SEARCH

SEARCH BY CITATION

Keywords:

  • Epilepsy;
  • Quality of life;
  • Resilience;
  • Socioeconomic status

Summary

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

Purpose: Epilepsy has been associated with reduced quality of life (QOL), but QOL outcomes are heterogeneous. Some people are able to maintain a good QOL despite poorly controlled epilepsy and others report poor QOL despite well-controlled epilepsy. Maintaining a good QOL in the face of adversity is embodied by the concept of resilience. We explored the factors associated with having a resilient outcome in people with epilepsy (PWE). Our definition of adversity included socioeconomic disadvantage as well as continuing seizures.

Methods: We analyzed data collected as part of the Standard and New Antiepileptic Drugs (SANAD) trial. At the end of 4-year follow-up, patients were classified into four groups on the basis of seizure control (good/poor) and socioeconomic status (advantaged/disadvantaged). We identified individuals with resilient and vulnerable outcomes and the factors associated with having them.

Key Findings: Seizure control was more important in determining QOL than material advantage, but socioeconomic status appeared to act as an additional protective or risk factor for QOL. Significant predictors of a resilient outcome were absence of depression and fewer adverse treatment effects at 4 years and good QOL at baseline. Significant predictors of a vulnerable outcome were fair/poor health perception, presence of depression, reduced sense of mastery, and more adverse treatment effects at follow-up.

Significance: Reducing the adverse effects of treatment, along with psychosocial interventions to increase self-mastery, reduce health concerns, treat depression, and promote positive adjustment to a diagnosis will likely improve the QOL of PWE despite less favorable clinical and socioeconomic circumstances.

Epilepsy is a disorder that is associated with reduced quality of life (QOL). Many people with epilepsy (PWE) must manage the psychological sequelae associated with having the condition, which can impact on their overall QOL. There is a large body of evidence exploring the risk factors for impaired QOL. For those with active epilepsy, QOL is associated with seizure frequency and seizure severity (Jacoby et al., 1996; Jacoby & Baker, 2008). Those with well-controlled seizures generally report better QOL than people with poorly controlled seizures (Baker et al., 1997; Leidy et al., 1999; Kobau et al., 2007). The majority of those in remission have near-normal functioning, with scores on QOL measures, employment rates, and educational status similar to those of the general population (Jacoby, 1992; Stavem et al., 2000; Aldenkamp et al., 2003; Hessen et al., 2006; Jacoby & Baker, 2008). There are also nonepilepsy risk factors for reduced QOL, such as anxiety and depression, other comorbid conditions, and social stigma (Jacoby & Baker, 2008).

However, QOL profiles are heterogeneous and PWE can maintain a good QOL despite poorly controlled epilepsy. Equally, PWE can report a poor QOL despite well-controlled epilepsy (Argyriou et al., 2004). The idea of maintaining good QOL in the face of adversity is embodied by the concept of resilience. Resilience can be conceptualized as the process of achieving unexpected positive outcomes in adverse conditions, as opposed to an individual trait (Luthar et al., 2000; Canvin et al., 2009).

Research on resilience originally stemmed from the developmental psychology literature, but has been increasingly used in health research to understand how adults adapt to chronic illness such as diabetes (Yi et al., 2008), chronic pain (Zautra et al., 2005), human immunodeficiency virus (HIV) and acquired immunodeficiency syndrome (AIDS) (Farber et al., 2000), cancer (Rowland & Baker, 2005) and severe physical injury (Quale & Schanke, 2010). In this article we investigate factors associated with having a resilient outcome in PWE. Exploring resilience, and the ability of PWE to overcome adversity, represents an important shift toward promoting good QOL away from the risk factors for negative outcomes. In this study our definition of adversity includes continuing seizures as well as socioeconomic disadvantage. We explore the physical, psychological, and social factors associated with reporting a good QOL in the context of poor seizure control and socioeconomic disadvantage (“resilient” outcome) and the factors associated with reporting a poor QOL in the context of good seizure control and socioeconomic advantage (“vulnerable” outcome). The data were collected as part of the Standard and New Antiepileptic Drugs (SANAD) trial (Marson et al., 2007a,b).

Methods

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

SANAD was a pragmatic, randomized, unmasked, parallel group clinical trial comparing the clinical and cost-effectiveness of standard and new antiepileptic drugs (AEDs). A full description of the study methods can be found in Marson et al. (2007a,b).

QOL assessment was also investigated, as part of the secondary outcomes, in those aged 5 years and older without any significant learning disability as judged by the randomizing clinician from the history and examination. All eligible adults (defined as aged 16 years and older), were asked to self-complete QOL questionnaires at entry to SANAD, at 3 months and annually thereafter, up to a maximum of 4 years post-randomization. Questionnaires were sent by mail, with a single mailed reminder being sent to nonresponders 3 weeks after the initial mailing and telephone contact after a further 3-week period to those failing to respond. All questionnaires were accompanied by a cover letter explaining the purpose of the QOL study and a reply-paid envelope. Patients declining to return either a baseline or 3-month questionnaire were not contacted again, the assumption being that those who declined to complete questionnaires at this early stage in the life of the study were “active” refusers who would be unlikely to do so later. However, all patients completing either baseline or 3-month assessments were then cued to receive questionnaires at all subsequent time points, thereby maximizing the amount of data available for analysis. SANAD received appropriate multicenter and local ethics and research committee approvals. All patients gave informed written consent to inclusions and to long-term follow-up.

Description of QOL battery

The QOL assessment for SANAD involved the use of a battery of previously validated generic and epilepsy-specific measures (Table 1). The NEWQOL (Newly Diagnosed Epilepsy Quality of Life) battery examined physical, psychological, social, and cognitive domains. All the measures have been extensively used and validated in previous studies by the Liverpool Epilepsy Research Group (Baker et al., 1994; Abetz et al., 2000). In addition, the battery included a revised 12-item version of the “impact of epilepsy scale” (Jacoby et al., 1993), and single items relating to education, employment, driving, and marital status.

Table 1.   Content of the NEWQOL battery
PhysicalPsychologicalCognitiveSocial
  1. ABNAS, Aldenkamp-Baker Neuropsychological Assessment Schedule.

  2. aFor purposes of this analysis, items transformed to a binary response (general health perception = excellent, very good, good versus fair, poor; seizure worry = yes versus no; felt stigma = yes versus no; social limitations = restricted versus not restricted).

  3. bFor purposes of this analysis, items transformed to a better versus same versus worse response.

General health perception (single item; score 0–4)a (Abetz et al., 2000)Seizure worry (2 items regarding past or future seizures) (transformed score of 0–60)a (Abetz et al., 2000)ABNAS (24-item scale; score of 0–72) (Aldenkamp et al., 1995)Social activities (9-item scale; score of 0–27) (Abetz et al., 2000)
Health transition (single item; score of 0–4)b (Abetz et al., 2000)Hospital Anxiety and Depression Scale(14-item scale; 7 items = anxiety, 7 items = depression; score of 0–21 for each domain) (Zigmond & Snaith, 1983) Social limitations (single item; score of 0–3)a (Abetz et al., 2000)
Adverse events profile (19-item scale; score of 19–76) (Abetz et al., 2000)Sense of mastery (7-item scale; score of 7–28) (Pearlin & Schooler, 1978) Felt stigma (3-item scale; score of 0–3)a (Abetz et al., 2000) Work limitations (5-item scale; score of 0–20) (Abetz et al., 2000)

A single item measure of global quality of life, the Terrible-Delighted Faces scale (Andrews & Withey, 1976), was used as the primary outcome measure. The scale represents both an affective and a cognitive evaluation of life quality (Bowling, 1991). Respondents are shown seven faces, with expressions ranging from ecstatic to devastated (including a neutral one), and are asked to say which of the expressions comes closest to the way they feel about life as a whole. For the purposes of this analysis, those who reported being ecstatic, very happy, or happy were classified as having good QOL and those who reported being neutral, unhappy, very unhappy, or devastated were classified as having poor QOL at 4-year follow-up.

Classification of outcome groups

To explore the role of resilience for QOL, the SANAD patients were classified into four groups at the end of 4-year follow-up:

  • 1
     Good seizure control and socioeconomically advantaged.
  • 2
     Poor seizure control and socioeconomically advantaged.
  • 3
     Good seizure control and socioeconomically disadvantaged.
  • 4
     Poor seizure control and socioeconomically disadvantaged.
Seizure control

Seizure control was measured by the number of seizures that patients self-reported in the previous year. “Good” seizure control was defined as no self-reported seizures in the previous year and “poor” control as at least one seizure in the previous year at 4-year follow-up. One seizure was chosen as indicative of poor control, since previous research suggests that even a single seizure is detrimental to QOL, and in the United Kingdom it has major implications for driving, known to be an important aspect of QOL for many PWE.

Socioeconomic status

Socioeconomic status was determined on the basis of a composite index factor designed to provide a more accurate reflection of actual deprivation. The composite factor comprised three variables: employment status (paid employment, in full-time education or training versus not economically active, e.g., unemployed, retired, or looking after home); occupational social class (nonmanual vs. manual), and indicators of deprivation of area of residence (more affluent vs. more deprived).

Occupational social class

Occupational social class was defined according to the UK Registrar General’s social class scheme (Office of Population Censuses and Surveys, 1991), which classifies individuals according to their occupations. Social class was based on present occupation, or previous occupation for people who were not currently employed. Respondents who could not be assigned a social class—for example, individuals currently without work who did not report their previous occupation or people in a full-time education program—were excluded from this analysis. For the purposes of this analysis, study participants were classified into manual or nonmanual professions.

Indicators of area deprivation

Indicators of area deprivation were defined using the Townsend Material Deprivation Index (Townsend et al., 1988). The Townsend Material Deprivation Index is a composite score assigned to geographical areas in England and Wales based on four variables: the percentage of economically active residents who are unemployed (excluding students); the percentage of private households who do not possess a car or van; the percentage of private households not owner-occupied; and the percentage of private households classified as overcrowded (more than one person per room). The variables combine to form an overall deprivation score ranking a particular area relative to others, based on data taken from the 2001 UK Census. The higher the score, the more deprived the area (mean = 0). For the purposes of the composite index factor, deprivation scores were obtained for participants using their residential postal codes at 4-year follow-up. Study participants were classified as falling above (more deprived) or below (more affluent) the mean.

Creation of the composite index factor

At 4-year follow-up, those who scored high (i.e., in paid employment/nonmanual/living in more affluent areas) for at least two of the variables were classified as being “socio-economically advantaged” and those who scored low (i.e., not economically active/manual/living in more deprived areas) for at least two of the variables were classified as being “socioeconomically disadvantaged.” We included only those of working age (18–64 years), as it is difficult to determine accurately the socioeconomic status of retired participants. In addition, for older adults, epilepsy is more often secondary to other health conditions, for example, stroke or multiple health problems, making assessment of epilepsy-specific effects on QOL difficult.

Statistical analysis

All analyses were conducted using SPSS version 16.0 (SPSS Inc, Chicago, IL, U.S.A.). Relative risks (RRs) of reporting a good QOL in each of the four groups were calculated relative to group one (good seizure control and socioeconomically advantaged). Focusing on groups one and four only, chi square, independent t, and Mann-Whitney tests were used to compare those who had a QOL as expected given their clinical and socioeconomic circumstances with those who had either a “resilient” outcome (good QOL despite clinical and socioeconomic disadvantage) or a “vulnerable” outcome (poor QOL despite clinical and socioeconomic advantage). Groups were compared on baseline and 4-year follow-up factors from clinical, physical, psychological, and social domains. Because depression may be an important predictor of vulnerable outcome, and potentially related to other factors, for example, sense of mastery, health perception, and cognitive complaints, a series of logistic regressions were performed controlling for presence/absence of depression, to ensure that negative scores for these factors were not simply a reflection of low mood. For the purposes of the logistic regression analyses, health transition, anxiety, cognitive function, and education were recoded to binary variables [health transition = better versus same/worse than 1 year ago; anxiety = case/borderline versus not anxious; cognitive function = high score ≥ 15 versus low score < 15 based on cut-off points defined in (Aldenkamp et al., 2002); education = no formal qualifications versus formal qualifications] and mastery was log transformed to meet the assumption of normality. Finally, to assess possible predictors of “resilient” and “vulnerable” outcomes, two separate multiple logistic regressions were performed. Any variables that were significantly associated with each outcome (p > 0.05) were selected for entry into the multiple logistic regressions. A forward stepwise selection procedure was used to determine the final models (criteria for entry p < 0.05 and for removal p > 0.01).

Results

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

Response rate

Of the 1,881 eligible adults who were invited to take part in the QOL study, 1,611 completed a questionnaire at baseline (response rate of 85.6%). At 4 years, 1,442 people were retained in the study and sent questionnaires, of whom 1,001 responded (response rate of 70%; 53.2% of originally eligible; 62.1% of baseline responders). For the purposes of this analysis, 617 patients could be classified into the four groups, representing 38.3% of baseline responders, 32.8% of those who were originally eligible, and 61.6% of 4 year responders. There were a number of reasons that patients could not be classified into any of the four groups: did not respond to global QOL item (n = 14) or number of seizures in the previous year (n = 6); were > age 65 years (n = 152); had not provided details about their employment history (n = 19) or occupational class (n = 146); and Townsend Material Deprivation index scores were not available (n = 47).

Characteristics

At entry into SANAD, patients had a mean age of 37 years (range 16–60 years). There were slightly more men than women. The majority (81.2%) was newly diagnosed and had never been previously treated. Seventy-two percent had partial epilepsy and approximately two-thirds (68.5%) had experienced more than four seizures at the time of completing the baseline questionnaire. Fifteen percent had a prior neurologic deficit or disorder recorded in their clinical notes (Table 2).

Table 2.   Baseline demographic characteristics
 Good seizure control/advantaged (n = 267)Poor seizure control/advantaged (n = 115)Good seizure control/disadvantaged (n = 106)Poor seizure control/disadvantaged (n = 129)p-ValueTotal (n = 617)
  1. Number of respondents missing information on: number of self-reported seizures at baseline (14), seizure classification (1). NA, numbers in cell too small for tests of significance.

AgeMean (SD)37.9 (12.2)35.2 (11.1)36.9 (13.6)39.7 (12.7)0.03737.6 (12.4)
GenderMale (n, %)141 (52.8)48 (41.7)63 (59.4)79 (61.2)0.011331 (53.6)
Female126 (47.2)67 (58.3)43 (40.6)50 (38.8)286 (46.4)
Treatment historyUntreated (n, %)227 (85.0)92 (80.0)84 (79.2)98 (76.0)NA501 (81.2)
Monotherapy36 (13.5)18 (15.7)17 (16.0)28 (21.7)99 (16.0)
2 year Remission4 (1.5)5 (4.3)5 (4.7)3 (2.3)17 (2.8)
Neurologic deficitYes (n, %)31 (11.6)19 (16.5)17 (16.0)25 (19.4)0.19892 (14.9)
No236 (88.4)96 (83.5)89 (84.0)104 (80.6)525 (85.1)
Seizure classificationPartial (n, %)193 (72.6)88 (76.5)66 (62.3)98 (76.0)445 (72.2)
Generalized33 (12.4)16 (13.9)17 (16.0)17 (13.2)0.13383 (13.5)
Unclassified40 (15.0)11 (9.6)23 (21.7)14 (10.9)88 (14.3)
No. of seizures at baseline1 (n, %)17 (6.4)5 (4.5)1 (1.0)4 (3.3)27 (4.5)
2–392 (34.6)19 (17.1)38 (36.9)14 (11.4)NA163 (27.0)
4+157 (59.0)87 (78.4)64 (62.2)105 (85.4)413 (68.5)

Likelihood of reporting a good QOL in each group

Table 3 illustrates the proportions of patients reporting good QOL in each group. The majority of patients (71%) reported a good QOL at 4 years. This figure was much higher (84%) for those who had good seizure control and were socioeconomically advantaged and much lower (49%) for those who had poor seizure control and were socioeconomically disadvantaged. A higher proportion of those who had good seizure control but were socioeconomically disadvantaged reported a good QOL compared to those who were socioeconomically advantaged but had poor seizure control (73% vs. 66%).

Table 3.   Four-year global QOL by group
Socioeconomic statusSeizure control4-year QOL (n, %)TotalRR (95% CI)
GoodPoor
AdvantagedGood224 (84)43 (16)2671.00
Poor76 (66)39 (34)1150.79 (0.68–0.91)
DisadvantagedGood77 (73)29 (27)1060.87(0.76–0.98)
Poor63 (49)66 (51)1290.58 (0.48–0.70)
Total 440 (71)177 (29)617 

All other groups were significantly less likely to report a good QOL relative to those who had good seizure control and were materially advantaged. This was particularly evident for those who were doubly disadvantaged in terms of their clinical and socioeconomic circumstances (RR 0.58; 95% CI 0.48–0.70).

Factors associated with a “ resilient” and “ vulnerable” outcome

Focusing only on group four (poor seizure control and socioeconomically disadvantaged), 63 patients were classified as having a resilient outcome. Comparing these with the 66 who had a poor outcome as expected given their double disadvantage, there were no statistically significant differences for sociodemographic (e.g., age, sex, marital status, personal income) or clinical characteristics (e.g., seizure type, presence of neurologic disorders) (Table 4). However, the two groups differed on several psychological, social, and physical factors. At 4 years, those in the resilient outcome group were more than twice as likely to perceive their general health as being good and three times more likely to say they felt their health was better than in the previous year compared with those who reported a poor as expected QOL. They also had significantly lower anxiety and depression scores on the Hospital Anxiety and Depression Scale (HADS). They were four times less likely to be classified as anxious, and only one patient was classified as experiencing depression compared with 46.2% in the expected poor outcome group. At 4 years, they reported experiencing significantly fewer adverse events and fewer cognitive problems. They also had a higher sense of mastery. Only 54.0% of patients in the resilient group felt socially restricted compared with 92.3% of those reporting a poor as expected QOL; and only half felt stigmatized compared with 81.5% of those reporting a poor QOL as expected. The resilient group also reported experiencing significantly less worry about past and future seizures, despite their seizures being active. In addition, those with a resilient outcome were more than twice as likely to report a good QOL at baseline (Table 5). After controlling for the absence of depression, the following remained significant predictors of a resilient outcome: absence of anxiety and felt stigma, a higher sense of mastery, lack of social restrictions, few cognitive complaints and adverse events, and a good QOL at baseline. In the multiple regression analysis, a good QOL at baseline was one of the most significant predictors along with absence of depression and fewer self-reported adverse events at 4 years (Table 6).

Table 4.   Clinical and physical factors associated with a resilient and vulnerable outcome
  Doubly disadvantagedp-ValueDoubly advantagedp-Value
Resilient (n = 63)Expected poor (n = 66)Vulnerable (n = 43)Expected good (n = 224)
  1. Number of respondents missing information on: seizure type (1), health perception (2). NA, numbers in cell too small for tests of significance.

Age (4 years)Mean (SD)42.4 (13.3)44.0 (12.2)0.48941.7 (10.7)41.5 (12.5)0.900
GenderMale (n, %)35 (55.6)44 (66.7)0.19525 (58.1)116 (51.8)0.445
Female28 (44.4)22 (33.3) 18 (41.9)108 (48.2) 
Seizure typePartial (n, %)47 (74.6)51 (77.3)0.92734 (81.0)159 (71.0)0.382
Generalised9 (14.3)8 (12.1) 3 (7.1)30 (13.4) 
Unclassified7 (11.1)7 (10.6) 5 (11.9)35 (15.6) 
Neurologic disorderYes (n, %)8 (12.7)17 (25.8)0.0616 (14.0)25 (11.2)NA
No55 (87.3)49 (74.2) 37 (86.0)199 (88.8) 
Health perception (4 years)Excellent-good (n, %)45 (71.4)18 (27.7)<0.00130 (69.8)216 (96.9)NA
Fair-Poor18 (28.6)47 (72.3) 13 (30.2)7 (3.1) 
Health compared to 1 year ago (4 years)Better (n, %)19 (30.2)7 (10.6)<0.0017 (16.3)53 (23.7)NA
Same41 (65.1)34 (51.5) 26 (60.5)167 (74.6) 
Worse3 (4.8)25 (37.9) 10 (23.3)4 (1.8) 
Table 5.   Psychological and social factors associated with a resilient and vulnerable outcome
  Doubly disadvantagedp-ValueDoubly advantagedp-Value
Resilient (n = 63)Expected poor (n = 66)Vulnerable (n = 43)Expected good (n = 224)
  1. aFor the purposes of this analysis, baseline QOL dichotomized into a binary response. AEP, Adverse Events Profile; NA, numbers in cells too small for tests of significance. Number of respondents missing information on: anxiety (3), depression (4), mastery (6), worry about past and future seizures (5), felt stigma (8), social restrictions (3), cognitive function (1), adverse events (3), personal income (11), marital status (2), driving status (8), education (8), baseline QOL (11).

  2. bWe note that 26 people reported still holding a UK driving license despite experiencing seizures in the previous 12 months; however, we have no evidence that they continued to drive .

Anxiety (4 years)Median (IQR)6 (5)12 (5)<0.0018 (5)3.25 (5.25)<0.001
Anxiety (4 years)Case (n, %)11 (17.5)47 (72.3)<0.00114 (32.6)6 (2.7)NA
Borderline12 (19.0)9 (13.8) 12 (27.9)28 (12.6) 
Not anxious40 (63.5)9 (13.8) 17 (39.5)188 (84.7) 
Depression (4 years)Median (IQR)3 (4)10 (6.5)<0.0017 (5)2 (2)<0.001
Depression (4 years)Case (n, %)1 (1.6)30 (46.2)<0.0018 (18.6)1 (0.5)NA
Borderline10 (15.9)17 (26.2) 11 (25.6)2 (0.9) 
Not depressed52 (82.5)18 (27.7) 24 (55.8)218 (98.6) 
Mastery (4 years)Median (IQR)21 (5.62)16 (5.25)<0.00118 (4)22 (6)<0.001
Worry about past seizures (4 years)Yes (n, %)49 (79.0)61 (92.4)0.02930 (71.4)115 (52.0)0.021
No13 (21.0)5 (7.6) 12 (28.6)106 (48.0) 
Worry about future seizures (4 years)Yes (n, %)49 (77.8)60 (90.9)0.03938 (88.4)155 (70.8)0.017
No14 (22.2)6 (9.1) 5 (11.6)64 (29.2) 
Felt stigma (4 years)Yes (n, %)30 (49.2)53 (81.5)<0.00115 (34.9)47 (21.5)0.058
No31 (50.8)12 (18.5) 28 (65.1)172 (78.5) 
Social limitations (4 years)Restricted (n, %)34 (54.0)60 (92.3)<0.00116 (37.2)39 (17.6)0.004
Not restricted29 (46.0)5 (7.7) 27 (62.8)183 (82.4) 
Cognitive function (4 years)Median (IQR)15 (28)40 (26)<0.00123 (23)7 (12.5)<0.001
AEP score (4 years)Mean (SD)37.95 (11.45)54.05 (9.09)<0.00141.67 (9.61)31.99 (18.45)<0.001
Personal income (4 years)Below national mean (n, %)61 (96.8)64 (98.5)NA21 (51.2)126 (58.3)0.399
Above national mean2 (3.2)1 (1.5) 20 (48.8)90 (41.7) 
Marital status (4 years)Married/living with partner (n, %)29 (46.0)36 (56.2)0.24932 (74.4)155 (69.2)0.494
Not married34 (54.0)28 (43.8) 11 (25.6)69 (30.8) 
Driving status (4 years)bLicence (n, %)15 (23.8)11 (18.3)0.45733 (76.7)181 (81.5)0.466
No licence48 (76.2)49 (81.7) 10 (23.3)41 (18.5) 
Education (baseline)None (n, %)20 (32.3)20 (31.7)NA7 (16.3)20 (9.1)NA
School-level38 (61.3)37 (58.7) 19 (44.2)101 (45.9) 
Further education4 (6.5)6 (9.5) 17 (39.5)99 (45.0) 
Global QOL (baseline)aGood (n, %)40 (64.5)16 (25.4)<0.00120 (46.5)176 (81.1)<0.001
Poor22 (35.5)47 (74.6) 23 (53.5)41 (18.9) 
Table 6.   Predictors of a “resilient” and “vulnerable” outcome
 βSE (β)eβ95% CIp-Value
Resilient outcome
 Constant3.851.62   
 Good QOL at baseline1.220.513.381.25–9.100.016
 Absence depression at 4 years1.1800.563.251.09–9.710.035
 Adverse events score at 4 years−0.190.030.900.85–0.95<0.001
Vulnerable outcome
 Constant16.665.25   
 Ln Mastery score at 4 years−6.871.720.000.00–0.03<0.001
 Case/borderline depression at 4 years2.420.7811.522.45–51.720.002
 Fair/poor health perception at 4 years2.100.728.121.98–33.270.004
 Adverse events score at 4 years0.050.0261.051.00–1.110.045

Focusing only on group one (good seizure control and socioeconomically advantaged), 43 patients were classified as having a vulnerable outcome. Comparing these with the 224 who had a good outcome as expected, there were no differences for sociodemographic (e.g., age, sex, marital status, personal income) or clinical characteristics (e.g., seizure type, presence of neurologic disorders) (Table 4). At 4-year follow-up, 30.2% of those with a vulnerable outcome perceived their general health to be fair or poor, compared with only 3.1% of those who had a good QOL outcome as expected from their double advantage. They were also more likely to say that their health was worse than in the previous year. Those with vulnerable outcomes had significantly higher anxiety and depression scores on the HADS and were more likely to be classified as having anxiety (32.6% vs. 2.7%) and depression (18.6% vs. 0.5%) than those with good QOL outcomes as expected. They also reported experiencing significantly more adverse events and more cognitive problems. They had a reduced sense of mastery and significantly more worry about past and future seizures, despite experiencing no seizures in the previous year. Those with vulnerable outcomes were twice as likely to feel restricted in their social life and more likely to feel stigmatized, although this was not statistically significant. They were also nearly three times more likely to report a poorer QOL at baseline (Table 5). After controlling for the presence of depression, the following remained significant predictors of a vulnerable outcome: fair/poor health perception, case/borderline anxiety, reduced sense of mastery, social restrictions, high number of cognitive complaints and adverse events, and a poor QOL at baseline. In the multiple logistic regression, the most significant predictors were having fair or poor health perception; being classified as either borderline or a case for depression; a reduced sense of mastery; and more self-reported adverse events at 4 years (Table 6).

Discussion

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

It is well-documented that epilepsy is associated with reduced QOL, but in this study we found that 71% of PWE reported a good overall QOL at 4 year follow-up. This is not surprising, given that 61% had been seizure free for at least the previous 12 months. Of those who were seizure free, 81% reported a good QOL. This supports previous work reporting that those in remission have near-normal levels of functioning (Jacoby, 1992; Stavem et al., 2000; Aldenkamp et al., 2003; Hessen et al., 2006; Jacoby & Baker, 2008; Hessen et al., 2009). In contrast, only 57% of those not seizure-free reported a good QOL, which is consistent with previous research showing that those with poorly controlled seizures report a worse QOL than those with well-controlled epilepsy (Baker et al., 1997; Leidy et al., 1999; Kobau et al., 2007).

In this study, we also considered the contribution of socioeconomic status to QOL, determined on the basis of a composite factor comprising individual employment status, occupational social class, and area-based deprivation. Seizure control was more important in determining QOL than material advantage; however, socioeconomic status appeared to act as an additional protective or risk factor for QOL. Those who were doubly advantaged in terms of their seizure control and socioeconomic status had the highest proportion of individuals with a good QOL. Those who were doubly disadvantaged in terms of their clinical and socioeconomic circumstances, were less likely to report a good QOL, with just over half reporting their overall QOL as poor.

We identified 63 individuals who reported a good QOL despite poor seizure control and socioeconomic disadvantage. We classified these individuals as having a resilient outcome, since they achieved positive and unexpected outcomes in the face of adversity (Canvin et al., 2009). Compared to those who reported a poor QOL outcome as expected given their combined clinical and socioeconomic situation, those with a resilient outcome differed on several psychological, physical, and social factors. The most significant predictors were absence of depression, fewer adverse treatment effects, and good QOL at baseline (the point of diagnosis of their epilepsy).

In addition to exploring factors associated with a resilient outcome, we explored which factors were associated with the reporting of poor QOL in the context of good seizure control and socioeconomic advantage. We identified 43 PWE who had a vulnerable outcome. Compared to those who had a good QOL as expected given their combined clinical and socioeconomic situation, these individuals also differed on several physical, psychological, and social factors. The most significant predictors of a vulnerable outcome were fair or poor health perception; being classified as borderline or case for depression; a reduced sense of mastery; and a higher score on the Adverse Events Profile (AEP).

It was not surprising that the absence or presence of depression was an important factor for having a resilient or vulnerable outcome. There is a growing literature associating depression with lower QOL in PWE, with some suggesting that it is a greater determinant of QOL than seizure-related variables (Boylan et al., 2004; Johnson et al., 2004; Loring et al., 2004). Similarly, the experience of adverse treatment effects was also related to both resilient and vulnerable outcomes. This is in line with previous research that has found the experience of adverse effects to negatively impact on quality of life (Gilliam et al., 2004; Auriel et al., 2009; Perucca et al., 2009). It is also of interest that lack of a sense of mastery and seizure worry were implicated, since both may be addressable through focused psychoeducational interventions.

Helping individuals to develop psychological strategies to cope in the face of adversity represents an important shift toward positive psychology, which has been largely neglected in PWE. Considering resilience as a process rather than an innate and immutable trait clarifies that there are factors that can be the focus of change. For example, supporting people to improve their mood state or gain a sense of mastery will likely enable individuals to develop resilience and be better able to deal with the challenges associated both with epilepsy and socioeconomic disadvantage.

Inevitably, there are limitations to this research. SANAD was designed as a randomized trial comparing QOL outcomes of different AEDs, rather than as a cohort study; however, it presented us with an opportunity to investigate the concepts of resilience and vulnerability in a large cohort of PWE. In order to explore the concept of resilience, we adopted a person-focused approach, grouping people with resilient and vulnerable outcomes and examining factors associated with those outcomes (Hildon et al., 2010). Based on our earlier work on QOL in epilepsy and on the nature of resilience, PWE were classified into four groups on the basis of seizure control, socioeconomic status, and their response to a measure of overall QOL at 4 years. Because of the way in which the groups were defined, we were able to analyze data from only a subset of responders (38.3% of baseline responders; 61.6% of those who responded at 4 years). Due to the exclusion of potential patients, for example, students, because they could not be categorized according to our specified criteria, the final sample available for analysis was small and so may not be representative of PWE more broadly. In addition, the majority of patients in this study were adults with new-onset epilepsy at the time of randomization into SANAD. We have highlighted the factors that drive resilience and vulnerability in this group, but these factors may be different for patients with chronic, long-standing, childhood-onset epilepsy; this issue would be deserving of further investigation. Finally, the use of overall QOL as an end-point does not fully capture the idea of resilience as a dynamic process. Some individuals with a resilient outcome in our model may have reported a poorer QOL earlier or may have a poorer QOL outcome later in the illness trajectory, in response to different stressors. Our ongoing qualitative work, which mirrors this quantitative work, aims to conceptualize resilience as a process. By interviewing 70 patients selected from each of our four groups, we aim to explore individual, familial, community, and wider societal influences promoting or undermining resilience in PWE.

Despite the limitations, this research has important clinical and social care implications, once more confirming the importance of going beyond seizure counting in assessment of outcomes for PWE. The results suggest that reducing the adverse effects of AED treatment, along with psychoeducational interventions to increase self-mastery, reduce health concerns, promote positive adjustment to a diagnosis, and treat depression will improve the QOL of PWE despite reduced socioeconomic and clinical circumstances. Such interventions need to be developed and their impact assessed in randomized controlled trials. A key aim of the associated qualitative work will be to identify if there are wider social support efforts beyond these that may further promote improved QOL for PWE.

Acknowledgments

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

Our thanks go to the patients who completed the questionnaires for this study. We would like to acknowledge the contributions of all members of the SANAD study group who assisted with the Quality of Life substudy from which these data were drawn. In particular we thank Julie Doughty, Barbara Eaton, and Virginia Swain for assistance with data collection and coding. We also thank Steven Lane for advice on the statistical analysis.

The SANAD study was funded by a grant from the Health Technology Assessment Programme of the NHS. JT was supported by a grant from Epilepsy Research UK. Epilepsy Action also contributed toward our linked qualitative work on resilience.

Disclosure

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

None of the authors has any conflict 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.

References

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