SEARCH

SEARCH BY CITATION

Keywords:

  • Epilepsy;
  • Quality of Well-Being Scale–Self-Administered;
  • EQ-5D;
  • Utility;
  • Validation study;
  • China

Summary

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure
  8. References
  9. Supporting Information

Purpose

Generic preference-based health-related quality of life (HRQoL) instruments are increasingly used to estimate the quality-adjusted life years (QALYs) in cost-effectiveness/utility studies. However, no such tool has been used and validated in epilepsy patients in China. This study was conducted to validate a generic preference-based HRQoL instrument, namely the Quality of Well-Being Scale–Self-Administered (QWB-SA) in Chinese patients with epilepsy.

Methods

Accepted translation procedures were followed to develop the Chinese QWB-SA. An epilepsy group (adults with established diagnosis of epilepsy) and a control group (adults without manifested cognitive problems) were recruited between July and October, 2012, from two tertiary hospitals in China. After giving informed consent, each subject completed both the QWB-SA and the EuroQol (EQ-5D) as well as provided sociodemographic data. Construct validity was examined by six (convergent) and two (discriminative) a priori hypotheses. Sensitivity was compared by ability to differentiate epilepsy-specific variable-based subgroups. Agreement between the QWB-SA and EQ-5D was assessed by intraclass correlation coefficient (ICC) and Bland-Altman plot.

Key Findings

One hundred forty-four epilepsy patients and 323 control subjects were enrolled, respectively. The utility medians (interquartile range, IQR) for the QWB-SA and EQ-5D were 0.673 (0.172), 0.848 (0.275) for epilepsy group and 0.775 (0.258), 1.000 (0.152) for control group, respectively. The difference in utilities between the two measures were significant (p < 0.0001). Construct validity was demonstrated by six a priori hypotheses. In addition, the QWB-SA was able to discriminate across different seizure frequency and antiepileptic drug (AED) treatment subgroups. Agreement between the QWB-SA and EQ-5D was demonstrated by ICC (0.725). Finally, the multiple linear regression analysis indicated that group and the EQ-VAS had influences on the utility difference of these two measures, whereas seizure frequency and number of AEDs were predictors of HRQoL as measured by the QWB-SA.

Significance

The QWB-SA is a valid and sensitive HRQoL measure in Chinese patients with epilepsy. Compared to the EQ-5D, the QWB-SA showed superiority in coverage of health dimensions, sensitivity, and ceiling effects. However, future study is still needed to ascertain its responsiveness.

Health-related quality of life (HRQoL) is a multidimensional concept that covers physical health, psychological state, and social relationship (Schipper et al., 1996), thereby describing a comprehensive picture of the individual's overall well-being. Another commonly used measure, quality-adjusted life year (QALY) is a composite metric that integrates HRQoL with the duration of life to provide a single comprehensive expression of health outcome. More specifically, QALY incorporates both quality and quantity of life into one score, thereby enabling the comparisons across diseases and populations. As such, QALY has become a standard measure of HRQoL in cost-effectiveness research in clinical medicine (Gold, 1996).

When assessing HRQoL of interested subjects, health care providers have the choice of using a generic or disease-specific instrument. Disease-specific measures are often more sensitive to subtle changes in the disease of interest, but may ignore changes in other areas of health or functioning. Given the unpredictability of interventions/medications on multiple body systems, it is essential to ascertain health in ways that can capture a subject's overall functioning and wellbeing (Gold, 1996). Hence, in practice, a generic instrument is usually applied together with a disease-specific instrument.

Epilepsy, as a chronic disorder, has considerable negative effect on people's day-to-day functioning (Baker, 1995). Apart from experiencing seizures and their detrimental impact on cognitive function (particularly memory), those with epilepsy may also experience adverse reactions to antiepileptic drugs (AEDs). In addition, epilepsy is also associated with psychological burden, including anxiety and depression (Wong & Lhatoo, 2000; Vingerhoets, 2006; Ramaratnam et al., 2008). In view of these factors, the traditionally assessed clinical outcomes that measure the treatment effect such as seizure frequency, seizure-free days might not be sufficiently comprehensive to reflect the total impact on the patient's well-being and perception about treatment effect. To capture the patient's own perception of treatment effect, a variety of validated HRQoL measures are available. For epilepsy, the three most commonly reported epilepsy-specific measures were Quality of Life Epilepsy Inventory (QOLIE-10, QOLIE-31, and QOLIE-89), and the two most commonly used generic measures were the Short-Form Questionnaire (SF-18 and SF-36) and World Health Organization Quality of life questionnaire (WHOQOL-BREF and WHOQOL-100; Taylor et al., 2011). Nevertheless, none of the aforementioned instruments could provide a utility score, thus hampering their subsequent uses in the cost-effectiveness/utility research.

Unlike the aforementioned generic instruments, Quality of Well-being Scale (Seiber et al., 2008) was the first instrument specifically designed to measure the quality of life for the estimation of QALYs. It is a preference-weighted instrument combing the three scales of functioning with a measure of symptoms and problems to produce a point-in-time expression of wellbeing that runs from 0 (for death) to 1.0 (for symptomatic full function). With the preference weights derived from a community sample, a unique aspect of QWB-SA version is that a person's utility score reflects a societal perspective on the value of that person's level of functioning and wellbeing (Seiber et al., 2008). The information obtained via QWB-SA would therefore be extremely beneficial for conducting cost-effectiveness/utility research.

Several generic preference-based HRQoL instruments are available in the Chinese versions. For instance, EuroQol (EQ-5D) and Short-form 6D (SF-6D) have been validated in certain Chinese populations (Zhao et al., 2010). However, both EQ-5D and SF-6D, focus only on the functioning aspects, whereas in contrast, QWB-SA has a functioning component complemented by a strong symptom component. Prior work by developers of QWB has demonstrated that on any particular day, nearly 80% of the general population is optimally functional, but less than half of the population experiences no symptoms (Seiber et al., 2008). Consequently, administration of QWB-SA could provide important supporting information that is not captured by EQ-5D or SF-6D.

Our research, therefore, intended to translate and validate the QWB-SA and investigate the psychometric properties of this Chinese version in Chinese patients with epilepsy. At the same time, the performance of QWB-SA was compared with another widely utilized generic preference-based HRQoL instrument: EQ-5D.

Methods

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure
  8. References
  9. Supporting Information

Study design and subjects recruitment

The cross-sectional study recruited participants from two tertiary hospitals in China: Renmin Hospital of Wuhan University, and the Fifth Hospital of Wuhan (Wuhan, Hubei, China) between July and October 2012. The study was approved by the institutional review board of the two study sites. After informed consent was received from each participant (age >16 years), a convenience sample of inpatients or outpatients with the diagnosis of epilepsy and a control group (without manifestation of cognitive problems) were recruited. Attending physicians or consultant neurologists/epileptologists were responsible for initially identifying patients with epilepsy. The diagnosis of epilepsy was based on the clinical history, symptoms, examinations, electroencephalography (EEG; epileptic discharges), neuroimaging (magnetic resonance imaging [MRI], computed tomography [CT]) with the consensus between two physicians (SQP and LX). Each subject was interviewed by a trained interviewer using standardized questionnaires containing QWB-SA and EQ-5D/visual analog scale (VAS). Other information including sociodemographic data (for both epilepsy patients and controls) and epilepsy specific data (for epilepsy patients) were collected simultaneously.

Instruments

QWB-SA

The QWB-SA includes five sections. The first section assesses the presence/absence of 19 chronic symptoms or problems (e.g., blindness, hearing loss), followed by assessment of 25 acute physical symptoms (e.g., headache, breathless, chest pain), and 14 mental health symptoms and behaviors (e.g., sadness, blue, frustration). The remaining sections are assessments of persons' mobility (including use of transportation), physical activity (e.g., walking and carrying stuff), and social activity (completion of role expectation like work, school). Each item in the QWB-SA is described as a health state to be rated on a 0–100 scale (Visual Analog Scale [VAS]) (Seiber et al., 2008). Each participant recalls the answers to the particular QWB-SA question within the last 3 days before the day of the survey. Once all the subjects have provided ratings, preference weight of the corresponding item is then estimated using the following formula (Anderson & Zalinski, 1988):

  • display math

To calculate the QWB-SA utility, each section is to be computed, namely, CPX (acute and chronic symptoms), MOB (self-care and mobility), PAC (physical activity), and SAC (self-care and usual activity). Scoring algorithm and preference weights are then provided by the University of California, San Diego (UCSD) Health Services Research Center. In our current study, the use of QWB-SA was authorized by the QWB-SA copyright holders.

EQ-5D/VAS

The EQ-5D comprises five dimensions: mobility, self-care, usual activity, pain/discomfort, and anxiety/depression. Each dimension has three response levels (no problems, some problems, severe problems). The EQ-5D descriptive system can theoretically generate 243 health states, with a utility score ranging from −0.59 to 1.00. The utility scoring algorithm adopted in our study was developed using Time Trade-Off (TTO) based preference scores from a United Kingdom general population (Dolan, 1997). EQ-VAS is a 20-cm vertical visual analog scale ranging from 100 (best imaginable health state) to 0 (worst imaginable health state) to represent the overall health of the day. Each respondent classifies and rates their health status on the day of the survey. The simplified Chinese version of EQ-5D/VAS is an official version authorized by the EuroQol Group. The validity of this version has been reported recently by Zhao et al. (2010).

Translation process

Forward and backward translation

Two Chinese physicians (LG and LX) translated the English version of QWB-SA into simplified Chinese independently. The two Chinese versions of QWB-SA were consolidated into one via thorough discussion of the two translators and the inputs of two professors (SQP and SCL). Next another two bilingual physicians who were blind to the original QWB-SA as well as the study design performed the back-translation process. Finally, the two back-translations were submitted to the developers for appraisal.

Culture adaptations

To evaluate the equivalence with the original version, the initial Chinese version of QWB-SA was sent to two consultant physicians and a pharmacist. Considering the number of motor vehicles per 1,000 people was only 83 (WebDataSource) but there are >520 million bicycles (including electric ones) in China (WebDataSource), and together with the similar physical and mental requirements for driving and riding, one key change was proposed by these experts to make the content of QWB-SA applicable to China.

Items 5b and 6c “drive a motor vehicle” were replaced by “ride a (electric) bicycle or do the housework.”

Pilot testing

Forty subjects including hospital general staff (n = 20), interns (n = 12), nurses (n = 4), and outpatients of a neurology clinic (n = 4) were interviewed to complete the draft Chinese QWB-SA. During the process, a couple of respondents neither drove a car nor rode a (electric) bicycle for commuting (Items 5b and 6c), thus “do the housework” was added. Furthermore, from the qualitative input of the pilot study, wording and phrasing were further refined accordingly to avoid confusion in understanding and were then integrated into the final simplified Chinese version.

Data analyses

Descriptive statistics

Descriptive statistics were used to characterize the sample and the distribution of QWB-SA and EQ-5D/VAS scores. Continuous variables were presented by mean, standard deviation/standard error (SE), median, and interquartile range (IQR) where applicable, whereas categorical variables were shown by the number and proportion of the entire sample in corresponding group. The differences between epilepsy and control groups were examined by analysis of variance (ANOVA) (if the distribution was normal) or Mann-Whitney U-test (if the distribution was abnormal) in the case of continuous variables, or chi-square test in the case of categorical variables.

Construct validity

To test the convergent validity, the associations between QWB-SA utility and EQ-5D/VAS were assessed at subscale and scale levels. According to the literature and clinical experience, six a priori hypotheses were tested with expected moderate to strong correlation coefficients (ρ):

  1. Correlation between QWB-SA utility score and EQ-5D utility score.
  2. Correlation between QWB-SA utility score and EQ-VAS.
  3. Correlation between QWB-SA acute and chronic symptoms (CPX) with EQ-5D pain/discomfort and anxiety/depression.
  4. Correlation between QWB-SA self-care and mobility (MOB) with EQ-5D mobility.
  5. Correlation between QWB-SA physical activity (PAC) with EQ-5D mobility and usual activity.
  6. Correlation between QWB-SA self-care and usual activity with EQ-5D usual activity and self-care.

Correlation coefficients were computed as Spearman's rank correlation coefficient (ρ), with ρ > 0.5 considered as strong correlation, 0.35–0.5 as moderate correlation, and 0.2–0.34 as weak correlation (Juniper et al., 1996).

The discriminative validity was assessed based on criterion validity. Abilities of QWB-SA and EQ-5D to discriminate between epilepsy and general populations as well as different levels of self-rating health status according to QWB-SA and EQ-VAS were examined. Specifically, patients with epilepsy were expected to have lower utility scores on both QWB-SA and EQ-5D than the general population. In addition, subjects with poorer self-rated health status would have lower utility scores as well. The five levels health statuses (excellent, very good, good, fair, poor) according to QWB-SA were adopted as the grouping factor. At the same time, EQ-VAS was also employed as an indicator for self-reported health status, and subsequently categorized into four subgroups: <65 (bad), 65–79 (fair), 80–89 (good), and 90–100 (excellent) (Barton et al., 2008).

Sensitivity of QWB-SA and EQ-5D

This analysis was undertaken for the epilepsy group only. Precisely, two-step analyses were performed to assess the sensitivity of the two measures toward epilepsy characteristics that are known to affect health and quality of life. For example, as shown in a review for determinants of HRQoL for patients with epilepsy, seizure frequency is negatively correlated with HRQoL (Taylor et al., 2011). If the measure could better differentiate HRQoL for patients with distinctive seizure characteristics, the measure might be considered as sensitive to this disease cohort. At first, correlations between sociodemographic or epilepsy-specific variables and HRQoL utility scores were assessed via Spearman's correlation coefficient with p-value < 0.1 to identify candidate predictors. Then, a series of one-way ANOVA analyses (or independent-samples t-test) were carried out to further investigate the different effect of epilepsy-specific variables on utilities. Relative efficiency (RE) statistics were also calculated to compare two utility instruments regardless of statistical significance. The RE statistic is the ratio between two F-ratios (or t-statistics) from the one-way ANOVA (or independent-samples t-test) for each measure, with higher RE suggesting stronger validity. Lastly, multiple linear regression (MLR) was run to investigate the candidate predictors that were ascertained to be significantly correlated with QWB-SA and EQ-5D in the univariate analysis.

Levels of agreement between QWB-SA and EQ-5D

The mean and median utility scores between these two instruments for the entire sample and within each group were compared. Both Wilcoxon's signed-rank test and Spearman's rank correlation were adopted to investigate the association between these two utility scores. In order to address the limitations of simple correlation, the agreement between utility scores of QWB-SA and EQ-5D was assessed by intraclass correlation coefficient (ICC; calculated with two-way random effects model based on absolute agreement and coefficient >0.7 indicates a strong agreement) and Bland-Altman plot. Bland-Altman plot was used to evaluate the agreement between two different instruments or measurements by investigating the existence of any systematic difference (e.g., fixed bias) between the measurements and to identify possible outliers (Bland & Altman, 1986). If no clinically important differences are observed within 95% confidence intervals (CIs), the two methods may be used interchangeably. The one-sample t-test was undertaken to compare the mean difference of utility scores with 0, with p-value > 0.05 implying the total agreement between QWB-SA and EQ-5D.

Factors affecting utility difference between QWB-SA and EQ-5D

In investigating the factors attributing to the differences between two instruments, multiple linear-regression was performed to test the socioeconomic characteristics that related to the variance with the difference in the utility set as the dependent variable. At the same time, groups (epilepsy or control group), age, levels of education, marital status, working status, QWB-SA self-rating health status, and EQ-VAS for global wellbeing were selected as the independent variables.

All the statistical analyses were performed on SPSS 20.0 (SPSS Inc, Chicago, IL, U.S.A.). p-value < 0.05 was considered as statistically significant.

Results

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure
  8. References
  9. Supporting Information

Characteristics of subjects

In total, 467 subjects completed both the QWB-SA and EQ-5D with 144 in the epilepsy group. There were statistically significant differences between the epilepsy and control groups in terms of age (p = 0.033), gender (p < 0.0001), working status (p = 0.029), and level of education (p < 0.0001; Table 1).

Table 1. Characteristics of subjects and distribution of QWB-SA and EA-5D utility scores
CharacteristicsEpilepsy group N = 144Control group N = 323p-Value
  1. a

    In total, these data were retrieved for only 141 epilepsy patients.

Age in years   
Mean ± SD33.11 ± 13.04436.15 ± 16.4060.033
Median ± IQR30.57 ± 22.0031.67 ± 27.00
Range16–6516–86
Gender (male)75 (52.1%)127 (40.7%)<0.0001
Han ethnicity (%)142 (98.6)308 (98.7)0.926
Marital status (%)   
Unmarried71 (49.3)123 (39.4)0.182
Married70 (48.6)184 (59.0)
Divorced2 (1.4)2 (0.6)
Widow/widower1 (0.7)3 (1.0)
Working status (%)   
Employed65 (45.1)175 (56.1)0.029
Unemployed69 (54.9)137 (43.9)
Year of education (%)   
≤6 years16 (11.1)16 (5.0)<0.0001
7–12 years106 (73.6)144 (44.5)
>12 years22 (15.3)163 (50.4)
Age of epilepsy onset (median ± IQR)a18.00 ± 13.50
Duration of epilepsy (median ± IQR)a6.00 ± 13.00
Seizure frequency (%)a   
<1/year6 (4.3)
1–11/year63 (44.7)
≥12/year72 (51.1)
Seizure types (%)a   
Simple partial7 (5.0)
Complex partial78 (55.3)
Absence18 (12.8)
Clonic27 (19.1)
Tonic–clonic11 (7.8)
Epilepsy syndromes (%)a   
Localization-related107 (75.9)
Generalized24 (17.0)
Unknown localization10 (7.1)
Antiepileptic treatment (%)a   
Monotherapy66 (46.8)
Polytherapy75 (53.2)
QWB-SA   
Mean ± SD0.657 ± 0.1350.802 ± 0.155<0.0001
Median ± IQR0.673 ± 0.1721.000 ± 0.152
Range0.261–0.9340.308–1.000
CPX   
Mean ± SD−0.315 ± 0.103−0.190 ± 0.144<0.0001
Median ± IQR−0.324 ± 0.116−0.225 ± 0.256
Range−0.531–−0.066−0.523–−0.000
MOB   
Mean ± SD−0.002 ± 0.008−0.002 ± 0.0130.794
Median ± IQR0.000 ± 0.0000.000 ± 0.000
Range−0.031–0.000−0.089–0.000
PAC   
Mean ± SD−0.006 ± 0.019−0.005 ± 0.0200.731
Median ± IQR0.000 ± 0.0000.000 ± 0.000
Range−0.072–0.000−0.072–0.000
SAC   
Mean ± SD−0.018 ± 0.038−0.001 ± 0.008<0.0001
Median ± IQR0.000 ± 0.0000.000 ± 0.000
Range−0.150–0.000−0.096–0.000
EQ-5D   
Mean ± SD0.828 ± 0.2060.923 ± 0.132<0.0001
Median ± IQR0.848 ± 0.2751.000 ± 0.152
Range0.079–1.0000.002–1.000
EQ-VAS   
Mean ± SD79.57 ± 16.41982.64 ± 13.9390.052
Median ± IQR80.00 ± 20.0085.00 ± 11.00
Range30–10010–100

Description statistics of QWB-SA and EQ-5D

For utility of QWB-SA, the mean (standard deviation, SD) was 0.657 (0.135) for epilepsy group and 0.802 (0.155) for control group, and the median (IQR) was 0.673 (0.172) for epilepsy group and 1.000 (0.152) for control group. For utility of EQ-5D, the mean (SD) for epilepsy group was 0.828 (0.206) and 0.923 (0.132) for control group, whereas the median (IQR) was 0.848 (0.275) for epilepsy group and 1.000 (0.152) for control group. Utility scores on QWB-SA and EQ-5D were significantly different between the two groups (p < 0.0001), whereas the EQ-VAS did not show a difference (p = 0.052). Two of four sections of QWB-SA, namely CPX (p < 0.0001) and SAC (p < 0.0001), were significantly different between epilepsy and control groups. More specifically, epilepsy patients tended to experience more problems in these two sections (Table 1).

Given the significant differences between the epilepsy and control groups in terms of age, gender, working status, and level of education, the adjusted means for QWB-SA and EQ-5D are presented in Table S1.

The Shapiro-Wilk normality test showed that QWB-SA and EQ-5D utility scores were not normally distributed (p < 0.0001). The distribution of QWB-SA self-rating health status and EQ-5D are presented in Table 2. Particularly, for QWB-SA self-rating health status, from Fair to Very good, a decreasing trend was observed in both groups, with more subjects having excellent health status in the control group (11.5% vs. 1.4%). For EQ-5D, higher celling effects were observed in domains of mobility (92.4%), self-care (92.4%), and usual activity (84.7%) for the epilepsy group (Table 2).

Table 2. Distribution of QWB-SA self-rated health status and EQ-5D results
GroupExcellent (%)Very good (%)Good (%)Fair (%)Poor (%)
Epilepsy2 (1.4)26 (18.1)53 (36.8)56 (38.9)7 (4.9)
Control37 (11.5)86 (26.6)94 (29.1)98 (30.3)8 (2.5)
EQ-5D (%)
LevelMobilitySelf-careUsual activityPain/discomfortAnxiety/depression
1     
Epilepsy133 (92.4)133 (92.4)122 (84.7)94 (65.3)75 (52.1)
Control314 (97.2)316 (97.8)312 (96.6)255 (78.9)270 (83.6)
2     
Epilepsy11 (7.6)9 (6.3)19 (13.2)49 (34.0)64 (44.4)
Control7 (2.2)7 (2.2)11 (3.4)65 (20.1)52 (16.1)
3     
Epilepsy02 (1.4)3 (2.1)1 (0.7)5 (3.5)
Control2 (0.6)003 (0.9)1 (0.3)

Construct validity

Convergent validity

Convergent validity was demonstrated by moderate to strong correlation coefficients (0.365–0.590, p < 0.0001) of all the six a priori hypotheses between QWB-SA and EQ-5D on both scale and subscale levels (Table 3). The univariate analyses indicated that age, education level, working status, EQ-VAS score, and QWB-SA self-rating health status all contributed to the differences in the utility scores of QWB-SA and EQ-5D for either or both group(s) (Table S2). For example, there was a gradual reduction for utilities of both QWB-SA and EQ-5D in the control group with increasing age and decreasing education level, but the same effect was observed in the epilepsy group with increasing age only (Fig. 1). In the epilepsy group, working status alone contributed to the variation in the utility, with employed epilepsy patients reporting higher utilities on both instruments (p = 0.005; Table S2).

Table 3. Correlation between QWB-SA and EQ-5D or EQ-VAS
QWB-SAEQ-5DUtilityEQ-VAS
MSCUAPDAD
  1. Underlined figures corresponded to the six a priori hypotheses being tested.

  2. CPX, acute and chronic symptoms; MOB, self-care and mobility; PAC, physical activity; SAC, self-care and usual activity; M, mobility; SC, self-care; UA, usual activity; PD, pain/discomfort; AD, anxiety/depression.

  3. p < 0.01 otherwise indicated.

CPX−0.220−0.174−0.273 −0.402 −0.554 0.587 0.411
MOB −0.365 −0.120−0.226−0.205−0.080 (p = 0.085)0.1780.190
PAC −0.533 −0.363−0.425−0.211−0.115 (p = 0.013)0.2950.224
SAC−0.373−0.447 −0.509 −0.242−0.1630.3030.183
Utility0.2580.2360.3260.4000.524

0.590

Pearson

0.569

0.415
image

Figure 1. QWB-SA utility scores by age groups.

Download figure to PowerPoint

Discriminative validity

There was no significant correlation between MOB and AD (p = 0.085). However, weak correlations were observed between CPX and SC (−0.174), PAC and AD (−0.211), MOB and AD (−0.205), PAC and AD (−0.115), and MOB and SC (−0.120), which indicated lower correlations between different constructs (Table 3).

Discriminative validity was also confirmed via the know-group hypotheses, with epilepsy patients generating lower utility scores. Furthermore, congruent with the decreasing EQ-VAS score or the decline in QWB-SA self-rating health status, utilities of QWB-SA and EQ-5D also declined simultaneously. QWB-SA score also showed convergent validity with self-rating health status (Table S2).

Sensitivity of QWB-SA

The Spearman's correlation coefficients identified seven variables significantly correlated with utility of QWB-SA and five factors with EQ-5D (Table S3). In addition, one-way ANOVA further indicated that there were significant differences in the utilities of QWB-SA according to varying seizure frequency and antiepileptic treatment. RE statistics also showed that after QWB-SA had stronger discriminative validity than the EQ-5D (except for ability to discriminate patients with different numbers of AEDs; Table S4). Subsequent MLR showed seizure frequency (p = 0.039) and AEDs treatment (mono vs. polytherapy; p = 0.035) as predictors of the utility on the QWB-SA. In contrast, there was no predictor for utility on the EQ-5D (Table S5).

Agreement between QWB-SA and EQ-5D

In our study, generally, subjects got higher utility scores on EQ-5D than QWB-SA in both groups. According to the one-sample t-test, a statistically significant difference was observed between utilities of QWB-SA and EQ-5D (p < 0.0001). However, when it came to the ICC, it also showed a strong correlation between these two measures, with ICC of 0.725 (95% CI 0.671, 0.771) for the entire sample. Again, a higher ICC was detected among patients with epilepsy (0.771, 95% CI 0.681, 0.835; Table S6).

Nevertheless, the minimal clinically important difference (MCID) for the EQ-5D is reported to range from 0.04 to 0.10 (Brazier et al., 2004; Le et al., 2013), and for the QWB-SA from 0.02 to 0.05 (Le et al., 2013). In our current study, the 95% CI of utility difference via the Bland-Altman analysis was −0.4508 to 0.1621 (Fig. 2), which demonstrated a difference between the two measures.

image

Figure 2. Bland-Altman plot of difference in utility scores between QWB-SA and EQ-5D.

Download figure to PowerPoint

Factors associated with the disagreement between QWB-SA and EQ-5D

When the difference between the QWB-SA and EQ-5D was modeled as a dependent variable, with age, education level, marital status, working status, different groups, QWB-SA self-rating health status, and EQ-VAS self-health rating scores modeled as independent variables, the results from multiple regression indicated that, except for EQ-VAS self-health rating scores (p = 0.018) and different groups (p = 0.002), other factors did not influence the disagreement between QWB-SA and EQ-5D. Furthermore, even with these significant factors, the magnitudes of the influence were very small, for example, the coefficients were −0.001 and 0.053 for EQ-VAS scores and different groups, respectively (Table 4). Therefore, no significant demographic variable or global quality of life (as measured by EQ-VAS and QWB-SA self-rating health status) was identified to contribute significantly to the different utilities between QWB-SA and EQ-5D; other nondemographic variables or the fixed bias between the two instruments might have caused such difference.

Table 4. Multiple linear regression analyses for utility differences between QWB-SA and EQ-5D
Independent variablesDependent variable (utility differences)
Coefficient (95% CI)p-Value
  1. Utility of QWB-SA is the subtrahend.

  2. Significance level is p < 0.05 for the bold values.

Age in years0.001 (−0.001, 0.002)0.891
Groups0.053 (0.020, 0.086) 0.002
Education−0.001 (−0.006, 0.005)0.835
Ethnic minority−0.003 (−0.124, 0.118)0.964
Marital status−0.020 (−0.058, 0.018)0.310
Working status−0.004 (−0.034, 0.026)0.781
QWB-SA self-rated health status−0.004 (−0.021, 0.012)0.601
EQ-VAS−0.001 (−0.002, 0.000) 0.016

Discussion

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure
  8. References
  9. Supporting Information

Given that QALYs have been widely adopted as the effectiveness outcome in cost-effectiveness/utility analysis studies, the utility generated from generic preference-based HRQoL instruments is an important determinant in making clinical as well as health care allocation decisions. In the case of epilepsy, which is the most common neurologic disorders affecting people of all ages (Hauser et al., 1991; Forsgren et al., 2005; Preux & Druet-Cabanac, 2005), no generic preference-based HRQoL measure has yet been validated in epileptic patients in China. With the increasing numbers of new antiepileptic drugs/devices/technologies being invented and introduced, cost-effectiveness/utility analysis will be needed to assess their cost-effectiveness. Hence, a validated HRQoL instrument that could calculate QALYs would be the most useful and greatly in demand. Our study is the first to translate and validate such an instrument (QWB-SA) in Chinese epilepsy patients.

Studies have been conducted previously to investigate the psychometric properties of generic preference-based HRQoL instrument in English-speaking patients with epilepsy. In general, EQ-5D/UK/US, 15D, SF-6D, HUI-2, and HUI-3 were shown to be reliable utility instruments in an epilepsy population (Stavem et al., 2001; Langfitt et al., 2006). In addition, compared to EQ-5D/VAS, the following instruments seemed to be more capable of discriminating between patients with different seizure controls and seizure severity: HUI-2 and HUI-3, SF-6D. This would suggest better psychometric advantages of the SF-6D over the other preference instruments for epilepsy patients. Although 15D and the assessment of Quality of life (AqoL) were sensitive to variability at the upper end of the HRQoL continuum as well, the studies were not targeted at epilepsy patients (Langfitt et al., 2006).

The construct (convergent, discriminative, sensitivity) validity of QWB-SA has been successfully demonstrated in our study. Most importantly, the sensitivity of the QWB-SA was demonstrated by its ability to discriminate between different seizure frequencies and antiepileptic treatment (mono vs. poly) groups, which is of clinical importance. In addition, seizure frequency and antiepileptic treatment were found to be predictors of HRQoL as measured by the QWB-SA rather than the EQ-5D. Lastly 77 (16.5%) versus 275 (58.9%) subjects on the QWB-SA and the EQ-5D scored 1.0 (perfect health), respectively, which suggested that the QWB-SA has fewer ceiling effects.

The utility of the QWB-SA was substantially lower than that of the EQ-5D in both epilepsy and control groups. It was worth noting that the disagreement on utility scores for these two instruments was not uncommon and had been observed by previous large sample studies (N = 3,844; Fryback et al., 2007; Bentley et al., 2011; Khanna et al., 2011). The means for EQ-5D and QWB-SA were reported to be 0.89 and 0.67, respectively (Fryback et al., 2007), whereas subjects with arthritis reported utilities ranging from 0.77 to 0.80 on EQ-5D, and from 0.56 to 0.59 on QWB-SA. The same difference was also observed in the utility scores of subjects without arthritis (Khanna et al., 2011). Furthermore, when the participants were categorized according to body mass index (BMI), the utility score for the EQ-5D was also higher than the QWB-SA among normal, overweight, and obese subjects (Bentley et al., 2011). There might be two explanations for this observation: first, unlike the EQ-5D, which utilizes the time trade-off to elicit the preference-weight for each health state, the QWB-SA adopts VAS, and the utility scores derived from VAS tend to be inherently lower than the TTO or Standard Gamble (Fryback et al., 2007). Second, the large acute and chronic symptom weight in the QWB-SA may cause the utility to be lower than the EQ-5D, as the latter does not include detailed symptoms. The difference in utility between EQ-5D and QWB-SA would raise a huge concern in future cost-effectiveness analysis, because the variation in utilities will definitely cause differences in the calculation of QALYs, and subsequently the incremental cost-effectiveness ratio (ICER). For example, in an analysis evaluating an antirheumatoid agent, it was reported that four kinds of HRQoL instruments (EQ-5D, HUI2, HUI3, and SF-6D) provided different QALYs and hence different ICERs (Marra et al., 2007). Hence, even if one AED generated obviously desirable ICER in indirect comparison with another AED, a decision could not be easily made because distinctive HRQoL measures with different sensitivities might have been utilized. Therefore, when conducting a cost-effectiveness analysis, the decision in choosing the ideal generic HRQoL measure has to balance the sensitivity and the generalizability of the instrument.

In our study, age- (Fig. 2) and education-by-group effects were observed on both the QWB-SA and the EQ-5D for the epilepsy or control populations. Generally, there was a downward trend in utility with increasing age (in both patient and control groups) and decreasing education level (in control population). However, our current results of the associations with age and education level observed in the epilepsy cohort were not in line with those of previous studies. According to a review of HRQoL determinants, age was not associated with HRQoL, whereas education level might be correlated although the conclusion was not consistent (Taylor et al., 2011). Nevertheless, the normative data of the QWB-SA reported a descending trend of utility with increasing age (Seiber et al., 2008). Therefore, the inherent attributes of the QWB-SA might be sensitive to identify changes in HRQoL affected by age, as the acute and chronic symptoms might occur more often in aged subjects, whereas other HRQoL measures such as the EQ-5D do not take the specific symptoms into account.

Furthermore, working status was another contributing factor of HRQoL for the epilepsy group. For both the QWB-SA and the EQ-5D, employed patients got higher scores even after age and level of education were controlled (e.g., the estimated QWB-SA utilities were 0.686 and 0.632 for employed and unemployed epilepsy patients, respectively). Still, the impact of employment status on the HRQoL of epilepsy patients was inconsistent across studies. Several studies showed unemployment associated with poorer HRQoL (Buck et al., 1999; Gilliam et al., 1999; Mollaoglu et al., 2004; Liou et al., 2005; Elsharkawy et al., 2009; Tlusta et al., 2009), whereas others reported no correlations (Choi-Kwon et al., 2003; Djibuti & Shakarishvili, 2003; Alanis-Guevara et al., 2005; Thomas et al., 2005; Mosaku et al., 2006; Tracy et al., 2007; Zhao et al., 2008; Giovagnoli et al., 2009). Even so, it should be noted that the sample size of three studies was <115, which indicated low statistical power (Thomas et al., 2005; Mosaku et al., 2006; Zhao et al., 2008). A recent study also reported that fully employed epileptic patients might have worse HRQoLs, owing primarily to the discrimination of and misconception about epilepsy in the work place (Mahrer-Imhof et al., 2012). So accordingly, the inconsistency in this result would necessitate future study to confirm.

As to the epilepsy-specific variables, in our multivariate analysis, seizure frequency was shown to be a predictor of HRQoL as measured by the QWB-SA. In addition to suggesting the better sensitivity of the QWB-SA over the EQ-5D, this is of clinical importance when evaluating the therapeutic effects of AEDs. If the HRQoL instrument is insensitive to changes in seizure frequency, the generated QALY and other clinical merits might be underestimated resulting in rejection of valuable therapy. Although numbers of AEDs were shown to be another predictor of utility by the QWB-SA in the present study, again, this association was not consistent across studies (Gilliam et al., 1999; Choi-Kwon et al., 2003; Johnson et al., 2004; Thomas et al., 2005; Tracy et al., 2007). Actually, it is well acknowledged that antiepileptic monotherapy may have several advantages compared to polytherapy in terms of better tolerability, improved adherence, fewer interactions, and lower cost (Guberman, 1998). In addition, adverse effects of AEDs have been shown to be positively associated with decreased HRQoL (Luoni et al., 2011). Therefore, it is reasonable to expect patients who are taking more than one AED to experience more toxic effects of medication, and consequently have poorer HRQoL. Yet, the correlation between numbers of AEDs and HRQoL requires future study to confirm.

The QWB-SA normative data (mean ± SD) reported that the utilities for clinical and control (general outpatient medical sample) cohorts were 0.599 ± 0.1629 to 0.648 ± 0.1257 and 0.602 ± 0.1323 to 0.67 ± 0.1286 for various age groups (range from 18 to >71 years) (Seiber et al., 2008). Studies were also conducted utilizing the QWB-SA to investigate HRQoL for different disease cohorts. For instance, a study that recruited inpatients and outpatients with depression found that the QWB-SA scores for inpatients were substantially lower than those for outpatients (0.383 ± 0.118 vs. 0.479 ± 0.115) (Pyne et al., 2003). Other reported QWB-SA utilities included family medicine controls (0.6427 ± 0.1349) and subjects with arthritis (0.4966 ± 0.1542) (Frosch et al., 2004); as well as presurgery cataract subjects (0.595 ± 0.134) (Rosen et al., 2005). The epilepsy data from our data set were comparable to the QWB-SA normative data as well as those from the general medical controls, although the utilities in our controls seemed to be higher than the controls from aforementioned studies. There might be several reasons underlying this. First of all, the controls from our data set were substantially younger (36.15 ± 16.406) as HRQoL would decline with increasing age (Seiber et al., 2008). Second, the participants were generally relatives/caregivers of patient group, medical school students, and hospital general staff, most may enjoy better health than subjects from outpatient medical samples or family medicine controls as included in the QWB-SA normative sample.

Nevertheless, several limitations should be noted. First of all, interrater reliability and responsiveness were not tested due to the cross-sectional design of our study. Admittedly, responsiveness is an important psychometric property of an HRQoL instrument, especially for epilepsy due to its chronic nature and unpredictability of seizures, thus requiring treatment adjustment from time to time. Second, heterogeneity existed between our two groups in terms of age, gender, level of education, and employment status. As identified by our study, the factors age, education, and employment might have associations with quality of life; the variation in these demographic data would somewhat introduce bias to the result. Nonetheless, even when age and education level were adjusted, utilities of the QWB-SA and the EQ-5D still showed differences between two groups. Third, the preference weights utilized to estimate the utilities of the QWB-SA and the EQ-5D were not originated from Chinese subjects (one from America, the other from United Kingdom). However, it was found that the preference scoring does not vary significantly, and the results are similar across different countries (Drummond et al., 2005). Nevertheless, future study to address the responsiveness of the Chinese QWB-SA and to ascertain the preference weights from societal perspective of China is still needed.

In conclusion, from the present study, the QWB-SA was shown to cover more dimensions of HRQoL, have better sensitivity, fewer ceiling effects, and less skewed distribution than the EQ-5D. Hence, it is potentially a more suitable HRQoL measure for patients with epilepsy in China.

Acknowledgments

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure
  8. References
  9. Supporting Information

The authors would like to acknowledge the authors (Seiber WJ, Groessl EJ, David KM, Ganiats TG, Kaplan RM) of the QWB-SA, University of California, San Diego, California, U.S.A.

Disclosure

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure
  8. References
  9. Supporting Information

None of the authors has any conflict of interest to disclose. All the authors confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this study is consistent with those guidelines.

References

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure
  8. References
  9. Supporting Information
  • Alanis-Guevara I, Pena E, Corona T, Lopez-Ayala T, Lopez-Meza E, Lopez-Gomez M. (2005) Sleep disturbances, socioeconomic status, and seizure control as main predictors of quality of life in epilepsy. Epilepsy Behav 7:481485.
  • Anderson NH, Zalinski J. (1988) Functional measurement approach to self-estimation in multiattribute evaluation. J Behav Decis Mak 1:191221.
  • Baker GA. (1995) Health-related quality-of-life issues: optimizing patient outcomes. Neurology 45:S29S34.
  • Barton GR, Sach TH, Avery AJ, Jenkinson C, Doherty M, Whynes DK, Muir KR. (2008) A comparison of the performance of the EQ-5D and SF-6D for individuals aged >or= 45 years. Health Econ 17:815832.
  • Bentley TG, Palta M, Paulsen AJ, Cherepanov D, Dunham NC, Feeny D, Kaplan RM, Fryback DG. (2011) Race and gender associations between obesity and nine health-related quality-of-life measures. Qual Life Res 20:665674.
  • Bland JM, Altman DG. (1986) Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1:307310.
  • Brazier J, Roberts J, Tsuchiya A, Busschbach J. (2004) A comparison of the EQ-5D and SF-6D across seven patient groups. Health Econ 13:873884.
  • Buck D, Jacoby A, Baker GA, Ley H, Steen N. (1999) Cross-cultural differences in health-related quality of life of people with epilepsy: findings from a European study. Qual Life Res 8:675685.
  • Choi-Kwon S, Chung C, Kim H, Lee S, Yoon S, Kho H, Oh J, Lee S. (2003) Factors affecting the quality of life in patients with epilepsy in Seoul, South Korea. Acta Neurol Scand 108:428434.
  • Djibuti M, Shakarishvili R. (2003) Influence of clinical, demographic, and socioeconomic variables on quality of life in patients with epilepsy: findings from Georgian study. J Neurol Neurosur Psychiatry 74:570573.
  • Dolan P. (1997) Modeling valuations for EuroQol health states. Med Care 35:10951108.
  • Drummond M, Sculpher M, Torrance G, O'Brien B, Stoddart G. (2005) Cost-utility analysis. Methods for the economic evaluation of health care programmes. Oxford University Press, New York.
  • Elsharkawy AE, May T, Thorbecke R, Koch-Stoecker S, Villagran A, Urak L, Pfafflin M, Pannek H, Pietila TA, Ebner A. (2009) Long-term outcome and determinants of quality of life after temporal lobe epilepsy surgery in adults. Epilepsy Res 86:191199.
  • Forsgren I, Beghi E, Ekman M. (2005) Cost of epilepsy in Europe. Eur J Neurol 12(Suppl. 1):5458.
  • Frosch DL, Kaplan RM, Ganiats TG, Groessl EJ, Sieber WJ, Weisman MH. (2004) Validity of self-administered quality of Well-Being scale in musculoskeletal disease. Arthritis Rheum 51:2833.
  • Fryback DG, Dunham NC, Palta M, Hanmer J, Buechner J, Cherepanov D, Herrington SA, Hays RD, Kaplan RM, Ganiats TG, Feeny D, Kind P. (2007) US norms for six generic health-related quality-of-life indexes from the National Health Measurement study. Med Care 45:11621170.
  • Gilliam F, Kuzniecky R, Meador K, Martin R, Sawrie S, Viikinsalo M, Morawetz R, Faught E. (1999) Patient-oriented outcome assessment after temporal lobectomy for refractory epilepsy. Neurology 53:687694.
  • Giovagnoli AR, da Silva AM, Federico A, Cornelio F. (2009) On the personal facets of quality of life in chronic neurological disorders. Behav Neurol 21:155163.
  • Gold M. (1996) Panel on cost-effectiveness in health and medicine. Med Care 34:DS197DS199.
  • Guberman A. (1998) Monotherapy or polytherapy for epilepsy? Can J Neurol Sci 25:S3S8.
  • Hauser WA, Annegers JF, Kurland LT. (1991) Prevalence of epilepsy in Rochester, Minnesota: 1940–1980. Epilepsia 32:429445.
  • Johnson EK, Jones JE, Seidenberg M, Hermann BP. (2004) The relative impact of anxiety, depression, and clinical seizure features on health-related quality of life in epilepsy. Epilepsia 45:544550.
  • Juniper E, Gordon H, Roman J. (1996) How to develop and validate a new health related quality of life instrument. In Spilker B (Ed) Quality of life and pharmacoeconomics in clinical trials. Lippincott-Raven Publisher, Philadelphia, PA, pp. 4956.
  • Khanna D, Maranian P, Palta M, Kaplan RM, Hays RD, Cherepanov D, Fryback DG. (2011) Health-related quality of life in adults reporting arthritis: analysis from the National Health Measurement Study. Qual Life Res 20:11311140.
  • Langfitt JT, Vickrey BG, McDermott MP, Messing S, Berg AT, Spencer SS, Sperling MR, Bazil CW, Shinnar S. (2006) Validity and responsiveness of generic preference-based HRQOL instruments in chronic epilepsy. Qual Life Res 15:899914.
  • Le QA, Doctor JN, Zoellner LA, Feeny NC. (2013) Minimal clinically important differences for the EQ-5D and QWB-SA in Post-traumatic Stress Disorder (PTSD): results from a Doubly Randomized Preference Trial (DRPT). Health Qual Life Outcomes 11:59.
  • Liou HH, Chen RC, Chen CC, Chiu MJ, Chang YY, Wang JD. (2005) Health related quality of life in adult patients with epilepsy compared with a general reference population in Taiwan. Epilepsy Res 64:151159.
  • Luoni C, Bisulli F, Canevini MP, De Sarro G, Fattore C, Galimberti CA, Gatti G, La Neve A, Muscas G, Specchio LM, Striano S, Perucca E; SOPHIE Study Group. (2011) Determinants of health-related quality of life in pharmacoresistant epilepsy: results from a large multicenter study of consecutively enrolled patients using validated quantitative assessments. Epilepsia 52:21812191.
  • Mahrer-Imhof R, Jaggi S, Bonomo A, Hediger H, Eggenschwiler P, Krämer G, Oberholzer E. (2012) Quality of life in adult patients with epilepsy and their family members. Seizure 22:128135.
  • Marra CA, Marion SA, Guh DP, Najafzadeh M, Wolfe F, Esdaile JM, Clarke AE, Gignac MA, Anis AH. (2007) Not all “quality-adjusted life years” are equal. J Clin Epidemiol 60:616624.
  • Mollaoglu M, Durna Z, Dener S. (2004) Quality of life in people with epilepsy in Sivas, Turkey. Neurol Psychiatry Brain Res 11:189194.
  • Mosaku KS, Fatoye FO, Komolafe M, Lawal M, Ola BA. (2006) Quality of life and associated factors among adults with epilepsy in Nigeria. Int J Psychiatry Med 36:469481.
  • Preux PM, Druet-Cabanac M. (2005) Epidemiology and aetiology of epilepsy in sub-Saharan Africa. Lancet Neurol 4:2131.
  • Pyne JM, Sieber WJ, David K, Kaplan RM, Rapaport MH, Williams DK. (2003) Use of the wuality of well-being self-administered version (QWB-SA) in assessing health-related quality of life in depressed patients. J Affect Disord 76:237247.
  • Ramaratnam S, Baker GA, Goldstein LH. (2008) Psychological treatments for epilepsy. Cochrane Database Syst RevCD002029.
  • Rosen PN, Kaplan RM, David K. (2005) Measuring outcomes of cataract surgery using the quality of well-being scale and VF-14 visual function index. J Cataract Refract Surg 31:369378.
  • Schipper H, Clinch J, Olweny CL. (1996) Quality of life studies: definitions and conceptual issue. In Spilker B (Ed) Quality of life and pharmacoeconomicsin clinical trials. Lippincott-Rave Publishers, New York, pp. 1124.
  • Seiber WJ, Groessl EJ, David KM, Ganiats TG, Kaplan RM. (2008) Quality of Well-being Self-Administered (QWB-SA) User's Manual. Available at: https://hoap.ucsd.edu/qwb-info/. Accessed January 23, 2013.
  • Stavem K, Bjornaes H, Lossius MI. (2001) Properties of the 15D and EQ-5D utility measures in a community sample of people with epilepsy. Epilepsy Res 44:179189.
  • Taylor RS, Sander JW, Taylor RJ, Baker GA. (2011) Predictors of health-related quality of life and costs in adults with epilepsy: a systematic review. Epilepsia 52:21682180.
  • Thomas SV, Koshy S, Nair CRS, Sarma SR. (2005) Frequent seizures and polytherapy can impair quality of life in persons with epilepsy. Neurol India 53:4650.
  • Tlusta E, Zarubova J, Simko J, Hojdikova H, Salek S, Vlcek J. (2009) Clinical and demographic characteristics predicting QOL in patients with epilepsy in the Czech Republic: how this can influence practice. Seizure 18:8589.
  • Tracy JI, Dechant V, Sperling MR, Cho R, Glosser D. (2007) The association of mood with quality of life ratings in epilepsy. Neurology 68:11011107.
  • Vingerhoets G. (2006) Cognitive effects of seizures. Seizure 15:221226.
  • WebDataSource. E-China: the bicycle kingdom is going electric. USAToday. Available at: http://usatoday30.usatoday.com/money/world/2009-07-25-china-electric-bike_N.htm. Accessed April 30, 2013.
  • WebDataSource. List of countries by vehicles per capita. Wikipedia. Available at: http://en.wikipedia.org/wiki/List_of_countries_by_vehicles_per_capita. Accessed April 30, 2013.
  • Wong IC, Lhatoo SD. (2000) Adverse reactions to new anticonvulsant drugs. Drug Saf 23:3556.
  • Zhao Y, Zhang Q, Tsering T, Sangwan, Hu X, Liu L, Shang H, Chen Q, Liu Y, Yang X, Wang W, Li S, Wu J, Sander JW, Zhou D. (2008) Prevalence of convulsive epilepsy and health-related quality of life of the population with convulsive epilepsy in rural areas of Tibet Autonomous Region in China: an initial survey. Epilepsy Behav 12:373381.
  • Zhao FL, Yue M, Yang H, Wang T, Wu JH, Li SC. (2010) Validation and comparison of EuroQol and short form 6D in chronic prostatitis patients. Value Health 13:649656.

Supporting Information

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure
  8. References
  9. Supporting Information
FilenameFormatSizeDescription
epi12324-sup-0001-TableS1-S6.docxWord document91K

Table S1. Adjusted means for QWB-SA and EQ-5D for epilepsy and control groups.

Table S2. Univariate analyses for QWB-SA and EQ-5D.

Table S3. Spearman's correlation coefficients between HRQoL scores and demographic variables (Epilepsy group).

Table S4. Results of one-way ANOVA according to epilepsy-specific variables.

Table S5. Multiple linear regression analysis for QWB-SA and EQ-5D.

Table S6. Intraclass Correlation Coefficient (ICC), Pearson correlation and Spearman rho between QWB-SA and EQ-5D.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.