Persistent behavioral, biological, and physiological sequelae in active-duty treatment-seeking U.S. special operations forces personnel: a cross-sectional study

Summary

Background

The breadth and constellation of sequelae in U.S. special operations forces (SOF) are underrepresented in the literature, especially following multiple traumatic brain injuries (TBIs), musculoskeletal injuries, and chronic stress. This study used multimodal markers of behavioral, biological, and physiological symptoms to evaluate 1) the breadth of sequelae experienced in this population, 2) their symptom structure, and 3) differential responses across participants that may improve clinical practice.

Methods

Participants (n = 222) included U.S. active-duty SOF seeking treatment for TBI-related sequelae, musculoskeletal injuries, and related comorbidities. Cross-sectional data included 31 variables across 17 subjective and objective measures. Symptom structure was evaluated using exploratory factor analysis (EFA). Differential symptom responses were evaluated using latent profile analysis (LPA) and latent class analysis (LCA).

Findings

Cardiometabolic (82.4%), pain (83.3%), cognitive (77.5%), endocrine (39.2%), headaches (80.2%), mental health (81.1%), sensory (73.4%), and sleep (92.8%) symptoms were common. Five factors were identified that accounted for 62.8% of the total variance in symptoms. One factor accounted for the most variance (23.2%) and included a broad reaction across multiple health domains. Two latent profiles were identified with 32.4% of the sample experiencing cumulative and global disruptions in pain, cognition, headaches, mental health, sensory function, and sleep. Dysphoric arousal (49.5%) was the predominant mental health phenotype, with fewer participants presenting with traditional forms of anxiety, depression, and PTSD.

Interpretation

This study provides emerging evidence of a cumulative and general burden of persistent posttraumatic behavioral, biological, and physiological sequelae in U.S. SOF personnel and provide insights into the differential experiences of these sequelae.

Funding

This study was not funded.

Research in context.

Evidence before this study

Despite increased risk for adverse health outcomes in U.S. special operations forces (SOF) compared to conventional military personnel, the prevalence, breadth, and integrative constellation of these symptoms remain poorly understood and underrepresented in the literature. We searched PubMed and Google Scholar between February 5, 2024 and April 17, 2026. Search terms included, “special operations,” “special forces,” “sequelae,” “health,” “suicide,” “injury,” “mental health.” We included qualitative and quantitative studies written in English. Studies identified largely evaluated U.S. special operations forces men with fewer studies evaluating non-U.S. special operations forces and women. Fewer than 50 studies presented health data from this discreet active-duty SOF population. Concurrently, this literature search yielded 15 studies regarding casualties, suicide, and suicidal behaviors in SOF, illustrating a critical gap between understanding the breadth of sequelae within this population and those sequalae's implicit impact on unmet healthcare needs and risk of deleterious health outcomes, including mortality. Some of the sequelae that were reported in SOF included behavioral (e.g., mental health, sleep), biological (e.g., endocrine), cognitive, and physical (e.g., pain, headaches, sensory dysregulation) symptoms. However, these sequelae have only been examined in isolation, as distinct rather than related outcomes. Despite the paucity of information. The quality of the literature search appeared adequate and the risk of bias appeared low.

Added value of this study

This study provides the most comprehensive and integrative examination of behavioral, biological, and physiological sequelae in active-duty and treatment-seeking U.S. SOF personnel. Accordingly, this study addresses a critical gap in the literature by approximating the breadth of multimodal symptoms experienced and identifying person-centered phenotypes that may be helpful in targeting clinical practices within this and related populations. This study also provides the first empirical examination of the emerging construct of “operator syndrome.” Altogether, this study extends previous clinical studies in SOF and evaluates seemingly disparate sequelae as potentially related sequelae using an integrative variable- and person-centered methodological approaches.

Implications of all the available evidence

Findings provide preliminary evidence of cumulative and systemic adverse health effects across multiple health systems. This study presents emerging evidence of persistent, cumulative, and systemic sequelae that may be reflective of “operator syndrome” and that may be underrecognized in current research and clinical practices. It also provides evidence of differential experiences of these symptoms within the SOF population, with some experiencing more and worse symptoms than others. The fact that the full scope of these sequelae was so prevalent, related across different health domains, and structurally cohesive across separate analyses is notable for this SOF population and those with similar exposures. In addition, this study provides an empirical blueprint future studies can employ to help replicate and extend these findings. Overall, current findings may help explain the sequelae experienced by SOF and other personnel with similar exposures following multiple mild TBIs, musculoskeletal injuries, traumatic stress, and increased allostatic load to inform more accurate, comprehensive, and effective clinical assessments and interventions.

Introduction

Following repetitive traumatic brain injury (TBI), multiple musculoskeletal injuries, and chronic exposures to high-tempo environments, and accompanied stress, United States (U.S.) active-duty special operations forces (SOF) and its operators carry a largely undisclosed and cumulative burden of related health consequences.1, 2, 3, 4 Despite increased reliance over approximately 20 years of the global war on terrorism (GWOT), ongoing operations (e.g., direct action, hostage rescue, reconnaissance, counterterrorism, counterinsurgency, civil affairs),⁵ growing public notoriety, and even adoration, little is known about the cumulative health burden SOF personnel face following unique occupational exposures.⁶ A cursory literature search yields fewer than 50 studies presenting quantitative health information from this discreet active-duty SOF population.1, 2, 3, 4^,7, 8, 9 Concurrently, a similar literature search yields at least 15 studies regarding casualties, suicide, and suicidal behaviors among SOF personnel, illustrating a critical gap between understanding the breadth of sequelae within this population and those sequalae's implicit impact on unmet healthcare needs and risk of deleterious health outcomes, including mortality.10, 11, 12, 13

The sequelae reported in SOF include behavioral (e.g., mental health, sleep), biological (e.g., endocrine), cognitive, and physical (e.g., pain, headaches, sensory dysregulation) symptoms stemming from unique occupational exposures and consequential complex injuries.1, 2, 3, 4^,7,8,14 The broader constellation of these symptoms has been described under the hypothesized construct of “operator syndrome”.^1,4 However, these sequelae have largely been examined in isolation, as distinct rather than related outcomes. Despite increased risk for adverse health outcomes in SOF personnel compared to conventional military personnel, the prevalence, breadth, and integrative constellation of these symptoms remain poorly understood.^2,15

The proposal of unique syndromes following war is not necessarily new.^14,16, 17, 18, 19 Advances in warfare (e.g., munitions, drones, improvised explosive devices) can produce previously unexplained symptoms without a known or singular cause, requiring adaptations to clinical practices.^18,19 For these reasons, the Polytrauma System of Care (PSC) was created during the course of GWOT within U.S. Veterans Health Administration (VHA) to develop novel, integrated, and increasingly effective treatments for complex patterns of multiple injuries, including TBI, other bodily injuries (e.g., musculoskeletal, amputation), and related physical, sensory, cognitive, or psychological symptoms.²⁰ In the case of active-duty SOF and related personnel (e.g., conventional military, first-responders), these symptoms are not necessarily new (e.g., persistent post-concussive symptoms, posttraumatic stress disorder [PTSD]-related effects).^14,21, 22, 23, 24, 25, 26, 27, 28 However, their seemingly common constellation in the treatment-seeking SOF population and complex interactions across multiple health domains may tax clinical decision-making in traditional discipline-specific settings (e.g., primary care, pain clinic, PTSD clinic) that may target one particular health domain without evaluating or treating others.^1,2 Although practitioners in these specialty clinics may appropriately evaluate and treat the relevant presenting problem (e.g., tendinitis, PTSD, hypertension), they may not evaluate or have the capacity to evaluate or treat the full spectrum of systemically-related conditions common to SOF personnel. These patients may benefit further from an integrative evaluation and treatment approach such as those offered within PSC.²⁰

Current study

This study used an integrative and transdiagnostic approach²⁷ to 1) evaluate the breadth and constellation of sequelae experienced by U.S. active-duty and treatment-seeking SOF personnel in a PSC program using multimodal markers of behavioral, biological, and physiological symptoms, 2) evaluate the latent factor and profile structure of these symptoms, and 3) identify latent mental health symptom presentations or phenotypes that may be helpful in targeting clinical practices within this population. It was hypothesized that 1) multiple markers/symptoms across all domains would be elevated; 2) at least two factors would be identified that included one factor with multiple elevated symptoms across all domains and at least one other factor that included more specific problem areas; and 3) at least two differential latent profiles and classes would be identified to capture unique experiences of these sequelae and mental health symptoms, in particular. This study provides a rare examination of health-related sequelae within an historically discreet active-duty SOF population to better understand the nuances of complex multiple injuries and to identify subpopulations that may require further integrated and personalized treatments.^27,29

Methods

Participants and procedures

Participants included active-duty SOF patients who participated in VHA Palo Alto's Intensive Evaluation and Treatment Program (IETP) between 2021 and 2025. IETP is a three-week residential rehabilitation program housed within PSC. To participate in IETP, patients must be active-duty or veteran SOF and have a history of at least mild TBI related to military duty. To date, only one veteran has completed IETP and was excluded to attain an active-duty sample. TBI was determined to be present when there was a plausible biomechanical mechanism of injury (e.g., head impact, proximal blast overpressure) and either neuroimaging evidence and/or clinical symptoms of brain injury, including loss of consciousness, posttraumatic amnesia, or altered mental status (e.g., dazed, disoriented, confused).³⁰ Mild TBI was defined as loss of consciousness ≤30 min, amnesia ≤24 h, or altered mental status ≤24 h.^30,31 Participants self-selected for participation in IETP for several potential reasons (e.g., cognitive difficulties, pain, headaches, mood dysregulation). To help ensure that SOF personnel who want treatment can receive it in IETP, the only inclusion criterion was a history of TBI. They were then evaluated for TBI history both by a military medical provider and IETP medical providers. Considering many SOF personnel incur significant brain injuries during the course of their career, no participant that has applied to IETP has been denied admission because a TBI was not documented.

Participants received interdisciplinary treatment from medical, psychology, neuropsychology, social work, nursing, occupational therapy, recreation therapy, physical therapy, speech pathology, nutrition, and other consults as needed (e.g., pain management, endocrinology, audiology). Measures were collected at baseline upon admission to IETP. Fasting venous blood serum samples were obtained by venipuncture between 0700 and 0800. Participants provided consent during pre-admission procedures for data collection and deidentified use of their data for research purposes. Study procedures were approved by VHA-affiliated Stanford University Institutional Review Board (Protocol #17671).

At the time of this article submission, 266 active-duty SOF patients have completed IETP at VHA Palo Alto. The first 44 participants of the program were excluded because their participation preceded implementation of current measures. This produced a final sample of 222 participants. Examination of participants excluded (n = 44) and included (n = 222) indicated no significant differences in anxious, depressive, posttraumatic stress, sleep, nor neurobehavioral symptoms, suggesting minimal selection bias in this final sample.

Participants were 98.6% male (3 female) aged between 25 and 56 years (Mean = 40.55) who included enlisted personnel (n = 163, 73.4%), warrant officers (n = 18, 8.1%), and officers (n = 41, 18.5%; see Table 1). Participants included personnel from the Navy Sea, Air, and Land (SEAL) Teams (n = 102, 45.9%), Navy Special Warfare Combat Crewman (SWCC; n = 26, 11.7%), Army Special Forces (n = 84, 37.8%), and other SOF personnel (n = 10, 4.5%). Military ranks ranged from E-5 to O-8, with most enlisted participants being E-7 or above (76.6%) and most officers being O-5 and above (63.4%). Sex, race, and ethnicity was queried during clinical interviews and categorized based on current U.S. Office of Management and Budget guidelines. Participants were 84.7% White (n = 188), 5.9% (n = 13) Hispanic, 5.3% (n = 12) Asian, 1.8% (n = 4) Black, 1.4% (n = 3) Native American, and 0.9% (n = 2) who identified as Other.

Table 1.

Demographics and sample characteristics (n = 222).

Variable	n	%
Age	M = 40.54	SD = 5.8
Sex
Female	3	1.35
Male	219	98.65
Race/Ethnicity
White	188	84.68
Hispanic	13	5.86
Asian	12	5.41
Black	4	1.80
Native American	3	1.35
Other	2	.90
Occupation
Navy SEAL	102	45.95
Navy SWCC	26	11.71
Army SF	84	37.84
Other SOF	10	4.50
Military rank
Enlisted	163	73.42
Warrant Officer	18	8.11
Officer	41	18.47

Note. M = mean; SD = standard deviation; SEAL = Sea, Air, and Land; SWCC = Special Warfare Combat Crewman; SF = Special Forces; SOF = special operations forces.

Measures

Thirty-one variables across 17 measures were collected at baseline using semi-structured clinical interview and self-reports (see Table 2 for measures and criteria used for cutoff scores). All symptoms were anchored and evaluated in relation to military exposures (e.g., TBI, psychological trauma, musculoskeletal injury). TBI was assessed using qualitative clinical interviews. Therefore, quantitative data including timing, frequency, and severity of TBIs were not available. A combination of both subjective (e.g., self-reports, perceived cognitive performance, pain perception) and objective (e.g., lipids panel, endocrine panel, cognitive testing) assessments were collected by practitioners from each discipline (e.g., psychology, physical therapy, medical). Please see Supplemental Materials for a description of measures. Notably, Bertec's Computerized Dynamic Posturography (CDP) was used to perform the Sensory Organization Test (SOT) to assess somatosensory, visual, and vestibular function.³² Bertec's CDP uses immersive virtual environments with dual-balance force plate technology to supplement assessment and targeted physical therapy interventions for participants experiencing dizziness, balance problems, and/or motion sensitivity. In addition, non-specific pain and somatization, pain interference, and pain related to psychological states were assessed.

Table 2.

Symptom domains, measures, and descriptives (n = 222).

Domain and measure	Score Mean, Median (SD)	Criteria met n (%)
Cardiometabolic disruption		183 (82.4%)
Body Mass Index (BMI; ≥30)	29.2, 28.7 (3.5)	77 (34.7%)
Systolic Blood Pressure (≥130)	131.6, 131.0 (12.4)	122 (55.0%)
Diastolic Blood Pressure (≥85)	82.1, 82.5 (8.3)	86 (38.7%)
Cholesterol Total (≥200 mg/dL)	201.6, 203.5 (38.8)	83 (37.4%)
Glucose Blood (≥100 mg/dL)	95.5, 95.0 (7.4)	60 (27.0%)
High-Density Lipoprotein (HDL) Cholesterol (≤35 mg/dL)a	51.6, 51.0 (11.3)	9 (4.1%)
Low-Density Lipoprotein (LDL) Cholesterol (≥130 mg/dL)	127.0, 125.5 (33.9)	66 (29.7%)
Triglycerides (≥150 mg/dL)	126.0, 107.0 (75.5)	34 (15.3%)
Chronic localized and non-specific pain		185 (83.3%)
Non-Specific Posttraumatic Somatic Distress (NSPSD; ≥6)	14.8, 14.0 (5.7)	145 (65.3%)
Pain Catastrophizing Scale (PCS; ≥30)	16.3, 14.0 (11.6)	34 (15.3%)
PROMIS Pain Interference (≥60)	61.7, 62.1 (6.4)	148 (66.7%)
Cognitive disruption		172 (77.5%)
Neuro-QoL (≤40)a	39.2, 39.0 (6.6)	132 (59.5%)
TOMAL Digits Forward (≤10)a	10.4, 11.0 (2.7)	28 (12.6%)
TOMAL Paired Recall (≤10)a	9.3, 10.0 (2.5)	43 (19.4%)
TOMAL Word Selective Reminding (WSR; ≤10)a	8.7, 9.0 (2.7)	61 (27.5%)
TOMAL WSR Delayed Recall (WSRD; ≤10)a	9.6, 10.0 (2.3)	31 (14.0%)
Endocrine disruptionb		87 (39.2%)
Follicular Stimulating Hormone (FSH; <.95 or >11.95 mIU/mL)a	4.1, 3.8 (2.8)	21 (9.5%)
Luteinizing Hormone (LH; <.6 or >12.1 mIU/mL)a	3.4, 3.2 (2.0)	18 (8.1%)
Testosterone (T; <400 ng/dL)a	570.5, 511.0 (233.5)	62 (27.9%)
Thyroid Stimulating Hormone (TSH; <.4 or >5.0 uIU/nL)a	1.9, 1.7 (.9)	3 (1.4%)
Headaches		178 (80.2%)
Headache Impact Test (HIT-6; ≥50)	57.9, 58.0 (8.4)	101 (45.5%)
Mental health disruption		180 (81.1%)
General Anxiety Disorder (GAD-7; ≥14)	10.7, 10.0 (5.3)	152 (68.5%)
Patient Health Questionnaire (PHQ-9; ≥10)	11.1, 11.0 (5.4)	159 (71.6%)
Posttraumatic Stress Disorder Checklist (PCL-5; ≥31)	32.6, 31.0 (18.5)	114 (51.4%)
Sensory disruption		163 (73.4%)
Dizziness Handicap Inventory (DHI; ≥10)	21.3, 17.0 (18.5)	91 (41.0%)
Total Sensory Organization Test (SOT; ≤70)a	67.9, 73.0 (14.4)	62 (27.9%)
SOT Somatosensory (≤90)a	94.1, 97.0 (8.8)	25 (11.3%)
SOT Vestibular (≤55)a	59.2, 67.0 (21.9)	46 (20.7%)
SOT Visual (≤74)a	75.1, 81.0 (19.3)	42 (18.9%)
Speech, Spatial, and Qualities of Hearing Scale (SSQ-12; ≤79)a	67.0, 66.0 (17.8)	140 (63.1%)
Sleep disruption		206 (92.8%)
Insomnia Severity Index (ISI; ≥8)	16.4, 17.0 (6.1)	206 (92.8%)

Note. Significance values and reference ranges shown in parentheses next to measures. Participants were classified as present on a domain if at least one of measure in that domain was present.

PROMIS = Patient-Reported Outcomes Measurement Information System; TOMAL = Test of Memory and Learning.

^aHigher numbers generally reflect desirable outcome.

^bExaminations of endocrine variables excluded 3 female due to sex differences.

Statistical analyses

Descriptive analysis and exploratory factor analysis (EFA) were conducted in SPSS Version 30. Missing data was less than 2.2% (n ≤ 5) across all variables. Although EFA used listwise deletion, latent profile analysis (LPA) and latent class analysis (LCA) used full-information maximum likelihood estimation to account for missing data. Studies of listwise deletion and imputation methods with simulated and real data demonstrate that these methods are effective when <5% of the data are missing as in this study.³³ Moreover, principal axis EFA is recommended when samples are smaller than 300 participants.³⁴ Variables were z-score adjusted to help limit the influence of skewness and kurtosis on subsequent statistical models. Participants met domain criteria if ≥1 measure from each domain surpassed the relevant cutoff score, indicating clinical significance.

Continuous z-score adjusted transdiagnostic variables, reflecting symptoms, were evaluated using a combination of variable-centered (EFA) and person-centered (LPA, LCA) statistical approaches that integrate multimodal data sources (i.e., self-report scales, vestibular testing, vital markers, biomarkers) to identify differential symptom responses within a broader set of heterogeneous sequelae (see Adams et al., 2023 for Middle Out Approach). EFA, LPA, and LCA provide complementary information regarding variable cohesion and unique individual variations within these broad variables, respectively.

EFA is a variable-centered statistical technique used to determine if variables from the selected domains covary together, potentially reflecting a shared underlying construct (factor), and, if so, what variables may be most relevant. Parallel analysis and Velicer's MAP criteria were used to generate random and observed datasets and eigenvalues to indicate the optimal number of factors to retain for the EFA considering several parameters including the number of variables, sample size, sampling error, and variable parameters.³⁵ The combination of these factor retention methods in addition to visual inspection of the scree plot is considered one of the most accurate methods for estimating optimal factor structures.^36,37 Principal axis EFA was then conducted to evaluate a five-factor structure on 31 variables. Because intercorrelations between symptom factors were expected, direct oblimin oblique rotation was used to allow modeling of intercorrelations between variables and factors.³⁴ Principal axis EFA is recommended for estimating nuanced communalities/factors between variables with samples <300 participants.³⁴ Kaiser-Meyer-Olkin (KMO) and Bartlett's test of sphericity were used to indicate adequate sample size and sufficient correlation between variables.³⁸ Factor loadings ≥.26 were considered clinically and statistically salient as recommended by the formula established by Norman & Streiner, 2014.³⁴

LPA and LCA were conducted in Mplus 8.11. LPA and LCA are person-centered statistical techniques used to determine if variables covary uniquely within homogenous subgroups of participants of a broader heterogeneous sample. Resulting latent profiles and classes reflect potential unique experiences of the related symptoms. Although LPA is applied to continuous data, LCA is applied to categorical data to help draw out unique and differential experiences of different types of symptoms, not just severity of symptoms. LCA and LPA used full information maximum likelihood (FIML) estimation to estimate missing data. Fifty random starts were used and FIML was replicated at least 20 times. Lower fit indices (Bayesian information criterion [BIC], sample-size adjusted BIC [ssBIC], Akaike information criterion [AIC]), higher entropy values, and statistically significant (p < .05) Lo–Mendell–Rubin adjusted likelihood ratio test (LMR-LRT) and bootstrapped likelihood ratio test (BLRT) indicated better model fit.³⁹ A combination of these indices, sufficient class size, theoretical coherence, parsimony, and clinical utility was used to select an optimal model.^39,40 LCA34, 35, 36, 37 was used to determine if unique mental health phenotypes could be identified that can be used to target clinical practices.^39,40 To address issues of conditional dependence between anxiety, depressive, and posttraumatic stress symptoms, the local independence assumption was relaxed by estimating parameters considering residual associations between indicators (parameterization = “RESCOV”) and more random starts.⁴¹

Role of the funder

There was no direct funding source for this study.

Results

All participants had a history of ≥mild TBI. Suicidal ideation (SI) was evaluated both during clinical interviews and with the PHQ-9 item 9. SI, indicated as a PHQ-9 item 9 score ≥1, was relatively uncommon and present for 5.9% (n = 13) of participants. Although quantitative data regarding onset of symptoms was not available in this study, qualitative reports collected during clinical interviews suggested that most participants experienced the onset of symptoms 1–15 years before their assessment in this study.

Symptom structure

Descriptives are reported in Table 2. Sleep (92.8%), pain (83.3%), cardiometabolic (82.4%), mental health (81.1%), headaches (80.2%), and cognitive (77.5%) disruptions were most prevalent. See Supplemental Table S1 for bivariate correlation between the 31 z-score adjusted variables. Variables were correlated, as expected, especially within-domain.

Variable-centered approach

EFA was performed on 31 z-score-adjusted variables of behavioral, biological, and physiological function using direct oblimin rotation. KMO = .683 indicated sufficient statistical power. Bartlett's test of sphericity (p < .001) suggested sufficient correlation between variables.³⁸ Parallel analysis, Velicer's MAP criteria, and inspection of the scree plot indicated a five-factor structure was optimal, which accounted for 62.84% of the variance in total variable outcomes (see Table 3 and Fig. 1 for factor loadings). Factor 1 accounted for the most variance (23.15%) and included significant (loading≥|.26|) variables from all domains. Variables with the highest Factor 1 loadings included greater mental health disruptions (.797–.880; GAD-7, PHQ-9, PCL-5); greater headaches (.746; HIT-6); greater pain (.694–.730; NSPSD, PCS, PROMIS), including non-specific pain and somatization, pain interference, and pain related to psychological states; dizziness (.677; DHI), lower hearing ability (−.665; SSQ-12); lower perceived cognitive performance (−.676; Neuro-Qol); greater sleep disruptions (.646; ISI); and lower somatosensory function (−.394 to −.665; SOT, Vestibular, Visual); with smaller but associated effects on HDL cholesterol (−.265) and FSH levels (−.264).

Table 3.

Exploratory factor analysis of z-score adjusted variables (n = 222).

Domain and measure	Factor
	1	2	3	4	5
% Accounted variance	23.15%	11.22%	9.66%	7.78%	5.82%
Cardiometabolic disruption
BMI	.064	.196	.347	−.194	.086
Systolic BP	.013	.492	.389	.039	.387
Diastolic BP	.046	.463	.434	.118	.318
Cholesterol Total (mg/dL)	−.101	.291	.719	.204	−.353
Glucose Blood (mg/dL)	.004	.186	.503	.204	.061
HDL Cholesterol (mg/dL)a	−.265	−.073	−.033	.535	.102
LDL Cholesterol (mg/dL)	.017	.241	.598	.195	−.490
Triglycerides (mg/dL)	.028	.342	.376	−.352	.098
Chronic Localized and Non-Specific Pain
NSPSD	.712	.174	−.098	−.088	−.036
PCS	.730	.208	−.062	.066	−.190
PROMIS pain interference	.694	.213	−.133	.110	−.221
Cognitive disruption
Neuro-QoLa	−.676	.210	.108	.026	−.001
TOMAL DF	−.177	.220	−.307	.273	−.164
TOMAL PR	−.155	.369	−.225	−.455	.036
TOMAL WSR	−.229	.532	−.232	−.179	−.013
TOMAL WSRD	−.213	.418	−.199	−.373	.211
Endocrine disruption
Total FSH (mIU/mL)a	−.264	−.091	−.015	.550	.341
Total LH (mIU/mL)a	−.086	−.064	−.113	.490	.438
Total T (ng/dL)a	−.037	−.269	−.276	.119	−.270
Total TSH (uIU/nL)a	−.033	.115	.206	−.021	.138
Headaches
HIT-6	.746	.282	−.019	.010	−.128
Mental health disruption
GAD-7	.797	.163	−.108	.272	.117
PHQ-9	.802	.109	−.101	.039	.211
PCL-5	.880	.167	−.058	.106	.014
Sensory disruption
DHI	.677	.086	−.131	.109	−.001
SOT Totala	−.537	.540	−.285	.210	−.149
SOT Somatosensorya	−.047	.281	−.183	−.153	.060
SOT Vestibulara	−.416	.593	−.398	.249	−.219
SOT Visuala	−.394	.619	−.295	.262	−.084
SSQ-12a	−.665	.146	−.104	−.116	.106
Sleep disruption
ISI	.646	.318	−.183	−.046	.174

Note. Bold font indicates significant factor loadings > |.26|.

BMI = body mass index; BP = blood pressure; HDL = high density lipoprotein; LDL = low density lipoprotein; NSPSD=Nonspecific Posttraumatic Somatic Distress Scale; PCS=Pain Catastrophizing Scale; PROMIS= Patient-Reported Outcomes Measurement Information System—Pain Interference Scale; NeuroQol = subjective cognitive performance; DF = digits forward; PR = paired recall; WSR = word selective reminding; WSRD = word selective reminding delayed recall; FSH = follicle stimulating hormone; LH = luteinizing hormone; T = testosterone; TSH = thyroid stimulating hormone; HIT-6 = Headache Impact Test; GAD-7 = Generalized Anxiety Disorder; PHQ-9 = Patient Health Questionnaire; PCL-5 = PTSD Checklist; DHI = Dizziness Handicap Inventory; SOT = Sensory Organization Test; SSQ-12 = Speech, Spatial, and Qualities of Hearing Scale; ISI = Insomnia Severity Index.

^aHigher numbers generally reflect desirable outcome.

Fig. 1.

EFA factor loading estimates by factor. ∗ Negative factor loading indicating lower score. Note. Bolder color indicates significant higher factor loadings > |.26|. BMI = body mass index; sBP = systolic blood pressure; dBP = diastolic blood pressure; Chol = cholesterol; HDL = high density lipoprotein; LDL = low density lipoprotein; Tri = triglycerides; NSPSD = Nonspecific Posttraumatic Somatic Distress Scale; PCS = Pain Catastrophizing Scale; PROMIS = Patient-Reported Outcomes Measurement Information System—Pain Interference Scale; NeuroQol = subjective cognitive performance; DF = digits forward; PR = paired recall; WSR = word selective reminding; WSRD = word selective reminding delayed recall; FSH = total follicle stimulating hormone; LH = total luteinizing hormone; T = total testosterone; TSH = total thyroid stimulating hormone; HIT-6 = Headache Impact Test; GAD-7 = Generalized Anxiety Disorder; PHQ-9 = Patient Health Questionnaire; PCL-5 = PTSD Checklist; DHI = Dizziness Handicap Inventory; SOT = Sensory Organization Test; SSQ-12 = Speech, Spatial, and Qualities of Hearing Scale; ISI = Insomnia Severity Index.

Factors 2–5 included subsets of these broader symptoms, often with cross-loadings between factors, suggesting interrelated effects. Variables with the highest Factor 2 loadings included cardiometabolic elevations, lower testosterone levels, greater sleep disruptions, while retaining higher levels of objective cognitive performance and sensory function. Factor 3 also included cardiometabolic elevations and lower testosterone levels, but with lower objective cognitive performance (digits forward) and somatosensory function, but generally better subjective pain and mental health reports. Factor 4 included several positives, including higher levels of FSH, LH, HDL cholesterol, lower triglycerides, and better visual function, that were associated with generalized anxiety and mixed performance on objective cognitive measures. Factor 5 included mixed physiological markers, including higher blood pressure, lower cholesterol and triglycerides, higher FSH, LH, and lower testosterone levels.

Person-centered approach

Next, all z-score adjusted variables were entered into an LPA as continuous indicators. One-to-five profiles were examined using an iterative process to identify differential but homogenous profiles of related symptoms within the larger heterogeneous sample (see Supplemental Table S2). A two-class LPA model was selected as the optimal solution (Fig. 2). Lower AIC, BIC, SSBIC, higher entropy (.929), significant likelihood ratio testing, and average posterior probabilities of participants being correctly (95.5%–99.1%) and incorrectly (2.0%–2.1%) assigned to each of the three classes indicated high classification accuracy, specificity, and sufficient statistical power. A three-profile solution (p = .417) failed to provide significantly improved fit over a two-profile solution (p < .0001).

Fig. 2.

LPA results of two-profile solution of health sequelae. Note. HA = Headaches; MH = Mental Health; Slp = Sleep. BMI = body mass index; sBP = systolic blood pressure; dBP = diastolic blood pressure; Chol = cholesterol; HDL = high density lipoprotein; LDL = low density lipoprotein; Tri = triglycerides; NSPSD = Nonspecific Posttraumatic Somatic Distress Scale; PCS = Pain Catastrophizing Scale; PROMIS = Patient-Reported Outcomes Measurement Information System—Pain Interference Scale; NeuroQol = subjective cognitive performance; DF = digits forward; PR = paired recall; WSR = word selective reminding; WSRD = word selective reminding delayed recall; FSH = total follicle stimulating hormone; LH = total luteinizing hormone; T = total testosterone; TSH = total thyroid stimulating hormone; HIT-6 = Headache Impact Test; GAD-7 = Generalized Anxiety Disorder; PHQ-9 = Patient Health Questionnaire; PCL-5 = PTSD Checklist; DHI = Dizziness Handicap Inventory; SOT=Sensory Organization Test; SSQ-12 = Speech, Spatial, and Qualities of Hearing Scale; ISI = Insomnia Severity Index.

The two LPA profiles differed by the type and severity of symptoms experienced (see Supplemental Table S3 for mean estimates). Profile 1 participants (n = 72, 32.4%) experienced greater pain, headaches, lower cognitive performance, endocrine imbalances, greater mental health symptoms, lower sensory function, and greater sleep disruptions. LPA results complemented EFA results and indicated that posttraumatic stress symptoms (z-score PCL-5 = 1.03), pain catastrophizing (z-score PCS = .99), and dizziness (z-score DHI = .95) carried the highest mean z-score estimates approximately one standard deviation from the sample mean. Anxious symptoms (z-score GAD-7 = .91), depressive symptoms (z-score PHQ-9 = .89), non-specific pain (z-score NSPSD = .83), pain interference (z-score PROMIS = .80), headaches (z-score HIT-6 = .75), subjective cognitive concerns (z-score NeuroQol = −.79), sleep difficulties (z-score ISI = .72), and hearing difficulties (z-score SSQ-12 = −.68) and also carried high mean z-score estimates greater than .5 standard deviations from the mean. Profile 2 participants (n = 150, 67.6%) experienced less pain, had fewer cognitive difficulties, more balanced endocrine panels, fewer mental health symptoms, and fewer sensory and sleep disruptions. Participants from both profiles experienced similar cardiometabolic disruptions. Nested binomial logistic regression indicated participants within different profiles did not differ by age (OR = .98, p = .44) or enlistment status (OR = 1.22, p = .70). However, when examining occupation, Army SF (OR = 1.93, p = .04) were significantly more likely and SWCC (OR = 2.26, p = .08) were marginally more likely than SEALs to belong to Profile 1 and experience worse symptoms (see Table 4).

Table 4.

LPA profiles and binomial logistic regression results.

Variable	Profile 1 (n = 150, 67.6%)	Profile 2 (n = 72, 32.4%)	OR	p
	n(%)	n(%)
Age	M = 40.76, SD = 6.2	M = 40.11, SD = 5.0	.98	.435
Enlistment status
Enlisted	118 (78.7%)	63 (87.5%)	1.22	.702
Officer	32 (21.3%)	9 (12.5%)	–	–
Occupation
SWCC	15 (10.0%)	11 (15.3%)	2.26	.076
Army SF	51 (34.0%)	32 (44.4%)	1.93	.041
Other	7 (4.7%)	4 (5.6%)	1.76	.397
SEAL	77 (51.3%)	25 (34.7%)	–	–

Note. M = Mean; SD = Standard deviation; OR = odds ratio. Officer status and SEAL are reference categories.

Closer examination of mental health

Considering mental health symptoms carried the greatest Factor 1 loadings, a closer examination of these symptoms as categorical indicators was conducted using latent class analysis (LCA). LCA and categorical indicators was used to help elucidate differences not only in severity of mental health symptoms but the type of symptoms experienced. Depressive, anxious, and posttraumatic stress symptoms were examined concurrently and dichotomized using criteria of “moderate” symptoms or above (score ≥2 on each measure).²⁷ Two items from the PHQ-9 (concentration, sleep) were excluded from the LCA due to overlap with the PCL-5, resulting in 34-items.

One-to-five classes were examined using the same iterative process and selection criteria as the LPA (see Supplemental Table S4). A three-class model was selected as the optimal solution (Fig. 3). Lower AIC, BIC, SSBIC, higher entropy (.961), significant likelihood ratio testing, and average posterior probabilities of participants being correctly (98.0%–99.1%) and incorrectly (.1%–2.1%) assigned to each of the three classes indicated high classification accuracy, specificity, and sufficient statistical power. A four-class solution (p = .078) failed to provide significantly improved fit over a three-class solution (p = .0003). Classes reflected three subgroups with varying symptom presentations: 1) high global symptoms, reflective of anxiety, depression, and PTSD (n = 53, 23.9%); 2) dysphoric arousal (n = 110, 49.5%); and 3) hyperarousal (n = 59, 26.6%; see Supplemental Table S5 for class probabilities). Dysphoric arousal was characterized by a lack of interest/motivation, diminished positive emotions, negative thought patterns, social withdrawal, fatigue, concentration issues, hypervigilance, trouble relaxing, and sleep problems. Hyperarousal was characterized by irritability, fatigue, concentration issues, and sleep problems, but lower overall symptoms. Nested multinomial logistic regression indicated participants within different classes did not differ by age, enlistment status, or occupation (e.g., SEAL, SWCC, Army SF; see Supplemental Table S6).

Fig. 3.

LCA results of three-class solution of mental health symptoms.

Discussion

This study provides the most comprehensive and integrative examination of behavioral, biological, and physiological sequelae in U.S. active-duty and treatment-seeking SOF personnel. Within this examination, evidence of a persistent, broad, cumulative, and systemic burden of transdiagnostic sequelae was identified as well as different experiences of these sequelae across participants. When further examining mental health symptoms, phenotypes of high global symptoms (i.e., anxiety, depression, PTSD), dysphoric arousal, and hyperarousal were identified, with dysphoric arousal experienced by the majority of participants. Altogether, findings present a comprehensive illustration of the breadth of sequelae in active-duty treatment-seeking SOF personnel and can preliminarily inform current approaches to clinical practices for SOF and potentially other populations with similar exposures.

An initial variable-centered statistical application (EFA) indicated that five factors accounted for approximately 63% of the total variance in health outcomes measured in this sample. Using semi-structured clinical interviews, symptoms were evaluated related to military exposures (e.g., TBI, psychological trauma, musculoskeletal injuries). Factor 1 accounted for most variance (23.15%) in health outcomes and indicated a cumulative and general effect largely composed of mental health disruptions, headaches, non-specific pain and somatization, pain interference, and pain related to psychological states (i.e., psychogenic), sensory disruptions, and sleep problems, but fewer cardiometabolic and endocrine disruptions.^2,21,42, 43, 44, 45, 46, 47, 48 Notably, all symptom domains were significantly represented across multiple factors with the exception of pain, which was only represented in Factor 1. Therefore, non-specific pain and somatization, pain interference, and pain related to psychological states may be a key components to Factor 1's symptom composition.

Factor 2 described symptom presentations composed of better sensory and cognitive function but more cardiometabolic disruptions. Factor 3 described symptom presentations composed of significant cardiometabolic disruptions, lower sensory function, and lower cognitive performance in working memory, but generally better subjective pain and mental health reports. Factor 4 described symptom presentations composed of better cardiometabolic function but a range of cognitive disruptions, generalized anxiety, and visual problems. Factor 5 described symptom presentations composed of higher blood pressure and endocrine disruptions. Cardiometabolic and endocrine disruptions had smaller loadings on all factors, indicating a typically smaller, but significant association with the broader sequelae studied here. However, it must be noted that the potential influence of pre-existing and related medical conditions (e.g., sleep apnea) and potential moderators like substance use, alcohol use, medications (e.g., statins), testosterone therapy, and supplements were not included in this study, which may have affected current symptoms, particularly cardiometabolic and endocrine function, and should be incorporated in future studies. Moreover, despite statistical significance, some factor loadings were smaller in magnitude and should be considered preliminary targets for future evaluation. Altogether, these factors may indicate subtle but important differences in patient experiences of these sequelae that may assist in efficient triage of clinical care. Although patient experiences resembling Factor 1 may require a higher level of integrated care, others may require more specialized care available in many outpatient and specialty clinics.

A person-centered statistical application (LPA) complemented EFA findings. Although the EFA results provided a variable-centered and granular view of measurement variance and responses, LPA results provided a person-centered overview of patient experiences and, in particular, two different patient experiences. Nearly one-third of participants experienced globally elevated disruptions in pain (i.e., non-specific pain and somatization, pain interference, pain related to psychological states), cognition, headaches, mental health, sensory function (i.e., vision, hearing, balance, dizziness), and sleep, with fewer disruptions in cardiometabolic function (Profile 1). Alternatively, two-thirds of participants experienced milder symptoms (Profile 2). In the globally elevated symptom profile, participants experienced much greater traumatic stress symptoms,h pain amplified by psychological states, and dizziness, with symptom severity approximately one standard deviation higher than the average treatment-seeking participant. Based on this sample's data, that would correspond to a significant 18.5 score increase on the PCL-5, an 11.6 score increase on the PCS, an 18.5 score increase on the DHI, respectively. For reference, PCL-5 scores ≥15 are suggestive of reliable and clinically significant differences.⁴⁹ Concurrently, these participants experienced greater non-specific pain, pain interference, headaches, depressive symptoms, anxious symptoms, subjective cognitive concerns, hearing difficulties, and sleep difficulties, with symptom severity approximately half a standard deviation higher than the average treatment-seeking participant.

Considering both EFA Factor 1 and LPA Profile 1, findings suggest a persistent, broad, and cumulative symptom response that may be significantly influenced by primary difficulties with traumatic stress, pain, particularly pain amplified by psychological states (i.e., psychogenic pain), and physical or sensory symptoms, particularly dizziness related to military exposures like multiple TBIs, psychological traumas, and musculoskeletal injuries. Non-specific pain, pain that interfered with functioning, depressed and anxious mood, cognitive concerns, sensory, and sleep disruptions were also prominent but less so than traumatic stress, pain, and sensory changes, indicating a potential secondary response that requires further study. Although symptoms were anchored to military exposures, due to the cross-sectional nature of this study and the systemic and multifactorial nature of the exposures and injuries studied here, direct causes, precipitants, or contributors (e.g., neurotoxic exposures) to these symptoms cannot be determined with certainty.

Sequelae identified here overlap with sequelae associated with persistent postconcussive symptoms (PPCS) following multiple mild TBIs (e.g., headaches/migraines, mood dysregulation, cognitive impairments, sensory disruptions),^21,42, 43, 44 the effects of traumatic stress and general psychological distress (i.e., allostatic load) related to previous and potentially ongoing stressors,45, 46, 47, 48 psychogenic pain, non-specific pain and somatization, and interfering pain following musculoskeletal injuries common to SOF operations,² and their cumulative systemic effects on sleep and multiple biological systems, including, cardiometabolic and endocrine disruptions.^26,28,50, 51, 52

Although most single incidents of mild TBI resolve within weeks-to-months, in cases of multiple mild TBIs, common to SOF, there can be persistent, disabling, and a broader scope of related and often non-specific symptoms.^9,28,31,42, 43, 44^,50 In addition, many of the events that cause TBI can also cause traumatic stress reactions and musculoskeletal injuries. Therefore, symptoms of PPCS, traumatic stress, allostatic load, and musculoskeletal injuries are intimately intertwined, but frequently not always integrated in clinical practice or research.^21,43,44 They are also associated with several general effects including somatization and non-specific pain,^24,53 autonomic dysregulation,^50,51 inflammation,⁵⁴ subjective cognitive concerns that may or may not be associated with objective cognitive impairments,^55,56 cardiometabolic dysregulation,25, 26, 27^,57 and endocrine dysregulation.^28,58, 59, 60 For example, hypopituitarism has been observed in up to 40% of individuals post-TBI and can result in hormonal disruptions across multiple biological systems that further disrupt cardiometabolic, neurological, and psychiatric functioning.^28,57,59 Without appropriate integrative clinical practices, these symptoms may appear unrelated and treated in isolation or as “unexplained medical symptoms.” If unmitigated, these symptoms can initiate a cascade of systemic processes that can produce more complicated health conditions over time (e.g., cardiovascular disease, hypogonadism, metabolic syndrome).^20,26 Current findings prompt a call for early, comprehensive, and integrative assessment and intervention practices. Notably, VHA's PSC offers a unique opportunity to use integrative methods to understand and treat the nuances of complex multiple injuries like those observed in SOF and to identify subpopulations that may require novel and personalized treatments.^27,29

Although EFA identified unique factors with subjective and objective cognitive performance disruptions, LPA results indicated that participants with globally elevated disruptions had lower scores on perceived, rather than objective cognitive performance. Current findings demonstrated an often–disparate relationship between subjective cognitive concerns and objective cognitive performance. Considering the broad and cumulative sequelae identified here, mood, traumatic stress, and persistent post-concussive symptoms related to mild TBI are often associated with increased cognitive concerns that may or may not be supported by objective cognitive performance.^55,56 Lower subjective appraisals of cognitive performance may be contingent on associated mental health symptoms, pain, headaches, sensory disruptions, and poor sleep, common to Factor 1 and related issues of PPCS, traumatic stress, and general psychological distress.^55,56 As opposed to decreased cognitive ability, patients may be experiencing transient disruptions to cognitive performance due to a range of multifactorial influences identified here. Despite discrepancies between subjective and objective cognitive performance measures, performance across sensory measures and multidimensional assessments of pain indicated global disruptions across multiple domains of functioning, subjective, and objective measures.

LPA results also indicated that Army SF and Navy SWCC were approximately twice as likely as Navy SEALs to experience globally elevated symptoms. Further investigation is needed to determine correlates and predictors of these two differential profiles. Although we cannot offer a conclusive explanation determining why Army SF and Navy SWCC may experience more and worse symptoms and Navy SEALs, SOF personnel are used differently given their specialty, operate in different environments, and on different tasks, thereby incurring potentially unique exposures.^9,61 It is possible that the unique exposures more common to Army SF and Navy SWCC personnel may contribute to more sequelae over time; however, there are too many potentially confounding variables (e.g., available resources, medical coverage, operational tempo, personal lifestyle factors) to offer an explanation with a high degree of confidence at this time. Larger samples with increasingly informative markers may lend additional insights into the validity of the profiles identified here, potential nuances between these profiles, and even more individualized profiles to inform etiological and phenomenological knowledge, clinical practice, and help clinicians better understand what interventions should be provided to whom and when.

Because mental health symptoms carried the greatest factor loadings for Factor 1, a tertiary person-centered statistical application (LCA) was conducted to determine if participants’ transdiagnostic mental health symptoms differed not only by symptom severity but by type of symptom presentation (i.e., phenotype). Most participants (76.1%) presented with dysphoric arousal (49.5%) or hyperarousal (26.6%) phenotypes, with fewer presenting with high global symptom elevations (23.9%), reflective of traditional forms of anxiety, depression, and PTSD.

Dysphoric arousal62, 63, 64 was the predominant mental health phenotype, characterized by dysphoria (i.e., lack of interest/motivation, blunted positive emotions, negative thought patterns, social withdrawal, fatigue) and hyperarousal (i.e., concentration issues, hypervigilance, trouble relaxing, sleep problems). Notably, dysphoric arousal has been identified as a precise and transdiagnostic phenotype that can be used to target interventions.62, 63, 64 Hyperarousal is one of four clusters of posttraumatic stress-related symptoms characterized by irritability, fatigue, hypervigilance, concentration issues, and sleep problems.⁶⁵ Although some participants only experienced hyperarousal mental health symptoms, hyperarousal was common to all three mental health phenotypes here. Hyperarousal has been associated with a range of deleterious health outcomes, including cardiometabolic and endocrine dysfunction highlighted here, and may be a key mechanism of systemic dysregulation that requires further study.^62,66

For both dysphoric arousal and hyperarousal, intrusive symptoms (e.g., thoughts/memories of the trauma [s], flashbacks, reactivity) and avoidance symptoms were less prevalent, helping to differentiate these symptom presentations from traditional PTSD. Consequently, despite having strong support in the literature across multiple populations,62, 63, 64^,67 dysphoric arousal and hyperarousal responses are not well-captured by current psychiatric diagnoses and may represent subthreshold presentations, illustrating a key gap between research and clinical practice and providing sufficient and necessary interventions. Intervention studies are required to determine if the identification of the dysphoric arousal phenotype may help target assessment and treatment practices to provide more precise and effective interventions. However, given the unique symptom constellation of dysphoric arousal, psychotherapies like cognitive behavior therapy (CBT) and cognitive processing therapy (CPT) that target negative thoughts, mood, and hyperarousal symptoms⁶⁸ may prove more effective than exposure-based therapies (e.g., prolonged exposure, eye movement desensitization and reprocessing) that target threat and fear conditioned responses stemming from intrusive reactions and avoidance of related stimuli.⁶⁹ SMART-CPT, which combines trauma-focused CPT with Cognitive Symptom Management and Rehabilitation Therapy (CogSMART) may be particularly advantageous for this population with comorbid traumatic stress and cognitive disruptions following multiple mild TBIs.⁷⁰ Therapies may often need to be completed concurrently or integrated with medication management and other therapies for sleep (e.g., CBT for Insomnia [CBT-I]),⁷¹ pain and headaches, particularly to reduce negative cognitions, like catastrophizing, regarding pain (e.g., medication management, CBT for Chronic Pain/Headaches),⁷² sensory disruptions (e.g., speech/language therapy, physical therapy to improve vestibular function, balance, dizziness), cardiometabolic disruptions, and endocrine disruptions.

Is there an “operator syndrome?” To date, the term, operator syndrome (OS), has been used to describe a hypothesized and emerging construct that captures the persistent, cumulative, and systemic burden of behavioral, biological, and physiological symptoms commonly experienced in SOF and related populations following exposure to multiple mild TBIs, multiple musculoskeletal injuries, and chronic traumatic stress or allostatic load.^1,4 OS is like many war syndromes that came before it that have been used to capture complex and interrelated symptoms following unique exposures and injuries following military operations, for which institutions like PSC were created.^14,16, 17, 18, 19, 20 Although the current study alone does not substantiate a novel medical syndrome of OS, it provides evidence of the persistent, cumulative, and systemic effects of these symptoms that may be underrecognized in current research and clinical practices.^4,7,9,14,67 It also provides evidence of differential experiences of these symptoms within the SOF population, with some experiencing more and worse symptoms than others. The fact that the full scope of these sequelae was so prevalent, related across different health domains, and structurally cohesive across separate analyses is notable for this SOF population and those with similar exposures.

The documented systemic effects of multiple mild TBIs, multiple musculoskeletal injuries, traumatic stress, and allostatic load are diffuse and cumulative as demonstrated in this brief review of the extant literature.^24,28,45,73 Likewise, the sequelae identified in this study are also broad and non-specific with multiple potential etiologies, comorbidities, and heterogeneous effects that differ across individuals.^21,23,28,45 Accordingly, critiques of the OS framework as being overly broad, non-specific, and, therefore, having little clinical utility are inherent to the complex injuries from which these sequelae stem and, therefore, unavoidable. Although these sequelae are well-documented across populations, the breadth and interactive complexities of these symptoms, as well as their persistent, concurrent, and cumulative effect in populations like SOF, who are more likely to incur these injuries due to their unique occupational and lifestyle demands,^2,3,8 is notable and underrepresented in the literature.^24,28,45,61 Altogether, current findings offer preliminary support for the framework of OS, which warrants further study to determine its utility and treatment implications beyond current clinical practices.

Beyond the personal gratification and understanding many patients gain from the term, OS, the potential utility of this framework may come from establishing an expectation for both patients and practitioners that multiple health domains may be systemically disrupted, which domains may be disrupted, and how disruptions in one domain may influence another. Rather than targeting one domain in discipline-specific and isolated modalities, multiple health domains should be evaluated and treated from an integrative approach. Therefore, the label of OS, even if only reflective of a conceptual framework, may help increase the practitioner's ability to effectively integrate and triage the complex breadth of transdiagnostic and potentially systemic sequelae for more people.

This study has many strengths, including the sample of 222 active-duty SOF personnel, who are underrepresented in the extant literature, the breadth of 31 variables across 17 measures, the use of semi-structured clinical interviews, the combination of objective and subjective assessments, the multidimensional assessment of three types of pain, and the integration of these behavioral, biological, and physiological indicators in statistical analyses using both variable- and person-centered techniques. Despite these strengths, there are limitations that have been reported throughout this study. In addition, although the sample of three women approximated the proportion of men-to-women in SOF,⁷⁴ findings are not representative of all SOF women. Moreover, this sample included treatment-seeking SOF personnel and may not be representative of all SOF personnel.

This study presented a detailed and integrative examination that captures the underrepresented sequelae experienced by U.S. active-duty and treatment-seeking SOF personnel. Findings indicated a broad, cumulative, and systemic burden of multiple behavioral, biological, and physiological sequelae in U.S. SOF personnel and provided insights into the differential experiences of these sequelae following multiple complex TBIs, musculoskeletal injuries, and chronic stress to inform more comprehensive, accurate, and effective clinical assessments and interventions.

Contributors

SA, CAP, and RO interviewed participants and collected data during the course of clinical care. SA designed and directed this study and team, conducted statistical analyses, and drafted all versions of this manuscript. BF, JS, CAP, CPP, QC, and OAH provided input on study design and methods. JS and OAH directed administrative and program aspects of this study. All authors read, edited, and approved the final manuscript. JS and CPP accessed and verified the underlying data. All authors confirm that they had full access to all the data in the study and accept responsibility for publication.

Data sharing statement

Data contain protected health information within U.S. VHA records and are not available for public release. Select deidentified data and/or a data dictionary may be made available upon individual request.

Declaration of interests

The authors have no conflicts of interest to declare.