and medical comor-
bidity and quality of life outcomes in childhood-onset epilepsy. Pedi-
atrics 2011;128:1532–1543.
Supporting Information
Additional Supporting Information may be found in the
online version of this article:
Appendix S1. Additional information on caregiver and
family measures.