Effects of Stress Inoculation Training (SIT) with Relaxation Breathing on Posttraumatic Stress Disorder and Other Mental Health Outcomes in the Military: A Longitudinal Study Laurel Hourani, PHD, MPH,* Stephen Tueller, PHD; ,2 Paul Kizakevich, MS3 Paul Kizakevich, MS; and COL Laura Strange, PHD, RN; Gregory F. Lewis, PHD; Belinda Weimer, MPH; Jessica Morgan, MS; Daryl Cooney; Jessica Nelson • 1RTI International, Research Triangle Park, NC;
Background ■■
■■
In a previous study, we developed and evaluated a pilot stress inoculation training (PRESTINT) program designed to teach relaxation breathing skills to minimize the negative mental health consequences of combat stress. This study extends the investigation of the effectiveness of PRESTINT on PTSD scores and perceived stress symptoms in a longitudinal randomized control trial.
PRESTINT Components ■■ ■■
■■
Design & Methodology: Predeployment
Educational materials: Trifold brochures, developed by RTI International using the Army’s SM resources, were provided to all participants.
Enter Consent Questionnaire
Protocol
BB training: A 20-minute group presentation provided (1) attentional retraining, including relaxed breathing with eyes open; and (2) relaxed abdominal breathing with eyes closed.
Classroom Setup
Multimedia stress environment (MSE). An MSE tested physiological reactivity and speed/accuracy performance and provided an opportunity to practice breathing skills. This involved 12-minute stressor scenarios with “mission objectives” to anticipate enemy engagement and respond to in-scene cues, sudden events (e.g., explosions), loud noise, and post-event chaos while using a game controller to react to these stimuli.
Procedures
Start PreMSE Intervention MSE HRV MSE • HRV data collection
Classroom Setup
■■
SM control group received a 20-minute didactic presentation on SM.
■■
PRESTINT group received a 20-minute relaxation breathing skills lesson with biofeedback (BB).
■■
Both groups received pre-and post-training MSE practice with HRV and reaction time data measurement.
■■
All participants were provided with MP3 players with their respective training recordings and asked to listen when they felt stressed.
■■
Interim email reminders about the relaxation techniques were sent every 2-3 months during the 1.5 year follow-up.
MSE
MSE projection and reaction time targets MSE projection and reaction time targets
Introduction ■■
Decreased heart rate variability (HRV) is associated with posttraumatic stress disorder (PTSD) and depression symptoms, but the potential influence of HRV biofeedback in stress relaxation training is not well understood (Tan et al., 2011).
■■
Three phases of stress inoculation training (SIT) are education, skills training, and practice in a simulated stressful environment (Meichenbaum & Cameron, 1989).
■■
Emerging technological advances (Serino et al., 2014; Vakili, Brinkman, Morina, & Neerincx, 2014) provide the ability to create virtual reality environments in which stressful stimuli and their resulting physiological arousal can be observed and used (Bouchard, 2014) to develop group-based simulation environments (Hourani et al., 2011).
■■
Relaxation breathing, referred to as Battle Breathing (BB), reduces autonomic arousal and has been shown to increase HRV (Leonaite & Vainoras, 2010; Lewis et al., 2015; Terathongkum & Pickler, 2004).
Physiological Measures*
■■
Evaluate the effectiveness of pilot stress inoculation training (PRESTINT) versus stress management (SM) only in reducing perceived stress levels and other mental health problems, including PTSD symptoms, generalized anxiety disorder (GAD) symptoms, and loss of control or aggression.
■■
Identify potential subgroups of effectiveness (e.g., among those without mental problems at baseline, among those who successfully increased their HRV during baseline training, among those interested in learning BB).
Method: Subjects ■■
■■
■■
Participants were selected from a convenience sample of platoons of the U.S. Army’s 82nd Airborne Division at Fort Bragg, NC. Platoons were randomized into either the experimental group, who received PRESTINT training, or the SM control group, who received a didactic presentation on SM. 267 participants (149 PRESTINT condition; 118 controls) were followed for an average of 1.5 years.
■■
Sociodemographics: There were no significant differences between PRESTINT and SM groups.
■■
Age (mean): 24 years
■■ ■■
Gender: 95% male, 5% female Paygrade: 9% E1-E3, 86% E4-E9, 5% W1-O6
■■
Marital status: 59% married/living as married, 41% single/divorced/widowed
■■
Deployment status: 50% deployed in past, 50% never deployed
PRESTINT Workstations PRESTINT Workstations
Results
Reaction time game controller Reaction time game controller
■■
At baseline, PRESTINT significantly increased their RSA after BB training and controls reported practicing more relaxation techniques (controlled for in multivariate regressions).
■■
At follow-up, controls had higher education levels than experimental respondents (controlled for in multivariate regressions).
■■
Perceived stress was measured using the Perceived Stress Scale (Cohen, Kamarck, & Mermelstein, 1983), a 10-item scale to assess a person’s perception of stress and control.
28.3% reported practicing either BB or other relaxation techniques (no differences between control and PRESTINT groups).
■■
PTSD was measured using the PTSD Checklist—Civilian Version (PCL-C). A cutoff score of 30 or greater indicates possible PTSD.
PRESTINT participants were more likely to meet criteria for possible PTSD than controls in unadjusted analyses.
■■
When education, practice, and other mental health problems at baseline variables were entered into multivariate models, there were no differences between PRESTINT and control groups.
■■
When adjusting for all covariates including potential interactions, only “currently taking medications for depression, anxiety, or sleeping problems prescribed by a doctor or other health professional” predicted possible PTSD, high perceived stress, GAD, and loss of control or aggression.
Survey Outcome Measures
■■
■■
802.11b Wireless Hub 802.11b Wireless Hub
MSE Computer MSE Computer
■■
Specific Aims
HR and HRV biofeedback HR and HRV biofeedback
■■
■■
Intermediate outcome measures of hyperarousal included a noninvasive physiologic index of autonomic arousal (i.e., heart rate and RSA) during exposure to the MSE collected using a Biocom system to store and analyze a time series of pulse interval values.
■■
Heart rate and HRV were extracted according to the Porges–Bohrer method to measure RSA and LF-HRV. HRV changes were also analyzed within a broad frequency band spanning the RSA and LF-HRV ranges due to the experimental demand to shift respiration. HRV measures reflect combined effects on RSA and LF-HRV.
■■
To screen for GAD, seven items adapted from the Patient Health Questionnaire (Spitzer, Kroenke, & Williams, 1999) were used. Using the standardized scoring algorithm, we used a cutoff of 10 to indicate GAD. Loss of control or aggression was measured with the item, “During the past year, have you had thoughts or concerns that you might lose control or hurt someone?” Response options were yes, no, and unsure.
*Obtained only at baseline
Figure 1: Descriptive Indices of PRESTINT and SM Groups Follow-Up Variable N
Control
Gender (baseline) of those who did follow-up Male Female Education (baseline) of those who did follow-up High school or less Some college, 2-year college or trade school College graduate Paygrade E1–E3 E4–E9 W1–O6 Marital status Married/living as married Single Returned from last deployment Deployed in the past Never deployed
118 n
Percent
Experimental 149 n Percent
110 8
93.20% 6.80%
143 6
96.00% 4.00%
253 14
94.80% 5.20%
40 60 17
34.20% 51.30% 14.50%
75 59 13
51.00% 40.10% 8.80%
115 119 30
43.60% 45.10% 11.40%
7 105 4
6.00% 90.50% 3.40%
17 122 8
11.60% 83.00% 5.40%
24 227 12
9.10% 86.30% 4.60%
76 42
64.40% 35.60%
80 68
54.10% 45.90%
156 110
58.60% 41.40%
56 62
47.50% 52.50%
77 72
51.70% 48.30%
Total 267 n
Percent
133 134
49.80% 50.20%
p-Value NS 0.003
NS
NS NS
Figure 2: Prevalence and Incidence of Primary and Secondary Outcomes Primary Outcomes N PTSD 43 ≥ 43 Stress (PSS) > 24 Anxiety (GAD) > 10 Loss of control or aggression Yes
Baseline Follow-Up Not High at Baseline New at Follow-Up Comparing Baseline to Follow-Up Total Control BB p-Value Total Control BB p-Value Total Control BB p-Value Control BB 891 422 469 267 118 149 230 100 130 p-Value p-Value 0.71 0.39 0.82 0.28 0.01 116 57 59 55 21 34 36 15 21 (13) (14) (13) (21) (18) (23) (16) (15) (16) 0.15 0.6 0.68 0.66 0.43 88 49 39 31 15 16 23 11 12 (10) (12) (8) (12) (13) (11) (10) (10) (9) 0.99 0.57 0.61 0.98 0.49 131 62 69 42 17 25 29 14 15 (15) (15) (15) (16) (15) (17) (13) (14) (12) 0.56 0.25 0.25 0.47 0.019 55 (15)
21 (13)
34 (17)
25 (10)
15 (13)
10 (7)
22 (9)
14 (13)
8 (6)
Figure 3: Multiple Regression Results for Primary Outcomes Stress (PSS)a β P −1.38 0.58 −2.28 0.1
PRESTINT (ref: control) PRESTINT (ref: control) X Practiced PRESTINT (ref: did not practice) Mental health problems, medication, 1.3 sleep problems, or seeing a counselor at baselinec (ref: no problems) Interested in learning stress reduction 1.9 (ref: uninterested) Ever practiced relaxation techniques prior −1.57 to the study (ref: did not practice) Downloaded additional relaxation 1.24 method to MP3 (ref: did not download) High school or less (ref: at least some −0.19 college/trade school) Currently taking medication for 5.96 depression, etc. (ref: no) PRESTINT (ref: control) X Currently taking −3.28 medication for depression, etc. (ref: no) PRESTINT (ref: control) X Mental health 0.53 problems, medication, sleep problems, or seeing a counselor at baselinec (ref: no problems) PRESTINT (ref: control) X Interested 2.24 in learning stress reduction (ref: uninterested) a
0.26
Figure 3: Multiple Regression of Secondary Outcomes PTSD 43b β P −0.72 0.36 −0.84 0.17 0.85
0.2
0.21
0.54
0.37
0.12
−0.23
0.59
0.37
0.31
0.56
0.83
−0.08
0.8
0.01
1.73
0.01
0.2
−0.64
0.55
0.74
0.63
0.43
0.28
Conclusions
0.6
0.45
Linear regression
b
Logistic regression (β can be exponentiated to obtain adjusted odds ratios)
c
At baseline, was participant indicated for depression, anxiety, PTSD, or had a prescription for depression, anxiety, or sleeping problems
BB (ref: control) BB (ref: control) X Practiced BB (ref: did not practice) Mental health problems, medication, sleep problems, or seeing a counselor at baselineb (ref: no problems) Interested in learning stress reduction (ref: uninterested) Ever practiced relaxation techniques prior to the study (ref: did not practice) Downloaded additional relaxation method to MP3 (ref: did not download) High school or less (ref: some college/ trade school) Currently taking medication for depression, etc. (ref: no) BB (ref: control) X Currently taking medication for depression, etc. (ref: no) BB (ref: control) X Mental health problems, medication, sleep problems, or seeing a counselor at baselinec (ref: no problems) BB (ref: control) X Interested in learning stress reduction (ref: uninterested)
Anxiety (GAD)a β P −0.44 0.59 −0.38 0.62 0.75
0.28
Loss of Control or Aggression (yes or unsure vs. no)a β P −0.44 0.55 −1.00 0.1 0.16
■■
PRESTINT did not have an overall effect on PTSD or perceived stress scores when controlling for other variables.
■■
Due to study limitations, including few participants deployed and few with exposure to traumatic events (experience where you thought you could be injured or killed) during the study period, PRESTINT was not supported as a general mental health prevention or SM strategy.
■■
An HRV increase in response with relaxation breathing training suggests future research is warranted into mental health effects of self-regulation techniques.
0.71
−0.54
0.46
0.24
0.69
−1.35
0.01
−0.70
0.08
0.05
0.93
1.03
0.03
−0.28
0.5
0.33
0.38
More Information
2.58
0.01
2.21
0.00
−0.99
0.45
−0.24
0.85
*Presenting author: Dr. Laurel Hourani 919.485.7719
[email protected]
−0.46
1.66
0.59
0.08
−0.30
0.25
0.61
0.73
a
Logistic regression (β can be exponentiated to obtain adjusted odds ratios)
b
At baseline, was participant indicated for depression, anxiety, PTSD, or had a prescription for depression, anxiety, or sleeping problems
RTI International 3040 E. Cornwallis Road Research Triangle Park, NC 27709 Presented at: The 2016 Military Health System Research Symposium, Kissimmee, Florida, August 2016 References available upon request.
www.rti.org RTI International is a registered trademark and a trade name of Research Triangle Institute.
