Epilepsy & Behavior 41 (2014) 83–90

Contents lists available at ScienceDirect

Epilepsy & Behavior journal homepage: www.elsevier.com/locate/yebeh

Diagnoses, procedures, drug utilization, comorbidities, and cost of health care for people with epilepsy in 2012☆,☆☆ A.N. Wilner a,⁎, B.K. Sharma b,1, A. Thompson b,2, A. Soucy b,3, A. Krueger b,4 a b

Lawrence and Memorial Hospital, New London, CT, USA Accordant Health Services, 4900 Koger Blvd, Greensboro, NC 27407, USA

a r t i c l e

i n f o

Article history: Received 4 June 2014 Revised 22 August 2014 Accepted 23 August 2014 Available online xxxx Keywords: Epilepsy Seizure Medical claims Procedures Drugs Comorbidities

a b s t r a c t Our objective was to identify the top MD-office, inpatient and outpatient diagnoses, procedures, drug classes, comorbidities, and cost of health care for people with epilepsy. We examined health insurance claims for 8388 persons with epilepsy (females = 52%, males = 48%; average age = 35 years; privately insured = 78%, and Medicaid-insured = 22%) from eight health insurance plans for the year 2012. All of the top three diagnoses for MDoffice place of service were either for other convulsions (780.39) or for epilepsy (345.90 and 345.40). Two of the top three primary diagnosis codes from the inpatient hospital and emergency department places of service were 780.39 and 345.90 for convulsions and epilepsy, respectively, while the third code was 786.50 for chest pain. The top three procedures from the MD-office setting were for immunizations (90471 and 90658) and blood counts (85025). The top three procedure codes from the outpatient hospital setting were 85025 for complete blood count, 80053 for comprehensive metabolic panel, and 80048 for basic metabolic panel. In the emergency department, the top three procedures were electrocardiogram (93010), computed tomography (70450), and chest Xray (71020). The top five drug classes among prescription drugs billed using an NDC code were (1) anticonvulsants, (2) analgesic-opioids, (3) antidepressants, (4) penicillins, and (5) dermatologicals. The mean monthly health plan paid cost for each patient with epilepsy in 2012 was $1028 (SD = $3181). Of this total, $761 (SD = $2988; 74%) was for medical, and $267 (SD = $760; 26%) was for prescription pharmacy claims. Fiftyeight percent (58%) of the patients had one or more of 29 prespecified comorbidities, while 42% had none. Monthly health-care costs increased markedly as the number of comorbidities increased. This information should help guide cost estimates and resource allocation in order to optimally care for people with epilepsy. © 2014 Elsevier Inc. All rights reserved.

1. Introduction Epilepsy is a group of disorders characterized by recurrent seizures [1]. In the United States (U.S.), there are more than two million people with epilepsy [2]. Estimated annual direct and indirect costs for epilepsy in the U.S. exceed $12.5 billion [2]. The health-care costs for individuals with partial-onset seizures [3] and epilepsy, in general, [4] are significantly higher than the healthcare costs for demographically matched controls. Health-care costs are higher for patients newly diagnosed with epilepsy than for patients ☆ Statistical analysis performed by: BK Sharma. ☆☆ Funding: Accordant Health Services, a CVS Caremark Company, Greensboro, NC. ⁎ Corresponding author at: Lawrence and Memorial Hospital, New London, CT 06320, USA. Tel.: +1 860 576 9424; fax: +1 888 959 5651. E-mail addresses: [email protected] (A.N. Wilner), [email protected] (B.K. Sharma), [email protected] (A. Thompson), [email protected] (A. Soucy), [email protected] (A. Krueger). 1 Tel.: +1 336 315 3756. 2 Tel.: +1 336 217 2909. 3 Tel.: +1 336 315 3788. 4 Tel.: +1 336 315 0737.

http://dx.doi.org/10.1016/j.yebeh.2014.08.131 1525-5050/© 2014 Elsevier Inc. All rights reserved.

with established chronic epilepsy [5]. Health-care resource utilization is also significantly greater for patients with uncontrolled epilepsy than for those with well-controlled epilepsy [6]. We previously determined that hospital utilization primarily drives the cost of care for people with epilepsy [4]. The average hospital charge per admission for patients with epilepsy has increased over the years [7]. However, outpatient costs eclipsed inpatient costs in a recent study [8]. For individuals with epilepsy, health-care resource utilization is higher for those with comorbidities than for those without [9]. The presence of a comorbidity, such as depression, may increase epilepsy severity and health-care utilization [10]. In a claims analysis of 6621 insured individuals with epilepsy, the presence of one comorbidity approximately tripled the cost of care [11]. The addition of a second or a third comorbidity continued to augment costs [11]. Comorbidities can also complicate diagnosis and adversely affect prognosis [12]. This current study was undertaken to assess resource utilization of a population of patients with epilepsy by determining their most frequent diagnoses, procedures, medications, and comorbidities. The information generated from this study could be useful for guiding cost estimates and resource allocation for people with epilepsy.

84

A.N. Wilner et al. / Epilepsy & Behavior 41 (2014) 83–90

2. Methods

2.4. Determining the age–gender distribution

2.1. Data sources and description

The percentages of females and males in the study population were determined. The age of patients with epilepsy was calculated as of July 1, 2012. Age grouping was done by deciles of age in years. The percentages of patients in each age group were calculated.

This descriptive study involved retrospective analysis of health insurance claims and membership data from eight health insurance plans. These health plans were located in different parts of the U.S. and contracted with Accordant Health Services (AHS), a CVS Caremark Company, to provide care management (CM) services for their members with epilepsy or other chronic conditions. Two health plans were located in the Northeast, two in the Midwest, two in the Southeast, one in the Southwest, and the other one in the Western region of the U.S. The total number of health-insured members in these eight health plans was 3.5 million. Health insurance claims and membership data for the year 2012 obtained from these health plans by AHS for providing CM services to patients with epilepsy were analyzed.

2.2. Identification of patients with epilepsy Patients with epilepsy were identified using the patented in-house AHS patient identification (PID) algorithm as in our previous research studies [4,11]. The AHS PID process identifies patients based on the International Classification of Diseases, Ninth Revision (ICD-9) diagnosis codes, pharmacy utilization, and date and place of service (POS). The ICD-9 diagnosis code for epilepsy (345.xx) and the diagnosis codes for convulsions (779.0, 780.3, and 780.39) were used to identify people with epilepsy for this study. To be included, patients had to have multiple occurrences of an epilepsy diagnosis, convulsion diagnosis, or antiepileptic drug (AED) claim. However, patients would not have been identified solely by an AED, nor would they have been identified by a single instance of an epilepsy/convulsion diagnosis without other supporting claims detail such as AED usage. To minimize false positives from the identification algorithm, any patient who indicated that they did not have epilepsy when contacted by an AHS nurse was excluded from the study (contacting patients is part of the usual AHS process and was not done for the sake of conducting this study). For the patients included in this study based on the claims-based algorithm, contact and disease confirmation was obtained from more than 60% of the population. Data for individuals who were deceased prior to data analysis for this study and for those indicating to AHS nurses that the health plan was their secondary insurance were excluded from the study.

2.5. Determining the prevalence of comorbidities The presence of comorbidity-specific diagnosis codes was examined in health insurance claims for 29 different comorbidities as in Wilner et al. [11]. For this study, comorbidities were defined as “distinct clinical entities” present in association with epilepsy (index disease) [13]. The diagnosis codes shown in up to three diagnosis code fields of the health insurance claims were considered for counting the comorbidities. The percentage of the total study population having the specified comorbidity condition was considered as the relative prevalence rate of comorbidities. The top 20 comorbidities were identified for the study cohort based on these relative prevalence rates. 2.6. Determining the health plan paid costs To calculate the mean health plan paid costs per member per month (PMPM, $), we summed the amount of the health plan paid cost for each patient and divided the resulting amount by the sum of the number of member months for the patient. The mean health plan paid cost PMPM ($) was calculated for the whole study cohort and by claims type, gender, age group, and by the subgroups with different numbers of comorbidities. The calculated mean health plan paid costs represent cost for comprehensive care; no attempt was made to distinguish the cost of treating seizures from the cost of comorbidities associated with epilepsy [14]. 2.7. Statistical analysis Association between variables was measured by calculating Pearson correlation coefficients using the correlation procedures of Statistical Analytical System (SAS; Version 9.1, SAS Institute, Cary, NC). The health plan paid cost PMPM data for females and males were analyzed using the General Linear Model procedure of SAS (uses f-test) to test if the means were statistically different. The mean health plan paid costs PMPM ($) are presented with standard deviations (SD). 3. Results

2.3. Determining most frequent diagnoses and procedures 3.1. Top diagnoses, procedures, and drugs Health insurance claims from the MD-office setting, the inpatient hospital setting, and the emergency department (ED) setting were analyzed separately to identify the most frequently occurring diagnosis codes. Health insurance claims with current procedural terminology (CPT) codes from the MD-office setting, hospital inpatient setting, ED setting, and hospital outpatient setting were analyzed to show the most frequently performed procedure codes after excluding CPTs for evaluation and management (E&M). These procedures included common tests such as complete blood count, metabolic panel, and electrocardiograms. Prescription pharmacy claims billed using the National Drug Code (NDC) were analyzed to show the top drug classes as well as the most frequently prescribed drug names for treatment and/or management of patients with epilepsy. The NDC is a unique drug product identifier in the U.S. for drugs for human use, in which different portions of the code indicate the manufacturer's/labeler's identifier, drug product code, and dosage form. The top diagnoses, procedures, drug classes, and drug names were determined based on the percentages of members having the specified codes or names. Only the top codes/descriptions (based on percentages of the patient population having those codes or descriptions) are presented.

The top diagnoses and procedures appear in Table 1. The top two diagnosis codes identified from the MD-office setting, inpatient hospital setting, outpatient hospital setting, and ED setting were (1) other convulsions (Dx code 780.39) and (2) epilepsy, unspecified, without mention of intractable epilepsy (Dx code 345.90). The third primary diagnosis among the top three was localization-related (focal) (partial) epilepsy and epileptic syndromes with complex partial seizures, without mention of intractable epilepsy (Dx code 345.40) for the MD-office setting. The third among the top three diagnoses was chest pain, unspecified (Dx code 786.50) for the inpatient hospital setting and for the ED setting. The third primary diagnosis among the top three was generalized convulsive epilepsy (Dx code 345.10) for the outpatient hospital setting. After excluding procedure codes for E&M and hospital discharge day management codes, the top three procedure codes identified from the MD-office setting were for immunizations (CPT codes 90471 and 90658) and blood counts (CPT code 85025). The top three physicians' claims billed for the inpatient setting were (1) radiologic examination of chest, single view (CPT code 71010); (2) ECG routine (CPT code

A.N. Wilner et al. / Epilepsy & Behavior 41 (2014) 83–90

85

Table 1 Top diagnoses and procedure codes based on percentages of members having those codes at specified item or place of service (POS). Item/POS

Rank

Code

Code description

% of membersa

Diagnosis codeb/MD-office

1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3

780.39 345.90 345.40 780.39 345.90 786.50 780.39 345.90 345.10 780.39 345.90 786.50 90471 85025 90658 71010 93010 70450 85025 80053 80048 93010 70450 71020

Other convulsions Epilepsy, unspecified, without mention of intractable epilepsy Localization-related (focal) (partial) epilepsy Other convulsions Epilepsy, unspecified, without mention of intractable epilepsy Chest pain, unspecified Other convulsions Epilepsy, unspecified, without mention of intractable epilepsy Generalized convulsive epilepsy Other convulsions Epilepsy, unspecified, without mention of intractable epilepsy Chest pain, unspecified Immunization administration, 1 vaccine Blood count; complete (CBC), automated Influenza virus vaccine, ≥3 years of age Radiological examination, chest, single view Electrocardiogram routine ECG Computed tomography, head or brain; without contrast material Blood count; complete (CBC) Comprehensive metabolic panel Basic metabolic panel (calcium, total) Electrocardiogram routine ECG Computed tomography, head or brain; without contrast material Radiological examination, chest, two views

18% 16% 11% 24% 14% 7% 17% 14% 5% 20% 13% 10% 15% 13% 12% 4% 3% 2% 17% 14% 8% 7% 6% 6%

Diagnosis codeb/inpatient hospital

Diagnosis codeb/outpatient hospital

Diagnosis codeb/emergency department

Procedure codec/MD-office

Procedure codec/inpatient hospital

Procedure codec/outpatient hospital

Procedure codec/emergency department

a b c

Percentage of members with a specific code among those who had such codes or descriptions. Determined after excluding diagnosis codes for signs and symptoms. Determined after excluding evaluation and management procedure codes.

93010); and (3) computed tomography, head or brain without contrast material (CPT code 70450) (Table 1). The top three procedure codes from the outpatient hospital setting were (1) blood count; complete (CBC) (CPT code 85025), (2) comprehensive metabolic panel (CPT code 80053), and (3) basic metabolic panel (CPT code 80048). The top three procedure codes from the ED setting were (1) electrocardiogram routine ECG (CPT code 93010); (2) computed tomography, head or brain without contrast material (CPT code 70450); and (3) radiological examination, chest two views, frontal and lateral (CPT code 71020). The top ten drug classes based on percent of members taking these drugs among all prescription drugs and the top ten drug names determined separately among all of the drug names found in claims billed during the analysis year are shown in Table 2. The top five drug names among the top 10 drug classes based on percentages of the study population using those drugs were the following:

Regardless of drug class, the top 10 drug names based on percentages of the study population using those drugs were (1) levetiracetam, (2) hydrocodone/acetaminophen, (3) azithromycin, (4) amoxicillin, (5) lamotrigine, (6) phenytoin sodium extended, (7) topiramate, (8) fluticasone propionate, (9) omeprazole, and (10) prednisone (Table 2).

Table 2 Top drug classes and drug names based on percentages of members having prescription (Rx) for those drug classes or drug names identified among prescriptions billed using the National Drug Code (NDC). Item/POS

• Anticonvulsants: levetiracetam, lamotrigine, phenytoin sodium extended, topiramate, and clonazepam; • Analgesic-opioids: hydrocodone/acetaminophen, oxycodone/ acetaminophen, tramadol HCl, acetaminophen codeine #3, and oxycodone HCl; • Antidepressants: sertraline HCl, citalopram hydrobromide, trazodone HCl, fluoxetine HCl, and amitriptyline HCl; • Penicillins: amoxicillin, amoxicillin/clavulanate, penicillin v potassium, ampicillin, and dicloxacillin; • Dermatologicals: triamcinolone acetonide, mupirocin, ketoconazole, clotrimazole/betamethasone, and clobetasol propionate; • Antianxiety agents: lorazepam, diazepam, alprazolam, hydroxyzine HCl, and hydroxyzine pamoate; • Ulcer drugs: omeprazole, ranitidine HCl, pantoprazole sodium, esomeprazole magnesium, and famotidine; • Analgesic-antiinflammatory: ibuprofen, naproxen, meloxicam, diclofenac sodium, and ketorolac tromethamine; • Antihypertensives: lisinopril, clonidine HCl, losartan potassium, lisinopril/hydrochlorothiazide, and guanfacine HCl; • Macrolides: azithromycin, clarithromycin, erythromycin base, clarithromycin ER, and erythromycin ethylsuccinate.

Rank NDC code Code description

1 2 3 4 5 6 7 8 9 10 Drug name/NDC Rx 1 2

–b –b –b –b –b –b –b –b –b –b –b –b

3 4 5 6 7 8 9 10

–b –b –b –b –b –b –b –b

Drug class/NDC Rx

Anticonvulsants Analgesic-opioids Antidepressants Penicillins Dermatologicals Antianxiety agents Ulcer drugs Analgesic-antiinflammatory Antihypertensives Macrolides Levetiracetam Hydrocodone/ acetaminophen Azithromycin Amoxicillin Lamotrigine Phenytoin sodium extended Topiramate Fluticasone propionate Omeprazole Prednisone

% of members with Rxa 81% 27% 23% 19% 18% 16% 16% 16% 16% 16% 22% 17% 16% 13% 13% 11% 8% 7% 7% 7%

a Percentage of members with a specific code among those who had such codes or descriptions. b Multiple codes may exist for the same descriptions.

A.N. Wilner et al. / Epilepsy & Behavior 41 (2014) 83–90

3.2. Age–gender distribution The study population (N = 8388) consisted of 52% females and 48% males (Table 3). The average age was 35 years (36 years for females and 34 years for males). The median age was 34 years (36 years for females and 33 years for males). The age range was 1 month to 94 years. By health insurance product type, 78% of the patients were commercially insured, while 22% were Medicaid-insured. Each decile of age group between 0 and 100 years contained individuals with epilepsy (Table 3). For example, age group 1 was composed of individuals 0 to 10 years of age, and age group 2 was composed of those 11 to 20 years of age. The percentages of patients in each age group varied, but there were fewer older patients (deciles = 7–10) (Table 3).

Fifty-eight percent (58%) of the patients had one or more of the 29 prespecified comorbidities, while the remaining 42% had none (Fig. 1). An increasing count of comorbidities was associated with a smaller number of patients (Fig. 1). Regardless of comorbidity type, the percentage of the study population having four or more comorbidities was 10%, and the percentage of the study population having seven or more comorbidities was 1%. The relative prevalence (%) of the top 20 comorbidities among the 29 specified conditions is presented in Fig. 2 in descending order of prevalence. Psychiatric diagnosis was the most prevalent comorbidity. Hypertension, hyperlipidemia, asthma, and headache were the other four of the top five comorbidities. The average number of comorbidities by age group based on deciles of a patient's age in years is shown in Fig. 3. The average number of comorbidities increased with advancing age group. Correlation analysis showed that comorbidity count for patients with epilepsy was positively and significantly associated with the age of patients with epilepsy (+0.35; p = 0.001). 3.4. Health-care costs Overall, the mean health plan paid cost PMPM ($) for the whole of the study cohort was $1028 (SD = $3181). Of the total paid PMPM amount, $761 (SD = $2988; 74%) was for medical, and $267 (SD = $760; 26%) was for pharmacy claims (Table 4). Table 3 Demographic characteristics of individuals with epilepsy health-insured by 8 different commercial health plans in the U.S.a Item

Group

N

Population characteristics

N Gender Product groupb Average age

Allb Female, %/male, % Commercial, %/medicaid, % Allb Female Male Allb Female Male 1 = 0 to 10 years 2 = 11 to 20 years 3 = 21 to 30 years 4 = 31 to 40 years 5 = 41 to 50 years 6 = 51 to 60 years 7 = 61 to 70 years 8 = 71 to 80 years 9 = 81 to 90 years 10 = over 90 years

8388 4386/4002 6551/1837 8388 4386 4002 8388 4386 4002 1219 1331 1154 1137 1282 1401 618 163 76 –c

100% 52%/48% 78%/22% 35 years 36 years 34 years 34 years 36 years 33 years 15% 16% 14% 14% 15% 17% 7% 2% 1% –c

Age group in deciles of age in yearsb

a b c

60 50

42

40 30

25

20

14

9

10 0

5

0

1

2

3

4

3

1

1

5

6

>=7

Number of comorbidities among 29 defined conditions

Fig. 1. Percentage distribution of members with epilepsy (N = 8388) by number of comorbidities among 29 different comorbidity conditions.

3.3. Comorbidities

Median age

Percentage of total members

86

Data analysis period was between January 1, 2012, and December 31, 2012. For both female and male combined. Data not shown if the associated n per subgroup shown was b10.

By gender, the mean health plan paid costs PMPM ($) for female patients with epilepsy and for male patients with epilepsy were $1006 (SD = $2597) and $1052 (SD = $3716), respectively (Table 4). The mean health plan paid costs PMPM ($) for females and males were statistically similar (p N 0.05). By age group, the mean health plan paid cost PMPM ($) was highest for the youngest age group of 0- to 10-year-old patients ($1321, SD = $4651) and for the 51- to 60-year-old patients ($1364, SD = $4589) (Table 4). However, the mean health plan paid costs PMPM ($) for the various age groups were not statistically different (p N 0.05). The correlation analysis showed that the health plan paid cost PMPM ($) was not significantly associated with age (r = +0.007; p = 0.50). By number of comorbidities, the mean health plan paid cost PMPM ($) was lowest for individuals with no comorbidities (Table 4). The mean cost PMPM increased for individuals having one comorbidity and then kept rising as the number of comorbidities increased, up to a comorbidity count of six, regardless of the comorbidity type. The mean cost PMPM for patients with four comorbidities was $2009. The mean cost PMPM was $3027 when there were more than four comorbidities, which was almost three times the mean cost PMPM for the whole study population. The mean health care cost PMPM for patients with seven or more comorbidities remained high but was quite variable because the number of patients for this comorbidity count subgroup was low. The correlation analysis showed that the health plan paid cost ($, PMPM) was positively and significantly associated with the number of comorbidities present in individuals with epilepsy (r = +0.23; p b 0.001). 4. Discussion 4.1. Top diagnosis codes We did not find a benchmark study to compare with our finding of top diagnoses and procedures for patients with epilepsy. Although over half (58%) of the study population had comorbidities, the top primary diagnosis codes (Table 1) for the MD-office settings, inpatient hospital settings, outpatient hospital settings, and ED settings included “other convulsions” (Dx code 780.39) and epilepsy (Dx code 345.90), indicating that patients with epilepsy seek and utilize health-care resources primarily for convulsions/seizures. 4.2. Top procedure codes The top CPT procedure codes identified from health insurance claims from the MD-office setting, inpatient hospital setting, outpatient hospital setting, and ED setting (Table 1) are consistent with the usual treatment of epilepsy. Physicians routinely order several laboratory tests or procedures, such as complete blood count (CBC), comprehensive metabolic panel, liver and thyroid function tests, electrocardiograms (ECG), and chest X-rays for patients presenting with seizures [15].

A.N. Wilner et al. / Epilepsy & Behavior 41 (2014) 83–90

87

25% 23% 20% 16% 14%

15%

11%

10%

8%

7%

5%

Osteoporosis

Renal failure

Sleep disorder

Dysphagia

COPD

Overweight and obesity

CAD

Cancer

Osteoarthritis

Migraine

GERD

Anemia

UTI

Hypothyroidism

Diabetes

Headache

Asthma

Hyperlipidemia

Hypertension

Psychiatric diagnosis

0%

6% 6% 5% 5% 4% 4% 4% 4% 3% 3% 2% 2% 2% 1%

CAD = coronary artery disease; COPD = chronic obstructive pulmonary disease; GERD = gastroesophageal reflux disease; UTI = urinary tract infection. Fig. 2. Relative prevalence rate (%) shown in y-axis for top 20 comorbidities (among 29 different comorbidity conditions defined) shown in x-axis for N = 8388 members with epilepsy. CAD = coronary artery disease; COPD = chronic obstructive pulmonary disease; GERD = gastroesophageal reflux disease; UTI = urinary tract infection.

Electroencephalograms (EEG), computed tomography (CT) brain scans, and brain magnetic resonance imaging (MRI) are also commonly ordered [15]. 4.3. Top medications Antiepileptic drugs (AEDs), also termed “anticonvulsants,” are the mainstay for epilepsy treatment [15]. Antiepileptic drugs were used by 81% of the patients. While one might expect that 100% of patients with epilepsy should be taking AEDs, some patients may have had “controlled” epilepsy and discontinued their medications, while others may have not filled their prescriptions. A sensitivity analysis conducted by excluding claims for certain AEDs, namely, topiramate, gabapentin, pregabalin, lamotrigine, and clonazepam which could be prescribed for treating different pain symptoms, indicated that 69% used other AEDs. The use of an AED for pain alone would not have resulted in the diagnosis of epilepsy in the PID algorithm. Two of the top 10 drug classes identified based on frequency of utilization are pain medications (analgesic-opioids and analgesicantiinflammatory), two are antimicrobial agents (penicillin and macrolides), and the other two are for psychiatric conditions (antidepressants and antianxiety agents). The reason for the high use of analgesic-opioids in a population with epilepsy is unclear as pain is not a characteristic symptom of epilepsy. Studies are needed to examine whether prevalence of opioid use is higher in people with epilepsy than in controls without epilepsy. If opioid use is increased in the epilepsy population, the possible causes of this increased utilization require investigation.

Average comorbidity count

3

2.68 2.00

2

4.4. Age–gender distribution Our population of health-insured individuals with epilepsy included 52% females and 48% males and all age groups from multiple geographic areas in the U.S. This observed gender distribution is consistent with the approximately 50:50 female/male epilepsy prevalence in unselected populations [16].

4.5. Comorbidities prevalent in patient populations with epilepsy The measured prevalence of medical and psychiatric comorbidities in people with epilepsy varies depending upon the study population,

Table 4 Health plan paid cost per member per month (PMPM) and standard deviation (SD) for overall study cohort and costs by claims type, gender, age group, and number of comorbidities.a Item

Group

N

Health plan paid cost mean PMPM ± SD ($)

Overall costs By claims type

All Medical claims Pharmacy claims Female Male 1 = 0 to 10 years 2 = 11 to 20 years 3 = 21 to 30 years 4 = 31 to 40 years 5 = 41 to 50 years 6 = 51 to 60 years 7 = 61 to 70 years 8 = 71 to 80 years 9 = 81 to 90 years 10 = over 90 years 0 1 2 3 4 5 6 7 8 9 10 N =11

8388 8388 8388 4386 4002 1219 1331 1154 1137 1282 1401 618 163 76 b10 3513 2133 1142 727 398 215 117 77 29 15 12 b10

1028 (SD = 3181) 761 (SD = 2988) 267 (SD = 760) 1006 (SD = 2597) 1052 (SD = 3716) 1321 (SD = 4651) 745 (SD = 1985) 892 (SD = 2765) 874 (SD = 2370) 934 (SD = 1877) 1364 (SD = 4589) 1029 (SD = 2067) 967 (SD = 1920) 1240 (SD = 2301) –c 419 (SD = 1037) 936 (SD = 3027) 1331 (SD = 3277) 1920 (SD = 5943) 2009 (SD = 4635) 2318 (SD = 3171) 3579 (SD = 7421) 3293 (SD = 3282) 3638 (SD = 4393) 3820 (SD = 2957) 3383 (SD = 1496) –c

By genderb By age group in deciles of age in years

2.82

2.23

By number of comorbidities

1.66 1.22

1

0

0.67

0.66

1

2

0.92

3

4

5

6

7

8

9

Age in deciles of years

Fig. 3. Average comorbidity count for patient age groups based on deciles of age in years. Patient count varies by age group. The count for age in decile 10 is not shown because this subgroup had b10 patients.

a b c

Claims data analyzed dates of service between January 1, 2012, and December 31, 2012. Average paid costs PMPM were similar (p N 0.05) for female and male. Data not shown if the associated n per subgroup shown was b10.

88

A.N. Wilner et al. / Epilepsy & Behavior 41 (2014) 83–90

methodology, definitions, and other factors. For example, data may be obtained by surveys or measured more directly by health insurance claims. Despite these differences, researchers consistently report an increase in medical and psychiatric comorbidities in people with epilepsy. For example, a recent report from the Centers for Disease Control and Prevention that focused on the prevalence of nonpsychiatric comorbidities based on data from the 2010 National Health Interview Survey generated a long list of medical disorders that were more prevalent in adults with epilepsy than in those without, including cardiovascular, respiratory, inflammatory, and other disorders such as headache, migraine, and certain other types of pain [17]. Results from the National Comorbidity Survey Replication, a face-to-face household survey of 5692 U.S. adults, indicated that people with epilepsy were more likely to report at least one comorbid physical disorder than those without epilepsy (93.6% vs. 77.8%, p b 0.001), with an odds ratio of 4.2 [18]. Further, people with epilepsy were more than twice as likely as people without epilepsy (41.2% vs. 20.2%) to have “high comorbidity,” defined as having four or more comorbid physical disorders (p b 0.001). In addition, mental disorders were reported more commonly by people with epilepsy (67.9% vs. 47.0%, p = 0.011), specifically posttraumatic stress disorder, panic disorder, conduct disorder, and drug abuse (odds ratio = 1.8–3.3, p = 0.002–0.043) [18]. A recent retrospective cohort study based on claims data from the Thomson Reuters MarketScan Commercial database revealed that the study population of 10,107 adult patients with epilepsy had a mean of 2.3 chronic conditions [8]. Depression and other mood disorders occurred in 6.3% of those with “stable” epilepsy, 10.9% of those with uncontrolled epilepsy, and 7% overall [8]. In a smaller study of 549 patients with refractory partial epilepsy, administrative claims data from the PharMetrics integrated medical and pharmacy claims database indicated that headache (10.6%), hypertension (10.0%), and depression (6.9%) were frequent comorbidities [19]. While it is not possible to directly compare these studies with ours due to differences in populations, methodology, and definitions, these studies support an increased prevalence of medical and psychiatric comorbidities in people with epilepsy. Our previous study analyzed health insurance claims for the year 2010 for 6621 commercially insured individuals from 10 health plans [11] and demonstrated that 50% of women and 43% of men (47% of the study population) had one or more of the 29 prespecified comorbidities. The health insurance claims data for the current study came from eight health plans that included four of the health plans from the earlier study [11]. Only 1969 (29.7%) patients in the current study cohort were also present in the previous study cohort. In the present study, 58% of the patients had one or more of the same 29 prespecified comorbidities. Examined by gender, 61% of females and 55% of males in the current study had at least one comorbidity. The current study demonstrates an 11% increase in the percentage of patients having one or more comorbidities compared with our prior study [11]. This difference could be related to the inclusion of Medicaid patients, a different mix of health plans, patient turnover in various health plans, improved coding that was more inclusive of comorbidities, and other factors as yet unidentified. The findings from our current study demonstrate a significant association between patient's age and number of comorbidities (+0.35; p = 0.001; Fig. 3), and between number of comorbidities and health-care costs (r = +0.23; p b 0.001). However, age per se had no significant association with health-care costs (Table 4). Comorbidities are important because their presence may complicate diagnosis and treatment, adversely affect quality of life, increase mortality [12,20], and increase cost [9,11]. The diagnosis and treatment of comorbidities is one of the key areas necessary to improve care for people with epilepsy identified by the Institute of Medicine's (IOM) 2012 report on epilepsy [21]. Psychiatric comorbidity plays an important role in epilepsy [22–24] and was the most frequent comorbidity detected in our previous study [11] as well as this one. Understanding psychiatric comorbidities is important for adequately managing patients with epilepsy [22].

When queried in surveys, people with epilepsy were more likely to complain of pain disorders than people without epilepsy [17,25]. Our findings that the analgesic-opioid drug class is the second most important drug class (27%) and that analgesic-antiinflammatory drugs (16%) are also among the top 10 most important drug classes in patients with epilepsy are consistent with these survey findings and do not have the limitations of self-reported data (Table 2). In the current study, pain was not considered as a separate comorbidity because of the difficulty in isolating pain from claims data. For example, pain may manifest in multiple common diagnoses such as arthritis and headache but may not be acknowledged with a separate claims code. 4.6. Health-care costs From an earlier study investigating the incremental cost of epilepsy, we [4] showed that annual health plan paid cost in 2009 was $11,232 for people with epilepsy, while the annual health plan paid cost for demographically matched controls without epilepsy was only $3026. The annual cost of $11,232 in 2010 brought to 2012 dollars by using Consumer Price Index's inflation adjustment calculator translated to a PMPM amount of $1002. Ivanova et al. [3] reported health-care costs PMPM to be $953 in the year 2010, which, after inflation adjustment to 2012 dollars, was $1003. Our PMPM cost estimate of $1028 for patients with epilepsy from the current study is very close to these two reported cost estimates [3,4]. A recent analysis of health-care utilization and cost of 1536 patients with uncontrolled epilepsy and 8571 patients with stable epilepsy revealed annual overall costs of $23,238 ($1936/month) for patients with uncontrolled epilepsy and $13,839 ($1153/month) for patients with stable epilepsy [8]. Both of these costs are higher than others in the literature but serve to illustrate the greater reliance on medical services and consequent increased cost for patients with uncontrolled epilepsy. For example, the mean inpatient hospitalization cost for patients with uncontrolled epilepsy was $6196 vs. $2818 for patients with stable epilepsy. When reporting “epilepsy-related medical costs,” inpatient hospitalization was the primary cost driver [8]. Our study did not differentiate patients with uncontrolled epilepsy from patients with controlled epilepsy. Our finding of higher health-care costs with an increasing number of comorbidities is consistent with observations from our earlier study [11]. An increasing number of comorbidities have also been associated with a marked increase in costs in a national random sample of 1,217,103 Medicare fee-for-service patients [26]. In this cross-sectional analysis, per capita expenditures increased from $211 for those without chronic conditions to $1154 for those with one chronic condition, $2394 for those with two chronic conditions, $4701 for those with three chronic conditions, and $13,973 for those with four or more types of chronic conditions [26]. Although this Medicare study was not restricted to epilepsy, it mirrors our observations of an increase in cost associated with an increase in the number of comorbidities. 4.7. Improving epilepsy care with lower cost Results from the current study suggest that people with epilepsy seek and utilize health-care resources primarily for convulsions/ seizures, pain, and infection. This study and our earlier study [11] demonstrate that comorbidities add to the cost of care for patients with epilepsy. Our previous research [4] indicated that inpatient hospital utilization is the top cost driver for health-care costs of people with epilepsy. The results of this study indicate that hospital admissions are primarily for the diagnoses of convulsions (780.39), epilepsy (345.90), and chest pain (786.50). Based on our findings, cost savings are likely to come from (1) controlling seizures and eliminating the need for hospital admissions and ED visits for seizures and (2) identifying and treating comorbidities.

A.N. Wilner et al. / Epilepsy & Behavior 41 (2014) 83–90

The top procedures shown in Table 1 were ranked based on the frequency of utilization. Reranking of the procedure codes based on contribution to total cost of care (procedure frequency multiplied by cost per procedure for that procedure) indicated that in addition to evaluation and management procedure codes in the MD-office setting (CPT codes 99214 and 99213), which were the most numerous patient procedures, the following five procedures had the largest impact on cost of care: 1) magnetic resonance imaging (CPT code for MRI 70553), 2) monitoring for localization of cerebral seizure focus (CPT code 95951), 3) nursing care in home (CPT code S9124), 4) ED visit evaluation and management code (CPT code 99285), and 5) ED visit evaluation and management code (CPT code 99284). 4.8. Study limitations The claims data we reviewed did not allow separation of patients by seizure type, syndrome, severity, etiology, or disease duration. Patients with only “other convulsions” (Dx code 780.39) and having no epilepsy diagnoses during the claims analysis period comprised 12% of the study cohort. It is possible that some of these patients did not have epilepsy. In addition, patients who had the diagnosis of epilepsy prior to but not during the claims analysis period could have been included in our study cohort. Consequently, patients with “other convulsions” who did not have epilepsy and patients with prior but not current epilepsy could account, at least in part, for the finding that only 81% of the patients with epilepsy were taking AEDs. It is likely that the AHS PID algorithm included some patients with nonepileptic seizures, which could have inflated the total population of patients with epilepsy and affected costs. Misdiagnosis of nonepileptic seizures as epilepsy commonly occurs in medical practice; it is estimated that 5–20% of people in the U.S. diagnosed with epilepsy actually have nonepileptic seizures [27]. The percentage of patients with psychogenic nonepileptic seizures (PNES) in our data set could not be determined, which is an inherent limitation of epilepsy studies that rely on claims data. Conversely, the AHS PID algorithm would have missed patients with controlled epilepsy who did not utilize medical services. We report comorbidities associated with epilepsy and make no attempt to infer causality or shared pathophysiological mechanisms [12,28,29]. The study relied on claims data generated for billing purposes, not scientific research. Coding errors and omissions limit the precision of analysis of this type of data. By study design, no matched-pair control group was used. Additional research that includes a control population could reveal the relative use of opioids, other pain medications, antibiotics, antidepressants, and other medications as well as patient services by people with epilepsy compared with the general population or those with other specific disease states. The cost of epilepsy has many facets, including direct, indirect, social, and emotional, but this study is restricted to measurable direct costs. The calculation of costs in this study is limited to those costs that could be tallied using claims data. This method excludes “out-of-pocket” costs that may have been paid by the patient directly when meeting the deductible portion of an insurance policy or when paying for items or services not covered by insurance. Consequently, the cost estimates in this paper are underestimates of the total costs per patient. Although these health plans were located in different geographic areas, they do not constitute a random or necessarily representative distribution of people with epilepsy in the U.S. Our study population consisted primarily of privately insured patients from eight U.S. insurers with a lesser component of Medicaid patients. In addition, patients with Medicare as their sole insurance were not included, which would have tended to exclude people older than 65 years as well as those with severe disabilities. Uninsured patients were not included. Consequently, these results may not be generalizable to all people with epilepsy. Nonetheless, this data set is one of the largest for reporting inpatient and outpatient diagnoses, procedures, medication use, and costs from health-care claims of people with epilepsy.

89

5. Conclusions In this descriptive study comprising 8388 people with epilepsy, “convulsions” and “epilepsy” were the primary indications for MD-office visits, hospital admission, outpatient hospital, and ED visits. The top three procedures from the MD-office setting were for immunization, blood counts, and influenza virus vaccination. The top three procedures from the inpatient hospital setting were chest X-ray, ECG, and head CT. The top three procedures from the outpatient hospital setting were complete blood count, comprehensive metabolic panel, and basic metabolic panel tests. The top three procedures from the ED setting were ECG, head CT, and chest X-ray. While not the most common procedures, magnetic resonance imaging, monitoring for localization of cerebral seizures, in-home nursing care, and ED visits were major contributors to the cost of care. The most common drug classes were anticonvulsants, analgesicopioids, antidepressants, penicillins, and dermatologicals. Fifty-eight percent (58%) of the patients had one or more of the 29 prespecified comorbidities, while 42% had none. Monthly health-care costs for patients with epilepsy increased markedly as the number of comorbidities increased. The mean health plan paid cost PMPM ($) for this patient population with epilepsy in 2012 was $1028 (SD = $3181). Of the total paid PMPM amount, $761 (SD = $2988; 74%) was for medical, and $267 (SD = $760; 26%) was for pharmacy claims. This information should help guide cost estimates and resource allocation in order to optimally care for people with epilepsy. 6. Ethical approval This research involved analyzing health insurance claims and membership data provided to Accordant Health Services by different health plans. No human subjects were contacted for the purpose of conducting this study. Member and health plan identifiers are not reported. Consequently, review by an Institutional Review Board was not pursued for this study. Disclosure AN Wilner is a Medical Advisory Board member for epilepsy for Accordant Health Services, a CVS Caremark Company. The coauthors are employees of Accordant Health Services. All authors have a relationship with Accordant Health Services where this study was conducted. Indexing keywords Epilepsy, Seizure, Medical claims, Procedures, Drugs, Comorbidities References [1] Fisher RS, van Emde Boas W, Blume W, Elger C, Genton P, Lee P, et al. Epileptic seizures and epilepsy: definitions proposed by the International League Against Epilepsy (ILAE) and the International Bureau for Epilepsy (IBE). Epilepsia 2005; 46(4):470–2. [2] Begley CE, Famulari M, Annegers JF, Lairson DR, Reynolds TF, Coan S, et al. The cost of epilepsy in the United States: an estimate from population-based clinical and survey data. Epilepsia 2000;41(3):342–51. [3] Ivanova JI, Birnbaum HG, Kidolezi Y, Qiu Y, Mallett D, Caleo S. Economic burden of epilepsy among the privately insured in the US. Pharmacoeconomics 2010;28(8): 675–85. [4] Wilner AN, Sharma BK, Soucy A, Krueger A. Health plan paid cost of epilepsy in 2009 in the U.S. Epilepsy Behav 2012;25:412–6. [5] Begley CE, Lairson DR, Reynolds TF, Coan S. Early treatment cost in epilepsy and how it varies with seizure type and frequency. Epilepsy Res 2001;47:205–15. [6] Manjunath R, Paradis PE, Parise H, Lafeuille MH, Bowers B, Duh MS, et al. Burden of uncontrolled epilepsy in patients requiring an emergency room visit or hospitalization. Neurology 2012;79:1908–16. [7] Vivas AC, Baaj AA, Benbadis SR, Vale FL. The health care burden of patients with epilepsy in the United States: an analysis of a nationwide database over 15 years. Neurosurg Focus 2012;32(3):E1. http://dx.doi.org/10.3171/2012.1.FOCUS11322. [8] Cramer JA, Wang ZJ, Chang E, Powers A, Copher R, Cherepanov D, et al. Healthcare utilization and costs in adults with stable and uncontrolled epilepsy. Epilepsy Behav 2014;31:356–62.

90

A.N. Wilner et al. / Epilepsy & Behavior 41 (2014) 83–90

[9] Cardarelli WJ, Smith BJ. The burden of epilepsy to patients and payers. Am J Manag Care 2010;16:S331–6. [10] Cramer JA, Blum D, Fanning K, Reed M, Epilepsy Impact Project Group. The impact of comorbid depression on health resource utilization in a community sample of people with epilepsy. Epilepsy Behav 2004:337–42. [11] Wilner AN, Sharma BK, Soucy A, Thompson A, Krueger A. Common comorbidities in women and men with epilepsy and the relationship between number of comorbidities and health plan paid costs in 2010. Epilepsy Behav 2014;32:15–20. [12] Gaitatzis A, Sisodiya SM, Sander JW. The somatic comorbidity of epilepsy: a weighty but often unrecognized burden. Epilepsia 2012;53(8):1282–93. [13] Feinstein AR. The pre-therapeutic classification of co-morbidity in chronic disease. J Chronic Dis 1970;23:455–68. [14] Pugliatti M, Beghi E, Forsgren L, Ekman M, Sobocki P. Estimating the cost of epilepsy in Europe: a review with economic modeling. Epilepsia 2007;48:2224–33. [15] Goldenberg MM. Overview of drugs used for epilepsy and seizures. Proc Natl Acad Sci U S A 2010;35:392–415. [16] Christensen J, Kjeldsen MJ, Andersen H, Friis ML, Sidenius P. Gender differences in epilepsy. Epilepsia 2005;46(6):956–60. [17] Kadima NT, Kobau R, Zack MM, Helmers S. Comorbidity in adults with epilepsy— United States, 2010. MMWR Morb Mortal Wkly Rep 2013;62(43):849–53. [18] Kessler RC, Lane MC, Shahly V, Stang PE. Accounting for comorbidity in assessing the burden of epilepsy among US adults: results from the National Comorbidity Survey Replication (NCS-R). Mol Psychiatry 2012;17:748–58. [19] Lee WC, Arcona S, Thomas SK, Wang Q, Hoffmann MS, Pashos CL. Effect of comorbidities on medical care use and cost among refractory patients with partial seizure disorder. Epilepsy Behav 2005;7:123–6.

[20] St. Germaine-Smith C, Liu M, Quan H, Wiebe S, Jette N. Development of an epilepsyspecific risk adjustment comorbidity index. Epilepsia 2011;52:2161–7. [21] Institute of Medicine. Epilepsy across the spectrum: promoting health and understanding. [Internet] Washington: The National Academies Press; 2012 [[Accessed 2014 May 14] at: http://www.ncbi.nlm.nih.gov/books/NBK91506/]. [22] Tellez-Zenteno JF, Patten SB, Jetté N, Williams J, Wiebe S. Psychiatric comorbidity in epilepsy: a population-based analysis. Epilepsia 2007;48(12):2336–44. [23] Gaitatzis A, Carroll K, Majeed A, Sander J. The epidemiology of the comorbidity of epilepsy in the general population. Epilepsia 2004;45(12):1613–22. [24] Fazel S, Wolf A, Langstrom N, Newton CR, Lichtenstein P. Premature mortality in epilepsy and the role of psychiatric comorbidity: a total population study. Lancet 2013; 382(9905):1646–54. [25] Ottman R, Lipton RB, Ettinger AB, Cramer JA, Reed ML, Morrison A, et al. Comorbidities of epilepsy: results from the Epilepsy Comorbidities and Health (EPIC) survey. Epilepsia 2011;52(2):308–15. [26] Wolff JL, Starfield B, Anderson G. Prevalence, expenditures, and complications of multiple chronic conditions. Arch Intern Med 2002;162:2269–76. [27] LaFrance WC, Benbadis SR. Avoiding the costs of unrecognized psychological nonepileptic seizures. Neurology 2006;66:1620–1. [28] Bonavita V, De Simone R. Towards a definition of comorbidity in the light of clinical complexity. Neurol Sci 2008;29:S99–S102. [29] Valderas JM, Starfield B, Sibbald B, Salisbury C, Ronald M. Defining comorbidity: implications for understanding health and health services. Ann Fam Med 2009;7: 357–63.