Abstract
Study Objectives:
The primary objective was to describe trends in the 2-year limited duration prevalence of narcolepsy from 2013–2016 in a large insured population with claims activity. Secondary objectives were to assess the prevalence of other sleep disorders and the frequency of diagnostic sleep testing.
Methods:
Nationwide medical/prescription claims (Symphony Health) were analyzed to estimate the annual prevalence per 100,000 persons of narcolepsy and other sleep disorders (obstructive sleep apnea, idiopathic hypersomnia, rapid eye movement sleep behavior disorder, periodic limb movement disorder) and the frequency of diagnostic sleep testing. Prevalence was adjusted to the age/sex distribution of the 2016 US census estimates.
Results:
The prevalence of narcolepsy per 100,000 persons increased 14% from 38.9 in 2013 to 44.3 in 2016. Obstructive sleep apnea prevalence increased 41% over the study period from 2,429 to 3,420 per 100,000. Large increases in prevalence were also seen for idiopathic hypersomnia (32%), periodic limb movement disorder (30%), and rapid eye movement sleep behavior disorder (64%). For each sleep disorder, prevalence was higher for those with commercial insurance versus Medicare/Medicaid, and markedly lower prevalence was observed for the Northeast compared with the Midwest, South, and Western US regions. The frequency of multiple sleep latency/maintenance of wakefulness testing declined by 20%, and polysomnography declined by 15%. Conversely, home sleep apnea testing increased by 117%.
Conclusions:
The prevalence of narcolepsy, obstructive sleep apnea, and the other sleep disorders increased appreciably over the 2013–2016 period. It remains to be determined whether the trends seen in our analyses are due to increased incidence or increased awareness of these conditions.
Citation:
Acquavella J, Mehra R, Bron M, Suomi JM-H, Hess GP. Prevalence of narcolepsy, other sleep disorders, and diagnostic tests from 2013–2016: insured patients actively seeking care. J Clin Sleep Med. 2020;16(8):1255–1263.
Keywords: idiopathic hypersomnia; periodic limb movement disorder; prevalence; narcolepsy; obstructive sleep apnea; REM sleep behavior disorder,; sleep testing
BRIEF SUMMARY
Current Knowledge/Study Rationale: The prevalence of narcolepsy and other sleep disorders has not been assessed recently on a nationwide basis in the United States. We conducted a descriptive analysis of prevalence and the use of sleep testing during the interval 2013–2016.
Study Impact: The results of this study provide important information for clinicians, payers, and patients. It is clear from our analyses that the prevalence of narcolepsy, obstructive sleep apnea, and the other sleep disorders is increasing, possibly indicative of enhanced recognition. Moreover, commercially insured persons are being diagnosed and treated more frequently than those on Medicare/Medicaid, and there are large regional differences in the frequency of testing and the diagnosis of these sleep disorders.
INTRODUCTION
Narcolepsy, a sleep disorder associated with morbidity and decrements in quality of life, is characterized by excessive daytime sleepiness, cataplexy, sleep paralysis, and hypnagogic hallucinations.1 It has an average time from onset of symptoms to diagnosis of 10 years or more,2 and symptoms usually appear in the mid-to-late teens or early adulthood. Patients may initially delay seeking medical help, and once under medical evaluation, may be assigned a variety of incorrect sleep disorder or other diagnoses. Expert opinion is that narcolepsy may be undiagnosed in about half of all affected patients, but that medical awareness of this condition is increasing1.
There are no recent US nationwide assessments of narcolepsy prevalence, although an estimate has been published for Minnesota for 1985 (56/100,000 persons).3 Longstreth et al4 reviewed multinational published studies, some based in the United States, and found a range of prevalences for narcolepsy with cataplexy (type 1) between 25 and 50 per 100,000 persons. As expected, prevalence was higher for narcolepsy without cataplexy (type 2). The published studies used different methodologies, most based primarily on questionnaire and interview data, contributing to the wide variation in prevalence estimates. Moreover, although narcolepsy has been reported to be almost twice as prevalent in men than in women3, other aspects of the descriptive epidemiology, beyond age and sex, remain to be established.
The purpose of this study was to estimate annual 2-year limited-duration narcolepsy prevalence in a US nationwide insured population of more than 60 million participants during the time period 2013–2016. In addition, for context, we estimated the 2-year limited-duration prevalence of other common sleep disorders during the same time period, including obstructive sleep apnea (OSA), idiopathic hypersomnia (IH), rapid eye movement sleep behavior disorder (RBD), and periodic limb movement disorder (PLMD). Compared with complete prevalence (all past and current cases who are still alive in the eligible population), this limited-duration prevalence focuses on patients who are diagnosed and actively using medical and pharmacy services. We also estimated the frequency of diagnostic sleep testing and ascertained the specialty of the providers ordering the tests. The analyses of data for such a large population were intended to enable the evaluation of prevalence with high statistical precision according to demographic factors, geographic areas, insurance type, and provider specialty—all important considerations for understanding the nationwide prevalence of narcolepsy and other sleep disorders.
METHODS
Design, data source, search strategy
We conducted a retrospective cohort study using Symphony Health data for the period 1 January 2013 through 31 December 2016. These data include both medical encounter claims (including diagnostic, therapeutic, and procedure information) and outpatient pharmacy claims. Medical encounters include practitioner care in both ambulatory and hospital settings for approximately 150 million unique patients on an annual basis, including patients from all 50 states, patients with various insurance types, as well as uninsured pharmacy patients. Among the types of insurance represented in the database are Medicare, a US federal program that generally covers patients 65 years and older, Medicaid, a state program(s) that covers low income citizens, and a wide variety of commercial insurers that provide coverage for all others. For Medicare patients, Medicare Advantage enrollees and enrollees with Medicare supplemental plans are included; however, enrollees with only Medicare Part A and B are not captured. Claims are linked longitudinally at the patient level through the generation and matching of a Health Insurance Portability and Accountability Act–compliant, unique, synthetic identifier that is permanent for each patient.
Cohort and case definitions
The prevalence cohort (denominator) included patients who met 4 eligibility criteria: (1) 1 or more private practitioner claim and 1 or more pharmacy claim at least 12 months before the start of any year’s prevalence calculation, (2) 1 or more private practitioner claim and 1 or more pharmacy claim at least 12 months after the start of the year of the prevalence calculation, (3) sex specified as male or female, and (4) a valid date of birth. Cases (numerators) were eligible patients in the prevalence cohort with 2 or more claims within 6 months with a diagnosis of narcolepsy or 1 narcolepsy claim and a multiple sleep latency test (MSLT) or maintenance of wakefulness test (MWT) within the prior 6 months. Both claims had to occur before the end of the previous or current calendar year for the specific annual prevalence calculation. This case definition provides for estimating 2-year limited-duration prevalence, not the complete prevalence of the disease (namely, all past and current cases who are still alive) and is best thought of as those under medical care for narcolepsy within 2-year time windows during the study period. For other sleep disorders, cases were those with 2 diagnostic claims within a 6-month period.
We identified relevant claims by searching the database using the International Classification of Diseases Ninth or Tenth Revision codes for diagnosis, and the Healthcare Common Procedure Coding System codes for sleep-disorder testing and treatments. Sleep tests included in our analyses were: MSLT/MWT, home sleep apnea test (HSAT), and polysomnography (PSG).
Data extraction, synthesis of results, and analysis
Prevalence was calculated by age and sex strata by dividing the total number of new and existing cases in each stratum in the calendar year by the total number of persons in the stratum in the calendar year, and multiplying by 100,000. Age categories included 14 years or younger, 15–17 years, 18–24 years, 25–34 years, 35–44 years, 45–54 years, 55–64 years, and 65 years or older. Overall prevalence was age- and sex-adjusted by summing the products of age- and sex-specific prevalences and the 2016 US Census5 proportions in each age and sex stratum. This weighting (or age-adjustment) of prevalence ensures that trends in annual prevalences are not influenced by changes in the age or sex distribution of the database population. Prevalences and their 95% confidence intervals are expressed as percentages or numbers of cases per 100,000 persons. The 95% confidence interval was calculated as follows: . The SE (standard error) was estimated to be the square root of [p (1 – p)/n], where p equals the prevalence and n is the number of cases in the prevalence calculation.6
Patient subgroups based on geographic region and payer type were analyzed for each calendar year. Geographic region was assessed according to US Census categories: Northeast, South, Midwest, and West. The predominant payer type was categorized hierarchically in rank order of Medicare, Medicaid, and commercial (third-party, nongovernmental payers). Because Medicare participants in the database population are subscribers to Medicare Advantage plans, they are considered to be a healthier subset of the overall Medicare population.
The percentage breakdown by provider specialty was assessed for each sleep test type. Specialty was based on the practitioners’ self-reported health-care provider taxonomy (specialty classification) when applying for a National Provider ID with the Centers for Medicare and Medicaid Services. For providers who reported more than 1 specialty, specialties were assigned hierarchically, in the following order: pediatric, neurology, sleep medicine, pulmonology, general practitioner, and other.
Data sets were merged and analyzed with the use of SAS (Statistical Analysis Software) version 9.4 (SAS Institute, Inc, Cary, NC).
RESULTS
Narcolepsy
Table 1 summarizes characteristics of the individuals included in our analyses by calendar year. The size of the study population ranged from approximately 62.9 million individuals in 2013 to 67.5 million individuals in 2015. The sex distribution of the population was approximately 40% male and 60% female for all calendar years. The age distribution was steady over the study period, with approximately 15% of the population younger than 18 years of age, 57% aged 18–64 years, and 28% aged 65 years and older. Approximately 63–65% of individuals had commercial insurance that included supplemental plans; 22% were Medicare subscribers, and those on Medicaid constituted between 11% and 14% of the study population. Geographically, 40% of the study population resided in the South, 13% in the West, and 24% each in the Northeast and Midwest.
Table 1.
Eligible database study population from 2013–2016.
2013 | 2014 | 2015 | 2016 | |||||
---|---|---|---|---|---|---|---|---|
Eligible Database Population | n | % | n | % | n | % | n | % |
Total | 62,895,565 | 66,446,545 | 67,519,490 | 66,060,231 | ||||
Males | 24,835,605 | 39 | 26,348,132 | 40 | 26,737,490 | 40 | 26,126,600 | 40 |
Females | 38,059,960 | 61 | 40,098,413 | 60 | 40,782,000 | 60 | 39,933,631 | 60 |
Age | ||||||||
≤14 years | 7,846,297 | 12 | 8,057,169 | 12 | 7,912,383 | 12 | 7,478,977 | 11 |
15–17 years | 1,686,640 | 3 | 1,777,561 | 3 | 1,762,905 | 3 | 1,678,652 | 3 |
18–24 years | 4,061,100 | 6 | 4,234,132 | 6 | 4,265,067 | 6 | 4,117,463 | 6 |
25–34 years | 5,564,915 | 9 | 5,862,189 | 9 | 6,115,502 | 9 | 6,026,099 | 9 |
35–44 years | 6,763,262 | 11 | 6,976,581 | 10 | 7,101,784 | 11 | 6,980,485 | 11 |
45–54 years | 9,077,937 | 14 | 9,441,987 | 14 | 9,571,509 | 14 | 9,203,704 | 14 |
55–64 years | 10,691,462 | 17 | 11,428,472 | 17 | 11,847,662 | 18 | 11,708,051 | 18 |
≥65 years | 17,203,952 | 27 | 18,668,454 | 28 | 18,942,678 | 28 | 18,866,800 | 29 |
Insurance type | ||||||||
Commercial | 41,169,585 | 65 | 42,604,426 | 64 | 42,618,908 | 63 | 41,388,022 | 63 |
Medicare | 13,768,889 | 22 | 14,906,183 | 22 | 14,990,366 | 22 | 14,637,684 | 22 |
Medicaid | 6,974,455 | 11 | 7,996,607 | 12 | 9,004,328 | 13 | 9,159,884 | 14 |
All other | 982,636 | 2 | 939,329 | 1 | 905,888 | 1 | 874,641 | 1 |
Census region | ||||||||
Northeast | 14,997,269 | 24 | 16,105,104 | 24 | 16,093,047 | 24 | 15,351,246 | 23 |
Midwest | 14,567,318 | 23 | 15,303,816 | 23 | 15,964,741 | 24 | 16,075,636 | 24 |
South | 25,158,186 | 40 | 26,477,731 | 40 | 26,979,188 | 40 | 26,304,636 | 40 |
West | 8,172,792 | 13 | 8,559,894 | 13 | 8,482,514 | 13 | 8,328,713 | 13 |
The age- and sex-adjusted prevalence of narcolepsy increased by 13.9% over the study period, from 38.9 per 100,000 persons in 2013 to 44.3 per 100,000 in 2016 (Table 2). The increase in age-adjusted prevalence was somewhat greater for women than for men (16% vs 11%) and, contrary to the higher prevalence for males than females reported in the literature,3 the male to female prevalence ratio was less than 1.0 in each calendar year (0.96, 0.96, 0.96, and 0.92, respectively). Age-specific prevalences peaked in the 18–24-, 25–34-, and 35–44-year age groups and declined thereafter (Figure 1). Prevalence increased in all age groups during the period 2013–2016, but most appreciably in the <14-, 35–44-, and 45–54-year age groups (18.9%, 21.1%, and 22.9%, respectively; Figure 1). Multiplying the 2016 age- and sex-specific prevalences for the eligible database population by the 2016 age- and sex-specific population estimates from the US census produced a national estimate for 2016 of approximately 142,000 prevalent cases.
Table 2.
Prevalence of narcolepsy (95% confidence interval) per 100,000 persons: 2013–2016.
2013 | 2014 | 2015 | 2016 | % Change 2013–2016 | |
---|---|---|---|---|---|
Overall* | 38.9 (38.6, 39.2) | 42.4 (42.1, 42.6) | 43.6 (43.4, 43.9) | 44.3 (44.0, 44.6) | 13.9 |
Male** | 38.1 (37.3, 39.0) | 41.5 (40.7, 42.4) | 42.8 (41.9, 43.6) | 42.5 (41.6, 43.3) | 11.4 |
Female** | 39.6 (39.0, 40.2) | 43.1 (42.4, 43.7) | 44.4 (43.7, 45.0) | 46.0 (45.3, 46.7) | 16.2 |
Insurance | |||||
Commercial* | 39.2 (38.6, 39.8) | 43.3 (42.6, 43.9) | 45.2 (44.5, 45.8) | 46.0 (45.3, 46.6) | 17.3 |
Medicare* | 57.7 (55.9, 59.5) | 56.8 (55.1, 58.5) | 55.8 (54.2, 57.5) | 53.9 (52.3, 55.5) | −6.6 |
Medicaid* | 33.9 (32.4, 35.5) | 35.4 (34.0, 36.8) | 37.0 (35.7, 38.3) | 37.6 (36.3, 38.8) | 10.6 |
Region | |||||
Northeast* | 26.8 (26.0, 27.7) | 28.5 (27.7, 29.4) | 28.4 (27.6, 29.3) | 28.2 (27.3, 29.1) | 5.2 |
Midwest* | 52.3 (51.2, 53.5) | 56.8 (55.6, 58.0) | 57.5 (56.3, 58.7) | 58.2 (57.0, 59.3) | 11.3 |
South* | 41.2 (40.4, 42.0) | 45.5 (44.7, 46.4) | 47.3 (46.4, 48.1) | 48.9 (48.0, 49.7) | 18.7 |
West* | 30.2 (29.0, 31.4) | 33.2 (31.9, 34.4) | 35.4 (34.1, 36.7) | 33.7 (32.5, 35.0) | 11.6 |
Age- and sex-adjusted to the 2016 census. **Age-adjusted to the 2016 census.
Figure 1. Annual prevalence of narcolepsy by age group, 2013–2016.
Percentages are change in prevalence from 2013 to 2016.
Analyses by type of insurance showed higher age- and sex-adjusted prevalence for Medicare beneficiaries than for commercial insurance or Medicaid subscribers, with marked disparities in trends during the study period. Adjusted prevalence declined by 6.6% for Medicare subscribers, but increased by 10.6% for those on Medicaid and 17.3% for those on commercial insurance. To some extent, because of the skewed age distributions for the Medicare (∼84% ≥65 years of age) and Medicaid (95%; <65 years of age) populations compared with the US Census, the age adjustment masks age-specific trends within these populations. For Medicare beneficiaries who were 65 years of age or older, narcolepsy prevalence tended to be lower than for their age peers with commercial insurance (eg, 19.8 vs 29.6 per 100,000, respectively, in 2016) and prevalence increased less than for the commercially insured (0.7% vs 6.7%). However, among the 16% of Medicare beneficiaries younger than 65 years of age, presumably those with disabilities, narcolepsy prevalence was appreciably higher than for their age peers with commercial insurance (eg, 76.4 vs 51.0, respectively, in 2016). In terms of growth, the prevalence for those younger than 65 years old with Medicare only increased 0.2% over the study period compared with 20.8% for those who were commercially insured. The age-specific data for Medicaid beneficiaries tended to show lower prevalence than for their age peers with commercial insurance. This difference in prevalence was more marked for those younger than 65 years of age (eg, 35.5 vs 56.0, respectively, in 2016) than for those 65 years of age or older (eg, 28.1 vs 29.6, respectively, in 2016).
Large disparities in prevalence were also seen by geographic region. For example, age- and sex-adjusted prevalence for 2016 was much higher in the Midwest and South (58.2 and 48.9 per 100,000 persons, respectively) than for the West and Northeast (33.7 and 28.2, respectively). Prevalence increased in all geographic regions over the study period, with appreciably greater increases in the South (18.7%), Midwest (11.3%), and West (11.6%) compared with the Northeast (5.2%).
Other sleep disorders
Among the other sleep disorders, OSA was by far the most prevalent, ranging from 2,429 per 100,000 in 2013 to 3,420 per 100,000 in 2016 (Table 3, Figure 2). Notably, the prevalence increased 40.8% over the study period. Prevalence was more than twice as high in men relative to women, and the relative increase over the study period was similar for both sexes. Persons with commercial insurance had appreciably higher prevalence (4,037 per 100,000 in 2016) than those on Medicare (2,780 per 100,000) and Medicaid (2,420 per 100,000), and the increase in prevalence over the study period was also greater for those with commercial insurance (44%) compared with those on Medicare (31%) and especially those on Medicaid (13%). Geographically, prevalence was highest in the Western region (4,710 per 100,000 in 2016) and lowest in the Northeast (2,881 per 100,000), and the trend of increasing prevalence was strongest in the West (50%).
Table 3.
Prevalence of OSA, PLMD, IH, and RBD (95% confidence interval) per 100,000 persons: 2013–2016.
2013 | 2014 | 2015 | 2016 | % Change 2013–2016 | |
---|---|---|---|---|---|
OSA | |||||
Overall* | 2,429.5 (2,427.5, 2,431.5) | 2,649.7 (2,647.7, 2,651.7) | 2,965.5 (2,963.5, 2,967.6) | 3,420.3 (3,418.1, 3,422.5) | 40.8 |
Male** | 3,379.3 (3,372.5, 3,386.0) | 3,684.8 (3,678.1, 3,691.6) | 4,122.8 (4,115.7, 4,129.9) | 4,747.4 (4,739.7, 4,755.0) | 40.5 |
Female** | 1,534.5 (1,531.0, 1,538.0) | 1,674.5 (1,670.9, 1,678.1) | 1,876.2 (1,872.4, 1,879.9) | 2,170.9 (2,166.8, 2,174.9) | 41.5 |
By insurance | |||||
Commercial* | 2,798.7 (2,793.8, 2,803.6) | 3,096.7 (3,091.7, 3,101.7) | 3,496.0 (3,490.7, 3,501.3) | 4,037.1 (4,031.3, 4,042.9) | 44.2 |
Medicare* | 2,115.6 (2,108.4, 2,122.7) | 2,199.0 (2,192.1, 2,205.9) | 2,416.4 (2,409.4, 2,423.5) | 2,780.0 (2,772.4, 2,787.5) | 31.4 |
Medicaid* | 2,141.4 (2,127.2, 2,155.6) | 2,192.6 (2,179.5, 2,205.7) | 2,272.2 (2,260.4, 2,284.1) | 2,419.6 (2,407.8, 2,431.3) | 13.0 |
By region | |||||
Northeast* | 2,049.1 (2,042.2, 2,056.1) | 2,194.0 (2,187.1, 2,200.9) | 2,472.5 (2,465.2, 2,479.8) | 2,880.6 (2,872.6, 2,888.7) | 40.6 |
Midwest* | 2,534.8 (2,527.0, 2,542.5) | 2,759.3 (2,751.5, 2,767.2) | 3,046.0 (3,038.0, 3,054.0) | 3,483.6 (3,475.1, 3,492.1) | 37.4 |
South* | 2,366.3 (2,360.6, 2,372.1) | 2,561.6 (2,555.9, 2,567.4) | 2,834.1 (2,828.1, 2,840.1) | 3,287.9 (3,281.4, 3,294.4) | 38.9 |
West* | 3,135.3 (3,123.9, 3,146.6) | 3,587.5 (3,575.6, 3,599.3) | 4,159.8 (4,147.0, 4,172.5) | 4,709.9 (4,696.2, 4,723.5) | 50.2 |
PLMD | |||||
Overall* | 33.4 (33.2, 33.7) | 38.6 (38.3, 38.8) | 41.6 (41.4, 41.9) | 43.4 (43.1, 43.6) | 29.6 |
Male** | 42.1 (41.4, 42.8) | 48.8 (48.1, 49.6) | 53.2 (52.4, 54.0) | 55.3 (54.5, 56.1) | 31.3 |
Female** | 25.5 (25.1, 26.0) | 29.1 (28.7, 29.6) | 31.0 (30.5, 31.4) | 32.3 (31.8, 32.8) | 26.7 |
By insurance | |||||
Commercial* | 36.7 (36.2, 37.3) | 42.4 (41.8, 42.9) | 46.9 (46.3, 47.5) | 49.6 (48.9, 50.2) | 34.9 |
Medicare* | 30.7 (30.0, 31.4) | 37.1 (36.3, 37.8) | 36.5 (35.7, 37.3) | 35.3 (34.6, 36.1) | 15.2 |
Medicaid* | 22.9 (21.4, 24.3) | 26.6 (25.2, 28.1) | 29.8 (28.5, 31.2) | 33.3 (31.9, 34.7) | 45.6 |
By region | |||||
Northeast* | 20.0 (19.4, 20.7) | 20.2 (19.5, 20.8) | 20.7 (20.0, 21.3) | 21.5 (20.9, 22.2) | 7.5 |
Midwest* | 42.2 (41.2, 43.2) | 47.3 (46.3, 48.3) | 48.4 (47.4, 49.4) | 52.3 (51.3, 53.4) | 24.1 |
South* | 34.1 (33.4, 34.8) | 40.0 (39.3, 40.7) | 43.1 (42.4, 43.9) | 46.2 (45.5, 47.0) | 35.5 |
West* | 40.3 (39.0, 41.5) | 53.2 (51.8, 54.6) | 64.2 (62.6, 65.7) | 57.7 (56.2, 59.2) | 43.3 |
IH | |||||
Overall* | 7.8 (7.7, 8.0) | 8.8 (8.7, 8.9) | 8.8 (8.7, 9.0) | 10.3 (10.2, 10.5) | 31.9 |
Male | 7.3 (6.9, 7.6) | 8.5 (8.1, 8.8) | 8.4 (8.1, 8.8) | 9.5 (9.1, 9.9) | 30.5 |
Female | 8.4 (8.1, 8.7) | 9.1 (8.8, 9.4) | 9.2 (8.9, 9.5) | 11.1 (10.8, 11.4) | 32.8 |
By insurance | |||||
Commercial* | 8.3 (8.1, 8.6) | 9.5 (9.3, 9.8) | 9.7 (9.4, 10.0) | 11.3 (11.0, 11.6) | 36.3 |
Medicare* | 7.2 (6.7, 7.8) | 7.0 (6.6, 7.5) | 8.1 (7.5, 8.6) | 7.9 (7.4, 8.4) | 9.2 |
Medicaid* | 5.7 (5.0, 6.3) | 7.3 (6.6, 7.9) | 6.8 (6.2, 7.4) | 6.9 (6.4, 7.5) | 22.6 |
By region | |||||
Northeast* | 3.6 (3.3, 3.9) | 3.8 (3.5, 4.1) | 3.9 (3.5, 4.2) | 4.5 (4.2, 4.9) | 24.1 |
Midwest* | 10.2 (9.7, 10.0) | 12.0 (11.4, 12.5) | 10.7 (10.2, 11.2) | 11.8 (11.3, 12.4) | 11.8 |
South* | 9.3 (9.0, 9.7) | 11.0 (10.6, 11.4) | 11.6 (11.2, 12.1) | 13.4 (12.9, 13.8) | 43.3 |
West* | 6.8 (6.3, 7.0) | 6.2 (5.7, 6.7) | 6.3 (5.8, 6.8) | 9.0 (8.4, 9.6) | 32.3 |
RBD | |||||
Overall* | 5.3 (5.2, 5.4) | 6.5 (6.4, 6.6) | 7.6 (7.5, 7.7) | 8.7 (8.6, 8.8) | 63.7 |
Male** | 8.5 (8.2, 8.8) | 10.2 (9.9, 10.6) | 11.9 (11.6, 12.2) | 13.5 (13.1, 13.9) | 59.4 |
Female** | 2.6 (2.5, 2.8) | 3.3 (3.1, 3.4) | 3.8 (3.6, 4.0) | 4.6 (4.4, 4.7) | 74.5 |
By insurance | |||||
Commercial* | 5.7 (5.4, 5.9) | 6.9 (6.6, 7.1) | 8.2 (7.9, 8.5) | 9.5 (9.3, 9.8) | 68.5 |
Medicare* | 5.5 (5.2, 5.7) | 7.1 (6.9, 7.4) | 8.9 (8.5, 9.2) | 10.8 (10.5, 11.2) | 98.2 |
Medicaid* | 3.0 (2.4, 3.6) | 4.1 (3.4, 4.7) | 4.5 (3.9, 5.1) | 5.3 (4.7, 6.0) | 77.3 |
By region | |||||
Northeast* | 3.8 (3.5, 4.0) | 4.2 (3.9, 4.4) | 4.8 (4.5, 5.1) | 5.9 (5.5, 6.2) | 55.6 |
Midwest* | 6.8 (6.5, 7.2) | 8.5 (8.0, 8.9) | 9.3 (8.9, 9.7) | 9.9 (9.4, 10.3) | 44.4 |
South* | 5.2 (5.0, 5.5) | 6.5 (6.2, 6.8) | 7.8 (7.5, 8.0) | 9.8 (9.5, 10.2) | 88.3 |
West* | 5.7 (5.2, 6.1) | 7.2 (6.8, 7.7) | 8.8 (8.2, 9.4) | 8.2 (7.7, 8.7) | 45.2 |
IH = idiopathic hypersomnia; OSA = obstructive sleep apnea; PLMD = periodic limb movement disorder; RBD = rapid eye movement sleep behavior disorder.
Age- and sex-adjusted to the 2016 census. **Age-adjusted to the 2016 census.
Figure 2. Annual prevalence of sleep disorders from 2013-2016.
OSA prevalence is provided as an inset due to scaling considerations. IH = idiopathic hypersomnia; OSA = obstructive sleep apnea; PLMD = periodic limb movement disorder; RBD = rapid eye movement sleep behavior disorder.
IH prevalence increased approximately 32% over the study period from 7.8 to 10.3 per 100,000 persons. The annual male to female ratio was approximately 0.9 over the study period, and the percentage increase in prevalence was similar for both sexes. Prevalence and the percentage increase in prevalence over the study period were appreciably higher for those with commercial insurance (11.3 per 100,000 in 2016, 36% increase during 2013–2016) than for those with either Medicare (7.9 per 100,000, 9% increase) or Medicaid (6.9 per 100,000, 23% increase). Geographically, as with the other sleep disorders, prevalence was highest in the Midwest (11.8 per 100,000 in 2016) and South (13.4 per 100,000), and particularly low in the Northeast (4.5 per 100,000). The percentage increase in prevalence was highest in the South (43%) and West (32%) and lower in the Northeast (24%) and Midwest (12%).
The prevalence of PLMD increased approximately 30% from 33.4 to 43.4 per 100,000 over the study period. Prevalence was approximately 70% higher for men than for women throughout the study period, and the trend of increasing prevalence was slightly higher for men than for women. Prevalence was higher for those with commercial insurance, and the trend of increasing prevalence was much greater for those with commercial insurance (35%) or with Medicaid (46%) than for Medicare beneficiaries (15%). Prevalence was lower by 50% or more in the Northeast than in the other regions, and the Northeast had the smallest percentage increase in prevalence over the study period (8% vs 24–43% in the other regions).
RBD prevalence increased approximately 64% over the study period from 5.3 to 8.7 per 100,000 persons. Prevalence was approximately three-fold higher for males than for females. Prevalence was comparable for those with commercial insurance (9.5 per 100,000 in 2016) and Medicare (10.8 per 100,000), while prevalence for Medicaid beneficiaries was approximately half that of the other 2 groups (5.3 per 100,000). The general pattern of appreciably lower prevalence for the Northeast region (5.9 per 100,000 in 2016) compared with the Midwest (9.9 per 100,000) and South (9.8 per 100,000) was found as well for RBD, and prevalence increased most during the study period in the South (88%).
Sleep testing
Among sleep tests, MSLT and MWT, which share a procedure code in claims data, had by far the lowest frequency per 100,000 persons compared with HSAT and PSG (Table 4). The MSLT/MWT declined approximately 20% in frequency during the study period from 19 to 15 per 100,000 persons. The decline was most marked in the Northeast (31.5%) and minimal in the West (2.4%). HSAT frequency more than doubled over the study period; the gain was largest in the Midwest (168.5%) and lowest in the West (72.4%). PSG declined approximately 15% from 2013 to 2016, the decline being highest in the Northeast (25%) and lowest in the West (2%).
Table 4.
Frequency of the MSLT, MWT, HSAT, and PSG per 100,000 persons by region: 2013–2016.
2013 | 2014 | 2015 | 2016 | % Change 2013–2016 | |
---|---|---|---|---|---|
MSLT/MWT | |||||
Overall | 19.1 | 18.0 | 16.6 | 15.2 | −20.5 |
Northeast | 16.1 | 14.1 | 11.5 | 11.1 | −31.5 |
Midwest | 26.4 | 24.7 | 22.2 | 20.5 | −22.4 |
South | 18.8 | 18.1 | 17.4 | 15.5 | −17.7 |
West | 11.9 | 12.7 | 13.1 | 11.7 | −2.4 |
HSAT | |||||
Overall | 92.9 | 126.6 | 154.8 | 202.2 | 117.5 |
Northeast | 93.4 | 133.5 | 175.9 | 199.3 | 113.3 |
Midwest | 65.5 | 111.6 | 148.1 | 175.8 | 168.5 |
South | 106.4 | 130.9 | 144.5 | 229.8 | 116.1 |
West | 100.6 | 126.1 | 155.6 | 173.4 | 72.4 |
PSG | |||||
Overall | 640.0 | 600.5 | 573.0 | 546.1 | −14.7 |
Northeast | 564.6 | 475.3 | 432.7 | 423.4 | −25.0 |
Midwest | 738.0 | 732.0 | 700.2 | 660.1 | −10.6 |
South | 669.4 | 620.4 | 584.2 | 556.2 | −16.9 |
West | 512.5 | 541.2 | 567.1 | 523.0 | 2.1 |
Values were age- and sex-adjusted to the 2016 census. HSAT = home sleep apnea test; MSLT = multiple sleep latency test; MWT = maintenance of wakefulness test; PSG = polysomnography.
Provider specialty
General practitioners (family practice and internal medicine physicians) were the most frequent providers to order MSLTs/MWTs, HSATs, and PSGs for each year of the study period (Table 5). In 2016, they ordered approximately 27% of MSLTs/MWTs, 37% of HSATs, and 35% of PSGs. However, on a relative basis, their annual proportion of tests ordered declined by 18%, 12%, and 13%, respectively, over the study period. Pulmonologists were the next most frequent prescribers for all 3 tests, and on a relative basis they increased their proportion of tests prescribed by 26% (MSLT/MWT), 21% (HSAT), and 16% (PSG), respectively, from 2013 to 2016. Sleep medicine specialists prescribed a relatively low proportion of sleep tests in each year of the study period, although their proportion of tests prescribed increased by 56% (MSLT/MWT), 40% (HSAT), and 44% (PSG), respectively.
Table 5.
Percentages of sleep tests ordered by medical specialty, 2013–2016.
2013 | 2014 | 2015 | 2016 | % Change 2013–2016 | |
---|---|---|---|---|---|
MSLT/MWT | |||||
General practitioner* | 33.0 | 30.8 | 29.2 | 26.9 | −18.4 |
Pulmonology | 18.3 | 20.5 | 21.7 | 23.1 | 26.0 |
Neurology | 14.3 | 15.0 | 13.8 | 13.2 | −8.0 |
Sleep medicine | 6.4 | 6.7 | 8.5 | 10.0 | 56.2 |
Other specialty | 10.7 | 10.4 | 9.9 | 9.9 | −7.0 |
Unknown specialty | 17.3 | 16.5 | 16.9 | 16.9 | −2.3 |
HSAT | |||||
General practitioner* | 42.8 | 40.2 | 37.3 | 37.5 | −12.4 |
Pulmonology | 12.9 | 14.7 | 15.9 | 15.6 | 20.7 |
Neurology | 3.0 | 3.7 | 3.9 | 4.0 | 32.0 |
Sleep medicine | 3.3 | 3.5 | 4.4 | 4.6 | 39.8 |
Other specialty | 13.9 | 13.8 | 13.0 | 13.1 | −6.3 |
Unknown specialty | 24.1 | 24.1 | 25.5 | 25.3 | 5.0 |
PSG | |||||
General practitioner* | 39.6 | 37.8 | 35.6 | 34.5 | −12.8 |
Pulmonology | 17.9 | 18.9 | 19.6 | 20.7 | 16.0 |
Neurology | 6.3 | 6.5 | 6.3 | 6.5 | 3.4 |
Sleep medicine | 4.6 | 5.2 | 6.2 | 6.6 | 44.5 |
Other specialty | 15.5 | 14.7 | 14.0 | 13.8 | −10.8 |
Unknown specialty | 16.2 | 16.9 | 18.4 | 17.9 | 10.0 |
HSAT = home sleep apnea test; MSLT = multiple sleep latency test; MWT = maintenance of wakefulness test; PSG = polysomnography. *General practitioner includes general practitioners, internal medicine physicians, and family practice physicians.
DISCUSSION
Narcolepsy is a debilitating sleep disorder that is widely recognized to be underdiagnosed,1 and for which limited US prevalence data are available. In our study of approximately one-fifth of the US population, we found age- and sex-adjusted prevalence to have increased by approximately 14%, from 39 to 44 cases per 100,000 persons, during the period 2013–2016. Prevalence was higher for those with commercial insurance and varied markedly by geographic region, being more than twice as high in 2016 in the Midwest than in the Northeast. Contrary to the existing, somewhat limited literature,3 we did not find narcolepsy to be twice as common for men than women. In fact, we found a slightly higher prevalence for women.
Estimating prevalence is uncertain for a disease like narcolepsy that is substantially underdiagnosed and for which the prediagnostic interval can be long. It is impossible to know whether the 14% increase we observed was due to increasing prevalence per se or to increased recognition by the medical community. This is illustrated somewhat by the geographic variation we observed. It is unlikely that the true prevalence of narcolepsy is twice as high in the Midwest as in the Northeast. Presumably, the disparate prevalence estimates were due, perhaps in large part, to regional variation in medical administrative or diagnostic practices. We also cannot exclude other factors such as environmental influences, gene–environment interactions, or cultural variations. The disparate narcolepsy prevalence increases in commercial versus the Medicare and Medicaid population are also noteworthy. Further investigation is warranted to explain these differences (eg, the potential role of differences in socioeconomic status), given that insurance-based disparities may be contributing to poorer health care and negative sleep health outcomes in specific populations, which has important public health implications.
We projected a nationwide prevalence of approximately 142,600 narcolepsy cases for 2016. One limitation of this methodology is that our prevalence estimates are 2-year limited-duration estimates and not complete prevalences (namely, everyone alive in our denominator with a current or historical narcolepsy diagnosis). In a sensitivity analysis, prevalence estimates increased 40% when we extended our lookback window for diagnoses from 2 years to 4 years. Accordingly, our results are best considered as the prevalence within an insured population of those with narcolepsy who are actively seeking medical care over a 2-year time period. The difference between our 2- and 4-year limited prevalence estimates illustrates that a substantial proportion of individuals with narcolepsy have long intervals without a related medical contact, as reflected in medical or pharmacy claims.
Among other sleep disorders, as anticipated, we found OSA to be the most prevalent by far. Consistent with prior reports,7 OSA prevalence was more than 2-fold higher for men than for women and increased 41% over the study period. Prevalence varied markedly by geographic region, being 63% higher in the West than in the Northeast. Prevalence for PLMD, IH, and RBD showed considerable increases over the study period of 30%, 32%, and 64%, respectively, with appreciable geographic variation and markedly lower prevalence estimates in the Northeast.
There were notable changes in the frequency of prescribed sleep tests over the study period. We found declines of 15–20% for PSG and MSLT/MWT, respectively, and an increase of 118% for the HSAT. Although ordering of PSG testing—of which the most common clinical indication is OSA evaluation—experienced an expected decline, this was overshadowed by the dramatic increase in the conduct of HSATs. Interestingly, this occurred in the context of a 40% increase in OSA diagnosis. Although we cannot state it definitively, these data suggest an enhanced ability to more efficiently diagnose OSA via portable sleep testing without substantial compromise in PSG testing. Furthermore, although a decline in MSLTs/MWTs was observed, this was paralleled by an increase the diagnosis of narcolepsy and IH. Sleep tests were most commonly ordered by general practitioners, although there was a trend over the study period for more procedures being ordered by pulmonologists and, to a lesser extent, by sleep medicine specialists.
There are a number of internal and external validity limitations of our study not previously mentioned. Regarding internal validity, foremost is the reliance on medical and pharmacy claims data that are primarily collected for reimbursement purposes. We did not have access to clinical records to validate our case identification procedures. With respect to external validity (or generalizability), our analyses focused on those who have medical insurance and pharmacy coverage. Prevalence may differ between those who are insured and those who are not insured. In addition, the Medicare beneficiaries included in our analyses participate in Medicare Advantage and are commonly considered to be a healthier subpopulation than those on traditional (part A and part B) Medicare.
Nonetheless, our findings should be informative for clinicians, payers, and patients. Despite the limitations of our analyses, it is likely that the recognition of narcolepsy, OSA, and the other sleep disorders is increasing; that seeking diagnosis and care for narcolepsy is not appreciably different in frequency for males and females; that those who are commercially insured are being diagnosed with sleep disorders and treated more frequently than those on Medicare/Medicaid; and that there are large regional differences in the frequency of testing and the diagnosis of narcolepsy and of other sleep disorders.
DISCLOSURE STATEMENT
All authors have contributed to the development of this manuscript, have reviewed it, and have approved its contents. This study was funded by Jazz Pharmaceuticals. M.B. was an employee of Jazz Pharmaceuticals during the conduct of this research. The other authors report no conflicts of interest.
ABBREVIATIONS
- HSAT
home sleep apnea test
- IH
idiopathic hypersomnia
- MSLT
multiple sleep latency test
- MWT
maintenance of wakefulness test
- OSA
obstructive sleep apnea
- PLMD
periodic limb movement disorder
- PSG
polysomnography
- RBD
rapid eye movement sleep behavior disorder
REFERENCES
- 1.Scammell TE. Narcolepsy. N Engl J Med. 2015;373(27):2654–2662. 10.1056/NEJMra1500587 [DOI] [PubMed] [Google Scholar]
- 2.Morrish E, King MA, Smith IE, Shneerson JM. Factors associated with a delay in the diagnosis of narcolepsy. Sleep Med. 2004;5(1):37–41. 10.1016/j.sleep.2003.06.002 [DOI] [PubMed] [Google Scholar]
- 3.Silber MH, Krahn LE, Olson EJ, Pankratz VS. The epidemiology of narcolepsy in Olmsted County, Minnesota: a population-based study. Sleep. 2002;25(2):197–202. 10.1093/sleep/25.2.197 [DOI] [PubMed] [Google Scholar]
- 4.Longstreth WT, Koepsell TD, Ton TG, Hendrickson AF, van Belle G. The epidemiology of narcolepsy. Sleep. 2007;30(1):13–26. 10.1093/sleep/30.1.13 [DOI] [PubMed] [Google Scholar]
- 5.US Census Bureau . Annual Estimates of the Civilian Population by Single Year of Age and Sex for the United States and States. Suitland, Maryland: Population Division; 2016 [Google Scholar]
- 6.Newcombe RG. Two-sided confidence intervals for the single proportion: comparison of seven methods. Stat Med. 1998;17(8):857–872. [DOI] [PubMed] [Google Scholar]
- 7.Punjabi NM. The epidemiology of adult obstructive sleep apnea. Proc Am Thorac Soc. 2008;5(2):136–143. 10.1513/pats.200709-155MG [DOI] [PMC free article] [PubMed] [Google Scholar]