ARTICLES

A Multisite Randomized Controlled Trial of

Mindfulness-Based Stress Reduction in the Treatment
of Posttraumatic Stress Disorder
Lori L. Davis, M.D., Charles Whetsell, Ph.D., Mark B. Hamner, M.D., James Carmody, Ph.D., Barbara O. Rothbaum, Ph.D.,
Rebecca S. Allen, Ph.D., A.B.P.P., Al Bartolucci, Ph.D., Steven M. Southwick, M.D., J. Douglas Bremner, M.D.

Objective: Posttraumatic stress disorder (PTSD) is often compared with PCGT, the MBSR group showed a statistically
difficult to treat, and many patients do not achieve full re- significant improvement in PTSD on the self-reported PTSD
mission. Complementary and integrative health approaches, Checklist for DSM-IV over the nine weeks. This difference
such as mindfulness meditation, are intended to be inte- was not maintained posttreatment, at week 16. Strengths
grated with evidence-based treatment. This study examined of the study include its large sample size, multisite design,
the efficacy of mindfulness-based stress reduction (MBSR) active control group, single-blind outcome ratings, fidelity
in the treatment of PTSD in U.S. military veterans. monitoring, large minority representation, and randomized
approach. The study was limited by its high attrition rate and
Methods: Veterans with a diagnosis of PTSD (N=214) were low representation of women.
randomly assigned to either 90-minute group MBSR or
present-centered group therapy (PCGT) for eight weeks. Conclusions: Both MBSR and PCGT appear to have beneficial
Follow-up assessments were obtained at baseline and weeks effects in treating PTSD in veterans, with greater improve-
3, 6, 9 (primary endpoint), and 16. ment observed in self-reported PTSD symptoms in the MBSR
group. No differences between groups were observed on the
Results: Both the MBSR and PCGT groups achieved signifi- CAPS-IV scale.
cant improvement in PTSD as measured by the Clinician-
Administered PTSD Scale for DSM-IV (CAPS-IV), with no
statistically significant differences between groups. However, PRCP in Advance (doi: 10.1176/appi.prcp.20180002)

Posttraumatic stress disorder (PTSD) involves an overgen- rates of response and remission to these interventions are
eralization of a conditioned fight-or-flight response to mixed (4, 5). Approximately one-third of PTSD patients
previously neutral stimuli. Fear conditioning involves a neu- discontinue medication or therapy prematurely because of
ronal circuit that once triggered, fires repeatedly, despite difficulty tolerating the treatment. Many individuals living
efforts to turn it off. This circuit becomes not only the path of with a PTSD diagnosis are looking for alternatives to medi-
least resistance, but also the final common pathway, firing cation and/or trauma-focused psychotherapy. Complemen-
automatically in response to stimuli that would be better tary and integrative health approaches, such as mindfulness
served by a more flexible response. Psychotherapy uses the meditation, have grown in popularity as low-risk interventions
neuroplastic capabilities of the nervous system to facilitate the used with evidence-based medical treatments (6). Mindfulness-
formation and strengthening of new neuronal circuits while based stress reduction (MBSR) is a technique taught in a series
weakening the connections among overlearned troublesome of classes that trains individuals to focus attention on thoughts,
circuits (1, 2). In the case of PTSD, the firing of the overlearned sensations, and feelings as they appear (7).
circuit is accompanied by cognitive patterns that mark the dis- Incorporated into MBSR exercises such as body scan,
order, such as believing the present situation to be unsafe, re- sitting meditation, mindful yoga, and use of mindfulness in
gardless of the circumstances. The habitual patterns of memory everyday life, mindfulness involves the intentional awareness
include cognitive representations of the trauma and accompa- of, and nonreactivity to, thoughts, sensations, and feelings
nying strong emotions embedded in the limbic component of as they arise. The focus of attention can be maintained on
the circuit, leading to rage and fear responses (3). these mental/sensory contents and concomitant emotional
Current treatments for PTSD include trauma-focused responses, or the focus may be deliberately redirected to a
psychotherapy and antidepressant medications, although the different emerging thought, sensation, or feeling or to a wider

prcp in Advance prcp.psychiatryonline.org 1

MINDFULNESS-BASED STRESS REDUCTION IN POSTTRAUMATIC STRESS DISORDER

field of awareness. This self-regulatory behavior represents forming a group practice to experience the full benefit of
an openness to and acceptance of mental and sensory mindfulness. Group social support enhances one’s sense of
experiences that can change one’s relationship with one’s coherence (17) and facilitates compassionate behavior (18).
experience. Rather than remaining preoccupied with the Recent work has suggested that feelings of empathy for
content of mental or sensory experiences, one recognizes social pain are associated with increased activation on
that events occurring in the field of awareness will, by their fMRI of the anterior insula and anterior cingulate cortex
nature, change. For example, if a distressing memory is (19), regions of interest in the pathophysiology of PTSD.
noticed, no attempt is made to change or suppress it; in- Another key component of PTSD is avoidance. In many
stead, these thoughts or feelings are noticed as one part of ways, mindfulness is the opposite of avoidance. Mindfulness
a broader range of experience in that moment. Thus, the meditation resembles an exposure situation in that “practi-
attention required to sustain the thought and its attendant tioners turn toward their emotional experience, bring ac-
distress is re-directed, preventing the escalation of neg- ceptance to bodily and affective responses and refrain from
ative thoughts into ruminative patterns. Mental space is engaging in internal reactivity toward emotional experience”
left for more creative and less habitual or conditioned (20). The MBSR practitioner does not try to avoid or push
responses, which may contribute to a greater sense of con- away whatever comes to mind but rather observes and fo-
trol in stressful situations (8–12). cuses on mental images without judgment. MBSR uses one
What is the rationale for using MBSR in the treatment of of the most important components of evidence-based treat-
PTSD and by what mechanism could MBSR intervene in ment for PTSD (i.e., exposure). However, MBSR does not
altering the habitual responses that comprise PTSD? The intentionally or actively elicit traumatic memories, which
therapeutic process in the treatment of PTSD depends on may make it more palatable. Cognitive therapies are directive
the simultaneous presence of vivid experience and nonreac- and ask that attention be given to challenging or even aversive
tivity. Simple distraction from the distressing cognitions and material within a timeframe dictated by the constraints of a
emotions provides only temporary relief, as the underlying scheduled psychotherapy session. Such direction of attention
circuits remain intact and ready to fire whenever triggered. may not be in accord with a patient’s willingness, interest, or
As with exposure therapy, in therapy for PTSD, it is essential capacity to engage in difficult material. If there is a mismatch
that the trauma be actively brought to mind at the time of the between the timing of the therapeutic intervention and the
treatment; however, trauma representation should be mod- patient’s readiness to process traumatic events, resistance
ulated by cognitive processes to avoid triggering the full arises (21). Cognitive therapy, therefore, is more likely to elicit
trauma response, which would only reinforce the fear circuit. greater resistance than mindfulness, which is inherently self-
In the case of MBSR, traumatic memories that arise may be directed. Although mindfulness practice is not trauma fo-
experienced within an accepting frame of mind and be less cused, it is trauma inclusive. Unpleasant experiences that
likely to induce the fear-conditioned stress response. It is need resolution arise naturally during the practice, as the
possible to experience aspects of traumatic recall, dysfunc- patient’s defenses relax. This relaxation of defenses is gen-
tional emotions, and sustaining cognitions without judg- erally coordinated with a readiness to attend to aversive
ments such as “I’ll never be free of this.” By fostering an stimuli in a nonjudgmental and accepting attitude. By not
acceptance of raw experience, without trying to change it, being trauma focused, mindfulness is less aversive; by being
MBSR decouples the experience from neuronal firings that trauma inclusive, it facilitates awareness and engagement
code for defensive maneuvers and judgments about the ex- with troubling aspects of PTSD.
perience (13). This acceptance allows the patient to be less In theory, mindfulness could be useful in the treatment of
reactive to the symptoms of PTSD (14). The unfinished PTSD, in teaching patients to stay in the moment and not
business of the trauma is less likely to preoccupy the thought dwell on past or future events beyond their control. This less
stream of the patient, and he or she is then free to pursue new reactive mode of coping may provide a way for people with
experiences in life. Mindfulness meditation assists in de- PTSD to feel a greater sense of control and be less avoidant,
creasing rumination. As such, mindfulness may enhance self- which may lead to improved quality of life and emotional
regulation and decrease emotional reactivity (15). well-being (8, 22). Indeed, participation in MBSR has been
What is the benefit of a nontrauma-focused interven- associated with reduced intrusive ideation, worry, anxiety,
tion, such as MBSR? As pointed out by Brewer et al. (16), and emotional distress and increased sense of control (17) and
mindfulness training targets one’s relationship with thoughts emotional well-being (12, 23, 24). This reduced distress has
and the process of thoughts and feelings arising, whereas been found to endure upon three-month, six-month, and
cognitive-behavioral therapy intends to change the content of four-year follow up. Of the randomized controlled trials
thoughts. Mindfulness incorporates cognitive reappraisal, comparing MBSR to treatment as usual, 11 demonstrated
which is an important component of cognitive-behavioral improved mental health symptoms with overall medium
treatments. In this case, the cognitive reappraisal involves effect sizes (i.e., MBSR had a clinically meaningful response
learning to judge a traumatic memory as neither bad nor compared with the control group) (25). MBSR has been
good. In addition, social support is known to buffer against shown to reduce symptoms of chronic pain, anxiety, de-
symptoms of PTSD. Many mindfulness experts recommend pression, and PTSD (24, 26, 27).

2 prcp.psychiatryonline.org prcp in Advance

 DAVIS ET AL.

In response to a growing consumer request for meditation Patients were excluded if they would be receiving concur-
to be offered to veterans with PTSD, in 2010 the Veterans rent cognitive-behavioral therapy, cognitive processing ther-
Health Administration (VHA) prioritized a need for addi- apy, or prolonged exposure therapy during the study.
tional research on the effectiveness of meditation in veterans
with PTSD. In this article, we present the results of a mul- Screening Procedures and Assessments
tisite randomized controlled trial on the use of MBSR for Baseline assessment included psychiatric evaluation; review
treatment of veterans with PTSD. of psychotropic medication and psychotherapy treatment
history; recording of demographic data and disability status;
inventory of general medical conditions; and review of re-
METHODS
cent physical examination and laboratory tests, including
Study Design urine screen for drugs of abuse. A trained clinical research
From January 2012 to September 2013, U.S. military veterans coordinator conducted the baseline Mini-International Neu-
diagnosed as having PTSD were randomly assigned to re- ropsychiatric Interview, a structured clinician-administered
ceive eight weeks of either MBSR or present-centered group inventory that assessed current and lifetime DSM-IV dis-
therapy (PCGT) at three clinical research sites in VA Medical orders (28). At baseline, the participants were asked whether
Centers located in the southeastern United States. We hy- they had a preference regarding treatment assignment (MBSR,
pothesized that MBSR would improve symptoms of PTSD PCGT, or none).
over a nine-week follow up compared with PCGT. The The CAPS-IV (29) was used to confirm diagnosis of PTSD
Clinician-Administered PTSD Scale for DSM-IV (CAPS-IV) and to evaluate changes in PTSD symptoms. Trauma expo-
served as the primary outcome to evaluate treatment efficacy, sure was based on the participants’ verbal history, supple-
and secondarily, to examine the treatment’s effects on PTSD mented by the CAPS-IV Life Events form, and included
symptom clusters and rates of response. Outcome assess- combat, noncombat, and/or sexual trauma events. Partici-
ments were obtained at baseline and weeks 3, 6, and 9 (pri- pants were instructed to focus on the worst incident of
mary endpoint). As an exploratory measure of short-term trauma, the one that had most likely resulted in the diagnosis
durability, these assessments were repeated posttreatment at of PTSD. CAPS-IV assessments were rated for the past week
week 16. All sites obtained local institutional review board and, if needed, the baseline CAPS-IV was repeated so that the
approval prior to engaging in human subjects research. baseline score was collected the week prior to starting the
The study was monitored by an independent data monitor- intervention. The majority of the CAPS-IV assessments were
ing committee. All participants received full explanation of conducted by a trained independent assessor at each site
the purpose, procedures, risks, benefits, and alternatives to who was blind to the treatment (single-blind assessment).
treatment and provided informed consent and privacy au- CAPS-IV interviews were audio recorded and submitted to
thorization prior to study entry. a CAPS-rating fidelity monitor for review.
The PTSD Checklist—Self-Report (PCL), a 17-item self-
Participants report scale, was used to evaluate PTSD symptoms. The Five
Veterans of any combat or noncombat era were included if Facet Mindfulness Questionnaire–Self-Report (FFMQ) was
they were able to provide informed consent; were 19 to used to assess five facets of mindful living (observing, de-
65 years of age (inclusive); had a diagnosis of PTSD; had a scribing, acting with awareness, nonjudging of inner expe-
CAPS-IV score of $45 for the week prior to randomization; rience, and nonreactivity to inner experience) (30–32). The
had no substance use disorders (except nicotine and/or nine-item self-report Patient Health Questionnaire was used
caffeine) for two weeks prior to randomization; had no di- to evaluate depression.
agnosis of bipolar I disorder, schizophrenia, or a schizo-
affective disorder; and were not actively considering suicide Interventions
or homicide. Participants were excluded if they had current The MBSR groups met for eight weekly, 90-minute sessions
psychotic symptoms that in the investigator’s opinion im- and a six-hour retreat prior to week 6. MBSR training in-
paired their ability to provide informed consent or made cluded the body scan meditation (a gradual moving of at-
participation unsafe, a severe cognitive disorder (e.g., de- tention through the body from feet to head accompanied by
mentia or severe traumatic brain injury), or a clinically sig- awareness of breathing and other bodily sensations while
nificant unstable or severe medical condition that would lying in a supine position), sitting meditation (a focusing on
contraindicate study participation or expose the participant the awareness of breathing, bodily sensations, thoughts, and
to undue risk. Patients taking psychotropic medications were emotions, practiced while sitting upright on a chair or
included if the medication had been taken for at least four cushion), and mindful stretching (exercises practiced with
weeks and if the dosage had been stable for two weeks prior awareness of breathing and intended to develop mindful
to randomization and remained stable throughout the study awareness during movement). Participants were given two
(dose reductions due to unwanted side effects were allowed). guided meditation CDs to practice at home. In-class didactic
Patients taking pain medication were included if their dosage material emphasized the systematic development of mindful
had been stable for the two weeks prior to randomization. awareness and its application in everyday life. The six-hour

prcp in Advance prcp.psychiatryonline.org 3

MINDFULNESS-BASED STRESS REDUCTION IN POSTTRAUMATIC STRESS DISORDER

TABLE 1. Baseline demographic and clinical characteristics of veterans with posttraumatic stress disorder (PTSD) included in the analysisa

MBSR PCGT MBSR PCGT

(N=96) (N=95) (N=96) (N=95)
Characteristic N % N % p Characteristic N % N % p
Sex .87 Period of service .76
Male 80 83.3 80 84.2 Korea 1 1.0 0 .0
Female 16 16.7 15 15.8 Vietnam 39 40.6 35 36.8
Race .95 Gulf War (I) 18 18.8 25 26.6
White 30 31.3 31 33.3 Gulf War (II), after 9/11 16 16.7 14 14.9
Black 63 65.6 59 63.4 Other 22 22.9 20 21.3
Other/unknown 3 3.1 5 5.2 Type of trauma .70
Education .71 Combat 51 54.8 48 51.1
Less than college 22 22.9 24 25.3 Military, noncombat 23 24.7 29 30.9
Some technical school or 36 37.5 37 39.4 Sexual 10 10.8 11 11.7
college Civilian 9 9.7 6 6.4
Technical school/associate 10 10.4 14 14.9 Concurrent DSM-IV disorders
degree Major depression 54 56.3 53 55.8 .95
College degree or higher 28 29.2 19 20.0 Dysthymia 9 9.4 7 7.4 .62
Marital status .58 Panic 33 34.4 27 28.4 .38
Single 13 8.3 17 17.9 Agoraphobia 46 47.9 38 40.0 .27
Married 43 44.8 35 36.8 Social phobia 17 17.7 24 25.3 .20
Separated 14 14.6 8 8.4 Obsessive-compulsive 8 8.3 5 5.3 .40
Divorced 25 26.0 32 33.7 Alcohol dependence (past 13 13.5 12 12.6 .85
Widowed 1 1.0 1 1.0 12 months)
Employment .06 Alcohol abuse (past 12 months) 3 3.1 6 6.3 .30
Unemployed 28 21.5 27 28.4 Substance dependence 9 9.4 9 9.5 .98
Employed 22 22.9 17 17.9 (past 12 months)
Retired 9 9.7 11 11.6 Substance abuse (past 12 months) 1 1.0 2 2.1 .55
Disabled 32 34.4 41 43.6 Eating disorders 6 6.3 5 5.3 .77
Receiving SSDI/SSI income 22 22.9 28 29.5 .44 M SD M SD p
Branch of military service .61
Age (years) 51.7 10.9 51.0 11.4 .67
Army 52 54.2 55 58.5
Service-connected disability (%) 34.0 35.2 35.8 35.4 .73
Navy 12 12.5 9 9.6
Medical service connection (%) 14.3 23.2 16.6 24.9 .52
Marines 14 14.6 9 9.6
Psychiatric service connection (%) 2.9 12.1 3.6 16.5 .77
Air Force 4 4.2 8 8.5
PTSD service connection (%) 15.7 28.1 18.1 29.6 .58
National Guard/Coast Guard 3 3.1 5 5.3
Duration of military service 7.4 6.9 8.6 7.8 .29
Combination 11 11.5 8 8.5
a
MBSR, mindfulness-based stress reduction; PCGT, present-centered group therapy; SSDI/SSI, Social Security Disability Insurance/Supplemental Security Income.

retreat included extended practice of the mindfulness body MBSR retreat reflect an uncontrolled difference in total
scan and mindful sitting, walking, stretching, and eating. treatment time. Both interventions provided an instruction
PCGT was selected as the comparison treatment because manual to guide the therapists in conducting the treatment.
of its well-established use as a control for the nonspecific
effects of a group-based intervention (33). The PCGT groups Adherence and Attendance
met for eight weekly, 90-minute sessions and a lunch gath- To minimize attrition, the number of assessments was limited
ering prior to week 8. PCGT facilitated the expectation of to decrease participant burden. The research coordinators
symptom reduction, normalization of PTSD symptoms through called the participants to remind them of their appointments.
education, decreased isolation, shared support, shared positive Assessment visits were scheduled at convenient times for the
experiences with other veterans with similar symptoms, ex- participants and often were paired with other appointments.
perience of an atmosphere of safety, and awareness and Participants were paid a small fee to offset out-of-pocket
objectivity of how PTSD affects one’s daily life. These groups expenses for attending the assessment visits but not for the
had a present focus on current events that avoided discussion MBSR or PCGT sessions. In keeping with the intent-to-treat
of traumas. PCGT treatment was psycho-educational and design, participants could remain in the study for assessments
included discussion of everyday problems of group members even if they dropped out of treatment.
and of how PTSD created or intensified these problems. The
PCGT participants were assigned to keep a journal and had a Fidelity Monitoring
two-hour lunch gathering to partially control for the MBSR To ensure consistent delivery of the MBSR and PCGT cur-
retreat. The duration differences in the PCGT lunch and ricula, supervisors held separate monthly teleconferences

4 prcp.psychiatryonline.org prcp in Advance

 DAVIS ET AL.

with the MBSR therapists and with the PCGT therapists. TABLE 2. Frequency counts of veterans attending group sessions
Curriculum delivery was reviewed and potential threats to and correlation of outcome with percentage of PTSD group
sessions attendeda
fidelity were discussed. Adherence to the MBSR and PCGT
curricula was assessed by an independent fidelity monitor Number of MBSR PCGT Correlation with
Groups Attended (N) (N) % attendance P
who listened to a random selection of 30% of the audio-
recorded sessions. The fidelity monitor rated the MBSR 0 11 12
1 8 2
sessions using an adapted version of the Mindfulness-Based
2 4 3
Cognitive Therapy Adherence Appropriateness and Quality 3 7 1
Scale (34, 35). The fidelity monitor used a similar scale to 4 7 7
evaluate the PCGT group (replacing MBSR with PCGT 5 7 8
techniques). If a therapist deviated from the treatment guide, 6 8 18
7 12 29
the fidelity monitor provided the information to the MBSR
8 18 27
or PCGT supervisors, who then worked with the therapist
to either remediate or be replaced. CAPS-IV endpoint
Week 9 .10 .22
Statistical Analyses Week 16 .07 .41
Using a blocking strategy and stratification based on site, FFMQ endpoint
we randomized the sample, using the Dallal software Week 9 2.08 .37
Week 16 2.08 .35
(Tufts University) in a 1:1 allocation. Analyses adhered to a
a
modified intent-to-treat principle, classifying participants CAPS-IV, Clinician-Administered PTSD Scale for DSM-IV; FFMQ, Five Facet
Mindfulness Questionnaire; MBSR, mindfulness-based stress reduction;
by randomized treatment condition and attendance in at PCGT, present-centered group therapy.
least one group therapy session. To examine the balance
across randomized treatment groups, we compared the
two groups (MBSR vs. PCGT control) on baseline de- stratification variable (site). Because the interventions were
mographic and clinical characteristics using t tests for in group formats, a random effect for each therapy group
continuous variables, Wilcoxon tests for ordinal variables, accounted for correlation of outcomes due to idiosyncratic
and chi-square tests for categorical variables. These anal- group therapy factors. Likelihood ratio (LR) tests examined
yses were used to identify potential confounding vari- the incremental contribution of the treatment by time in-
ables to be used as covariates in subsequent analyses. Only teraction. The decision rule called for rejection of the null
baseline variables that differed significantly between hypothesis of no treatment effect if this interaction was
treatment groups and were correlated at 0.30 or higher statistically significant (two-tailed a = 0.05). In addition, LR
with the outcome (CAPS-IV scores) were included as tests compared model fit that included a treatment-by-
covariates. Unless stated otherwise, each statistical test site interaction and a first-order autoregressive covariance
was conducted with a two-tailed alpha of 0.05. We cal- structure (37).
culated medians, means, and standard deviations, and The secondary outcomes (mindfulness, depression, PTSD
change from baseline for the CAPS-IV (primary outcome) symptom clusters, and PTSD response rates) were ana-
and all other scales by treatment condition over time. We lyzed in separate three-level mixed-effects linear regression
calculated the within-treatment condition rate of response, models by using the strategy described above for the pri-
defined as a $30% decrease in CAPS-IV score. Both between- mary outcome. The categorical outcome measure (response
treatment condition and within-treatment condition effect status $30% decrease in CAPS-IV score from baseline) was
sizes were calculated: Cohen’s d for continuous outcomes and examined with mixed-effects logistic regression analysis.
the number needed to treat (NNT) for response rates. The The Hochberg multiplicity adjustment (38) was used for
effect size conveys a description of the magnitude of change analyses of all secondary outcomes, with a familywise alpha
that is independent of sample size (36). A 95% confidence level of 0.05.
interval accompanies each effect size to guide interpretation.
For all analyses, week 9 was considered the end of the acute
RESULTS
phase of treatment.
A three-level mixed-effects linear regression analysis Participants
was used to compare MBSR and PCGT treatment on total The CONSORT diagram (online supplement) shows the
CAPS-IV score during the trial. The data structure involved number of participants who provided informed consent
repeated measures over time nested within a participant, (N=254), were randomly assigned to a study condition
who in turn, was nested within a therapy group. Each model (N=214), included in the analysis (N=191), completed the
included up to four repeated assessments of the CAPS-IV week 9 assessment visit (N=142), and completed the week
as the dependent variable (baseline and weeks 3, 6, and 9). 16 assessment visit (N=130). Reasons for not being randomly
The models included a random intercept, a random slope, assigned (16%) included having a subthreshold CAPS-IV
and fixed effects for treatment condition, time, and the score (N=15), alcohol or drug use disorder (N=4), bipolar I

prcp in Advance prcp.psychiatryonline.org 5

MINDFULNESS-BASED STRESS REDUCTION IN POSTTRAUMATIC STRESS DISORDER

TABLE 3. Change from baseline for primary and secondary outcomes for veterans assigned to mindfulness-based stress reduction
(MBSR) or to present-centered group therapy (PCGT)a
Week 3 Week 6 Between-group differences Week 16
Scale M SD M SD pb d M SD 95% CI M SD pb
CAPS-IV (total score) .53 2.26 26.4 24.8 214.78, 1.97 .97
MBSR 212.9 19.9 213.6 24.8 218.3 30.6
PCGT 29.3 17.9 217.3 21.2 218.2 25.1
CAPS-B (re-experiencing) .81 2.16 21.5 9.6 24.75, 1.71 .92
MBSR 24.2 8.4 23.6 8.8 26.2 10.9
PCGT 24.6 7.6 26.6 8.1 26.4 9.2
CAPS-C (avoidance and emotional .42 .17 22.0 11.6 25.90, 1.95 .98
numbing)
MBSR 25.2 9.2 26.1 11.7 27.5 18.9
PCGT 23.1 9.3 27.1 10.8 27.5 11.3
CAPS-D (hyperarousal) .35 2.36 22.9 8.4 25.74, .08 .85
MBSR 28.5 7.3 23.9 8.8 24.6 9.1
PCGT 21.6 7.1 23.5 6.9 24.3 8.8
PTSD Checklist (self-report) .57 2.14 21.1 13.8 26.71, 2.9 .68
MBSR 24.3 12.0 26.3 13.3 26.6 14.4
PCGT 22.8 9.5 26.7 11.5 25.5 15.7
FFMQ (self-report) .40 .21 3.1 14.8 21.86, 8.11 .25
MBSR 1.9 12.2 3.9 14.6 6.3 20.5
PCGT 21.0 10.6 .5 9.9 2.6 14.9
PHQ-9 (self-report) .80 2.25 21.5 6.5 23.71, 0.64 .81
MBSR 21.1 5.5 22.0 5.7 23.0 7.4
PCGT 2.9 4.7 22.5 5.9 22.6 8.0
a
CAPS-IV, Clinician-Administered PTSD Scale for DSM-IV (total score range 0–136, CAPS-B range 0–40, CAPS-C range 0–56, CAPS-D range 0–40; higher
score indicates more severe PTSD symptoms); FFMQ, Five Facet Mindfulness Questionnaire (range 39–195; higher score indicates greater mindfulness);
PHQ-9, nine-item Patient Health Questionnaire (range 0–27; higher score indicates more severe depression).
b
p values are from the mixed model procedure.

disorder (N=5), not returning for random assignment (N=6), missed week 9 but continued in the study and attended the
withdrawing or moving away (N=7), or other (N=3). Reasons week 16 assessment.
for not being included in the analysis (N=23) were being lost Reasons for not completing the week 16 assessment (36%)
to follow-up (N=5), withdrawing consent (N=4), or entering in the MBSR group included being lost to follow-up (N=7)
addiction rehabilitation treatment (N=2) in the MBSR group and moving (N=3). Due to unavailability of a blinded rater,
and being lost to follow-up (N=7), withdrawing consent one CAPS-IV assessment was missing. Reasons for not com-
(N=3), and relocating (N=2) in the PCGT group. pleting the week 16 assessment (27%) in the PCGT group
In the MBSR group, 71 remained in the study; however, included being lost to follow-up (N=6), moving (N=4), and
only 65 (68%) attended the week 9 assessment. Reasons unknown (N=2). Due to unavailability of a blinded rater, one
for not completing the week 9 assessment visit (32%) CAPS-IV assessment was missing.
in the MBSR group included being lost to follow-up
(N=14), withdrawing (N=5), moving (N=2), investigator Baseline Demographic and Clinical Characteristics
withdrawing (N=1), noncompliance (N=2), and other No significant differences were observed between groups
(N=1). Due to absence of a blinded rater, the CAPS-IV as- in baseline demographic and clinical characteristics (Table 1).
sessment was not completed for three participants who On average, the sample was 84% male, one-third Caucasian,
attended the week 9 assessment, leaving 62 in the CAPS- and two-thirds African American or another racial-ethnic
IV primary end-point analysis. Six participants missed minority. Average age was 51, with a broad distribution of
week 9 but continued in the study and attended the week marital status. Fifty-six percent had served in the Army, 16%
16 assessment. had served in the military post-9/11, 48% had experienced
In the PCGT group, 81 remained in the study; however, combat-related trauma, 76% had more than a high school
only 77 (81%) attended the week 9 assessment. Reasons for education, and 80% were currently not working.
not completing the week 9 assessment visit (19%) in the
PCGT group included being lost to follow-up (N=8), in- Adherence to Group Therapy Sessions and
vestigator judgment that it was in the participant’s best Outcomes Assessments
interest to exit (N=2), work (N=2), unrelated adverse event Of participants who attended at least one group therapy
(N=1), and being detained in jail (N=1). Four participants session (analyzed sample), fewer MBSR participants than

6 prcp.psychiatryonline.org prcp in Advance

 DAVIS ET AL.

TABLE 4. Primary and secondary outcomes for veterans assigned to mindfulness-based stress reduction (MBSR) or to present-centered
group therapy (PCGT)a
Baseline Week 3 Week 6 Week 9 Week 16
Scale M SD M SD M SD M SD pb M SD
CAPS-IV (total score) .58
MBSR 84.3 19.5 71.1 24.3 68.6 26.9 59.1 27.1 64.8 30.8
PCGT 80.7 16.7 70.4 23.9 63.1 23.9 64.5 28.2 62.9 28.3
CAPS-B (re-experiencing) .73
MBSR 23.4 8.2 18.9 9.7 19.0 9.4 14.8 10.3 17.0 10.8
PCGT 23.0 6.9 18.3 8.9 16.1 8.9 16.5 9.8 16.8 9.9
CAPS-C (avoidance and emotional .42
numbing)
MBSR 34.6 8.4 29.2 11.4 27.9 12.3 24.9 12.5 26.5 13.8
PCGT 32.4 9.7 28.7 10.9 25.2 13.9 25.7 13.8 25.3 13.6
CAPS-D (hyperarousal) .49
MBSR 26.3 6.3 22.9 6.6 21.7 8.3 19.5 7.8 21.3 8.9
PCGT 25.2 5.4 23.4 8.2 21.8 7.2 22.3 7.9 20.8 8.5
PTSD Checklist (self-report) .04
MBSR 63.1 12.2 58.7 13.9 56.7 15.8 53.8 17.4 56.2 16.5
PCGT 58.7 12.2 55.3 12.4 57.9 14.2 52.6 14.4 53.8 15.9
FFMQ (self-report) .48
MBSR 109.9 21.0 110.9 20.1 114.7 20.6 116.9 20.2 117.2 23.6
PCGT 112.1 16.2 111.0 15.7 113.6 15.1 114.6 16.8 116.1 16.7
PHQ-9 (self-report) .54
MBSR 27.7 6.7 26.7 6.3 25.0 6.5 23.8 7.6 24.0 7.6
PCGT 26.8 6.9 25.6 6.2 24.0 6.1 24.5 7.4 24.5 7.4
a
CAPS-IV, Clinician-Administered PTSD Scale for DSM-IV (total score range 0–136, CAPS-B range 0–40, CAPS-C range 0–56, CAPS-D range 0–40; higher score
indicates more severe PTSD symptoms); FFMQ, Five Facet Mindfulness Questionnaire (range 39–195; higher score indicates greater mindfulness); PHQ-9,
nine-item Patient Health Questionnaire (range 0–27; higher score indicates more severe depression).
b
p values are from the mixed model procedure.

PCGT participants completed the outcomes assessments between groups (30.7% MBSR vs. 27.3% PCGT, p=0.662,
(68% vs. 81%, respectively, at week 9; 64% vs. 73%, re- between-group NNT=30).
spectively, at week 16; online supplement), but the difference Statistically significant correlations were observed be-
did not reach statistical significance. Table 2 shows the tween reduction in CAPS-IV scores and improvement in
distribution of attendance at each group therapy session. FFMQ scores in each group (MBSR, –0.509, p=0.01; PCGT,
Although the MBSR group had more participants attending –0.337, p=0.001). To check the comparability of the direction
fewer group sessions compared with the PCGT group, no and intensity of the combined CAPS-IV and FFMQ result
significant differences were observed in the number of group between the two groups, the difference in the CAPS-IV and
sessions attended between treatment arms (p=0.905). The FFMQ scores within each treatment group was calculated
participants’ adherence to treatment (attendance) did not and the correlation of the difference was then determined to
correlate with CAPS-IV or FFMQ outcomes at week 9 or 16. be 0.927 (representing the correlation of the direction and
intensity of the combined CAPS-IV and FFMQ results be-
Outcome Measures tween the two groups, which was high and statistically sig-
No statistically significant differences were observed be- nificant, p=0.023). These results show that the relationship
tween the MBSR and PCGT groups in terms of the primary between the CAPS-IV and FFMQ was comparable between
(CAPS-IV) or secondary outcomes, except for the PCL the two groups.
(Tables 3 and 4). A statistically greater improvement was
observed in PTSD based on the self-reported PCL in the Adverse Events
MBSR group compared with the PCGT group (Table 4). Both treatments were well tolerated. In the MBSR group, five
The participants’ baseline treatment preference (MBSR vs. unrelated serious adverse events occurred: one psychiatric
PCGT vs. none) had no moderating effect on CAPS-IV scores inpatient admission for suicidal ideation of a participant who
at week 9 (p=0.734) or 16 (p=0.741). Rates of response, de- was randomly assigned to MBSR but had not yet attended a
fined as CAPS-IV reduction $30%, did not statistically treatment session, one participant with a wisdom tooth in-
differ between groups (45.2% MBSR vs. 37.7% PCGT, fection that resulted in medical admission, and three par-
p=0.293, between-group NNT=12). Rates of remission, ticipants with psychiatric inpatient admissions for suicidal
defined as CAPS-IV score #45, did not statistically differ ideation. In the PCGT group, two unrelated serious adverse

prcp in Advance prcp.psychiatryonline.org 7

MINDFULNESS-BASED STRESS REDUCTION IN POSTTRAUMATIC STRESS DISORDER

events occurred: one participant with a psychiatric inpatient we used three sites instead of one. The multisite approach
admission for suicidal ideation and one participant with may have increased variability in patient selection, treatment
medical hospitalization for hypotension. delivery, and CAPS-IV scores. Our study had substantial
minority representation: two-thirds were African American
compared with only 8% in the Polusny et al. study. To balance
DISCUSSION
the treatment interventions, we limited the MBSR groups
In this multisite randomized controlled study of veterans to 90-minute sessions to match the recommended PCGT
diagnosed as having PTSD, the MBSR and PCGT inter- session time, whereas Polusny and colleagues held MBSR
ventions improved PTSD symptoms over time, with no groups for 2.5 hours and PCGT groups for 90 minutes, thus
significant differences between groups on the CAPS-IV providing more experiential practice and reinforcement
scale. Although the results of the self-reported PTSD for the MBSR groups. Compared with Polusny et al., our study
Checklist differed significantly between groups at week 9, had higher attrition rates overall, most notably in the MBSR
this difference was not maintained at week 16. Although group (10% vs. 32%, respectively).
PCGT was selected as a control group, it has many active Our results were included in a meta-analysis (41) of
therapeutic elements, including group cohesion, validation nine randomized controlled trials of PTSD, which found an
from other veterans, health education, and psychosocial overall effect size of –0.34 (p,0.001, 95% CI=–0.48, –0.18) for
support, and these attributes may explain why the two mindfulness-based meditation. All but one of these trials had
treatment groups yielded similar results. For example, an active control group. In addition to these studies, a pilot
PCGT was shown to have similar positive outcomes as a study conducted to examine the effects of MBSR versus
group-based exposure therapy in a VA multisite study of PCGT among veterans with PTSD found that MBSR was
male Vietnam veterans with PTSD (33). A larger sample size associated with changes in functional response to exposure to
may be needed to better differentiate responses between the stressor of Iraq combat-related slides and sounds in the
the two treatments. anterior cingulate, parietal cortex, and insula (27). This small
It is not clear why a significant difference between treat- comparison study showed a reduction in both CAPS-IV and
ments was found on the self-reported PCL versus the FFMQ scores in the MBSR group but not in the PCGT group,
CAPS-IV. Perhaps veterans in the MBSR group felt greater with the effects maintained at six months.
efficacy and control, which led to a perception of clinical
improvement that was not noted by investigators charting
symptoms. Mindfulness is an approach that may improve ap- CONCLUSIONS
praisals of secondary control (39) and acceptance of thoughts In conclusion, MBSR did not have a significant advantage
and emotions (40). over PCGT in our sample. The overall small effect sizes of
Strengths of our study include its multisite design, active mindfulness-based meditation should be viewed with caution
control group (i.e., PCGT rather than waitlist control), in the context of larger effect sizes of trauma-focused be-
blinded primary outcome ratings, fidelity monitoring, large havioral psychotherapies. As with all complementary and
minority representation, and randomized approach. This integrative health approaches, mindfulness-based medita-
study was limited by its high attrition rate and brief MBSR tion should be a supplement to, not a replacement for, trauma-
group sessions. We shortened MBSR from its usual 160- or focused behavioral psychotherapies. Additional studies are
180-minute format to match the 90-minute duration of the needed to better understand the effects of MBSR for the
PCGT groups. Other researchers may want to include a treatment of PTSD.
psychoeducation overview prior to participants committing
to an eight-week MBSR treatment as a means to reduce at- AUTHOR AND ARTICLE INFORMATION
trition and may want to provide access to MBSR practice
Research and Development Service, Veterans Affairs Medical Center,
groups during long-term follow-up, as would be provided in Tuscaloosa, AL (Davis); Department of Psychiatry and Behavioral Sciences,
real-world settings. University of Alabama School of Medicine, Birmingham (Davis); private
A similarly designed study by Polusny et al. (26) found practice, Birmingham (Whetsell); Mental Health Service, Ralph H. Johnson
a greater decrease on the self-reported PCL for the MBSR VA Medical Center, Charleston, SC (Hamner); Department of Psychiatry,
Medical University of South Carolina, Charleston (Hamner); Department
treatment group compared with the PCGT group, as well
of Medicine, Division of Preventive and Behavioral Medicine, University
as significant improvements in CAPS-IV and quality-of-life of Massachusetts Medical School, Worcester (Carmody); Department of
scores at week 17. Our study differed from the Polusny et al. Psychiatry and Behavioral Sciences, Emory University, Atlanta (Rothbaum,
study in several ways, including sample size, number of sites Bremner); Alabama Research Institute on Aging and Department of
and therapists, demographic characteristics of the study Psychology, University of Alabama, Tuscaloosa (Allen); Department of
Biostatistics, University of Alabama at Birmingham (Bartolucci); Veterans
sample, length of the MBSR group sessions, response to
Affairs National Center for PTSD, Connecticut VA Healthcare System, West
PCGT, primary outcome measure, and attrition rates. These Haven, and Department of Psychiatry, Yale University, New Haven, CT
factors may explain the differences in the findings. Our study (Southwick); Mental Health Service, Atlanta VA Medical Center, Decatur,
enrolled more participants (N=214 vs. N=116) who scored GA (Bremner).
approximately 15 points higher on baseline CAPS-IV, and Send correspondence to Dr. Davis (lori.davis@va.gov).

8 prcp.psychiatryonline.org prcp in Advance

 DAVIS ET AL.

This work was supported by the Department of Veterans Affairs Office of 13. Lazar SW, Kerr CE, Wasserman RH, et al: Meditation experience is
Research and Development, Clinical Science Research and Development associated with increased cortical thickness. Neuroreport 2005; 16:
(Award 5-101-CX000687-02). 1893–1897
The authors gratefully acknowledge the collaborations of Beverly Thorn, 14. Cozolino L: The Neuroscience of Psychotherapy: Healing the Social
Ph.D., and Mary Fran Burnette, L.C.S.W.; the MBSR therapists; the PCGT Brain, 2nd ed. New York, WW Norton, 2010
therapists; the clinical research assistants; the late Andrew Leon, Ph.D.; 15. Astin JA: Stress reduction through mindfulness meditation: effects
and the technical and editorial assistance of Katherine Thompson. on psychological symptomatology, sense of control, and spiritual
experiences. Psychother Psychosom 1997; 66:97–106
These views represent the opinions of the authors and are not necessarily
16. Brewer JA, Sinha R, Chen JA, et al: Mindfulness training and stress
those of the Department of Veterans Affairs or the U.S. Government.
reactivity in substance abuse: results from a randomized, controlled
Clinical Trial Registration (ClinicalTrials.gov): NCT01532999 stage I pilot study. Subst Abus 2009; 30:306–317
Dr. Davis reports consulting fees from Bracket, Lundbeck, Otsuka, and 17. Weissbecker I, Salmon P, Studts JL, et al: Mindfulness-based stress
Tonix and research funding from Allergan, Forest, Merck, the Substance reduction and sense of coherence among women with fibromyalgia.
Abuse and Mental Health Services Administration, Tonix, and the U.S. J Clin Psychol Med Settings 2002; 9:297–307
Department of Veterans Affairs. Dr. Rothbaum reports equity in Virtually 18. Meyer EC, Frankfurt SB, Kimbrel NA, et al: The influence of
Better, Inc.; research funding from the Brain and Behavior Research mindfulness, self-compassion, psychological flexibility, and post-
Foundation (National Alliance for Research on Schizophrenia and Af- traumatic stress disorder on disability and quality of life over time in
fective Disorders), Department of Defense, McCormick Foundation, war veterans. J Clin Psychol 2018; 74:1272––1280. doi: 10.1002/jclp.
National Institute of Mental Health, and the Wounded Warrior Project; 22596
royalties from American Psychiatric Association Publishing, Guilford, 19. Laneri D, Krach S, Paulus FM, et al: Mindfulness meditation reg-
Emory University, and Oxford University Press; and an advisory board ulates anterior insula activity during empathy for social pain. Hum
payment from Genentech. Dr. Bremner reports research funding from the Brain Mapp 2017; 38:4034–4046
Defense Advanced Research Projects Agency; National Heart, Lung, and 20. Tang YY, Hölzel BK, Posner MI: The neuroscience of mindfulness
Blood Institute; and National Institute of Mental Health. The other authors meditation. Nat Rev Neurosci 2015; 16:213–225
report no financial relationships with commercial interests. 21. Miller W, Rollnick S: Motivational Interviewing: Helping People
Change. New York, Guilford Press, 2013
Received February 15, 2018; revision received June 22, 2018; accepted
22. Teasdale JD, Segal Z, Williams JM: How does cognitive therapy
June 27, 2018; published online September 13, 2018.
prevent depressive relapse and why should attentional control
(mindfulness) training help? Behav Res Ther 1995; 33:25–39
REFERENCES 23. Shapiro SL, Lopez AM, Schwartz GE, et al: Quality of life and breast
1. Siegel DJ: The Mindful Therapist: A Clinician’s Guide to Mindsight cancer: relationship to psychosocial variables. J Clin Psychol 2001;
and Neural Integration. New York, WW Norton, 2010 57:501–519
2. Doidge N: The Brain That Changes Itself: Stories of Personal Triumph 24. Gu J, Strauss C, Bond R, et al: How do mindfulness-based cognitive
From the Frontiers of Brain Science. New York, Viking Press, 2007 therapy and mindfulness-based stress reduction improve mental
3. Averill LA, Averill CL, Kelmendi B, et al: Stress response modula- health and wellbeing? A systematic review and meta-analysis of
tion underlying the psychobiology of resilience. Curr Psychiatry Rep mediation studies. Clin Psychol Rev 2015; 37:1–12
2018; 20:27 25. Hempel S, Taylor SL, Marshall NJ, et al: Evidence Map of Mind-
4. Watts BV, Schnurr PP, Mayo L, et al: Meta-analysis of the efficacy of fulness. VA Evidence-Based Synthesis Program Reports. Project
treatments for posttraumatic stress disorder. J Clin Psychiatry 2013; 05-226. Washington, DC, US Department of Veterans Affairs, Oct
74:e541–e550 2014
5. Lee DJ, Schnitzlein CW, Wolf JP, et al: Psychotherapy versus 26. Polusny MA, Erbes CR, Thuras P, et al: Mindfulness-based stress
pharmacotherapy for posttraumatic stress disorder: systemic review reduction for posttraumatic disorder among veterans: a randomized
and meta-analyses to determine first-line treatments. Depress Anxiety clinical trial. JAMA 2015; 314:456–465
2016; 33:792–806 27. Bremner JD, Mishra S, Campanella C, et al: A pilot study of the ef-
6. Clarke TC, Black LI, Stussman BJ, et al: Trends in the use of fects of mindfulness-based stress reduction on post-traumatic stress
complementary health approaches among adults: United States, disorder symptoms and brain response to traumatic reminders of
2002–2012. Natl Health Stat Rep 2015; 10:1–16 combat in Operation Enduring Freedom/Operation Iraqi Freedom
7. Carmody J: Re-conceptualizing mindfulness: the psychological combat veterans with post-traumatic stress disorder. Front Psy-
principles of attending in mindfulness practice and their role in chiatry 2017; 8:157
well-being; in Handbook of Mindfulness: Theory, Research, and 28. Sheehan DV, Lecrubier Y, Sheehan KH, et al: The Mini-International
Practice. Edited by Brown KW, Creswell JD, Ryan RM. New York, Neuropsychiatric Interview (MINI): the development and valida-
Guilford Press, 2015, pp 62–78 tion of a structured diagnostic psychiatric interview for DSM-IV
8. Teasdale JD, Segal ZV, Williams JM, et al: Prevention of relapse/ and ICD-10. J Clin Psychiatry 1998; 59(Suppl 20):22–33
recurrence in major depression by mindfulness-based cognitive 29. Blake DD, Weathers FW, Nagy LM, et al: The development of a
therapy. J Consult Clin Psychol 2000; 68:615–623 Clinician-Administered PTSD Scale. J Trauma Stress 1995; 8:
9. Carmody J: Evolving conceptions of mindfulness in clinical settings. 75–90
J Cogn Psychother 2009; 23:270–280 30. Baer RA, Smith GT, Lykins E, et al: Construct validity of the five facet
10. Ryan RM, Kuhl J, Deci EL: Nature and autonomy: an organizational mindfulness questionnaire in meditating and nonmeditating sam-
view of social and neurobiological aspects of self-regulation in be- ples. Assessment 2008; 15:329–342
havior and development. Dev Psychopathol 1997; 9:701–728 31. Baer RA, Smith GT, Hopkins J, et al: Using self-report assessment
11. Carmody J, Baer RA: Relationships between mindfulness practice methods to explore facets of mindfulness. Assessment 2006; 13:
and levels of mindfulness, medical and psychological symptoms 27–45
and well-being in a mindfulness-based stress reduction program. 32. Carmody J, Reed G, Kristeller J, et al: Mindfulness, spirituality,
J Behav Med 2008; 31:23–33 and health-related symptoms. J Psychosom Res 2008; 64:393–
12. Carmody J, Baer RA, Lykins ELB, et al: An empirical study of the 403
mechanisms of mindfulness in a mindfulness-based stress reduction 33. Schnurr PP, Friedman MJ, Foy DW, et al: Randomized trial of
program. J Clin Psychol 2009; 65:613–626 trauma-focused group therapy for posttraumatic stress disorder:

prcp in Advance prcp.psychiatryonline.org 9

MINDFULNESS-BASED STRESS REDUCTION IN POSTTRAUMATIC STRESS DISORDER

results from a Department of Veterans Affairs cooperative study. 37. Lachin JM: Statistical considerations in the intent-to-treat principle.
Arch Gen Psychiatry 2003; 60:481–489 Control Clin Trials 2000; 21:167–189
34. Barber JP, Liese BS, Abrams MJ: Development of the Cognitive 38. Hochberg Y: A sharper Bonferroni procedure for multiple tests of
Therapy Adherence and Competence Scale. Psychother Res 2003; significance. Biometrika 1988; 75:800–803
13:205–221 39. Heckhausen J, Schulz R: Optimisation by selection and compen-
35. Segal ZV, Teasdale JD, Williams JM, et al: The mindfulness-based sation: balancing primary and secondary control in life span de-
cognitive therapy adherence scale: inter-rater reliability, adherence velopment. Int J Behav Dev 1993; 16:287–303
to protocol and treatment distinctiveness. Clin Psychol Psychother 40. Hayes SC, Strosahl KD, Wilson KG: Acceptance and Commitment
2002; 27:131–138 Therapy, 2nd ed. New York, Guilford Press, 2012
36. Leon AC, Shear MK, Portera L, et al: Effect size as a measure of 41. Gallegos AM, Crean HF, Pigeon WR, et al: Meditation and yoga for
symptom-specific drug change in clinical trials. Psychopharmacol posttraumatic stress disorder: a meta-analytic review of randomized
Bull 1993; 29:163–167 controlled trials. Clin Psychol Rev 2017; 58:115–124

10 prcp.psychiatryonline.org prcp in Advance