https://orcid.org/0000-0002-5894-8289
Abstract
Epidemiologic studies of obstructive sleep apnea (OSA) and insomnia in the U.S. military are limited. The primary aim of this study was to report and compare OSA and insomnia diagnoses in active duty the United States military service members.
Data and service branch densities used to derive the expected rates of diagnoses on insomnia and OSA were drawn from the Defense Medical Epidemiology Database. Single sample chi-square goodness of fit tests and independent samples t-tests were conducted to address the aims of the study.
Between 2005 and 2019, incidence rates of OSA and insomnia increased from 11 to 333 and 6 to 272 (per 10,000), respectively. Service members in the Air Force, Navy, and Marines were diagnosed with insomnia and OSA below expected rates, while those in the Army had higher than expected rates (p < .001). Female service members were underdiagnosed in both disorders (p < .001). Comparison of diagnoses following the transition from ICD 9 to 10 codes revealed significant differences in the amounts of OSA diagnoses only (p < .05).
Since 2005, incidence rates of OSA and insomnia have markedly increased across all branches of the U.S. military. Despite similar requirements for overall physical and mental health and resilience, service members in the Army had higher rates of insomnia and OSA. This unexpected finding may relate to inherent differences in the branches of the military or the role of the Army in combat operations. Future studies utilizing military-specific data and directed interventions are required to reverse this negative trend.
This study provides a comprehensive epidemiologic overview of the two most common sleep disorders in U.S. military personnel. Our findings contribute to the growing body of literature that sleep disorders are increasingly diagnosed in what should otherwise be a healthy population. The finding that service members in the Army have the highest rates of insomnia and obstructive sleep apnea (OSA) is notable and concerning given the similarities in service members throughout the Department of Defense. Additionally, women in the military may be underdiagnosed for OSA and insomnia.
Introduction
Numerous studies in civilian populations have reported increasing rates and prevalence of obstructive sleep apnea (OSA) and insomnia [1–3]. As is the case in the general population, sleep disturbances have become increasingly recognized in U.S. military personnel. Using the Pittsburgh Sleep Quality Index, Troxel and colleagues [4] reported 48.6% of military personnel had poor sleep. This high rate of sleep disturbances likely contributes to the increasing incidence of sleep disorders that had been reported in the U.S. Army [5]. This is important for society in general, because when military personnel separate from active duty and become Veterans, they join a group with one of the highest reported rates of insomnia and OSA [6, 7]. At present, little research has been conducted to characterize the prevalence or incidence of the two most common clinically significant sleep disorders—insomnia and OSA—among active-duty military personnel in all branches of the U.S. military (i.e. Army, Navy, Air Force, Marines, Coast Guard, and the recently established Space Force). There are inherent differences in the branches of the military in terms of their deployments, missions, and operational exposures which could impact their sleep and subsequent development of sleep disorders. Thus, empirical evaluation of insomnia and OSA in U.S. military personnel is warranted.
The majority of epidemiological studies evaluating sleep disorders have been conducted in U.S. Army personnel and likely do not represent all military personnel in the Department of Defense (DoD). Two recent studies assessed insomnia prevalence in large cohorts of U.S. Army Soldiers and reported similar rates of 19.9% and 22.76 [8, 9]. However, a more recent study on the prevalence of insomnia in U.S. Navy Sailors and Marines reported a prevalence of only 3.4% [10]. Regarding OSA, there are no studies that have systematically evaluated the prevalence of this sleep disorder within the armed services. In a study of 29,247 service members, Seelig and colleagues [11] used one question to evaluate for the prevalence of OSA, asking if a participant had a diagnosis of sleep apnea, with 2.7% having a positive response. A more recent study by Caldwell and colleagues [5] examined medical records of U.S. Army personnel and reported OSA incidence rates increased from 4.3 (per 1,000) in 2003 to 30.1 (per 1,000) in 2011, a 600% increase. These findings support a high prevalence of insomnia as well as a rapidly increasing incidence of OSA in the Army, but not necessarily in the other branches of the military.
Recent reports on the epidemiology of OSA and insomnia in the U.S. military are limited as they focus on one branch of service, used non-validated measures, or simply report count data with little statistical analyses conducted to contextualize diagnoses. There are clinical studies that have reported sleep disorder diagnoses in members of all branches of the military. These studies, however, were relatively small in numbers, and only reported on military personnel who had clinically significant sleep disorders [12–14]. Additionally, there are studies which establish associations between comorbid disorders such as posttraumatic stress disorder (PTSD) and/or traumatic brain injury (TBI), which are relatively unique to the military and veteran populations, but they do not provide a basis for understanding the overall rates of sleep disorders in the DoD [10, 15, 16]. Specifically, there is one study which evaluated medical encounters related to insomnia and OSA in all branches of the U.S. military, but not the respective sleep diagnoses [17]. This study demonstrated “epidemic like increases” in the clinical care rendered for these sleep disorders, with encounters for OSA substantially higher than those for insomnia.
These issues create a substantial gap in the knowledge about rates of OSA and insomnia across all branches of the DoD. The primary objective of this study was to report on and compare the incident diagnoses of insomnia and OSA across all branches of the U.S. military to provide a better understanding of clinically significant sleep disorders in active-duty military personnel. A secondary objective was to determine if the change from ICD-9 to ICD-10 impacted the diagnostic rates of insomnia and OSA in military personnel.
Methods
Incidence rates of OSA and insomnia in active-duty U.S. military service members were derived from count data provided from the Defense Medical Epidemiology Database (DMED). The DMED is a subset database of the Defense Medical Surveillance System and records individual instances of medical diagnoses wherein an ICD code is placed in an active duty service member’s electronic medical record during any visit to a military medical facility. All DMED data are validated by DoD medical records. Each diagnosis of OSA and insomnia utilized for this study was available in a de-identified form and provided as grouped counts by standard military classifications (see Table 1). These classifications include (1) military pay grade (junior enlisted [E-1 to E-4], noncommissioned officers [E-5 to E-9], junior commissioned officers [O-1/WO1 to O-3/CW3], and senior commissioned officers [O-4/CW4 to O-6/CW5]); (2) service branch (i.e. Army, Navy, Air Force, Marine Corps); (3) sex (i.e. male/female); (4) age (i.e. <20, 20–24, 25–29, 30–34, etc.); (5) marital status (i.e. married, single, or other); and (6) race (i.e. white, black, or other). Data were not available on the U.S. Space Force, which was recently established in 2019 or in Coast Guard personnel as the DMED tracks medical diagnoses in Department of Defense personnel. The Coast Guard is part of the Department of Homeland Security.
Demographics of United States military service members diagnosed with obstructive sleep apnea and insomnia between 2005 and 2019
| Disorder representation in the present sample | Residuals | Average change in diagnoses from 2005 to 2019 | |||||||
|---|---|---|---|---|---|---|---|---|---|
| OSA | Insomnia | ||||||||
| Demographic | Military composition | OSA | Insomnia | Raw | Standardized | Raw | Standardized | OSA | Insomnia |
| Sex | |||||||||
| Female | 15% | 8% | 12% | –28292.7 | –116.4 | –10742.3 | –44.5 | 40% | 44% |
| Male | 85% | 92% | 88% | 28292.7 | 48.8 | 10742.2 | 18.7 | 40% | 34% |
| Age, years | |||||||||
| <20 | 7% | 0% | 1% | –26868.5 | –161.0 | –21740.6 | –131.3 | 46% | 69% |
| 20–24 | 32% | 8% | 14% | –96461.3 | –270.3 | –69474.9 | –196.2 | 39% | 43% |
| 25–29 | 23% | 14% | 17% | –35060.1 | –115.9 | –23900.1 | –79.6 | 43% | 45% |
| 30–34 | 15% | 15% | 15% | 519.7 | 2.1 | –756.4 | –3.12 | 43% | 36% |
| 35–39 | 12% | 24% | 20% | 45033.8 | 206.1 | 32031.9 | 147.7 | 41% | 34% |
| 40+ | 10% | 39% | 32% | 112837.0 | 565.7 | 83840.1 | 423.7 | 38% | 29% |
| Marital status | |||||||||
| Married | 55% | 79% | 73% | 95269.1 | 204.6 | 67411.7 | 146.0 | 39% | 33% |
| Non-married | 45% | 21% | 27% | –95269.1 | –226.2 | –67411.8 | –161.4 | 42% | 41% |
| Race | |||||||||
| White | 69% | 61% | 61% | –30706.0 | –58.9 | –30182.6 | –58.3 | 39% | 35% |
| Black | 17% | 24% | 25% | 25750.1 | 99.5 | 29822.8 | 116.1 | 39% | 36% |
| Other | 14% | 15% | 14% | 4955.9 | 21.1 | 359.7 | 1.5 | 46% | 40% |
| Service branch | |||||||||
| Army | 38% | 48% | 50% | 38617.0 | 99.8 | 45553.9 | 118.6 | 48% | 44% |
| Navy | 24% | 22% | 19% | –9714.9 | –31.5 | –17914.8 | –58.7 | 46% | 35% |
| Air Force | 24% | 23% | 23% | –2233.9 | –7.2 | –3982.8 | –13.0 | 29% | 25% |
| Marine Corps | 14% | 7% | 8% | –26668.1 | –113.5 | –23656.3 | –101.5 | 56% | 51% |
| Military pay grade | |||||||||
| Junior Enlisted | 44% | 16% | 25% | –111243.7 | –267.2 | –75255.1 | –182.2 | 43% | 45% |
| Non-commissioned officers | 39% | 64% | 59% | 100575.0 | 256.6 | 77149.5 | 198.4 | 40% | 34% |
| Junior commissioned officers | 10% | 7% | 6% | –12525.8 | –63.1 | –13801.3 | –70.1 | 40% | 37% |
| Senior commissioned officers | 7% | 13% | 10% | 23194.3 | 139.6 | 11906.9 | 72.2 | 36% | 27% |
| Disorder representation in the present sample | Residuals | Average change in diagnoses from 2005 to 2019 | |||||||
|---|---|---|---|---|---|---|---|---|---|
| OSA | Insomnia | ||||||||
| Demographic | Military composition | OSA | Insomnia | Raw | Standardized | Raw | Standardized | OSA | Insomnia |
| Sex | |||||||||
| Female | 15% | 8% | 12% | –28292.7 | –116.4 | –10742.3 | –44.5 | 40% | 44% |
| Male | 85% | 92% | 88% | 28292.7 | 48.8 | 10742.2 | 18.7 | 40% | 34% |
| Age, years | |||||||||
| <20 | 7% | 0% | 1% | –26868.5 | –161.0 | –21740.6 | –131.3 | 46% | 69% |
| 20–24 | 32% | 8% | 14% | –96461.3 | –270.3 | –69474.9 | –196.2 | 39% | 43% |
| 25–29 | 23% | 14% | 17% | –35060.1 | –115.9 | –23900.1 | –79.6 | 43% | 45% |
| 30–34 | 15% | 15% | 15% | 519.7 | 2.1 | –756.4 | –3.12 | 43% | 36% |
| 35–39 | 12% | 24% | 20% | 45033.8 | 206.1 | 32031.9 | 147.7 | 41% | 34% |
| 40+ | 10% | 39% | 32% | 112837.0 | 565.7 | 83840.1 | 423.7 | 38% | 29% |
| Marital status | |||||||||
| Married | 55% | 79% | 73% | 95269.1 | 204.6 | 67411.7 | 146.0 | 39% | 33% |
| Non-married | 45% | 21% | 27% | –95269.1 | –226.2 | –67411.8 | –161.4 | 42% | 41% |
| Race | |||||||||
| White | 69% | 61% | 61% | –30706.0 | –58.9 | –30182.6 | –58.3 | 39% | 35% |
| Black | 17% | 24% | 25% | 25750.1 | 99.5 | 29822.8 | 116.1 | 39% | 36% |
| Other | 14% | 15% | 14% | 4955.9 | 21.1 | 359.7 | 1.5 | 46% | 40% |
| Service branch | |||||||||
| Army | 38% | 48% | 50% | 38617.0 | 99.8 | 45553.9 | 118.6 | 48% | 44% |
| Navy | 24% | 22% | 19% | –9714.9 | –31.5 | –17914.8 | –58.7 | 46% | 35% |
| Air Force | 24% | 23% | 23% | –2233.9 | –7.2 | –3982.8 | –13.0 | 29% | 25% |
| Marine Corps | 14% | 7% | 8% | –26668.1 | –113.5 | –23656.3 | –101.5 | 56% | 51% |
| Military pay grade | |||||||||
| Junior Enlisted | 44% | 16% | 25% | –111243.7 | –267.2 | –75255.1 | –182.2 | 43% | 45% |
| Non-commissioned officers | 39% | 64% | 59% | 100575.0 | 256.6 | 77149.5 | 198.4 | 40% | 34% |
| Junior commissioned officers | 10% | 7% | 6% | –12525.8 | –63.1 | –13801.3 | –70.1 | 40% | 37% |
| Senior commissioned officers | 7% | 13% | 10% | 23194.3 | 139.6 | 11906.9 | 72.2 | 36% | 27% |
Note: Military composition percentages provided by the DMED. Bolded standardized residuals are those that exceed a conservative ±3 and are highlighted as they represent areas that may be of interest for future research/interventions. In the DMED, marital status is provided as married, single, or other. As there is no way to meaningfully discriminate between “single” or “other,” this variable was dichotomized to improve interpretation. “Junior enlisted” includes E-1 to E-4; “non-commissioned officers” includes pay grades E-5 to E-9; “junior commissioned officers” include pay grades O-1/WO1 to O-3/CW3; “senior commissioned officers” include pay grades O-4/CW4 to O-6/CW5. Raw counts of diagnoses: OSA = 393,857; insomnia = 385,537.
Abbreviations: OSA, obstructive sleep apnea; DMED, Defense Medical Epidemiology Database.
Demographics of United States military service members diagnosed with obstructive sleep apnea and insomnia between 2005 and 2019
| Disorder representation in the present sample | Residuals | Average change in diagnoses from 2005 to 2019 | |||||||
|---|---|---|---|---|---|---|---|---|---|
| OSA | Insomnia | ||||||||
| Demographic | Military composition | OSA | Insomnia | Raw | Standardized | Raw | Standardized | OSA | Insomnia |
| Sex | |||||||||
| Female | 15% | 8% | 12% | –28292.7 | –116.4 | –10742.3 | –44.5 | 40% | 44% |
| Male | 85% | 92% | 88% | 28292.7 | 48.8 | 10742.2 | 18.7 | 40% | 34% |
| Age, years | |||||||||
| <20 | 7% | 0% | 1% | –26868.5 | –161.0 | –21740.6 | –131.3 | 46% | 69% |
| 20–24 | 32% | 8% | 14% | –96461.3 | –270.3 | –69474.9 | –196.2 | 39% | 43% |
| 25–29 | 23% | 14% | 17% | –35060.1 | –115.9 | –23900.1 | –79.6 | 43% | 45% |
| 30–34 | 15% | 15% | 15% | 519.7 | 2.1 | –756.4 | –3.12 | 43% | 36% |
| 35–39 | 12% | 24% | 20% | 45033.8 | 206.1 | 32031.9 | 147.7 | 41% | 34% |
| 40+ | 10% | 39% | 32% | 112837.0 | 565.7 | 83840.1 | 423.7 | 38% | 29% |
| Marital status | |||||||||
| Married | 55% | 79% | 73% | 95269.1 | 204.6 | 67411.7 | 146.0 | 39% | 33% |
| Non-married | 45% | 21% | 27% | –95269.1 | –226.2 | –67411.8 | –161.4 | 42% | 41% |
| Race | |||||||||
| White | 69% | 61% | 61% | –30706.0 | –58.9 | –30182.6 | –58.3 | 39% | 35% |
| Black | 17% | 24% | 25% | 25750.1 | 99.5 | 29822.8 | 116.1 | 39% | 36% |
| Other | 14% | 15% | 14% | 4955.9 | 21.1 | 359.7 | 1.5 | 46% | 40% |
| Service branch | |||||||||
| Army | 38% | 48% | 50% | 38617.0 | 99.8 | 45553.9 | 118.6 | 48% | 44% |
| Navy | 24% | 22% | 19% | –9714.9 | –31.5 | –17914.8 | –58.7 | 46% | 35% |
| Air Force | 24% | 23% | 23% | –2233.9 | –7.2 | –3982.8 | –13.0 | 29% | 25% |
| Marine Corps | 14% | 7% | 8% | –26668.1 | –113.5 | –23656.3 | –101.5 | 56% | 51% |
| Military pay grade | |||||||||
| Junior Enlisted | 44% | 16% | 25% | –111243.7 | –267.2 | –75255.1 | –182.2 | 43% | 45% |
| Non-commissioned officers | 39% | 64% | 59% | 100575.0 | 256.6 | 77149.5 | 198.4 | 40% | 34% |
| Junior commissioned officers | 10% | 7% | 6% | –12525.8 | –63.1 | –13801.3 | –70.1 | 40% | 37% |
| Senior commissioned officers | 7% | 13% | 10% | 23194.3 | 139.6 | 11906.9 | 72.2 | 36% | 27% |
| Disorder representation in the present sample | Residuals | Average change in diagnoses from 2005 to 2019 | |||||||
|---|---|---|---|---|---|---|---|---|---|
| OSA | Insomnia | ||||||||
| Demographic | Military composition | OSA | Insomnia | Raw | Standardized | Raw | Standardized | OSA | Insomnia |
| Sex | |||||||||
| Female | 15% | 8% | 12% | –28292.7 | –116.4 | –10742.3 | –44.5 | 40% | 44% |
| Male | 85% | 92% | 88% | 28292.7 | 48.8 | 10742.2 | 18.7 | 40% | 34% |
| Age, years | |||||||||
| <20 | 7% | 0% | 1% | –26868.5 | –161.0 | –21740.6 | –131.3 | 46% | 69% |
| 20–24 | 32% | 8% | 14% | –96461.3 | –270.3 | –69474.9 | –196.2 | 39% | 43% |
| 25–29 | 23% | 14% | 17% | –35060.1 | –115.9 | –23900.1 | –79.6 | 43% | 45% |
| 30–34 | 15% | 15% | 15% | 519.7 | 2.1 | –756.4 | –3.12 | 43% | 36% |
| 35–39 | 12% | 24% | 20% | 45033.8 | 206.1 | 32031.9 | 147.7 | 41% | 34% |
| 40+ | 10% | 39% | 32% | 112837.0 | 565.7 | 83840.1 | 423.7 | 38% | 29% |
| Marital status | |||||||||
| Married | 55% | 79% | 73% | 95269.1 | 204.6 | 67411.7 | 146.0 | 39% | 33% |
| Non-married | 45% | 21% | 27% | –95269.1 | –226.2 | –67411.8 | –161.4 | 42% | 41% |
| Race | |||||||||
| White | 69% | 61% | 61% | –30706.0 | –58.9 | –30182.6 | –58.3 | 39% | 35% |
| Black | 17% | 24% | 25% | 25750.1 | 99.5 | 29822.8 | 116.1 | 39% | 36% |
| Other | 14% | 15% | 14% | 4955.9 | 21.1 | 359.7 | 1.5 | 46% | 40% |
| Service branch | |||||||||
| Army | 38% | 48% | 50% | 38617.0 | 99.8 | 45553.9 | 118.6 | 48% | 44% |
| Navy | 24% | 22% | 19% | –9714.9 | –31.5 | –17914.8 | –58.7 | 46% | 35% |
| Air Force | 24% | 23% | 23% | –2233.9 | –7.2 | –3982.8 | –13.0 | 29% | 25% |
| Marine Corps | 14% | 7% | 8% | –26668.1 | –113.5 | –23656.3 | –101.5 | 56% | 51% |
| Military pay grade | |||||||||
| Junior Enlisted | 44% | 16% | 25% | –111243.7 | –267.2 | –75255.1 | –182.2 | 43% | 45% |
| Non-commissioned officers | 39% | 64% | 59% | 100575.0 | 256.6 | 77149.5 | 198.4 | 40% | 34% |
| Junior commissioned officers | 10% | 7% | 6% | –12525.8 | –63.1 | –13801.3 | –70.1 | 40% | 37% |
| Senior commissioned officers | 7% | 13% | 10% | 23194.3 | 139.6 | 11906.9 | 72.2 | 36% | 27% |
Note: Military composition percentages provided by the DMED. Bolded standardized residuals are those that exceed a conservative ±3 and are highlighted as they represent areas that may be of interest for future research/interventions. In the DMED, marital status is provided as married, single, or other. As there is no way to meaningfully discriminate between “single” or “other,” this variable was dichotomized to improve interpretation. “Junior enlisted” includes E-1 to E-4; “non-commissioned officers” includes pay grades E-5 to E-9; “junior commissioned officers” include pay grades O-1/WO1 to O-3/CW3; “senior commissioned officers” include pay grades O-4/CW4 to O-6/CW5. Raw counts of diagnoses: OSA = 393,857; insomnia = 385,537.
Abbreviations: OSA, obstructive sleep apnea; DMED, Defense Medical Epidemiology Database.
The DMED provides two primary options for drawing rates of diagnoses: (1) an initial diagnosis of OSA or insomnia, and (2) any visit in which an ICD-9 or ICD-10 code was entered into the service member’s electronic medical record for the first time. During the timeframe examined in this study, ICD-10 codes were implemented. A recent report indicated the use of ICD-10 codes saw an increased utilization of “unspecified codes” thereby potentially leading to conflated rates of diagnoses [18]. To avoid potential overestimation of incidence, comparison of ICD-9 and ICD-10 diagnostic rates were conducted and only primary OSA and insomnia diagnosis data (ICD-9/10: 327.23/G47.33 and ICD-9/10: 327.00/F51.01 and G47.0 respectively) were drawn and analyzed. These codes were drawn solely on the first instance in which the service member was diagnosed as having OSA or insomnia. We examined OSA and insomnia incidence by all six available military classifications (i.e. sex, marital status, branch of service, age, military pay grade, and race). Data was accessed on June 25, 2020.
There are some limitations to the DMED database. It only allows for comparisons at the population level, and research data are limited to only active-duty military service members. Single sample chi-square goodness of fit tests were conducted to assess which demographic subgroups were over- or under-represented with OSA and insomnia diagnoses relative to their respective representation across the various demographics available with the DMED. The single sample Chi-Square test examines the distribution of cases in a single categorical variable (e.g. new-onset cases of OSA by service branch) and determines how that distribution follows a known or hypothesized distribution (e.g. service member distribution in the various branches of the military). Significance of the one sample chi-square goodness of fit test indicates that the observed proportions differ significantly from those in the known or hypothesized distributions. Where appropriate assessments of residuals were conducted to elaborate on the individual demographic subcategory contributions to the chi-square values (see Table 1) [19]. Negative standardized residuals indicate individual cells are populated by fewer diagnoses than expected based on known densities. Similarly, positive residuals indicate cell counts were greater than expected based on known densities. To compare OSA and insomnia specific diagnoses, independent samples t-tests were used to compare diagnostic rates of OSA and insomnia from available ICD-9 and ICD-10 codes.
The expected counts of diagnosis by each demographic variable were provided by the DMED database at the time of data extraction (see Table 1). They are not based on known clinical distribution but rather represent the average percentage of service members represented within each variable (or category) between 2005 and 2019. For instance, when extracting data, we found that 7% of all service members in the active force were under the age of 20 between 2005 and 2019. Therefore, we expected service members under the age of 20 to account for 7% of all OSA or insomnia diagnoses throughout this study.
Based on DMED data, the mean population of service members each year was calculated to be 1,354,451, with a low of 1,283,679 in 2017 and a high of 1,418,894 in 2010. All military classification variables were analyzed and are reported based on the cohort groupings by which they are available within the DMED database (i.e. E-1 to E-4, E-5 to E-9, O-1 to O-3, O-4 to O-6, etc.). The present evaluation is based on data from 2005 to 2019. These years provide a comparison and extension of previously reported data [17] with the difference that our study only used the first time a diagnosis of insomnia or OSA was used, whereas the previous study used encounters which could result in a service member having multiple visits counted as not an incident diagnosis. This study was reviewed by the Institutional Review Board at the University of Texas Health Science Center at San Antonio (protocol number: HSC20190664N) and was deemed to be not human subjects research.
Results
Between 2005 and 2019, the incidence rates of OSA (per 10,000 service members) in the United States DoD ranged from 11.8 in 2005 to 333.8 in 2016 (see Figure 1). Similarly, rates of insomnia increased across time from 5.7 in 2005 to 272.4 in 2015. The results of the single-sample chi-square tests showed the observed and expected frequencies significantly differed across all six military classification variables for both disorders (see Table 1). The service members most often diagnosed with OSA were married (79%), in the enlisted pay grade of E-5 to E-9 (64%), white (61%), males (92%), in the Army (48%), and 40 years of age or older (39%). Regarding insomnia, those most often diagnosed (see Table 1) were married (73%), in the enlisted pay grade of E-5 to E-9 (59%), white (61%), males (88%), in the Army (50%), and 40 years of age or older (32%). The average change in diagnoses over time for both disorders are listed in Table 1.
Incidence of obstructive sleep apnea (N = 393,857) and insomnia diagnoses (N = 385,537) from 2005 to 2019 are represented by incidence of cases per 10,000 service members. The incidence rates by each data point on the graph range from 6 (insomnia, 2005) to 333 (OSA, 2016).
Obstructive sleep apnea
Race
Statistically significant differences were noted among service members by race (χ2(2, N = 393,858) = 13,817.94, p < .001). Minority service members presented with higher rates of OSA diagnoses (n = 92,706 and n = 60,096, for black and other non-white races respectively) than expected (n = 66,955 and n = 55,140) based on known densities. Conversely, white service members, presented with fewer observed cases (n = 241,056) than expected (n = 271,762).
Age
Statistically significant differences were noted by age groups (χ2(5, N = 393,857) = 474,964.66, p < .001). Among active duty service members, those aged <20 years old, 20–24 years old, and 25–29 years old presented with fewer observed cases (n = 980, n = 30,846, and n = 56,442, respectively) than expected (n = 27,848, n = 127,307, and n = 91,502, respectively). Service members aged 30–34 presented with approximately as many observed cases (n = 60,195) as expected (n = 59,675). Service members aged 35 or older (groups: 35–39 and 40+) presented with more observed cases (n = 92,774 and n = 152,620 respectively) than expected (n = 47,740 and n = 39,783 respectively).
Military pay grade
Statistically significant differences were also observed among military pay grades (χ2(3, N = 393,858) = 160,768.04, p <.001). Early career (e.g. junior enlisted [E-1 to E-4] and junior commissioned officers [O-1/WO1 to O-3/CW3]) service members presented with fewer OSA diagnoses (n = 62,045 and n = 26,858, respectively) than expected (n = 173,288 and n = 39,383, respectively). Conversely, senior enlisted personnel (non-commissioned officers [E-5 to E-9] and senior commissioned officers [O-4/CW4 to O-6/CW5]) presented with more observed cases (n = 254,172 and n = 50,763, respectively) than expected (n = 153,596 and n = 27,568, respectively).
Marital status
Statistically significant differences were noted between married and nonmarried military personnel (χ2 (1, N = 393,858) = 93,108.48, p < .001). Married service members, presented with more observed cases of OSA (n = 311,891) than expected (n = 216,621). Conversely, non-married service members, presented with fewer observed cases (n = 81,967) than expected (n = 177,236).
Sex
Statistically significant differences were noted between males and females (χ2(1, N = 393,858) = 15,940.39, p < .001). Male service members presented with more observed cases (n = 363,072) than expected (n = 334,779). Conversely, female service members were diagnosed at lower rates, presenting with fewer observed cases (n = 30,786) than expected (n = 59,078).
Service branch
Statistically significant differences were noted between the military service branches (χ2(3, N = 393,858) = 23,913.07, p < .001). Service members in the Air Force, Navy, and Marine Corps, presented with fewer observed cases (n = 92,292, n = 84,811, and n = 28,472, respectively) than expected (n = 94,925, n = 94,525, and n = 55,140, respectively). Alternatively, service members in the Army, presented with more observed cases (n = 188,283) than expected (n = 149,666).
Insomnia
Race
Statistically significant differences were noted among service members by Race (χ2(2, N = 387,595) = 16,906.75, p < .001). Black service members presented with higher rates of diagnosed cases of insomnia (n = 95,714) than expected (n = 65,891) based on known densities. Minority service members who self-identified as a race other than black presented with (n = 54,623) approximately as many cases as expected (n = 54,263). Conversely, white service members presented with fewer observed cases (n = 237,258) than expected (n = 267,440).
Age
Statistically significant differences were noted by age groups (χ2(5, N = 387,594) = 263,506.46, p < .001). Across the DoD, active duty service members aged <20 years old, 20–24 years old, 25–29 years old, and 30–34 years old, presented with fewer observed cases (n = 5,665, n = 55,808, n = 66,147, and n = 57,970, respectively) than expected (n = 27,405, n = 125,282, n = 90,047, respectively, and 58,726). Service members aged 35 or older (groups: 35–39 and 40+) presented with more observed cases (n = 79,013 and n = 122,991, respectively) than expected (n = 46,981 and n = 39,150, respectively).
Military pay grade
Statistically significant differences, for insomnia diagnoses were also observed within military pay grades (χ2(3, N = 387,573) = 82,727.59, p < .001). Over the course of this study, early career (e.g. junior enlisted [E-1 to E-4] and junior commissioned officers [O-1/WO1 to O-3/CW3]) presented with fewer insomnia diagnoses (n = 95,277 and n = 24,956) than expected (n = 170,532 and n = 38,757). Conversely, senior enlisted personnel (noncommissioned officers [E-5 to E-9]) and senior commissioned officers (O-4/CW4 to O-6/CW5) presented with more observed cases (n = 228,303 and n = 39,037, respectively) than expected (n = 151,153 and n = 27,130, respectively).
Marital status
Statistically significant differences were noted between married and nonmarried military personnel (χ2 (51, N = 387,595) = 47,371.58, p < .001). Married service members presented with more observed cases (n = 280,589) than expected (n = 213,177). Conversely, non-married service members, presented with fewer observed cases (n = 107,006) than expected (n = 174,417).
Sex
Statistically significant differences were noted between males and females (χ2(1, N = 387,595) = 23,335.08, p < .001). Male service members presented with more observed cases (n = 340,198) than expected (n = 329,455). Conversely, female service members presented with fewer observed cases (n = 47,397) than expected (n = 58,139).
Service branch
Statistically significant differences were noted between the military service branches (χ2(3, N = 387,595) = 28,023.00, p < .001). Service members in the Air Force, Marine Corps, and Navy presented with fewer observed cases (n = 89,040, n = 30,607, and n = 75,108, respectively) than expected (n = 93,022, n = 54,263, and n = 93,022, respectively). Alternatively, service members in the Army presented with more observed cases (n = 192,840) than expected (n = 147,286).
Comparison of diagnoses by ICD-9 and ICD-10
Following the conversion from ICD-9 (2005–2015) to ICD-10 codes (2016–2019), significant differences (t(13) = –2.53, p = .01; d = 1.40) were observed in the amount of OSA diagnoses. Similarly, significant differences were observed in the incidence rates of diagnoses of OSA (t(13) = –2.91, p = .006; d = 1.61). There were no significant differences between diagnoses of insomnia (t(13) = 0.08, p = .46; d = 0.04) or incidence rates of insomnia (t(13) = –0.77, p = .22; d = 0.42) following the conversion from ICD-9 to ICD-10 codes. The transition from ICD-9 to ICD-10 resulted in a significant and unexpected decline in OSA diagnoses which did not occur regarding insomnia diagnoses.
Discussion
Between 2005 and 2019, a total of 393,857 new cases of OSA and 385,537 newly diagnosed cases of insomnia, with incidence rates ranging from 11.8 to 333.8 for OSA and 5.7 to 272.4 for insomnia (per 10,000, respectively) occurred throughout the DoD. During the study period, new diagnoses of both sleep disorders markedly increased in every demographic. The increase in OSA and insomnia diagnoses found in this study, which evaluated the largest branches of the armed forces, provide an enhanced understanding and needed overview of the different rates of these two most common sleep disorders in the different branches of the military in the current era, building upon the previous studies by Caldwell et al. [5, 18]. While there is a perception in the lay press that service members may pursue the diagnosis of sleep apnea because of the potential for medical disability compensation [20], our study offers potential insight into this. The findings that insomnia and OSA had similar, marked increases in incidences over the study period even though the Veteran’s Administration does not provide similar medical disability compensation for insomnia, suggests that this is not OSA specific. In fact, it appears that service members have increasing rates of sleep disorders and are seeking appropriate clinical care.
Regarding OSA, service members in the Army had diagnostic rates higher than expected, conversely, service members in the Air Force, Navy, and Marines had rates lower than expected. There are several factors that could potentially contribute to this finding. As compared to the Marines, for example, the Army has the highest percentage of service members who are overweight [21], female representation is lower in the Marines compared to the Army, and the Army has a higher percentage of minority members compared to the other service branches [22].
In addition, soldiers in the Army have ready access to medical resources at Army installations with large military treatment facilities, many of which include sleep centers where screening for sleep disorders is more likely to occur. In contrast, Marines rely on Navy medical assets which may not be present at the base where they are stationed. Additionally, the Army had the first service-wide education program regarding military appropriate sleep practices, as part of the Performance Triad [23]. The enhanced awareness of sleep and sleep disorders, by both soldiers and their medical providers, may also explain the higher incidence in the Army. Overall, these findings align with what is known regarding OSA in the general population. Obesity is a well-established risk factor [3] and men have a higher prevalence of OSA compared to women, particularly under the age of 40 [24]. Minorities have high rates of sleep disorders, to include OSA [25–27], and access to sleep medicine resources is a limiting factor in the diagnosis of sleep apnea [28].
Deployments, which were not evaluated in this study, are associated with an increased risk of both insomnia and OSA [5]. During the period of the study, the number of service members deployed peaked in 2008 with soldiers serving in the Army typically having lengthier deployments to Iraq and Afghanistan at 21 months than those in the other services [29], which ranged from 12 to 16 months. While the length and nature deployments likely contributed to the marked increase in sleep disorders, especially from 2005–2012, other factors may have contributed. For instance, increased rates of comorbid PTSD and TBI [16], which are both associated with insomnia and OSA, as well enhanced screening for, and increased access to sleep medicine resources in the military health system, were other factors that likely contributed to the observed increases in diagnostic rates.
Similar results were found for insomnia diagnoses where service members in the Army had diagnostic rates higher than expected; however, insomnia diagnoses were lower than expected in the other service branches. Previous research has found a high prevalence of insomnia (20%) in service members in the Army [8, 9]. Yet, an earlier study in marines and sailors reported similarly high rates of insomnia symptoms, with 33% having insomnia symptoms [30]. The reasons regarding the higher rates in the Army and lower rates in the other branches examined in our study are not readily apparent.
One prior study analyzed sedative-hypnotic prescriptions for insomnia and reported the majority of prescriptions went to service members in the Army (59.4%), compared to the Air Force (20.5%), Navy (11.1%), and Marines (7.5%) [31]. In conjunction with our study, these findings suggest Army-specific factors contributed to the higher than expected rate of insomnia as opposed to more general, military-related factors such as duty-related aspects (i.e. long work days, shift work, frequent moves, etc.) or deployments which occur in all services. Recent work reported that the service members in the Air Force are the most likely to obtain 7–8 h sleep [32]. This contrasts with sleep duration in service members in the Army, the majority of whom report a sleep duration of ≤6 h [33, 34]. It may be that the cultural views of sleep, especially in the Army, harken back to perceiving sleep as a weakness [35] and thus contribute to the higher rate of insomnia in the Army.
Of additional note is the finding that female service members were relatively less frequently diagnosed with insomnia than their male counterparts. In civilian, as well as veteran populations, women have a higher prevalence of insomnia [36–38]. Currently, there is limited research on the prevalence of insomnia in female service members. Taylor and colleagues [9] found a similar prevalence of insomnia in male Army service members at 19.9% compared to 19.1% in females using the insomnia severity index. With the Brief Insomnia Questionnaire, Klingaman and colleagues [8] reported female service members in the Army had a higher rate of insomnia of 28.4% compared to 22% in their male counterparts. As our findings are from medical diagnoses across the armed services rather than self-report symptoms reported in a sample of Army personnel, we suggest that symptoms in active-duty servicewomen are present but not diagnosed. Further, insomnia is increasingly diagnosed in women veterans [39], thus the lower rate of insomnia diagnoses found in our study suggests that female service members are potentially underdiagnosed for this sleep disorder. This is a potentially alarming finding and one that warrants further study.
The present findings also extend the current literature showing African Americans are at greater risk for OSA [17, 28, 40–42], but conflicts with a previous report indicating African Americans are at a decreased risk for insomnia [43]. We are unaware of any studies evaluating differences in OSA by race in active-duty military personnel. However, Wallace and colleagues [44] reported that African American veterans with OSA had lower positive airway pressure usage and greater difficulties initiating sleep than did Hispanic or white veterans. Previous studies of OSA have noted the younger age and lower body mass indices in active duty military populations [16, 45]. Longstanding OSA that was not clinically suspected while on active duty may have contributed to these findings. The higher rate of OSA diagnoses in African Americans serving in the military further suggests the need for enhanced access to screening and evaluation for OSA in minorities within military, veteran, and civilian populations.
This study extends and further develops the findings of a previous report of incidence rates of OSA and insomnia in active-duty military personnel [17]. However, the present study has several limitations. First, the study relies on accurate diagnoses and accurate ICD input into the service member’s electronic medical record. When examining differences in diagnoses made following the transition from ICD-9 and ICD-10 codes, a very large effect size was found for OSA, but not cases of insomnia. Thus, it could be that as overseas deployments have slowed in recent years, the service members who joined in the early 2000s began to seek care for these concerns around the time of the ICD transition. Alternatively, there is always a concern for clerical errors when working with large data sets [46]. Finally, the nature of the DMED database does not allow researchers to consider comorbid diagnoses or the time to diagnose from the presentation of initial symptoms.
Conclusion
The present data show that since 2005, incidence rates of OSA and insomnia have drastically increased across all branches of the military. This is particularly concerning as military personnel are otherwise healthy individuals and these sleep disorders developed while they were on active duty. The nature of our study does not allow us to determine the etiology of these sleep disorders. However, as the study period covers the longstanding conflicts in Iraq and Afghanistan, deployments and exposures likely contributed to the onset of insomnia, if not sleep apnea as well. Further consideration should be given to interventions targeting appropriate sleep practices, based on military-specific data, as well as enhanced screening for insomnia in female service members and OSA in African American service members.
Disclosure Statements
Financial Disclosure: The authors have no financial disclosures to declare.
Non-financial Disclosure: The authors have no conflict of interest to declare. The views expressed herein are solely those of the authors and do not reflect an endorsement by or the official policy or position of the Department of Defense, the Department of Veterans Affairs, or the U.S. Government.
Conflict of interest statement. Dr. Mysliwiec is a paid consultant for CPAP Medical, Bluegrass Oxygen, NOCTEM, and Sleep Care Inc. He has previously consulted for Ebb Therapeutics and Nightware.
References
Comments