Prevalence and cost of insomnia in a state Medicaid fee-for-service population based on diagnostic codes and prescription utilization

Sleep Medicine 11 (2010) 462–469

Contents lists available at ScienceDirect

Sleep Medicine journal homepage: www.elsevier.com/locate/sleep

Original Article

Prevalence and cost of insomnia in a state Medicaid fee-for-service population based on diagnostic codes and prescription utilization Anuja N Roy a,*, Michael Smith b,1 a b

Global Health Economics & Outcomes Research, Takeda Pharmaceuticals Inc., One Takeda Parkway, Deerfield, IL 60015, USA Department of Pharmacy, Clinical & Administrative Sciences, College of Pharmacy, University of Oklahoma, 4502 E. 41st St., Suite 1H09, Tulsa, OK 74135, USA

a r t i c l e

i n f o

Article history: Received 19 June 2009 Received in revised form 16 September 2009 Accepted 26 September 2009 Available online 3 April 2010 Keywords: Insomnia Prevalence Medicaid Prescription utilization Costs Comorbidities

a b s t r a c t Objectives: The aims of this research were to estimate prevalence of insomnia, describe the utilization patterns of physician office services and prescription medications for insomnia, and estimate related costs in a Medicaid population. Methods: A cross-sectional descriptive analysis using data from the West Virginia (WV) Medicaid fee-forservice paid claims records for the year 2003 was conducted. Recipients with a diagnosis related to insomnia or a prescription claim for an FDA-approved drug for insomnia or trazodone were selected as the study sample. Costs were from the perspective of WV Medicaid. Results: The overall prevalence of insomnia was 74.3 per 1000 recipients. Adults 45–64 years of age, females, and whites had the highest prevalence and office visit rates for insomnia among demographic groups. A majority of dollars spent on insomnia treatment was for prescription drugs. Zolpidem and trazodone accounted for 88% of prescription claims; however, 84% of the total dollars paid for prescriptions was for zolpidem. Conclusions: Among the WV Medicaid population, rates of insomnia and office visit use for insomnia varied by demographic groups. There was greater use of zolpidem and trazodone than benzodiazepine drugs. This study provides baseline estimates that can be used for ongoing surveillance of insomnia. Ó 2010 Elsevier B.V. All rights reserved.

1. Introduction Insomnia is the most widespread sleep complaint globally [1]. In terms of frequency, it ranks highest among the different types of sleep difficulties in the United States [2]. In the scientific literature, the term ‘‘insomnia” has referred to a variety of conditions. Sleep experts have described the condition as a distinct sleep disorder associated with characteristic mental and/or organic pathology [3–6]. Major classification systems including the International Classification of Diseases (ICD-10) Classification of Mental and Behavioral Disorders, the International Classification of Sleep Disorders, and the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) have proposed detailed definitions for insomnia. According to a 2005 National Institutes of Health (NIH) State-ofthe-Science Statement, insomnia may be defined as complaints of disturbed sleep in the presence of adequate opportunity and circumstance for sleep. Such sleep disturbance may be associated with difficulty falling asleep (sleep initiation), difficulty remaining asleep (sleep maintenance), or early morning awakening [7]. Whatever the nosology, all definitions rely on the symptomatic * Corresponding author. Tel.: +1 224 554 6614; fax: +1 224 554 7983. E-mail address: [email protected] (A.N Roy). 1 Tel.: +1 918 660 3571; fax: +1 918 660 3580. 1389-9457/$ - see front matter Ó 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.sleep.2009.09.012

presentation of insomnia, and a standard diagnostic definition is largely absent from the literature [8–10]. As such, an array of definitions has been used for the term ‘‘insomnia.” This coupled with the fact that patients may present to a variety of clinicians has resulted in the condition being understood differently and inconsistently among health care professionals [11]. A recent systematic review of several epidemiologic studies concluded that depending upon the definition and methods used prevalence rates of insomnia vary between 4% and 33% [10]. A majority of cases (75%) occur comorbid with other illnesses (e.g., a medical, psychiatric, circadian, or sleep disorder) [12]. Such comorbid insomnia is frequently assumed to be ‘‘secondary” to the accompanying illness and is assumed to resolve itself with treatment of the accompanying disorder [13,14]. Individuals with insomnia can, however, experience a variety of consequences resulting from their condition. These include impaired mood and daytime functioning, cognitive deficits including poor memory, confusion, and decreased concentration [15,16], higher absenteeism from work [17], greater health care utilization [18], and an increased risk for accidents [19], all of which can have a substantial impact on healthcare costs and lost productivity. The direct economic costs of insomnia in the United States were estimated to be $13.9 billion in 1995 [20]. An estimate including both direct and indirect costs, varied between $92.5 billion and $107.5 billion [21].

A.N Roy, M. Smith / Sleep Medicine 11 (2010) 462–469

In spite of the significant social, economic, and humanistic burden imposed by insomnia, it is an under-reported [22], under-recognized [23,24], under-diagnosed, and under-treated condition [23,24]. Given the importance of a good night’s sleep in optimal daytime functioning it is recommended that insomnia, whether accompanying other illnesses or whether stand-alone, be appropriately treated [13,25]. Available treatments are broadly classified into two categories: non-pharmacological and pharmacological. Non-pharmacological interventions consist primarily of cognitive behavioral therapies. These methods focus primarily on factors that are presumed to be responsible for causing and maintaining the insomnia. Therefore, they seek to modify such factors and educate patients about healthier sleep habits. The pharmacological agents approved for treatment of insomnia in the United States are benzodiazepine and non-benzodiazepine agents [26]. Studies have found the prevalence of insomnia to be higher in individuals with lower incomes [27–29] and in those with lower education [30,31]. Nevertheless, a review of the literature did not reveal studies examining prevalence, medical services, and prescription use for insomnia specifically within these vulnerable sub-populations. Therefore, the purpose of the current study is to describe the distribution of insomnia among persons enrolled in a state Medicaid program, to assess related medical services and prescription drug utilization, and to estimate associated costs. The specific objectives are to estimate the prevalence of insomnia among recipients of a state Medicaid program, to describe the patterns of use of physician office services and patterns of prescription use for insomnia, and to report the costs incurred by Medicaid for insomnia-related treatment/care. 2. Methods 2.1. Data source This study used computerized West Virginia (WV) Medicaid administrative paid claims records for medical services and prescription medications with a date of service between January 1 and December 31, 2003 for recipients enrolled in a fee-for-service payment system. The data also included enrollment information on all enrollees. Medicaid is a co-operative venture jointly funded by the federal and state governments to assist states in providing medical assistance to individuals and families with low incomes and resources in the United States. Data were collected after a proposal for the study was approved by the Institutional Review Board (IRB) and the Health Insurance Portability and Accountability Act (HIPAA) Compliance Board of West Virginia University (WVU). Data were password-protected and de-identified in compliance with WVU IRB & HIPAA confidentiality regulations that require the protection of patient health information.

463

bance, unspecified), 307.47 (Other dysfunctions of sleep stages or arousal from sleep), 780.59 (Other sleep disturbances), 307.42 (Persistent disorder of initiating or maintaining sleep), 307.41 (Transient disorder of initiating or maintaining sleep), 307.40 (Non organic sleep disorder), 307.48 (Repetitive intrusions of sleep), 780.56 (Dysfunctions associated with sleep stages or arousal from sleep), 780.55 (Disruptions of 24-h sleep-wake cycle); or (2) at least one prescription claim for a drug approved by the FDA for the indication of insomnia (zolpidem, zaleplon, flurazepam, quazepam, temazepam, estazolam, triazolam); or (3) at least one prescription claim for trazodone. A few facts peculiar to insomnia necessitated the use of both diagnostic codes and prescription claims for more complete sample identification. First, insomnia is seldom the reason for a visit to the physician [32], and there was concern we would be unable to identify an accurate representation of recipients with insomnia using only diagnosis codes. Second, when present along with other illnesses, insomnia may be considered a symptom rather than a stand-alone disorder that remits with treatment of the underlying medical or psychiatric illness. Therefore, insomnia is often mentioned by patients seeking treatment for other conditions but may be regarded by physicians as a casual secondary complaint for which no diagnosis is assigned [33]. Hypnotics may, however, be prescribed to alleviate the symptoms, although a hypnotic prescription may not always be associated with the diagnosis of insomnia [34]. Indeed each year millions of hypnotic prescriptions are written without a diagnosis for insomnia [35]. Additionally, observation of prescribing practices in medical care has shown that trazodone is most often used for its sleep promoting properties rather than for its antidepressant properties (or for other potential benefits). Therefore, it is important to consider that there is a high probability in favor of insomnia as the target problem when trazodone is prescribed. As such we believed that using both medical service claims identified by the above mentioned diagnostic codes and prescription claims for medications with first-line indications for insomnia would enable a more adequate approach to sample identification for the purposes of the current study. Such extraction criteria have also been used earlier and reported in the literature [36]. 2.3. Study design This study used a cross-sectional design to conduct a retrospective descriptive analysis of the prevalence of insomnia among recipients of WV Medicaid enrolled in a fee-for-service plan, utilization of medical services and prescription drugs for insomnia (including trazodone), and related costs to WV Medicaid in 2003. Unique identification numbers of persons who met our inclusion criteria were used to extract all medical services and prescription claims as well as demographic data for each recipient in the sample.

2.2. Study sample

2.4. Data analysis

A total of 296,131 persons were enrolled in West Virginia Medicaid in a fee-for-service payment system in 2003. Of these, 289,235 recipients used their benefits to receive services; of these recipients, 254,794 were under 65 years of age and comprised the target population for the current study (we limited the sample to recipients less than 65 years of age because the primary health care payer for the elderly in the US is Medicare). From this population, those recipients who were selected as the study sample had: (1) at least one medical service claim for hospitalization or emergency department (ED), or office visit with one of the following International Classification of Diseases, 9th Edition, Clinical Modification [ICD-9-CM] diagnosis codes for sleep difficulties recorded in any of nine diagnosis fields (780.52 (Insomnia), 780.50 (Sleep distur-

The period prevalence of insomnia was calculated as the unique number of recipients identified by our criteria based on either the presence of a diagnosis or a prescription medication as defined under the ‘‘Study sample” section. Using this criteria led to a prevalence estimate of 7.4%. But if the prevalence calculation was restricted to only those recipients with the presence of a diagnosis of insomnia, a rate of 1.9% resulted. The use of medical services was based on the frequency of visits with any diagnosis of insomnia. The total number of prescription claims for insomnia-related drugs (defined as those with an FDAapproved indication for insomnia or claims for trazodone) was used to report the utilization of prescription medications by the sample, as well as to calculate the proportion of recipients treated

464

A.N Roy, M. Smith / Sleep Medicine 11 (2010) 462–469

by various insomnia-related drug classes. The average days supply of an insomnia-related medication for each recipient in the study sample and average days supply per prescription claim were also calculated. Costs were reported as the dollars paid by WV Medicaid for insomnia-related medical service visits and prescription utilization among the sample. The total prescription drug costs and costs per prescription claim across various insomnia-related medications were also estimated. All analyses were performed using Microsoft Visual Basic FoxPro (v 6.0), SPSS (v 13.0) and Microsoft Excel (MS Office 2003).

45 and 64 years of age had the highest rate of insomnia (176.6 per 1000) which was almost 2.4 times the overall rate of prevalence. Recipients who were female had a higher rate of insomnia (88.5 per 1000) compared to recipients who were male (56.4 per 1000). Thus females had 1.6 times the prevalence of insomnia compared to males. The rate of insomnia among recipients who were white (76.1 per 1000) was 1.6 times that among recipients who were black (47.2 per 1000). The total numbers, proportions and rates of insomnia among recipients by demographic groups are presented in Table 2. 3.2. Medical services utilization

3. Results 3.1. Sample characteristics and prevalence rates Of the 254,794 recipients who comprised our target population, 18,941 (7.4%) unique recipients met our criteria of being included in the sample. The sample consisted of 4733 recipients (defined as Group 1) who were identified by insomnia-related diagnostic codes, 9782 who were identified by a claim for a prescription drug with an FDA-approved indication for insomnia, and 8,005 recipients who were identified by a claim for trazodone. Of the 9782 recipients who had a prescription claim for a drug with an FDA-approved indication for insomnia, 8056 (defined as Group 2) had no medical claim with an insomnia diagnosis. Of the 8005 recipients who had a trazodone prescription, 6152 (defined as Group 3) did not have either a medical claim with a diagnosis for insomnia or a prescription claim for a medication with an FDA-approved indication for insomnia. The distribution of recipients identified with insomnia by the three different criteria is presented in Table 1. Analysis of demographic characteristics and prevalence rates of all recipients identified with insomnia revealed that the majority of the sample was female (66.1%), white (96.0%) and between 21 and 64 years of age (87.3%). There were similar proportions of recipients in the age groups of 21–44 years (45.8%) and 45–64 years (41.5%). Again, while the overall rate of insomnia including all the recipients identified using our three different selection criteria was 74.3 per 1000 recipients of WV Medicaid, recipients between

A total of 132 hospitalizations, 253 ED visits, and 6194 office visits with a diagnosis of insomnia were made by the sample during 2003. Since office visits accounted for a substantial proportion of medical services utilized by the sample (94% of visits), office service utilization, specifically, is reported here. The overall rate of insomnia-related office visits among recipients of WV Medicaid benefits was 24.3 visits per 1000 recipients. Adults between 45 and 64 years of age had the highest rate of office visits (60.4 visits per 1000) compared to other age groups. The rate of office visits among females was higher (26.6 visits per 1000) compared to males (21.4 visits per 1000). The rate of office visit use was greater among whites (24.9 visits per 1000) than blacks (15.1 visits per 1000). The total numbers, proportions and rates of office visit utilization by age, gender, and race groups are presented in Table 3. 3.3. Comorbid states As mentioned earlier, insomnia often tends to be comorbid with other disorders. The most frequent concurrent diagnoses among our study sample are shown in Table 4. 3.4. Prescription utilization A total of 44,266 prescription claims for drugs with an FDA-approved indication for insomnia and 37,007 claims for trazodone was paid by WV Medicaid in 2003. A substantial majority (93.9%)

Table 1 Demographic characteristics of recipients of West Virginia Medicaid with insomnia identified using three different criteria. Demographic variables

*

Group 1*

Group 2**

Group 3***

Total

****

v2*****

p

N

(%)

N

(%)

N

(%)

N

(%)

Age 0–20 21–44 45–64 Unknown

741 2076 1916 –

(15.7) (43.9) (40.5)

359 3901 3796 –

(4.5) (48.4) (47.1)

1295 2706 2151 –

(21.1) (44.0) (35.0)

2395 8683 7863

(12.6) (45.8) (41.5) (0)

921.4 37.6 215.16

<0.001 <0.001 <0.001

Gender Male Female Unknown

1827 2901 5

(38.6) (61.3) (0.1)

2497 5546 13

(31.0) (68.8) (0.2)

2051 4081 20

(33.3) (66.3) (0.3)

6375 12528 38

(33.7) (66.1) (0.2)

77.7 76.0

<0.001

Race White Black Other Unknown

4558 144 26 5

(96.3) (3.0) (0.6) (0.1)

7767 227 49 13

(96.4) (2.8) (0.6) (0.2)

5866 219 47 20

(95.4) (3.6) (0.7) (0.3)

18191 590 122 38

(96.0) (3.1) (0.6) (0.2)

11.5 6.5

0.003

Total

4733

(100.0)

8056

(100.0)

6152

(100.0)

18,941

100.0

N = Recipients of West Virginia Medicaid enrolled in a fee-for-service plan with at least one medical claim with a diagnosis of insomnia. N = Recipients of West Virginia Medicaid enrolled in a fee-for-service plan with a prescription claim for a drug approved by the FDA for insomnia who did not have a medical claim with a diagnosis for insomnia. *** N = Recipients of West Virginia Medicaid enrolled in a fee-for-service plan with a prescription claim for trazodone who did not have either a medical claim with a diagnosis for insomnia or a prescription claim for a drug approved by the FDA for insomnia. **** Total N = Recipients of West Virginia Medicaid enrolled in a fee-for-service plan with a diagnosis of insomnia, or a prescription claim for a drug approved by the FDA for insomnia, or a prescription claim for trazodone. ***** Chi-square statistics were conducted to test for differences in proportion of recipients among groups by demographic variables. **

A.N Roy, M. Smith / Sleep Medicine 11 (2010) 462–469 Table 2 Prevalence of Insomnia Among Recipients of West Virginia Medicaid by demographic groups (2003). Prevalence rate per 1,000 recipients 

Demographic variables

Group 1

Group 2

Group 3

Total

Age 0–20 21–44 45–64

5.2 30.4 43.0

2.5 57.2 85.2

9.1 39.7 48.3

16.9 127.2 176.6

Gender Male Female

16.2 20.5

22.1 39.2

18.1 28.8

56.4 88.5

Race White Black Other à

19.1 11.5 –

32.5 18.1 –

24.6 17.5 –

76.1 47.2

Total

18.6

31.6

24.1

– 74.3

 

Rates based on number of recipients identified with insomnia divided by the number of recipients of WV Medicaid in a fee-for-service system in each demographic group (Total N = 254,794). à Rate could not be calculated due to small sample size.

of the 18,941 recipients included in the study sample had at least one such prescription claim. On average, each prescription claim equaled a 28-day supply of medication, and each recipient received a 127-day supply of the drug. On average, there were nearly 4.3 prescription claims per recipient during the year. Nevertheless, if only those recipients who actually filled a prescription for an insomnia-approved drug or trazodone are considered (i.e., 17,787), the average number of prescription claims per recipient was 4.6 per year. Considering drugs with an FDA-approved indication for insomnia specifically, the largest proportion of claims by pharmacotherapy category was for the non-benzodiazepine benzodiazepine receptor agonists (80.2%). Benzodiazepines accounted for the remaining 19.8% of claims. Zolpidem alone accounted for 78.5% of the claims followed by generic temazepam (17.9%). The distribution of prescription claims by specific agents and the proportions of recipients who used various agents for insomnia among the study sample are shown in Table 5. Table 3 Insomnia-related office visit utilization among recipients of West Virginia Medicaid by demographic groups (2003). Demographic variables

*

Office/Clinic visits

*

Office/Clinic visit rate per 1000 recipients  

N

(%)

Age 0–20 21–44 45–64 Unknown

802 2701 2691 0

(12.9) (43.6) (43.4) (0.0)

5.6 39.6 60.4 –

Gender Male Female Unknown

2422 3769 3

(39.1) (60.8) (0.1)

21.4 26.6 –

Race White Black Otherà Unknown

5961 189 41 3

(96.1) (3.1) (0.7) (0.1)

24.9 15.1 – –

Total

6194

(100.0)

24.3

N = Total number of office/clinic visits with a diagnosis of insomnia during 2003 that were reimbursed in a fee-for-service payment system. Rates based on number of office/clinic visits with a diagnosis of insomnia divided by the number of recipients of WV Medicaid in a fee-for-service system in each demographic group (Total N = 254,794). à Rate could not be calculated due to small sample size.  

465

3.5. Costs to Medicaid WV Medicaid reimbursed a total of $611,246 for office visits associated with a diagnosis of insomnia during 2003. On average, this amounted to $99 per office visit. This cost could have included services to treat conditions other than insomnia since insomnia was not the primary diagnosis on all visits included in the analysis. The total prescription cost for all drugs was $2.84 million. The total prescription cost to WV Medicaid for drugs approved by the FDA for insomnia equaled approximately $2.53 million, at an average cost of $57 per prescription claim. The highest cost per prescription claim for drugs with an FDAapproved indication for insomnia was for zolpidem ($68.78 per claim). Zolpidem accounted for 42.7% of all prescription claims for insomnia (including trazodone) and 84.3% of the total dollars paid by WV Medicaid for prescriptions to treat insomnia. As a drug class category, non-benzodiazepine gamma amino butyric acid agents (i.e., newer insomnia drugs zolpidem and zaleplon) accounted for 86% of prescription dollars. Claims for trazodone and benzodiazepines (mostly temazepam) accounted for the remaining 11.0% and 3.0%, respectively, of the total dollars paid by Medicaid for insomnia-related prescriptions. The dollars reimbursed by WV Medicaid for both insomnia-related office visit services and prescription medications totaled approximately $3.45 million. The average cost for insomnia-related care per recipient was approximately $182. A majority of the dollars (82.3%) was for prescription drugs to treat insomnia. Table 5 shows the distribution of total prescription dollars and average dollars per claim by agent.

4. Discussion In our study we found the rates of office visits for insomnia to be higher for women than men. These findings are similar to and supported by other studies [37,38]. As demonstrated by Ohayon’s recent review [39], however, a meaningful comparison of our findings with prevalence rates reported in the literature was rendered difficult by the fact that published prevalence estimates of insomnia in the United States vary depending upon the definition of insomnia employed and the population surveyed. Several researchers accept a 10% prevalence of chronic, clinically ‘‘non-trivial” insomnia which is very similar to the 10.2% prevalence estimate reported by Ford and Kamerow [40–44]. The lower prevalence estimate found in our study (7.4%) compared to some of the existing literature may have been due to several reasons. It is believed that while only 5% of the patients with insomnia visit their physician to specifically discuss their sleep problem, almost 26% will mention it when visiting for a different purpose, and 69% do not mention it at all [30]. Since not all patients suffering from insomnia consult a physician or actively seek treatment, our case-identification criteria involving actual diagnosis and utilization data may have led to a downward bias in estimate when compared to the Ford study which depended on less restrictive case-identification using patient self-reports. Our estimate of a 7.4% prevalence rate of insomnia, however, is similar to findings from epidemiological surveys which utilized diagnosis-linked criteria for case-identification that found prevalence rates of 5.6% and 7.0% [45,46]. These studies, among the few that provide prevalence of insomnia diagnoses according to the DSM-IV criteria, were conducted on general population samples taken from France and Italy as part of an international study on sleep disorders. Another issue to consider in our study is the use of certain prescription drugs used for the treatment of insomnia as one criterion for sample selection. Prescription medication use for insomnia has been found to be nominal (5.3%) [47], and research has shown that

466

A.N Roy, M. Smith / Sleep Medicine 11 (2010) 462–469

Table 4 Comorbidities Among Recipients of West Virginia Medicaid Identified with Insomnia (2003). Type of comorbidity

*

Unknown cause of morbidity/mortality Chest pain Lumbago Acute upper respiratory infections NOS Depressive disorder NEC () Headache Acute bronchitis Diabetes mellitus Anxiety State NOS (Not Otherwise Specified) Backache NOS Shortness of breath Acute Pharyngitis Chronic airway obstruction Benign Hypertension Esophageal Reflux Hyperlipidemia Urinary Tract Infection Recurrent Depressive Psychosis Asthma NOS Malaise and fatigue Generalized abdominal pain Recurrent Depressive Psychosis (Severe) Insomnia NEC Attention Deficit Hyperactivity Disorder

ICD-9 code

Group 1 (N = 4733) à

Group 2à (N = 8056)

Group 3à(N = 6152)

Chisquare

Pvalue

Total sampleà

799.9 786.50 724.2 465.9 311 784.0 466.0 250.00 300.00 724.5 786.05 462 496 401.1 530.81 272.4 599.0 296.32 493.90 780.79 789.07 296.33 780.52 314.01

29.8 19.4 18.9 13.2 13.1 12.8 12.6 11.6 10.9 10.5 10.5 9.9 9.8 9.4 9.3 8.3 8.2 7.5 7.3 7.1 6.4

23.6 20.1 16.2 9.4 9.2 10.7 10.1 12.0 6.6 8.3 9.7 4.9 9.8 8.5 7.2 6.2 7.9 7.6 4.7 5.4 6.8 12.9

27.4 14.4 13.1 10.0 12.6 9.2 9.7 8.9 5.9 6.3 6.7 6.4 7.0 6.2 6.1 4.6 7.1 16.8 4.8

68.7 29.6 8.5 20.5 75.9 3.2 5.1 13.8 52.1 17.8 16.9 76.9 13.7 10.3 8.5 21.9 1.6 315.9 19.5

<0.001 <0.001 0.014 <0.001 <0.001 0.207 0.078 0.001 <0.001 <0.001 <0.001 <0.001 0.001 0.006 0.014 <0.001 0.458 <0.001 <0.001

5.0 9.4

8.2

0.016

26.4 18.1 15.8 10.6 11.3 10.8 10.6 10.9 7.5 8.2 8.9 6.6 8.9 8.0 7.4 6.2 7.7 12.8 5.4 5.5 6.1 7.3 5.4 4.4

21.7 8.8

NEC, not elsewhere classified; NOS, not otherwise specified. Based on the primary ICD-9 CM code from all medical services claims reimbursed in a fee-for-service system for recipients of West Virginia Medicaid in our sample identified with insomnia. à Proportion of recipients of West Virginia Medicaid enrolled in a fee-for-service plan identified with insomnia who had a medical service claim with a primary diagnosis of the specific condition. Percents based on the total sample identified with insomnia using the three sample selection criteria. (N = 18,941). *

Table 5 Utilization and costs of prescription medications for insomnia by recipients of West Virginia Medicaid (2003). Drug

#Of Claims (%)*

FDA-Approved Medications for Insomnia Non-Benzodiazepines Zolpidem 34,740 (42.7%) Zaleplon 780 (1.0%) Benzodiazepines Estazolam Flurazepam Temazepam Triazolam

318 (0.4%) 116 (0.1%) 8107 (10.0%) 205 (0.3%)

Medication without an FDA-approved indication for Insomnia Trazodone 37,007 (45.5%)

Total costs (%)**

Average cost/claimà

8508 (43.7%) 425 (2.2%)

$2,389,317 (84.3%) $49,192 (1.7%)

$68.78 $63.07

97 (0.5%) 64 (0.3%) 2281 (11.7%) 93 (0.5%)

$6,890 (0.2%) $669 (0.0%) $72,857 (2.6%) $3933 (0.1%)

$21.67 $5.77 $8.99 $19.19

7983 (41.0%)

$312,630 (11.0%)

$8.45

#Treated (%)

 

*

Number of prescription claims for the specified product during 2003.   Number of recipients of West Virginia Medicaid enrolled in a fee-for-service plan identified with insomnia who had a prescription claim for the specified product during 2003. Percents based on the total sample identified with insomnia (N = 18,941). ** Total amount reimbursed by WV Medicaid for prescription claims for the specified product. à Average amount reimbursed equals total costs divided by number of claims.

approved hypnotics comprise less than half of prescription medications used in insomnia [20]. Non-prescription medications such as nighttime cough and cold remedies, antihistamines, alcohol, and various other popular herbal remedies are often used as sleep aids. A survey in the United States reports 13.3% of adults between 18 and 45 years of age as having used alcohol as a sleep aid in the past year [47]. Nevertheless, we did not include use of these agents in our case definition of insomnia because of the potential of misclassifying individuals with the condition since these agents are used for various purposes other than sleep. One of the most frequently used prescription drugs to treat insomnia, especially in recent years, has been the antidepressant medication trazodone. Trazodone in sub-therapeutic doses is reported to be widely used for promoting sleep [22,48], although the drug is not approved for such indication by the FDA. As such, we included prescriptions for trazodone as a sample identification

criterion in the current study. We also showed how recipients who had a claim for trazodone (in the absence of a diagnosis or claim for an FDA-approved insomnia drug) influence the prevalence of insomnia by reporting a range of rates based on groups formed by different inclusion criteria. While insomnia may exist independently (primary insomnia), quite often it occurs in association with other illnesses. Cardiovascular disease (including ischemic heart disease, nocturnal angina, congestive heart failure), chronic obstructive pulmonary disease, bronchial asthma, dyspnea, gastrointestinal disease (such as peptic ulcer disease and gastroesophageal reflux disease), and hypertension are some of the medical comorbidities specifically mentioned [42,49]. These illnesses can cause pain, immobility, and difficulty breathing during the day which may lead to difficulties in falling asleep at night. While pain may disrupt sleep, poor sleep may lower pain threshold and contribute to increased daytime perception

A.N Roy, M. Smith / Sleep Medicine 11 (2010) 462–469

of pain. As such, pain and insomnia have been suggested to go hand in hand [50,51]. The three subgroups of our sample based on the three different selection criteria (i.e., diagnosis, prescription of an FDA-approved hypnotic, and prescription of trazodone) showed consistent comorbidity diagnoses. In testing for significant differences in the proportions of recipients having various comorbidities among the three groups, some statistically significant differences were found. But some of these differences may be attributed to the large sample sizes involved rather than actual differences in the patterns of comorbid diagnoses among the groups. In light of such clinical understanding, our findings of the comorbidities found among our sample with insomnia seem logical. Chest pain and lumbago are both conditions associated with pain that might interfere with a good night’s sleep. Diagnoses of chronic airway obstruction, shortness of breath, acute bronchitis and upper respiratory tract infections may all be associated with difficulty breathing that disrupt sleep and therefore are highly likely concomitant conditions. We found that various proportions of recipients among the three groups making up our sample had medical claims with a primary diagnosis for some of these respiratory problems. Sleep disorders have also been reported in patients with diabetes. The occurrence of obesity as a common risk factor for both diabetes and obstructive sleep apnea and disruption of central nervous system regulation leading to periodic breathing (a form of respiratory dysfunction) have been suggested as possible explanations. Sleep and breathing disorders are hypothesized to be associated with type 2 diabetes [52]. Our finding of diabetes as a comorbidity therefore seems consistent with findings from the literature. Among the most frequently encountered psychiatric comorbidities in any insomnia population, depression and anxiety disorders are very common [53]. In our sample of recipients with insomnia, depression and anxiety were among the comorbid diagnoses encountered. Insomnia is closely linked to depression, and depression is believed to be the single most frequently linked comorbidity with insomnia [54]. In our sample, however, a diagnosis of depression did not account for the largest proportion of recipients by comorbid disorder. This could likely be due to a few reasons. Insomnia has historically been considered to be a symptom of depression leading to the common perception that treatment of depression would resolve insomnia in depressed individuals [43]. As such, a simultaneous diagnosis of depression and insomnia is not likely to be very common. Only recently have epidemiologic data suggested that insomnia may antedate or occur simultaneously with depression; patients responding to antidepressant therapy may have residual insomnia and may require specific interventions to treat it [43,54]. Indeed the recent NIH Consensus Statement emphasizes that insomnia in conjunction with other illnesses not be referred to as ‘‘secondary” to avoid under-treatment of co-occurring insomnia. It is also pertinent to note that we selected our sample based on the presence of a diagnosis of insomnia recorded in any of nine diagnosis fields for each medical service claim or through a prescription claim for a drug approved by the FDA for insomnia or trazodone, and analysis of comorbidities was based on an analysis of primary diagnoses. As such, conditions that were secondary or tertiary diagnoses (which may very well have been the case with depression) are not reflected in the list of comorbidities reported here and may explain why we did not find depression as the most frequently occurring comorbid diagnosis among our study sample. In addition to the comorbid disorders causing insomnia problems themselves, side effects due to drug treatment of the disorders also may contribute to insomnia. A variety of drugs are reported to cause increased wakefulness and poor quality sleep.

467

These include corticosteroids, beta-blockers, calcium-channel blockers, diuretics, and bronchodilators [49]. It is understandable, therefore, that various treatments for hypertension and obstructive airway diseases (where prescriptions for the above mentioned drugs are very common) may be associated with insomnia. In regard to prescription use among our sample, trazodone accounted for the most (45.5%) claims for insomnia followed by zolpidem (42.7%). Zolpidem, however, constituted the highest proportion (84.3%) of drug expenditure. Zolpidem also accounted for the most prescription claims among drugs with an FDA-approved indication for insomnia followed by temazepam and zaleplon. These results appear to be in fair corroboration with a study of national drug utilization in insomnia, where newer hypnotics (zolpidem or zaleplon) and temazepam were the most prescribed FDAapproved hypnotics (after trazodone) for insomnia [55]. Another study reporting the utilization of pharmacotherapies for insomnia in 2002 also found zolpidem to be the most commonly used drug (after trazodone) for treating insomnia. A steady decrease in the use of older benzodiazepines following the introduction of the newer non-benzodiazepine hypnotics over time has also been reported [26] and may be reflected in our data in that zolpidem was found to be the most extensively used FDA-approved hypnotic in our study sample. Nevertheless, another study using data from the National Ambulatory Medical Care Survey (NAMCS) found benzodiazepines to be more frequently prescribed than the newer non-benzodiazepines for insomnia (e.g., zolpidem) [36]. This could likely be due to two reasons. First, ours was a study of a sample from the Medicaid population, whereas the NAMCS data reports on medical care services and prescription utilization of a nationally representative sample from non-institutionalized office-based settings primarily involved in direct patient care. Patients visiting such physicians may have a variety of insurance arrangements (including public, private, and managed care). Out-of-pocket payments may have influenced the choice of drugs in such settings. It is likely that since many insurance plans do not cover prescription drugs (or have higher co-payments for newer drugs), less expensive benzodiazepines were more frequently used in the NAMCS population. Such cost considerations are less likely to be an issue with our sample since the nominal co-pay that is part of the drug benefit of WV Medicaid is not mandatory at time of dispensing if the recipient is unable to pay it, and most prescription drugs are covered by Medicaid. As such, zolpidem may have been more frequently prescribed than the older benzodiazepines in our sample of a Medicaid population.

5. Limitations There are two major limitations of this study pertaining to our dataset. First, the WV Medicaid fee-for-service data analyzed in this study includes claims for both severely disabled and non-disabled recipients. This could have resulted in an upwardly biased estimate of the true extent of healthcare utilization and costs attributed to this disease among the WV Medicaid population. Second, encounter data for recipients enrolled in Medicaid managed care plans were not available for this study, so the medical services use that occurred in managed care organizations were not accounted for in this study. Thus, the study results cannot be generalized to the true WV Medicaid population. In the presence of the above mentioned limitations, readers are cautioned that the results of this study must be interpreted carefully. Since our study used an administrative claims database, certain general limitations of secondary data should also be considered. Such limitations include lack of availability of clinical information and no option of re-collecting missing data or rectifying data entry

468

A.N Roy, M. Smith / Sleep Medicine 11 (2010) 462–469

and coding errors. In identifying part of our sample based on prescription drug use, we assumed insomnia was the primary reason for such prescription. Using a retrospective cross-sectional design made it difficult to conclude whether all prescriptions for insomnia were associated with a diagnosis since a diagnosis may have occurred before our span of data for this study. Additionally, Medicare data for dually eligible recipients over 64 years of age were not available in this study. Therefore, rates for those over 64 years of age have not been reported in this study. Our inclusion criteria used to identify our study sample (diagnosis or claim for a drug approved by the FDA for insomnia or prescription claim for trazodone) may not have captured all recipients with insomnia, and therefore our reported rates may underestimate the true rate of insomnia among enrollees of WV Medicaid in the fee-for-service population. For example, our results do not account for recipients who used pharmacotherapy treatments for chronic disorders that also may have provided therapeutic benefit for co-occurring insomnia (e.g., use of low dose sedating antidepressants in treating depression). But we feel that by limiting the inclusion of such possible cases of insomnia, we decreased the risk of misclassifying recipients with insomnia and inflating our prevalence rates. Our inclusion criteria enabled us to identify probable cases of insomnia and produce conservative prevalence estimates of the condition. 6. Conclusions We found an overall prevalence rate of insomnia to be 74.3 per 1000 recipients of WV Medicaid in a fee-for-service system. Rates of insomnia and related office visit use varied by demographic groups. Recipients 45–64 years of age, females, and whites had the highest prevalence of insomnia and also the highest office visit use for insomnia by age, gender, and race groups. Among prescription claims for medications used to treat insomnia, trazodone and zolpidem accounted for the greatest proportions of claims (45.5% and 42.7%, respectively). Trazodone, however, accounted for only 11% of the prescription costs for insomnia-related drugs, while zolpidem accounted for 84.3%. Our study provides baseline estimates of the distribution of insomnia among the WV Medicaid fee-forservice population. Our results can be the basis for ongoing surveillance of this condition, especially as the definition of insomnia continues to evolve. Acknowledgements The authors would like to thank and acknowledge the following for their assistance with this study: Marsha Morris, JD, RN, Peggy King, RPh, Vicki Cunningham, RPh, of the West Virginia Department of Health and Human Resources Bureau for Medical Services; Steve Small, MS, RPh, of the West Virginia Rational Drug Therapy Program; and Nancy Atkins, MSN, RNC, NP, of WellPoint, Inc. Partial support for this project was provided by grant number 1 P20 HS015390 from the Agency for Healthcare Research and Quality. The primary author, Anuja Roy, Ph.D., is currently affiliated with Takeda Pharmaceuticals North America, Deerfield, Illinois. This study was conducted while she was a graduate student at West Virginia University School of Pharmacy. Michael Smith, Ph.D., was a faculty member at West Virginia University School of Pharmacy at the time this study was conducted. References [1] Doghramji PP. Recognizing sleep disorders in a primary care setting. J Clin Psychiatry 2004;65(Suppl. 16):23–6. [2] Drake CL, Roehrs T, Roth T. Insomnia causes, consequences, and therapeutics: an overview. Depress Anxiety 2003;18:163–76.

[3] Leger D. Public health and insomnia: economic impact. Sleep 2000;23(3):S69–76. [4] Ancoli-Israel S, Roth T. Characteristics of insomnia in the United States: results of the 1991 National Sleep Foundation survey. I. Sleep 1999;22(Suppl. 2):S347–53. [5] Ford DE, Kamerow DB. Epidemiologic study of sleep disturbances and psychiatric disorders: an opportunity for prevention? JAMA 1989;262:1479–84. [6] Sateia MJ, Doghramji K, Hauri PJ, Morin CM. Evaluation of chronic insomnia: an American Academy of Sleep Medicine review. Sleep 2000;23:243–308. [7] National Institutes of Health. State-of-the-science conference statement on manifestations and management of chronic insomnia in adults; June 13–15, 2005. Sleep 2005;28:1049–57. [8] Doghramji K. The epidemiology and diagnosis of insomnia. Am J Manag Care 2006;12:S214–20. [9] Morin CM. The nature of insomnia and the need to refine our diagnostic criteria. Psychosom Med 2000;62:483–5. [10] Ohayon MM. Epidemiology of insomnia: what we know and what we still need to learn. Sleep Med Rev 2002;6(2):97–111. [11] Roth T, Roehrs T, Costa e Silva JA, et al. Public health and insomnia: consensus statement regarding its status and needs for future actions. Sleep 1999;22(Suppl. 3):S417–20. [12] Roth T, Roehrs T. Insomnia: epidemiology, characteristics, and consequences. Clin Cornerstone 2003;5(3):5–15. [13] Culpepper L. Secondary insomnia in the primary care setting: review of diagnosis, treatment, and management. Curr Med Res Opin 2006;22(7):1257–68. [14] Walsh JK. Pharmacologic management of insomnia. J Clin Psychiatry 2004;65(Suppl. 16):41–5. [15] National Center on Sleep Disorders Research Working Group on Insomnia, NHLBI. Insomnia assessment and management in primary care. Sleep 1999;22(Suppl. 2):S402–12. [16] Krystal AD, Roth T. Definitions, measurements, and management in insomnia. J Clin Psychiatry 2004;65(Suppl. 8):5–7. [17] Leigh P. Employee and job attributes and predictors of absenteeism in a national sample of workers: the importance of health and dangerous working conditions. Soc Sci Med 1991;33:127–31. [18] Weyerer S, Dilling H. Prevalence and treatment of insomnia in the community: results from the upper bavarian field study. Sleep 1991;14:392–8. [19] Balter MB, Uhlenhut EH. New epidemiologic findings about insomnia and its treatment. J Clin Psychiatry 1997;171:382–8. [20] Walsh JK, Engelhardt C. The direct economic costs of insomnia in the United States for 1995. Sleep 1999;22(Suppl. 2):S386–93. [21] Stoller MK. Economic effects of insomnia. Clin Ther 1994;16(5):873–97. [22] Benca RM. Diagnosis and treatment of chronic insomnia: a review. Psychiatry Serv 2005;56(3):332–43. [23] Winkelman J, Pies R. Current patterns and future directions in the treatment of insomnia. Ann Clin Psychiatry 2005;17(1):31–40. [24] Rosenberg RP. Sleep maintenance insomnia: strengths and weaknesses of current pharmacologic therapies. Ann Clin Psychiatry 2006;18(1):49–56. [25] Roth T. Prevalence, associated risks, and treatment patterns of insomnia. J Clin Psychiatry 2005;66(Suppl. 9):10–3. [26] Morin KA. Strategies for treating chronic insomnia. Am J Manag Care 2006;12:S230–45. [27] Bixler EO, Kales A, Soldatos CR, Kales JD, Healy S. Prevalence of sleep disorders in the Los Angeles metropolitan area. Am J Psychiatry 1979;136(10):1257–62. [28] Ohayon MM, Caulet M, Guilleminault C. Complaints about nocturnal sleep: how a general population perceives its sleep and how this relates to the complaint of insomnia. Sleep 1997;20:715–23. [29] Newman AB, Enright PL, Manolio TA, Haponik EF, Wahl PW. Sleep disturbance, psychosocial correlates, and cardiovascular disease in 5201 older adults: the cardiovascular health study. J Am Geriatr Soc 1997;45(1):1–7. [30] Ancoli-Israel S, Roth T. Characteristics of insomnia in the United States: results of the 1991 National Sleep Foundation Survey. I. Sleep 1999;22(Suppl. 2):S347–53. [31] Kim K, Uchiyama M, Okawa M, Liu X, Ogihara R. An epidemiological study of insomnia among the Japanese general population. Sleep 2000;23:41–7. [32] Nadolski N. Getting a good night’s sleep: diagnosing and treating insomnia. Plast Surg Nurs 2005;25(4):167–73. [33] Pete Fullerton DS. The economic impact of insomnia in managed care: a clearer picture emerges. Am J Manag Care 2006;12:S246–52. [34] McCall WV, Fleischer Jr AB, Feldman SR. Diagnostic codes associated with hypnotic medications during outpatient physician-patient encounters in the United States from 1990–1998. Sleep 2002;25(2):221–3. [35] Hays RD, Stewart AL. Sleep measures. In: Stewart AL, Ware JE, editors. Measuring functioning and well-being. The medical outcomes study approach. Durham and London: Duke University Press; 1992. p. 235–59. [36] Balkrishnan R, Rasu RS, Rajagopalan R. Physician and patient determinants of pharmacologic treatment of sleep difficulties in outpatient settings in the United States. Sleep 2005;28(6):715–9. [37] Radecki SE, Brunton SA. Management of insomnia in office-based practice. National prevalence and therapeutic patterns. Arch Fam Med 1993;2(11):1129–34. [38] Morlock RJ, Tan M, Mitchell DY. Patient characteristics and patterns of drug use for sleep complaints in the United States: analysis of National Ambulatory Medical Survey data, 1997–2002. Clin Ther 2006;28(7):1044–53.

A.N Roy, M. Smith / Sleep Medicine 11 (2010) 462–469 [39] Ohayon M. Epidemiology of insomnia: what we know and what we still need to learn. Sleep Med Rev 2002;6(2):97–111. [40] Walsh JK. Clinical and socioeconomic correlates of insomnia. J Clin Psychiatry 2004;65(Suppl. 8):13–9. [41] Roth T. New developments for treating sleep disorders. J Clin Psychiatry 2001;62(Suppl. 10):3–4. [42] Ancoli-Israel S. The impact and prevalence of chronic insomnia and other sleep disturbances associated with chronic illness. Am J Manag Care 2006;12:S221–9. [43] Drake CL, Roehrs T, Roth T. Insomnia causes, consequences, and therapeutics: an overview. Depress Anxiety 2003;18:163–76. [44] Ford DE, Kamerow DB. Epidemiologic study of sleep disturbances and psychiatric disorders: an opportunity for prevention? JAMA 1989;262:1479–84. [45] Ohayon MM. Prevalence of DSM-IV diagnostic criteria of insomnia: distinguishing between insomnia reltade to mental disorders from sleep disorders. J Psychiatry Res 1997;31:333–46. [46] Ohayon M, Smirne S. Prevalence and consequences of insomnia disorders. Sleep Med 2002;3(2):115–20. [47] Johnson EO, Roehrs T, Roth T, Breslau N. Epidemiology of alcohol and medication as aids to sleep in early adulthood. Sleep 1998;21(2):178–86.

469

[48] Walsh JK, Schweitzer PK. Ten-year trends in the pharmacological treatment of insomnia. Sleep 1999;22:371–5. [49] Thase ME. Correlates and consequences of chronic insomnia. Gen Hosp Psychiatry 2005;27:100–12. [50] Benca RM, Ancoli-Israel S, Moldofsky H. Special considerations in insomnia diagnosis and management: depressed, elderly, and chronic pain populations. J Clin Psychiatry 2004;65(Suppl. 8):26–35. [51] Chokroverty S. Diagnosis and treatment of sleep disorders caused by comorbid disease. Neurology 2000;54(Suppl. 1):S8–S15. [52] Ancoli-Israel S. The impact and prevalence of chronic insomnia and other sleep disturbances associated with chronic illness. Am J Manag Care 2006;12(Suppl. 8):S221–9. [53] Hajak G. Insomnia in primary care. Sleep 2000;23(Suppl. 3):S54–63. [54] Kupfer DJ. Pathophysiology and management of insomnia during depression. Ann Clin Psychiatry 1999;11(4):267–76. [55] Morlock RJ, Tan M, Mitchell DY. Atient characteristics and patterns of drug use for sleep complaints in the United States: analysis of national ambulatory medical survey data, 1997–2002. Clin Ther 2006;28(7):1044–53.