Epilepsy is a common neurologic disorder in the United States, with an estimated prevalence of 3.4 million cases, and approximately 150,000 new cases of epilepsy are diagnosed annually.1,2 Epilepsy is characterized by recurring seizures that involve either a part of the body (ie, partial-onset seizure) or the whole body (ie, generalized seizure).3 Partial-onset seizures are the most common type of seizures in patients with epilepsy.1 Episodes of partial-onset seizures may be accompanied by other symptoms, including loss of consciousness or other cognitive dysfunction, such as mental slowness, attention deficit, speech problems, memory difficulties, and hallucinations.1,4 Patients with epilepsy face high rates of comorbidities, including cardiovascular, respiratory, and metabolic disorders, as well as an increased risk for early mortality.5 Patients with this disease are also known to suffer from social stigma that results in lower self-esteem, reduced social support, discrimination, and unemployment,6-10 which often contributes to psychiatric issues that reduce quality of life and may pose a greater burden than the seizures themselves.6,11
Although epilepsy has a profound adverse impact on patients’ lives, the unpredictability of seizures also often affects the family members or caregivers of patients with epilepsy in various ways.12 First, the variability of seizure control among patients may have an impact on the caregiver’s feelings of anxiety about the patient.12-14 Caregivers may also have responsibilities regarding the patient in providing medication reminders, monitoring treatment adherence, and offering general and emotional support that can often vary by the type (ie, partial or generalized) and severity of seizures.12-14 In addition, families of patients with epilepsy have reported dissatisfaction with restricted lives and with increased levels of anxiety and depression.12,15 Furthermore, caregivers may bear a financial burden either directly through financial assistance to the patient or housing for the patient,14 or indirectly because of their own work-related absences or reduced productivity as a result of caregiving responsibilities.16,17
Finally, a patient’s antiepileptic drug (AED) regimen (ie, adjunctive or monotherapy AED regimen and varying treatment-related burden may impose a compounding effect on the magnitude of overall caregiver burden.
Little objective research has examined the different burdens on caregivers of patients who are receiving various AED treatment regimens. Evidence that expands the understanding of differences in caregiver burden between caregivers of patients who receive AED monotherapy or AED adjunctive therapy is needed, because this may play an underrecognized role in treatment choice and decision-making.
The objective of this study was to examine the potential differences in the annual direct healthcare costs, indirect costs (ie, payments made for absences resulting from sick leave, short- or long-term disability, and workers’ compensation), and lost time because of absences among employed caregivers of dependent spouses with partial-onset seizures who received monotherapy or adjunctive AEDs.
This retrospective cohort study was based on administrative claims from an integrated employer database. Healthcare claims, inpatient utilization, pharmaceutical expenditure data, time lost from work absences, and payments made to employees were obtained from the Human Capital Management Services (HCMS) integrated employer database between January 1, 2001, and June 30, 2014. At the time of this study, the HCMS database represented multiple employers in retail, service, manufacturing, and financial industries across the United States, and included information on 2 million employees plus their spouses and eligible dependents.
Data were deidentified to comply with the Health Insurance Portability and Accountability Act and the contractual obligations between HCMS and their employer-contributors. As a retrospective study using deidentified patient-level data that did not affect patient care, Institutional Review Board approval was not required.
The caregivers were identified as employed spouses of patients with partial-onset seizures, which was defined by any primary, secondary, or tertiary claims containing International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes for localization-related epilepsy and epileptic syndromes with complex partial seizures (ICD-9-CM code 345.4x) or with simple partial seizures (ICD-9-CM code 345.5x). Caregivers with epilepsy were excluded from the cohorts.
The 2 cohorts consisted of caregiver-patient pairs with information in the HCMS database. The caregiver-patient pairs were allocated into the monotherapy or adjunctive therapy cohorts based on the patient’s mode of therapy (Figure 1).
Adjunctive therapy was defined as more than 90 days of concomitant use of ≥2 AEDs. The index dates in the monotherapy cohort were the date of the first AED prescription for each caregiver-patient pair. The index dates for the adjunctive therapy cohort were the date of the first additional therapy for each caregiver-patient pair. Each pair was required to have 1 or more years of continuous data in the database after their index date.
The annual caregiver healthcare utilization and costs were calculated for the direct health benefit costs, including the medical and prescription costs for the caregiver, as well as for the indirect costs that result from the caregiver’s workplace absences over a 12-month follow-up period. The indirect costs that resulted from absences were calculated from payments for sick leave, short-term disability, long-term disability, and workers’ compensation. The workers’ compensation costs also included medical claims paid under the workers’ compensation benefit.
The likelihood of point-of-service use, the per-person direct medical costs, and the number of visits for each service among caregivers were assessed based on the available point-of-service (ie, doctor’s office, inpatient hospital, outpatient hospital or clinic, emergency department, laboratory, and other) data. Logistic regression that controlled for age, tenure, marital status, caregiver race, work-related components (ie, exempt status, full- and part-time status, and salary), location, and Charlson Comorbidity Index (CCI) score was used to model the likelihood of using each point of service. The per-person medical costs, absence payments, and absence time were modeled using sequential 2-part regression analysis that controlled for age, tenure, marital status, caregiver race, exempt status, full- and part-time status, salary, location, and CCI score.
All costs were inflation-adjusted to December 2017 US dollars; the medical costs used the medical cost component of the Consumer Price Index, the prescription costs used the prescription component, and all the other costs used the overall component of the Consumer Price Index.18
The baseline demographic and job-related variables were compared between the cohorts using t-tests for continuous variables and chi-square tests for binary variables.
Separate 2-part regression models were developed to assess the differences in cost and absence time between the caregiver cohorts of patients with partial-onset seizures who were receiving monotherapy or adjunctive therapy. In the first part of the model, logistic regression was used to predict the likelihood of caregivers having any costs or absence time. In the second part of the model, generalized linear models with a gamma distribution and a log-link function calculated the costs or absence time for caregivers with more than zero costs or absences.
All models controlled for differences in age, tenure (years with current employer), sex, marital status, race, exempt and nonexempt status (exempt employees were not paid on an hourly basis or for overtime work), full- and part-time status, salary, caregiver’s CCI score, and geography. The indirect costs were summed over all the absence claims during the 12-month postindex period. The differences were considered significant if P <.05.
Approximately 65% of the 367 caregiver-patient pairs were in the monotherapy cohort and 35% were in the adjunctive therapy cohort. Caregivers in both cohorts had similar age, duration of employment, sex, race, and mean CCI score; however, caregivers in the monotherapy cohort had significantly higher annual salaries than those in the adjunctive therapy cohort. The patients’ age and sex were similar between the 2 cohorts (Table 1).
The medical treatment costs were significantly lower for caregivers in the monotherapy cohort than in the adjunctive therapy cohort ($2751 vs $5269; P= .0002), but the caregiver’s prescription drug costs were similar in the 2 cohorts (Table 2).
Caregivers in the monotherapy cohort were absent from work for a total of 2.7 days annually compared with 5.1 days annually for caregivers in the adjunctive cohort, who had significantly more absences as a result of sick leave (Figure 2).
The number of days lost as a result of short-term disability was 0.3 days in the monotherapy cohort and 0.7 days in the adjunctive therapy cohort, although this difference was not statistically significant. The caregivers in the monotherapy and adjunctive therapy cohorts incurred $912 and $1192, respectively, of total indirect costs (Table 2). Workers’ compensation cost was significantly higher in the monotherapy cohort, whereas the sick leave cost was significantly higher in the adjunctive therapy cohort. The difference in short-term disability costs was not significant between the 2 groups.
The rate of inpatient hospital services among caregivers was 1.3% in the monotherapy cohort and 9.9% in the adjunctive therapy cohort (P= .0016; Table 3). No significant difference was observed between the cohorts in the likelihood of using other services.
Significant differences in direct healthcare costs were observed among caregivers for doctor’s office ($1062 with monotherapy vs $1420 with adjunctive therapy), inpatient hospital ($46 vs $1696, respectively), and “other” services ($25 vs $71, respectively). There were significant differences associated with the caregivers in the monotherapy and adjunctive therapy cohorts in the number of doctor’s office visits (5.68 vs 7.23, respectively) and inpatient hospital visits (0.04 vs 0.29, respectively).
To our knowledge, this is the first attempt to analyze and quantify the impact of patients’ differing AED therapy regimens for partial-onset seizures and their associated burden on caregivers’ direct and indirect healthcare costs and absenteeism. Although the study does not establish causation, it demonstrates a strong association between patients’ AED treatment burden and caregivers’ direct and indirect costs.
In this study, caregivers of patients who were receiving adjunctive therapy incurred 1.7 times greater direct costs and 2.2 times higher indirect costs than caregivers of patients receiving monotherapy. Of the total direct costs incurred by caregivers, the direct medical care costs accounted for more than 65% of the direct costs for caregivers in both cohorts.
When analyzed by point of service, caregivers in the adjunctive therapy cohort had a greater likelihood of receiving inpatient hospital care than those in the monotherapy cohort. This was further illustrated by the significantly higher cost and the number of hospital care services received by caregivers of patients in the adjunctive therapy cohort. Similarly, the indirect costs related to absences were higher among caregivers of patients in the adjunctive therapy cohort.
Although the burden among caregivers of patients with partial-onset seizures has been hypothesized to be significant, little is known about the potential healthcare resource utilization and costs among such caregivers. The database used in this study represented employers from various industries. Although the direct and indirect healthcare costs have increased in both study cohorts, it is plausible that factors such as industry sector, job level, and employer benefit coverage may influence caregiver healthcare utilization to varying degrees, which warrants examination in future studies.
Previous studies have focused on the relationship between caregiving and emotional health and quality of life12,14,15; however, the impact of caregiving on physical health has been less clear. The findings from this study show that caregivers of patients with epilepsy who are receiving adjunctive therapy have a greater likelihood of receiving hospital care, which results in increased healthcare resource utilization and costs. Research on caregivers and patients with other chronic illnesses suggests that health problems associated with caregiving are related to the degree of patient dependence.19
In epilepsy, the burden on a caregiver’s general health may also be related to the patient’s seizure type and severity.12 The greater medical costs observed with patients receiving adjunctive therapy in this study (data reported elsewhere20) may reflect greater patient dependency on their caregivers, which adversely affects the health of the caregivers. The study, however, had surprisingly similar prescription costs among caregivers of patients receiving adjunctive therapy with AEDs and caregivers of patients who received monotherapy. Although the reasons for the similarities in prescription drug costs among the caregivers in both groups are unclear, similar baseline CCI scores suggest that caregivers may have comparable comorbid and disease characteristics, which can potentially lead to parallel pharmacy utilization needs and patterns.
This study assessed payroll, disability claims, and workers’ compensation claims data to objectively measure indirect costs and absence days, rather than inferring this information from medical claims data or surveys. Furthermore, work absence data were based on filed claims and were not subject to recall bias. This research included the 2-part regression methodology that has been shown previously to effectively account for nonnormal distributions of outcome variables.21,22
As with all research, this analysis is limited mainly by the retrospective claims data design. The coding for epilepsy or partial-onset seizures was not verifiable, given the absence of clinical diagnostic variables in claims. Furthermore, patients with seizure activity that was coded other than with ICD-9 codes 345.4x or 345.5x or those who did not seek medical care were excluded. Given the absence of disease severity, level of seizure control, and other relevant clinical diagnostic information within administrative claims, it is possible that these variables play a role in caregiver burden.
In addition, the identification of caregiver-patient pairs depended on marital status identifiers. Cohabitation of the pairs could not be confirmed, and the patients might have received care from persons other than their spouse. Therefore, the impact on the caregiver might have been underestimated. Nevertheless, the cost differences between the cohorts were high after adjusting for potential confounders, which indicates a potentially strong association between patients’ disease severity and caregivers’ costs and burden.
Furthermore, patients in the adjunctive therapy group might have had additional AEDs added to their regimen during the 12-month follow-up period. Neither the number of drugs nor the addition of drugs to patients in the adjunctive therapy cohort was evaluated in the analysis; both situations would potentially imply an increased need for care and time, which can impose a greater burden on their caregivers.
Finally, the study population may not be representative of the real-world experience of caregivers of patients with partial-onset seizures across the entire socioeconomic spectrum, because caregivers in this study received health benefits and have compensation levels that are associated with moderate or higher socioeconomic status groups, whereas epilepsy is more prevalent among lower socioeconomic status groups.23,24 Future research should evaluate epilepsy patient caregiver burden in a wider range of socioeconomic statuses.
The present analysis found that caregivers of patients with partial-onset seizures who received adjunctive therapy had approximately twice the direct healthcare costs of, and used more sick days than, caregivers of patients who received monotherapy. Although the prevalence of partial-onset seizures is low, research with representatively large cohorts would be important to understand the relationship among the specific attributes of AED treatment burden that adversely affect caregiver burden. In addition, individual employer plans vary widely in benefit design and as such, there may be differences based on employer plans, as well as between industry sectors, that may have an additive impact on the magnitude of caregiver burden. Therefore, the caregiver burden in different employer sectors and among the unemployed data in a real-world setting should be studied further.
Adding concerns about the health of the caregiver to the existing family burden may have a further impact on the caregiver and the family. Understanding the types of preventive services that could lead to proactive care for the caregiver may improve preventive services and caregiver outcomes and reduce the overall healthcare costs.
Future studies should include an evaluation of the impact of programs such as counseling, diet, and exercise, on the healthcare needs of the caregiver. Such evidence may guide the development of potential self-management strategies for patients that would mitigate the overall burden on caregivers of patients with partial-onset seizures.
The authors wish to thank Fulton F. Velez, MD (formerly with Sunovion Pharmaceuticals) and Patricia Segarini, PhD (formerly with Percolation Communications), for medical editorial support.
Funding for this study was provided by Sunovion Pharmaceuticals.
Author Disclosure Statement
Mr Brook and Mr Smeeding received research support from, and Dr Rajagopalan was an employee of, Sunovion Pharmaceuticals during this study.
Mr Brook is VP and Head of Retrospective Research, the JeSTARx Group, Newfoundland, NJ; Dr Rajagopalan was VP and Head, Global HEOR & Market Access, Sunovion Pharmaceuticals, Marlborough, MA, during this study; Mr Smeeding is President, the JeSTARx Group, Dallas, TX.
- Shafer PO. About epilepsy: the basics. Epilepsy Foundation. January 2014. www.epilepsy.com/learn/about-epilepsy-basics. Accessed October 23, 2018.
- Zack MM, Kobau R. National and state estimates of the numbers of adults and children with active epilepsy — United States, 2015. MMWR Morb Mortal Wkly Rep. 2017;66:821-825.
- World Health Organization. Epilepsy. Fact sheet. February 8, 2018. www.who.int/news-room/fact-sheets/detail/epilepsy. Accessed February 9, 2018.
- Aldenkamp AP, Baker G, Mulder OG, et al. A multicenter, randomized clinical study to evaluate the effect on cognitive function of topiramate compared with valproate as add-on therapy to carbamazepine in patients with partial-onset seizures. Epilepsia. 2000;41:1167-1178.
- Centers for Disease Control and Prevention. Comorbidity in adults with epilepsy — United States, 2010. MMWR Morb Mortal Wkly Rep. 2013;62:849-853.
- de Boer HM, Mula M, Sander JW. The global burden and stigma of epilepsy. Epilepsy Behav. 2008;12:540-546.
- Leaffer EB, Hesdorffer DC, Begley C. Psychosocial and sociodemographic associates of felt stigma in epilepsy. Epilepsy Behav. 2014;37:104-109.
- Mula M, Sander JW. Psychosocial aspects of epilepsy: a wider approach. BJPsych Open. 2016;2:270-274.
- Smith G, Ferguson PL, Saunders LL, et al. Psychosocial factors associated with stigma in adults with epilepsy. Epilepsy Behav. 2009;16:484-490.
- Bautista RE, Shapovalov D, Shoraka AR. Factors associated with increased felt stigma among individuals with epilepsy. Seizure. 2015;30:106-112.
- Whatley AD, DiIorio CK, Yeager K. Examining the relationships of depressive symptoms, stigma, social support and regimen-specific support on quality of life in adult patients with epilepsy. Health Educ Res. 2010;25:575-584.
- Westphal-Guitti AC, Alonso NB, Vaz Pedroso Migliorini RC, et al. Quality of life and burden in caregivers of patients with epilepsy. J Neurosci Nurs. 2007;39:354-360.
- Walker ER, Bamps Y, Burdett A, et al. Social support for self-management behaviors among people with epilepsy: a content analysis of the WebEase program. Epilepsy Behav. 2012;23:285-290.
- Walker ER, Barmon C, McGee RE, et al. A dyadic model of living with epilepsy based on the perspectives of adults with epilepsy and their support persons. Epilepsy Behav. 2015;53:1-9.
- Ellis N, Upton D, Thompson P. Epilepsy and the family: a review of current literature. Seizure. 2000;9:22-30.
- Riechmann J, Strzelczyk A, Reese JP, et al; for the EpiPaed study group. Costs of epilepsy and cost-driving factors in children, adolescents, and their caregivers in Germany. Epilepsia. 2015;56:1388-1397.
- Thompson R, Kerr M, Glynn M, Linehan C. Caring for a family member with intellectual disability and epilepsy: practical, social and emotional perspectives. Seizure. 2014;23:856-863.
- US Bureau of Labor Statistics. Consumer Price Index. www.bls.gov/cpi/. Accessed September 28, 2017.
- Miravitlles M, Peña-Longobardo LM, Oliva-Moreno J, Hidalgo-Vega Á. Caregivers’ burden in patients with COPD. Int J Chron Obstruct Pulmon Dis. 2015;10:347-356.
- Brook RA, Smeeding JE, Young JW, et al. Comparisons of the health benefit costs among monotherapy and adjunctive therapy patients with epilepsy and their employed spouses. Poster presented at the annual meeting of the College of Psychiatric and Neurologic Pharmacists. April 19-22, 2015; Tampa, FL.
- Kleinman NL, Brook RA, Ramachandran S. An employer perspective on annual employee and dependent costs for pediatric asthma. Ann Allergy Asthma Immunol. 2009;103:114-120.
- Kleinman NL, Sadosky A, Seid J, et al. Costs, work absence, and adherence in patients with partial onset seizures prescribed gabapentin or pregabalin. Epilepsy Res. 2012;102:13-22.
- Beghi E, Hesdorffer D. Prevalence of epilepsy—an unknown quantity. Epilepsia. 2014;55:963-967.
- Tang DH, Malone DC, Warholak TL, et al. Prevalence and incidence of epilepsy in an elderly and low-income population in the United States. J Clin Neurol. 2015;11:252-261.