Epilepsia, 52(5):925–931, 2011 doi: 10.1111/j.1528-1167.2011.03098.x
FULL-LENGTH ORIGINAL RESEARCH
Employment after anterior temporal lobectomy *Katherine Zarroli, yzJoseph I. Tracy, yzMaromi Nei, yxAshwini Sharan, and yzMichael R. Sperling *Jefferson Medical College, Thomas Jefferson University, Philadelphia, Pennsylvania, U.S.A.; yJefferson Comprehensive Epilepsy Center, Thomas Jefferson University, Philadelphia, Pennsylvania, U.S.A.; zDepartment of Neurology, Thomas Jefferson University, Philadelphia, Pennsylvania, U.S.A.; and xDepartment of Neurological Surgery, Thomas Jefferson University, Philadelphia, Pennsylvania, U.S.A.
SUMMARY Purpose: To explore the effect of anterior temporal lobectomy on employment and define demographic and clinical predictors of postoperative employment in a large cohort with a prolonged observational period. Methods: Subjects had an anterior temporal lobectomy for refractory epilepsy. All had an assessment period of 4 years or more with documentation of demographic factors, employment status, and seizure frequency prospectively registered in a database at surgery and at each contact after surgery. McNemar chi-square and a Wilcoxon matched pairs test were used to compare employment status before and after surgery. A multiple logistic regression assessed independent predictors of postoperative employment status based on preoperative employment status. Key Findings: Three hundred sixty-nine patients were evaluated. Employment levels were higher and unemployment levels were lower after surgery (McNemar v2 = 3.96; p = 0.047). Working before surgery (Wald’s v2 = 22.69, p < 0.0001) and having a greater percent of seizure-free years (Wald’s v2 = 34.43, p < 0.0001) were asso-
People with epilepsy have higher unemployment and underemployment rates than the general population (Elwes et al., 1991; Smeets et al., 2007). Cognitive, behavioral, and psychiatric impairments along with social barriers associated with epilepsy all probably contribute to this high rate of unemployment (Harden et al., 2004; Clarke et al., 2006; Bautista & Wludyka, 2007; Smeets et al., 2007). Seizure control appears to be an important factor influencing employment, and patients with uncontrolled seizures are more likely to be unemployed than patients whose seizures Accepted March 1, 2011. Address correspondence to Michael R. Sperling, Thomas Jefferson University, 900 Walnut St., Suite 200, Philadelphia, PA 19107, U.S.A. E-mail:
[email protected] Wiley Periodicals, Inc. ª 2011 International League Against Epilepsy
ciated with being employed after surgery. Of 131 patients who were unemployed or homemakers before surgery, 67 (51.1%) became employed postoperatively, with a younger age at surgery, a younger age of epilepsy onset, and driving a motor vehicle associated with gaining employment. Of 172 patients who were working at baseline, 27 (15.7%) became unemployed or homemakers after surgery. Gender was the only variable associated with loss of employment, with women being more likely to become homemakers (v2 = 14.98, d.f. = 6, p = 0.02). Most students were working after surgery, with seizure control influencing outcome. Significance: Anterior temporal lobectomy is followed by reduced unemployment and underemployment, with elimination of seizures, relative youth, and operating a motor vehicle serving as the main driving forces for improvement. This is important information for patients and physicians who contemplate surgery as it helps define reasonable expectations, and provides further objective evidence for benefits beyond purely medical outcomes after epilepsy surgery. KEY WORDS: Epilepsy surgery, Temporal lobectomy, Employment, Seizure, Driving.
are controlled (Chaplin et al., 1998; Smeets et al., 2007). Hence, medical interventions that improve seizure control should improve social functioning in epilepsy and promote the integration of patients into the workforce. When seizures can no longer be controlled with medication, brain surgery is typically considered. Surgery is an effective means of controlling seizures in patients with epilepsy (Sperling et al., 1996; Wiebe et al., 2001) and often leads to long-term seizure control (Elsharkawy et al., 2009a; Sperling et al., 2008). Many patients report that going back to work is an important reason to have surgery (Taylor et al., 2001). However, whether epilepsy surgery alters employment has been the subject of study with differing results. Some studies have reported a lack of improvement in number of patients working after surgery. Augustine et al. (1984) followed 32 adult patients
925
926 K. Zarroli et al. after cortical resection for 1–10 years and did not find a change in unemployment level following surgery. Fraser et al. (1993) also found that unemployment levels were unchanged in patients after varying follow-up periods. Reeves et al. (1997) found that more patients lost employment than gained employment after temporal lobectomy. In contrast, other studies have reported employment gains following surgery. Jones et al. (2002) compared employment outcomes in 61 patients who had an anterior temporal lobectomy and 23 patients who were eligible for surgery but chose to proceed with medical management. Surgery resulted in a significant improvement in employment status. Surgical intervention was also found to lower unemployment levels in 86 patients who underwent temporal lobectomy for refractory epilepsy (Sperling et al., 1995). A recent study observing a cohort of 299 patients reported modest employment gains 2 years after surgery (Chin et al., 2007). Certain clinical features have been associated with increased likelihood of postoperative employment. Seizure control (Augustine et al., 1984; Sperling et al., 1995; Reeves et al., 1997; Chin et al., 2006, 2007; Kral et al., 2007; Elsharkawy et al., 2009a,b), driving (Reeves et al., 1997), race (Chin et al., 2007), and age at surgery (Sperling et al., 1995) have all been associated with being employed after surgery. It is important to understand the social effects of surgery, yet the literature is replete with contradictory findings, small sample sizes, and relatively short observational periods. Therefore, further investigation is required. The present study was designed to overcome prior limitations. We explored the effect of anterior temporal lobectomy on employment in a large cohort and required a period of prolonged observation, a minimum of 4 years, for inclusion. We further investigated demographic and clinical predictors of postoperative employment.
Methods Subjects Data were obtained from patients with refractory temporal lobe epilepsy who had an anterior temporal lobectomy at Graduate Hospital or Thomas Jefferson University Hospital in Philadelphia, Pennsylvania between April 1986 and January 2006. Clinical and psychosocial data have been prospectively collected since 1986. Criteria for surgery have been published previously (Sperling et al., 1996). Patients with a Full Scale IQ 50%) following surgery, and Group 3 patients had a minority of seizure-free years (£50%) after surgery. We chose to look at seizure-free years rather than percent reduction in seizure frequency, reasoning that unemployment might only be reduced if extended periods of time, perhaps years, passed without any potentially disabling seizures. Data analysis Analyses were performed using STATISTICA (StatSoft, Tulsa, OK, U.S.A.) and Microsoft Excel (Microsoft, Redmond, WA, U.S.A.).
927 Employment after Epilepsy Surgery Employment outcomes Means, standard deviation, and ranges were used to summarize overall group demographics. Chi-square and oneway analysis of variance (ANOVA) was used to compare baseline characteristics between seizure control groups. McNemar chi-square and a Wilcoxon matched pairs test were used to compare employment status before and after surgery for the entire cohort and for each seizure control group. A multiple logistic regression was run to determine if seizure control group and preoperative employment status were independent predictors of postoperative employment status. Predictors of change in employment Subgroup analyses were carried out in patients who were unemployed and employed at baseline. Univariate analyses and multiple logistic regressions were run to determine variables associated with loss of employment in patients employed at baseline and variables associated with gaining employment in patients unemployed at baseline.
Results Three hundred sixty-nine patients were eligible for inclusion in this study, and their characteristics are summarized in Table 1. Table 2a shows employment status before and after surgery (at the time of the last evaluation) for the entire cohort. Table 2b shows employment status before and after surgery, for a subset of 303 patients with students and retired persons excluded. A separate analysis was performed excluding students and retired persons, since their inclusion could potentially bias the results, as students would graduate and change employment status (perhaps with a bias to start working) and retired persons were not eligible for the workforce principally due to age. In this subpopulation, there was a substantial decline in unemployment: 124 patients (40.9%) were unemployed before surgery, whereas only 73 patients (24.1%) were unemployed after surgery. Of note, 11 women chose to become homemakers after surgery primarily to raise a Table 1. Patient characteristics Total (n = 369) Gender Educational level (% with high school diploma or less) Full Scale IQ (range) Age at last visit (range) Age at surgery (range) Presurgical duration of epilepsy (range) Age of epilepsy onset (range) Length of postsurgical follow-up (range)
188 M, 181 F 52.7%
92.8 45.6 35.9 20.2
± ± ± ±
13.0 (70–135) 11.6 years (22–80) 11.1 years (10–68) 11.3 years (0–62)
15.7 ± 12.5 years (0.01–64) 9.7 ± 4.7 years (4–23)
family; had these individuals been considered as unemployed, there still would be a significant decline in unemployment. After surgery, more patients were gainfully employed (full- or part-time) and fewer patients were not gainfully employed (unemployed or homemaker) than before surgery (Table 2b; McNemar v2 = 3.96; p = 0.047). A secondary analysis assessed whether surgical intervention altered employment status when individual work categories were considered; employment levels were ranked (fulltime = 1, part-time = 2, homemaker = 3, unemployed = 4) and a Wilcoxon matched pairs test was run. A difference remained between preoperative and postoperative employment status (z = 3.574, p < 0.0001). Employment outcomes based on postoperative seizure control Gender, educational level, age at surgery, age of epilepsy onset, age at last patient visit, length of postsurgical followup, presurgical duration of epilepsy, and full scale IQ were compared between Group 1 (seizure-free in each year following surgery), Group 2 (>50% of seizure-free years following surgery), and Group 3 (£50% of seizure-free years following surgery). There were no significant differences between groups for any of these variables adjusting for multiple comparisons. Students and retired persons were excluded from this analysis. Employment status before surgery and at the most recent evaluation was compared for each seizure control group (Fig. 1). Categories were ranked (full-time = 1, parttime = 2, homemaker = 3, unemployed = 4). Patients in Group 1 and Group 2 had a difference between preoperative and postoperative employment statuses (Wilcoxon matched pairs test; Group 1: z = 4.068, p < 0.001; Group 2: z = 1.992, p = 0.046), with higher employment rates postoperatively. Employment status did not change for patients in Group 3. Therefore, seizure control independently predicted employment status. Table 2. Employment status before and after surgery (a) all subjects (n = 369), and (b) with students and retired persons excluded (n = 303) Before surgery
After surgery
(a) All subjects Full-time 163 44.2% Full-time Part-time 24 6.5% Part-time Homemaker 7 1.9% Homemaker Unemployed 134 36.3% Unemployed Student 35 9.5% Student Retired 6 1.6% Retired Total 369 Total (b) Excluding subjects who were students or retired Full-time 150 49.5% Full-time Part-time 22 7.3% Part-time Homemaker 7 2.3% Homemaker Unemployed 124 40.9% Unemployed Total 303 Total
192 46 18 81 8 24 369
52.0% 12.5% 5.0% 22.0% 2.2% 6.5%
171 41 18 73 303
56.4% 13.5% 5.9% 24.1%
Epilepsia, 52(5):925–931, 2011 doi: 10.1111/j.1528-1167.2011.03098.x
928 K. Zarroli et al. Table 3. Predictors of becoming gainfully employed in patients who were not working before surgery
Postoperative driving status Age at surgery Age of epilepsy onset Full-scale IQ Gender Side of operation
Log reg estimate
Wald’s v2
p-value
1.64 )0.08 )0.06 0.04 0.33 0.13
12.66 6.62 6.44 2.69 0.51 0.09
0.00a 0.01a 0.01a 0.10 0.48 0.76
Model statistics: v2 = 49.60, d.f. = 6, p < 0.0001. a Significance levl, p < 0.05.
Figure 1. Change in employment status and postoperative seizure control. Graph illustrating change in employment status after anterior temporal lobectomy in relation to seizure outcome group. The number of patients with a change in employment category (full-time, part-time, homemaker, unemployed) after surgery is shown for each seizure control group. Group 1 = seizure-free in each year following surgery; Group 2 = majority of seizurefree years (>50%) following surgery; and Group 3 = minority of seizure-free years (£50%) following surgery. After surgery, the change in employment classification is significant for Groups 1 and 2, whereas there is no significant change in employment in Group 3 (see text for detail). Epilepsia ILAE
We next examined predictors of working after surgery. Preoperative employment status and seizure control group were independent predictors of postoperative employment status in a bivariate logistic regression (v2 = 67.68, d.f. = 2, p < 0.0001). For this analysis, employment status was categorized as employed (full-time or part-time) or unemployed (unemployed or homemaker). Being employed before surgery was associated with being employed following surgery (Wald’s v2 = 22.69, p < 0.0001) as was a better seizure outcome group (Wald’s v2 = 34.43, p < 0.0001). Predictors of a change in employment in patient subsets We next examined predictors of change in employment status after surgery in two subgroups of patients: (1) patients who were not gainfully employed (unemployed or homemaker) before surgery and (2) patients who were gainfully employed (full-time or part-time) before surgery. Univariate analysis and multiple logistic regressions were run to determine predictors of change in employment following surgery. Unemployed before surgery Of 131 patients who were unemployed or homemakers at baseline, 67 (51.1%) were gainfully employed (either fulltime or part-time) at their last evaluation; a complete dataset Epilepsia, 52(5):925–931, 2011 doi: 10.1111/j.1528-1167.2011.03098.x
for analysis was available for 130 of these subjects. A younger age at surgery, a younger age of epilepsy onset, and driving a motor vehicle after surgery were all associated with gainful employment after surgery (v2 = 49.57, d.f. = 6, p < 0.0001). Table 3 shows details. Gender, full scale IQ, and side of operation were not predictors of gaining employment. Seizure control was not included in this multivariate analysis. Because driving status is dependent on seizure status, seizure control (percent of seizure-free years) was evaluated separately in a univariate analysis. A greater percent of seizure-free years was associated with working after surgery (v2 = 18.52, p < 0.0001). Employed before surgery Of 172 patients who were working at baseline, 27 (15.7%) became unemployed or homemakers after surgery. Gender was the only variable associated with loss of employment (v2 = 14.98, d.f. = 6, p = 0.02); women were more likely to stop working, some having decided to become homemakers (no men reported becoming homemakers). Had homemaker been considered to be productively employed, then gender would not have had a significant impact on employment after surgery. None of the other variables, including age at surgery, age of epilepsy onset, postoperative driving status, full scale IQ, and side of operation were predictive of becoming unemployed. In a separate univariate analysis, having a lower percent of seizure-free years was found to be associated with unemployment following surgery (v2 = 5.28, p = 0.02). Students Of the 37 patients who were students prior to surgery, most became employed (22 full-time, five part-time) after surgery. Sixteen students who became employed were seizure-free since surgery and six students had a majority (>50%) of seizure-free years. Eight students became unemployed and two patients remained students following surgery. Five of the eight students who became unemployed had a minority (
