Abstract
Background
Despite the significant health and economic burden associated with OSA among civilians, little is known about this burden among active-duty military personnel.
Research Question
What is the health and utilization burden of OSA among active-duty service members in the United States?
Study Design and Methods
Data were derived from the Military Data Repository (2016-2021). Participants included active-duty service members aged < 65 years with 12 months of continuous enrollment prior to and following a new OSA diagnosis and no evidence of prior OSA or OSA treatment. They were matched 1:1 on demographic, clinical, and military characteristics to those without OSA. OSA and medical and psychiatric comorbidities were defined based on International Classification of Diseases, 10th Revision, codes. The impact of newly diagnosed OSA on psychiatric and medical outcomes was examined by using time-to-event models. The impact on 12-month health care resource utilization was examined by using generalized linear models.
Results
A total of 59,203 service members with OSA were matched to 59,203 service members without OSA. Participants were 83% male and 65% White, with most < 44 years old (81%). OSA was associated with an increased risk for all physical and psychological health outcomes; relative to those without OSA, service members with OSA exhibited a fourfold increased risk for posttraumatic stress disorder (hazard ratio, 4.41; 95% CI, 4.04-4.82). In terms of utilization, OSA was associated with an additional 170,511 outpatient, 66 inpatient, and 1,852 emergency department encounters per year.
Interpretation
Our findings show that among US active-duty military personnel, OSA is associated with substantially increased risk for adverse physical and psychological health outcomes, as well as utilization burden over 12 months. Screening, triage, and treatment efforts could have broad impact in this population.
Key Words: economics, health outcomes, military, sleep, sleep apnea
FOR EDITORIAL COMMENT, SEE PAGE 850
Take-Home Points.
Study Question: What is the health and utilization burden of OSA among active-duty military personnel in the United States?
Results: Relative to matched comparison patients without OSA, military personnel with newly diagnosed OSA exhibited an increased risk for adverse physical and psychological health outcomes as well as dramatically increased health care resource utilization across multiple points of service.
Interpretation: OSA screening, triage, and treatment efforts could have broad impact among active-duty military personnel in the United States.
The modern military operates under an unrelenting work tempo, nontraditional work schedules, and 24-hour worldwide operations. Consequently, insufficient and disturbed sleep are ubiquitous among service members.1,2 At least 42% of active-duty service members (ADSMs) sleep ≤ 5 hours per night, while only one-third obtain the recommended ≥ 7 hours per night for adults.2, 3, 4, 5 In addition to insufficient sleep, clinical sleep disorders such as OSA are prevalent and present significant costs, negatively affecting health and military readiness.1,2,6
Research has highlighted a growing awareness of OSA in the US military health system (MHS). For example, Moore et al7 examined trends in OSA diagnoses among active-duty military personnel from all branches and found a staggering increase in OSA diagnoses between 2005 and 2019 (ie, from 11 to 333 cases per 10,000 personnel). These trends in OSA are supported by other studies among active-duty personnel from the Army8 and all military branches.9 Caldwell et al8 analyzed a comprehensive sample of active-duty Army personnel and found a 600% increase in OSA diagnoses from 2003 to 2011. Furthermore, they identified that comorbid psychological (ie, posttraumatic stress disorder [PTSD]) and physical (ie, hypertension, gastroesophageal reflux disease, diabetes, obesity) conditions mediated the association between deployment history and OSA. Clearly, OSA is a major concern in the US military population.
In civilian populations, OSA is a well-established risk factor for adverse physical (eg, cardiovascular disease,10, 11, 12 diabetes,13, 14, 15 premature death16, 17, 18, 19, 20, 21) and psychological (eg, depression,22, 23, 24, 25, 26 anxiety,27 substance misuse28) health outcomes, as well as an increased economic burden.29,30 From a military perspective, the associations between OSA and trauma, including PTSD and traumatic brain injury (TBI), are especially notable.31, 32, 33, 34, 35, 36 Given the heightened incidence of OSA in the military, it is somewhat surprising that no prior study has quantified the health or utilization burden of OSA within the MHS. Understanding these impacts is vital not only for advancing our knowledge regarding the burden of OSA within a very large health care system but also for informing evidence-based policy and decision-making at both the strategic and operational levels within the military. For example, greater awareness of the burden of OSA within the MHS could support increased access to care via provider education; patient screening, triage, testing, and treatment; expansion of existing clinical programs; and development of new clinical programs, including OSA telehealth, to name several approaches.
To address this important gap in knowledge, the purpose of the current study was to quantify the health and utilization burden of OSA in the MHS from the military perspective. We hypothesized that relative to active-duty military personnel without OSA, individuals with OSA display increased risk for adverse physical and psychological health outcomes, as well as increased health care resource utilization (HCRU).
Study Design and Methods
Data Source
The current study adhered to Strengthening the Reporting of Observational Studies in Epidemiology reporting guidelines. This retrospective cohort study was conducted using data from the Military Data Repository (MDR) for 2016 to 2021. The MDR includes encounter, procedure, pharmacy, and durable medical equipment information for active-duty military personnel, military dependents, National Guard, and Reserves treated within the MHS, including both direct on-base care within military treatment facilities as well as the private sector TRICARE network.
This study used fully de-identified data and received an exempt determination from the Institutional Review Board at the Walter Reed Army Institute of Research (protocol number 2985).
Participants
Two cohorts of US-based ADSMs aged 17 to 64 years were created. The OSA cohort comprised those newly diagnosed with OSA. In addition to active-duty status, we required 24 months of continuous enrollment, including 12 months’ continuous enrollment in TRICARE Prime both prior to and following the date of first OSA diagnosis, which was considered the index date. Next, a comparison cohort was created of beneficiaries without any diagnosis of OSA or insomnia during the study period. For this comparison cohort, a random index date was selected such that there was ≥ 12 months of continuous enrollment both prior to and following that date.
OSA
OSA was defined as receipt of 2 or more International Classification of Diseases, 10th Revision, G47.33 codes in any position on an outpatient or inpatient claim following a 12-month clean period without any prior OSA diagnosis or treatment (ie, positive airway pressure therapy, oral appliance, upper airway surgery). Because an initial “provisional” diagnosis of OSA is often assigned to document the medical necessity for diagnostic sleep testing, we used this relatively conservative operational definition to maximize specificity of our operational definition of OSA.
Outcomes
Outcomes were measured during the first year following the index date and included new diagnoses of physical and psychological health conditions (Table 1) and HCRU. Physical health and psychological health conditions were defined by using codes from the International Classification of Diseases, 10th Revision. All-cause HCRU was measured as total counts of encounters according to point of service (outpatient, inpatient, and emergency department [ED]) over the 12 months following the index date. OSA-specific HCRU was the total counts for OSA-related care. Non-OSA HCRU was calculated as total all-cause encounters minus OSA-related encounters.
Table 1.
Participant Demographic, Military, and Clinical Characteristics at Index Date, Stratified According to OSA Status (N = 118,406)
| Characteristic | Non-OSA |
OSA |
Total |
SMD |
|---|---|---|---|---|
| (n = 59,203) | (n = 59,203) | (N = 118,406) | ||
| Demographic | ||||
| Age, y | 0.036 | |||
| 18-24 | 2,600 (4.4) | 2,719 (4.6) | 5,319 (4.5) | |
| 25-34 | 14,499 (24.5) | 15,278 (25.8) | 29,777 (25.1) | |
| 35-44 | 30,518 (51.5) | 30,173 (51.0) | 60,691 (51.3) | |
| 45-54 | 10,724 (18.1) | 10,209 (17.2) | 20,933 (17.7) | |
| 55-64 | 862 (1.5) | 824 (1.4) | 1,686 (1.4) | |
| Female sex | 4,480 (7.6) | 4,264 (7.2) | 8,744 (7.4) | 0.014 |
| CCI score | 0.076 | |||
| 0 | 54,634 (92.3) | 54,626 (92.3) | 109,260 (92.3) | |
| 1 | 2,751 (4.6) | 3,737 (6.3) | 6,488 (5.5) | |
| ≥ 2 | 549 (0.9) | 749 (1.3) | 1,298 (1.1) | |
| Missing | 1,269 (2.1) | 91 (0.2) | 1,360 (1.1) | |
| Race | 0.012 | |||
| White | 38,258 (64.6) | 38145 (64.4) | 76,403 (64.5) | |
| Black | 11,580 (19.6) | 11610 (19.6) | 23,190 (19.6) | |
| Asian/PI | 5,176 (8.7) | 5,216 (8.8) | 10,392 (8.8) | |
| AI/AN | 817 (1.4) | 767 (1.3) | 1,584 (1.3) | |
| Othera | 3,010 (5.1) | 3,117 (5.3) | 6,127 (5.2) | |
| Unknown | 356 (0.6) | 346 (0.6) | 702 (0.6) | |
| Missing | 6 (0.0) | 2 (0.0) | 8 (0.0) | |
| Ethnicity | 0.017 | |||
| Hispanic | 8,807 (14.9) | 8,526 (14.4) | 17,333 (14.6) | |
| None | 16,363 (27.6) | 16,210 (27.4) | 32,573 (27.5) | |
| Othera | 33,139 (56.0) | 33,538 (56.6) | 66,677 (56.3) | |
| Unknown | 894 (1.5) | 929 (1.6) | 1,823 (1.5) | |
| Service branch | 0.040 | |||
| Army | 22,425 (37.9) | 23,344 (39.4) | 45,769 (38.7) | |
| Air Force | 15,139 (25.6) | 15,210 (25.7) | 30,349 (25.6) | |
| Coast Guard | 3,516 (5.9) | 3,177 (5.4) | 6,693 (5.7) | |
| Marine Corps | 3,513 (5.9) | 3,409 (5.8) | 6,922 (5.8) | |
| NOAA | 12 (0.0) | 14 (0.0) | 26 (0.0) | |
| Navy | 14,163 (23.9) | 13,650 (23.1) | 27,813 (23.5) | |
| PHS | 435 (0.7) | 399 (0.7) | 834 (0.7) | |
| Enlisted | 45,704 (77.2) | 45,806 (77.4) | 91,510 (77.3) | 0.004 |
| Region | 0.225 | |||
| EN Central | 302 (0.5) | 641 (1.1) | 943 (0.8) | |
| ES Central | 1,066 (1.8) | 686 (1.2) | 1,752 (1.5) | |
| Mountain | 1,426 (2.4) | 1,928 (3.3) | 3,354 (2.8) | |
| Pacific | 7,860 (13.3) | 6,719 (11.3) | 14,579 (12.3) | |
| South Atlantic | 15,144 (25.6) | 15,630 (26.4) | 30,774 (26.0) | |
| WS Central | 4,115 (7.0) | 6,294 (10.6) | 10,409 (8.8) | |
| Missing | 29,290 (49.5) | 27,305 (46.1) | 56,595 (47.8) | |
| Physical comorbidities | ||||
| Asthma | 1,778 (3.0) | 2,192 (3.7) | 3,970 (3.4) | 0.039 |
| COPD | 538 (0.9) | 654 (1.1) | 1,192 (1.0) | 0.020 |
| CVDb | 585 (1.0) | 950 (1.6) | 1,535 (1.3) | 0.055 |
| Diabetes | 639 (1.1) | 793 (1.3) | 1,432 (1.2) | 0.024 |
| Fibromyalgia | 9,180 (15.5) | 10,879 (18.4) | 20,059 (16.9) | 0.077 |
| Fracture | 155 (0.3) | 171 (0.3) | 326 (0.3) | 0.005 |
| Hyperlipidemia | 6,963 (11.8) | 8,220 (13.9) | 15,183 (12.8) | 0.064 |
| Hypertension | 71 (0.1) | 126 (0.2) | 197 (0.2) | 0.023 |
| TBI | 456 (0.8) | 680 (1.1) | 1,136 (1.0) | 0.039 |
| Psychological comorbidities | ||||
| Adjustment disorder | 4,599 (7.8) | 6,067 (10.2) | 10,666 (9.0) | 0.087 |
| Alcohol/SUD | 2,367 (4.0) | 2,764 (4.7) | 5,131 (4.3) | 0.033 |
| Anxiety | 5,837 (9.9) | 8,249 (13.9) | 14,086 (11.9) | 0.126 |
| Depression | 2,756 (4.7) | 3,988 (6.7) | 6,744 (5.7) | 0.090 |
| PTSD | 1,486 (2.5) | 2,669 (4.5) | 4,155 (3.5) | 0.109 |
| Other mood disordersc | 6,601 (11.1) | 9,139 (15.4) | 15,740 (13.3) | 0.127 |
Data are presented as No. (%). AI/AN = American Indian or Alaska Native; CCI = Charlson Comorbidity Index; CVD = cardiovascular disease; EN = East North; ES = East South; MSKI = musculoskeletal injury; NOAA = National Oceanic and Atmospheric Administration; PHS = Public Health Service; PI = Pacific Islander; PTSD = posttraumatic stress disorder; SMD = standard mean difference; SUD = substance use disorder; TBI = traumatic brain injury; WS = West South.
Other indicates inclusion in a category other than the primary responses above.
Includes acute myocardial infarction, atrial fibrillation, ischemic heart disease, and peripheral procedures.
Includes bipolar disorder and mood disorders not specified elsewhere.
Covariates
Age, sex, race, ethnicity, service branch (Army, Navy, Air Force, Marines, Space Force, Coast Guard, and Public Health Service), military rank (enlisted vs officer), and region were extracted from the MDR. To maintain privacy, age was measured in categories (18-24, 25-34, 35-44, 45-54, and 55-64 years). Baseline Charlson Comorbidity Index category (0, 1, and ≥ 2) was included. In addition, all physical health and psychological health comorbidities (Table 1) diagnosed during the 12 months before the index date were considered present at baseline.
Analytic Plan
We constructed a 1:1 matched group of beneficiaries without OSA or insomnia. First, a propensity model was constructed including sex, age, race, ethnicity, service branch, military rank, Charlson Comorbidity Index score, and 12-month history of all comorbidities listed in Table 1; this model was used to predict the probability of OSA (propensity score) for each individual. This propensity score was next used to identify the nearest-neighbor individual (non-OSA) for each ADSM with OSA, breaking ties randomly and matching without replacement. To assess the effectiveness of the propensity score matching, a Love plot was constructed. Standardized mean differences (SMDs) representing the postmatching difference in distribution of each characteristic between individuals with and without OSA were used to assess covariate balance. This 1:1 matched cohort was used for all subsequent analyses.
To test the hypothesis that OSA increases risk for adverse physical health and psychological health outcomes, we estimated a series of time-to-event models, 1 for each outcome. For each physical and psychological health outcome, individuals with a preindex diagnosis of that outcome were excluded, along with their corresponding match. We calculated time to event in days from the index date. Individuals were censored at occurrence of the outcome of interest or at 365 days, whichever occurred first. Cox proportional hazards models were estimated, testing the proportional hazards assumption using Schonfield residuals. If the proportional hazards assumption was not met, the Akaike information criterion was used to assess nonparametric time-to-event model specifications (exponential, loglog, and log normal), and the outcome was re-analyzed by using the best alternative specification. All models included any covariates that were not balanced (SMD > 0.2) following matching. For Cox models, hazard ratios (HRs) and 95% CIs are reported. For nonparametric models, time ratios with 95% CIs are reported.
Poisson models were used to test for differences in counts of HCRU between ADSMs with and without OSA. These models included covariates with SMDs > 0.2 following matching and also controlled for the number of encounters of the same type (ie, outpatient, inpatient, ED) during the 12 months prior to the index date. Rate ratios (RtRs) and 95% CIs are reported. To determine the estimated annual HCRU attributable to OSA, we divided the difference in encounters between beneficiaries with OSA and those without OSA by 4 years (ie, the number of years of follow-up examined in this study).
Results
Participants
The final sample included 59,203 ADSMs with OSA meeting study criteria. These individuals were matched as described to 59,203 enrollees without OSA or insomnia. The Love plot of covariate balance prior to and following matchings (e-Fig 1) showed that the matching was effective in improving covariate balance. Characteristics of the matched sample are listed in Table 1. Overall, the sample was 82.6% male and 64.5% White. The majority (77.3%) were enlisted. Common physical comorbidities included fibromyalgia (16.9%) and hyperlipidemia (12.8%). Common psychological comorbidities included anxiety (11.9%) and adjustment disorder (9.0%). All SMDs except region were < 0.2, and thus no characteristics were included as covariates in the models.
Figure 1.
Adjusted HRs based on time-to-event models, reflecting risk for psychological and physical health outcomes based on OSA status. HRs reflect the increased risk associated with OSA for each individual outcome, with higher HRs indicating greater risk. HR = hazard ratio; MSKI = musculoskeletal injury; PTSD = posttraumatic stress disorder; SUD = substance use disorder; TBI = traumatic brain injury.
Health Burden of OSA
In the Cox proportional hazards models, OSA was associated with a significantly increased risk of all physical and psychological outcomes (Table 2). Among physical health outcomes, the greatest increases in risk were observed for TBI (HR, 3.27; 95% CI, 2.78-3.85) and cardiovascular disease (HR, 2.32; 95% CI, 1.99-2.71). Among psychological health outcomes, the greatest increases in risk were observed for PTSD (HR, 4.41; 95% CI, 4.04-4.82), anxiety (HR, 3.35; 95% CI, 3.16-3.55), and other mood disorders (HR, 3.34; 95% CI, 3.16-3.53). The proportional hazards assumption was violated for approximately one-half of the studied outcomes. Thus, we reran these models using alternate specifications as described earlier; the results were consistent with those from the Cox models (e-Table 1). Results are presented in Figure 1.
Table 2.
Summary of TTE Models to Determine the Association of OSA and Psychological and Physical Health Outcomes, including TTE, Adjusted HRs, and 95% CIs
| Outcome | No.a | Non-OSA |
OSA |
HR (95% CI) | ||
|---|---|---|---|---|---|---|
| No. (%) | TTE (d)b | No. (%) | TTE (d)b | |||
| Psychological | ||||||
| Adjustment disorder | 50,566 | 1,302 (2.6) | 159.5 | 3,849 (7.6) | 152 | 3.04 (2.85-3.23) |
| Alcohol/SUD | 55,072 | 927 (1.7) | 167 | 1,589 (2.9) | 151 | 1.73 (1.59-1.87) |
| Anxiety | 47,891 | 1,549 (3.2) | 168 | 4,986 (10.4) | 142 | 3.35 (3.16-3.55) |
| Depression | 53,316 | 921 (1.7) | 166 | 2,671 (5.0) | 151 | 2.95 (2.74-3.18) |
| Other mood disordersc | 46,772 | 1,684 (3.6) | 159 | 5,384 (11.5) | 145 | 3.34 (3.16-3.53) |
| PTSD | 55,297 | 595 (1.1) | 160 | 2,577 (4.7) | 148 | 4.41 (4.04-4.82) |
| Physical | ||||||
| Asthma | 55,747 | 451 (0.8) | 145 | 949 (1.7) | 122 | 2.11 (1.89-2.37) |
| Cardiovascular diseased | 57,769 | 233 (0.4) | 179 | 539 (0.9) | 154 | 2.32 (1.99-2.71) |
| COPD | 58,152 | 296 (0.5) | 175 | 555 (1.0) | 155 | 1.88 (1.63-2.16) |
| Diabetes | 57,868 | 219 (0.4) | 170 | 462 (0.8) | 178 | 2.11 (1.80-2.48) |
| Fibromyalgia | 44,840 | 2,580 (5.8) | 164 | 5,610 (12.5) | 145 | 2.26 (2.16-2.37) |
| Fracture | 58,913 | 61 (0.1) | 176 | 139 (0.2) | 166.5 | 2.10 (1.55-2.85) |
| Hyperlipidemia | 47,918 | 2,070 (4.3) | 159.5 | 3,699 (7.7) | 148 | 1.82 (1.73-1.92) |
| Hypertension | 59,008 | 63 (0.1) | 182 | 113 (0.2) | 129 | 1.79 (1.32-2.44) |
| MSKI | 6,848 | 1,899 (27.7) | 142 | 2,839 (41.5) | 122 | 1.66 (1.57-1.76) |
| TBI | 58,183 | 186 (0.3) | 162.5 | 606 (1.0) | 131 | 3.27 (2.78-3.85) |
HR = hazard ratio; MSKI = musculoskeletal injury; PTSD = posttraumatic stress disorder; SUD = substance use disorder; TBI = traumatic brain injury; TTE = time to event.
Number of matched pairs after excluding pairs with baseline diagnosis of outcome of interest.
Median.
Includes bipolar disorder and mood disorders not specified elsewhere.
Includes acute myocardial infarction, atrial fibrillation, ischemic heart disease, and peripheral procedures.
Utilization Burden of OSA
Table 3 presents the total number of non-OSA-related HCRU encounters in the matched sample based on point of service (outpatient, inpatient, and ED). Relative to those without OSA (n = 59,203), ADSMs with newly diagnosed OSA (n = 59,203) experienced significantly higher non-OSA-related encounters in outpatient (RtR, 1.76; 95% CI, 1.76-1.77), inpatient (RtR, 1.14; 95% CI, 1.07-1.22), and ED (RtR, 1.42; 95% CI, 1.39-1.45) settings. In terms of the overall utilization burden specifically attributable to OSA, the annualized excess numbers of total outpatient, inpatient, and ED encounters were n = 170,511, n = 66, and n = 1,852, respectively. Results are presented in Figure 2.
Table 3.
Total Number of Non-OSA-Related Encounters During the 12 Months Following OSA Diagnosis or Matched Index Date and Adjusted RtRs and 95% CIs of Association Between OSA and Non-OSA-Related Encounters According to Point of Service
| Encounter Type | No. of Encounters |
Effect: RtR (95% CI) | |
|---|---|---|---|
| Non-OSA (n = 59,203) | OSA (n = 59,203) | ||
| Outpatient | 787,496 | 1,469,539 | 1.76 (1.76-1.77) |
| Inpatient | 1,543 | 1,807 | 1.14 (1.07-1.22) |
| ED | 14,969 | 22,375 | 1.42 (1.39-1.45) |
RtRs were estimated by using Poisson models. All models controlled for the number of encounters of the same type (ie, outpatient, inpatient, ED) during the 12 months prior to the index date. All P values < .001. ED = emergency department; HCRU = health care resource utilization; RtR = rate ratio.
Figure 2.
Average number of inpatient, outpatient, and ED encounters over 12 months, based on OSA status. ED = emergency department.
Discussion
To our knowledge, this study represents the largest and most comprehensive analysis to date examining the burden of OSA within the US military. Relative to matched individuals without OSA, ADSMs with OSA had significantly increased risks for adverse physical and psychological health outcomes. They also showed increased HCRU across all points of service, resulting in an additional 170,511 outpatient, 66 inpatient, and 1,852 ED encounters each year. Notably, the associations between OSA and key military-relevant outcomes, including subsequent TBI and musculoskeletal injuries (MSKIs) within 12 months, are novel in the literature. These findings suggest several directions for clinical care, military health policy, and future research.
It is interesting to consider the current results in the context of past literature, particularly as it pertains to military-relevant outcomes. For instance, OSA and PTSD frequently co-occur, and our finding that newly diagnosed OSA was associated with an approximately fourfold greater risk for incident PTSD within 12 months builds upon prior work by quantifying this association among military beneficiaries, who by nature of their work are at increased risk for trauma exposure.31 Similarly, in addition to associations with PTSD, insufficient and disturbed sleep are recognized as important treatment targets to improve outcomes following TBI.32,33 Our findings are among the first to show an association between OSA and subsequent TBI. ADSM with newly diagnosed OSA were nearly 3 times as likely to experience a subsequent TBI and significantly more likely to experience subsequent MSKI within 12 months compared with the non-OSA cohort. A potential mechanism for these associations could be OSA-induced daytime sleepiness, which has been associated with an increased risk for motor vehicle crashes as well as workplace accidents and errors.29,30 Although speculative, it is possible that military personnel (many of whom obtain insufficient sleep) are particularly vulnerable to OSA-related sleepiness. Future research should aim to replicate this finding and determine the strength of the association between OSA and subsequent TBI among civilians and whether OSA treatment can mitigate this risk.
From a clinical perspective, these findings underscore the importance of recognizing OSA as a critical risk factor for a wide array of physical and psychological health outcomes. Not only is OSA often underdiagnosed, but within the MHS there is also a well-documented shortage of trained sleep specialists, which limits access to care. As a result, MHS beneficiaries with OSA are often deferred off-base to the contracted private sector TRICARE network, incurring substantial costs borne by the MHS.37 Our findings suggest a need for increased clinical attention to OSA, including provider education; patient screening, triage, and delivery of evidence-based clinical care; and expansion of existing clinical programs or development of new clinical programs. Given its scalability, telehealth holds particular promise for increasing access to OSA care within the MHS.38, 39, 40 At the same time, in terms of military health policy, as with all modern health systems, the MHS is particularly focused on value as determined by quality and cost of care. Indeed, in the modern health economic climate of increasing costs on the one hand and limited resources on the other, the economic burden of disease and pursuit of cost-effective care have never been more important. Thus, the increased utilization burden associated with OSA will be of significant interest to military health policymakers responsible for ensuring the future health and readiness of military personnel.
Although most OSA research has been conducted among civilians, and a number of studies have examined OSA among military veterans, the current study is the first major effort to quantify the burden of OSA among active-duty military personnel. It is notable that, compared with civilian populations, military personnel tend to be younger, less obese, and with fewer medical comorbidities such as hypertension, cardiovascular disease, and stroke. By contrast, relative to civilians, miliary personnel experience higher levels of psychological health conditions, including PTSD, TBI, and depression. Similarly, evidence identifies MSKI as a major driver of adverse outcomes, including disability, among ADSMs.41
The current study has several strengths. First and foremost, to the best of our knowledge, it is the first study to quantify the health and utilization burden of OSA in the US military health system, providing valuable novel insights not only to inform future research but also to help guide military economic decision-makers regarding allocation of scarce resources. Second, our large sample included 100% of active-duty military personnel meeting eligibility criteria, enabling high generalizability, adequate statistical power, and comprehensive assessment of our military population of interest. Third, rigorous and conservative analytic methods were used throughout.
Nevertheless, our findings must be interpreted in light of several limitations. Most importantly, the MDR does not include clinical characteristics such as OSA severity, objective sleep quality, daytime symptoms including sleepiness, or other clinical variables of interest. This limitation is especially important given differences in OSA pathophysiology between military personnel and civilians. For example, active-duty military personnel with OSA frequently experience the low-arousal threshold OSA phenotype that is characterized by less severe oxygen desaturation and greater sleep fragmentation.42 Future research should incorporate objective sleep characteristics and associated outcomes (eg, via linked data sets). Similarly, our administrative data source does not include behavioral and lifestyle factors (eg, exercise, smoking) that are associated with both OSA as well as psychological and physical health. Future research should include such metrics, as well as anthropometric measures such as BMI, to support evaluation of the healthy user effect.43 Third, our analysis of utilization burden included robust assessment of HCRU across numerous points of service but was limited to the payer/MHS perspective. Future research should assess OSA burden from the patient (eg, health-related quality of life) and employer/military operational (eg, readiness) perspectives.
Regarding sleep in the workplace more broadly, we were unable to assess such factors as job role, shift schedule, and other important employment factors that should be examined in future studies. Fourth, although our large sample includes 100% of the military population of interest, it is unclear how well our findings will generalize to retirees, veterans, or civilians (including military family members) with no military connection. Fifth, as a burden of illness study, our study did not examine the effects of treatment, which should be assessed in future work. Finally, despite our efforts to control for a wide range of potential confounding variables, including demographic, military, and clinical characteristics, residual confounding may still exist. Finally, the observational study design used was unable to determine causality.
Interpretation
Among active-duty military personnel, OSA is associated with dramatically increased risk for adverse psychological and physical health outcomes and increased HCRU. Future research should examine the impact of OSA treatment on improving both health and economic outcomes, including key military-relevant outcomes such as PTSD, TBI, and MSKI, within this vulnerable population.
Funding/Support
This study was funded by the Military Health System Research Program [DoD HT94022210006].
Financial/Nonfinancial Disclosures
The authors have reported to CHEST the following: E. M. W.’s institution has received research funding from the AASM Foundation, Department of Defense, Merck, National Institutes of Health/National Institute on Aging, ResMed, the ResMed Foundation, and the SRS Foundation. E. M. W. has served as a scientific consultant to Axsome Therapeutics, dayzz, Eisai, EnsoData, Idorsia, Merck, Nox Health, Primasun, Purdue, and ResMed; and is an equity shareholder in WellTap. None declared (V. F. C., J. H., B. S., C. T., S. G. W., J. K. W., W. F., T. N., J. S. A.).
Acknowledgments
Author contributions: E. M. W. is the guarantor of the manuscript. J. H. and W. F. had full access to all the data in the study. J. H. takes full responsibility for the integrity and accuracy of the data analyses. Individual contributions were as follows: concept and design, E. M. W. , V. F. C., B. S., J. H., and J. S. A.; acquisition, analysis, or interpretation of data: E. M. W. , V. F. C., B. S., J. H., W. F., and J. S. A.; drafting of the manuscript: E. M. W. , V. F. C., B. S., J. H., and J. S. A.; statistical analysis: J. H.; obtained funding: E. M. W. , V. F. C., J. S. A., and B. S.; administrative, technical, or material support: E. M. W., V. F. C., J. H., C. T., W. F., and J. S. A.; and supervision: E. M. W. , V. F. C., and W. F. All authors performed critical revision of the manuscript for important intellectual content.
Role of sponsors: The sponsor had no role in the design of the study, the collection and analysis of the data, or the preparation of the manuscript.
Other contributions: The authors thank Bernardo G. Buenviaje, PhD, of J-5 for his support of this effort, Allison Y. Hong, MHSA, MPH, of Kennell and Associates for her data analytic support and file creation, and Christine Johnston, BA, of the University of Maryland, Baltimore for her research support.
Additional information: The e-Figure and e-Table are available online under “Supplementary Data.”
Footnotes
A portion of these results were presented at Sleep, September 24-27, Seville, Spain.
Supplementary Data
e-Figure.
References
- 1.Capaldi V.F., Balkin T.J., Mysliwiec V. Optimizing sleep in the military: challenges and opportunities. Chest. 2019;155(1):215–226. doi: 10.1016/j.chest.2018.08.1061. [DOI] [PubMed] [Google Scholar]
- 2.Good C.H., Brager A.J., Capaldi V.F., Mysliwiec V. Sleep in the United States military. Neuropsychopharmacology. 2020;45(1):176–191. doi: 10.1038/s41386-019-0431-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Mysliwiec V., McGraw L., Pierce R., Smith P., Trapp B., Roth B.J. Sleep disorders and associated medical comorbidities in active duty military personnel. Sleep. 2013;36(2):167–174. doi: 10.5665/sleep.2364. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Troxel W.M., Shih R.A., Pedersen E.R., et al. Sleep in the military: promoting healthy sleep among U.S. servicemembers. Rand Health Q. 2015;5(2):19. [PMC free article] [PubMed] [Google Scholar]
- 5.Luxton D.D., Greenburg D., Ryan J., Niven A., Wheeler G., Mysliwiec V. Prevalence and impact of short sleep duration in redeployed OIF soldiers. Sleep. 2011;34(9):1189–1195. doi: 10.5665/SLEEP.1236. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Devine J.K., Collen J., Choynowski J.J., Capaldi V. Sleep disturbances and predictors of nondeployability among active-duty Army soldiers: an odds ratio analysis of medical healthcare data from fiscal year 2018. Mil Med Res. 2020;7(1):10. doi: 10.1186/s40779-020-00239-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Moore B.A., Tison L.M., Palacios J.G., Peterson A.L., Mysliwiec V. Incidence of insomnia and obstructive sleep apnea in active duty United States military service members. Sleep. 2021;44(7) doi: 10.1093/sleep/zsab024. [DOI] [PubMed] [Google Scholar]
- 8.Caldwell J.A., Knapik J.J., Shing T.L., Kardouni J.R., Lieberman H.R. The association of insomnia and sleep apnea with deployment and combat exposure in the entire population of US army soldiers from 1997 to 2011: a retrospective cohort investigation. Sleep. 2019;42(8) doi: 10.1093/sleep/zsz112. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Caldwell J.A., Knapik J.J., Lieberman H.R. Trends and factors associated with insomnia and sleep apnea in all United States military service members from 2005 to 2014. J Sleep Res. 2017;26(5):665–670. doi: 10.1111/jsr.12543. [DOI] [PubMed] [Google Scholar]
- 10.Konecny T., Kuniyoshi F.H.S., Orban M., et al. Under-diagnosis of sleep apnea in patients after acute myocardial infarction. J Am Coll Cardiol. 2010;56(9):742–743. doi: 10.1016/j.jacc.2010.04.032. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Mehra R., Benjamin E.J., Shahar E., et al. Association of nocturnal arrhythmias with sleep-disordered breathing: the Sleep Heart Health Study. Am J Respir Crit Care Med. 2006;173(8):910–916. doi: 10.1164/rccm.200509-1442OC. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Peppard P.E., Young T., Palta M., Skatrud J. Prospective study of the association between sleep-disordered breathing and hypertension. N Engl J Med. 2000;342(19):1378–1384. doi: 10.1056/NEJM200005113421901. [DOI] [PubMed] [Google Scholar]
- 13.Punjabi N.M., Shahar E., Redline S., Gottlieb D.J., Givelber R., Resnick H.E. Sleep-disordered breathing, glucose intolerance, and insulin resistance: the Sleep Heart Health Study. Am J Epidemiol. 2004;160(6):521–530. doi: 10.1093/aje/kwh261. [DOI] [PubMed] [Google Scholar]
- 14.Reichmuth K.J., Austin D., Skatrud J.B., Young T. Association of sleep apnea and type II diabetes: a population-based study. Am J Respir Crit Care Med. 2005;172(12):1590–1595. doi: 10.1164/rccm.200504-637OC. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Drager L.F., Lopes H.F., Maki-Nunes C., et al. The impact of obstructive sleep apnea on metabolic and inflammatory markers in consecutive patients with metabolic syndrome. PLoS One. 2010;5(8) doi: 10.1371/journal.pone.0012065. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Marshall N.S., Wong K.K., Cullen S.R., Knuiman M.W., Grunstein R.R. Sleep apnea and 20-year follow-up for all-cause mortality, stroke, and cancer incidence and mortality in the Busselton Health Study cohort. J Clin Sleep Med. 2014;10(4):355–362. doi: 10.5664/jcsm.3600. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Punjabi N.M., Caffo B.S., Goodwin J.L., et al. Sleep-disordered breathing and mortality: a prospective cohort study. PLoS Med. 2009;6(8) doi: 10.1371/journal.pmed.1000132. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Cowie M.R., Linz D., Redline S., Somers V.K., Simonds A.K. Sleep disordered breathing and cardiovascular disease: JACC State-of-the-Art Review. J Am Coll Cardiol. 2021;78(6):608–624. doi: 10.1016/j.jacc.2021.05.048. [DOI] [PubMed] [Google Scholar]
- 19.Redline S., Azarbarzin A., Peker Y. Obstructive sleep apnoea heterogeneity and cardiovascular disease. Nat Rev Cardiol. 2023;20(8):560–573. doi: 10.1038/s41569-023-00846-6. [DOI] [PubMed] [Google Scholar]
- 20.Mehra R., Chung M.K., Olshansky B., et al. Sleep-disordered breathing and cardiac arrhythmias in adults: mechanistic insights and clinical implications: a scientific statement from the American Heart Association. Circulation. 2022;146(9):e119–e136. doi: 10.1161/CIR.0000000000001082. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Yeghiazarians Y., Jneid H., Tietjens J.R., et al. Obstructive sleep apnea and cardiovascular disease: a scientific statement from the American Heart Association. Circulation. 2021;144(3):e56–e67. doi: 10.1161/CIR.0000000000000988. [DOI] [PubMed] [Google Scholar]
- 22.Wickwire E.M., Albrecht J.S. Occult, undiagnosed obstructive sleep apnea increases risk of depression among a national sample of older adult Medicare beneficiaries. Sleep Health. 2023;9(4):532–536. doi: 10.1016/j.sleh.2023.05.008. [DOI] [PubMed] [Google Scholar]
- 23.Sharafkhaneh A., Giray N., Richardson P., Young T., Hirshkowitz M. Association of psychiatric disorders and sleep apnea in a large cohort. Sleep. 2005;28(11):1405–1411. doi: 10.1093/sleep/28.11.1405. [DOI] [PubMed] [Google Scholar]
- 24.Gupta M.A., Simpson F.C. Obstructive sleep apnea and psychiatric disorders: a systematic review. J Clin Sleep Med. 2015;11(2):165–175. doi: 10.5664/jcsm.4466. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Haddad J.M., Chen E.S. Identifying psychiatric comorbidities for obstructive sleep apnea in the biomedical literature and electronic health record. AMIA Jt Summits Transl Sci Proc. 2017;2017:502–511. [PMC free article] [PubMed] [Google Scholar]
- 26.Stubbs B., Vancampfort D., Veronese N., et al. The prevalence and predictors of obstructive sleep apnea in major depressive disorder, bipolar disorder and schizophrenia: a systematic review and meta-analysis. J Affect Disord. 2016;197:259–267. doi: 10.1016/j.jad.2016.02.060. [DOI] [PubMed] [Google Scholar]
- 27.Garbarino S., Bardwell W.A., Guglielmi O., Chiorri C., Bonanni E., Magnavita N. Association of anxiety and depression in obstructive sleep apnea patients: a systematic review and meta-analysis. Behav Sleep Med. 2020;18(1):35–57. doi: 10.1080/15402002.2018.1545649. [DOI] [PubMed] [Google Scholar]
- 28.Xu C., Acevedo P., Wang L., et al. Sleep apnea and substance use disorders associated with co-occurrence of anxiety disorder and depression among US adults: findings from the NSDUH 2008-2014. Brain Sciences. 2023;13(4):661. doi: 10.3390/brainsci13040661. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Wickwire E.M., Albrecht J.S., Towe M.M., et al. The impact of treatments for OSA on monetized health economic outcomes: a systematic review. Chest. 2019;155(5):947–961. doi: 10.1016/j.chest.2019.01.009. [DOI] [PubMed] [Google Scholar]
- 30.American Academy of Sleep Medicine . 2016. Hidden health crisis costing America billions. Underdiagnosing and undertreating obstructive sleep apnea draining healthcare systemMountain View, CA: Frost & Sullivan. [Google Scholar]
- 31.McCall C.A., Watson N.F. A narrative review of the association between post-traumatic stress disorder and obstructive sleep apnea. J Clin Med. 2022;11(2):415. doi: 10.3390/jcm11020415. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Wickwire E.M., Schnyer D.M., Germain A., et al. Sleep, sleep disorders, and circadian health following mild traumatic brain injury in adults: review and research agenda. J Neurotrauma. 2018;35(22):2615–2631. doi: 10.1089/neu.2017.5243. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Wickwire E.M., Williams S.G., Roth T., et al. Sleep, sleep disorders, and mild traumatic brain injury. What we know and what we need to know: findings from a National Working Group. Neurotherapeutics. 2016;13(2):403–417. doi: 10.1007/s13311-016-0429-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Gold J.M., Gano A.L., McKinney G.L., Wickwire E.M. Primary care management of sleep disturbances associated with concussion/mild traumatic brain injury in service members and veterans. Prim Care Companion CNS Disord. 2024;26(5):23nr03691. doi: 10.4088/PCC.23nr03691. [DOI] [PubMed] [Google Scholar]
- 35.Leng Y., Byers A.L., Barnes D.E., Peltz C.B., Li Y., Yaffe K. Traumatic brain injury and incidence risk of sleep disorders in nearly 200,000 US veterans. Neurology. 2021;96(13):e1792–e1799. doi: 10.1212/WNL.0000000000011656. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Haynes Z.A., Stewart I.J., Poltavskiy E.A., et al. Obstructive sleep apnea among survivors of combat-related traumatic injury: a retrospective cohort study. J Clin Sleep Med. 2022;18(1):171–179. doi: 10.5664/jcsm.9530. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Eliason M., Jardine D., Pelchy K., Meyer C., McIntyre N. Sleep-disordered breathing in the active-duty military population and its civilian care cost. Fed Pract. 2017;34(2):32–36. [PMC free article] [PubMed] [Google Scholar]
- 38.Abdelwadoud M., Collen J., Edwards H., et al. Engaging stakeholders to optimize sleep disorders' management in the U.S. military: a qualitative analysis. Mil Med. 2022;187(7-8):e941–e947. doi: 10.1093/milmed/usab341. [DOI] [PubMed] [Google Scholar]
- 39.Wickwire E.M., Abdelwadoud M., Collen J., et al. Active duty service members, primary managers, and administrators' perspectives on a novel sleep telehealth management platform in the U.S. military healthcare system. Mil Med. 2022;187(9-10):e1201–e1208. doi: 10.1093/milmed/usac006. [DOI] [PubMed] [Google Scholar]
- 40.Wickwire E.M., Collen J., Capaldi V.F., et al. Virtual first: implementation of a novel sleep telehealth platform in the United States military. Front Sleep. 2024;3 [Google Scholar]
- 41.Molloy J.M., Pendergrass T.L., Lee I.E., Chervak M.C., Hauret K.G., Rhon D.I. Musculoskeletal Injuries and United States Army Readiness Part I: overview of injuries and their strategic impact. Mil Med. 2020;185(9-10):e1461–e1471. doi: 10.1093/milmed/usaa027. [DOI] [PubMed] [Google Scholar]
- 42.Mysliwiec V., Pruiksma K.E., Matsangas P., et al. Sex differences in US military personnel with insomnia, obstructive sleep apnea, or comorbid insomnia and obstructive sleep apnea. J Clin Sleep Med. 2024;20(1):17–30. doi: 10.5664/jcsm.10774. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Platt A.B., Kuna S.T., Field S.H., et al. Adherence to sleep apnea therapy and use of lipid-lowering drugs: a study of the healthy-user effect. Chest. 2010;137(1):102–108. doi: 10.1378/chest.09-0842. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.