Journal of Behavioral Medicine, Vol. 8, No.

2, 1985

The Clinical Use of Mindfulness Meditation

for the Self-Regulation of Chronic Pain

J o n K a b a t - Z i n n , ',2 Leslie L i p w o r t h , 3 and R o b e r t B u r n e y 4

Accepted for publication: May 1, 1984

Ninety chronic pain patients were trained in mindfulness meditation in a

lO-week Stress Reduction and Relaxation Program. Statistically significant
reductions were observed in measures o f present-moment pain, negative body
image, inhibition o f activity by pain, symptoms, mood disturbance, and psy-
chological symptomatology, including anxiety and depression. Pain-related
drug utilization decreased and activity levels and feefings o f self-esteem in-
creased. Improvement appeared to be independent of gender, source o f refer-
ral, and type o f pain. A comparison group o f pain patients did not show
significant improvement on these measures after traditional treatment pro-
tocols. A t follow-up, the improvements observed during the meditation train-
ing were maintained up to 15 months post-meditation training for all measures
except present-moment pain. The majority o f subjects reported continued
high compliance with the meditation practice as part o f their daily lives. The
relationship o f mindfulness meditation to other psychological methods for
chronic pain control is discussed.
KEY WORDS: meditation; pain; self-regulation; coping; stress.

'Stress Reduction and Relaxation Program, Division of Preventive and Behavioral Medicine,
Department of Medicine, University of Massachusetts Medical Center, Worcester, Massachusetts
01605.
21o whom correspondence should be addressed.
3Department of Family and Community Medicine, Universily of Massachusetts Medical Center,
Worcester, Massachusetts 01605.
"Pain Control Center, Department of Anesthesiology, University of Massachusetts Medical
Center, Worcester, Massachusetts 01605.
163
0160-7715/85/0600 0163504.50/0 9 1985 Plenum PublishingCorporation
164 Kabat-Zinn, Lipworth, and Burney

INTRODUCTION

Achieving improvement in the quality of life for individuals suffering

from chronic pain presents a profound dilemma for the clinician, perhaps
reflecting a deep dilemma in the paradigm of medicine itself (Cassel, 1982;
McCue, 1982). In spite of the modern armamentarium for directly treating
persistent pain with analgesics, narcotics, and surgery, reliable relief from
chronic pain in many cases remains an elusive goal. With this has come the
recognition of the need to intervene to treat the psychological dimensions
of chronic pain (Melzack and Wall, 1970; Sternback, 1978). This view has
led to a range of psychological interventions, many emphasizing strategies of
self-regulation. The latter have included biofeedback, relaxation training, hyp-
nosis, and cognitive-behavioral therapies (see Melzack and Wall, 1983; Turk
et al., 1983). Studies employing psychological modalities in the treatment
of chronic pain have been recently reviewed and critically discussed by Turner
and Chapman (Turner and Chapman, 1982a,b). All the modalities reviewed
have proven useful in certain circumstances, and all have particular limita-
tions in perspective (Turner and Chapman, 1982a,b).
In an even more recent development, meditative practives from orien-
tal traditions such as Zen Buddhism, Vipassana, and Yoga, without their
original religious, cultural, and ideological forms, have been introduced into
some therapeutic settings as strategies for self-regulation (Shapiro, 1980;
Deatherage, 1975; Shapiro and Giber, 1978; Kutz et al., 1985a,b) and have
also become the focus of systematic research efforts (Benson, 1975; Burns
and Ohayv, 1980; Davidson, 1976; Maliszewski, 1981; Walsh, 1977, 1978,
1983; Woolfolk, 1975). It has been suggested that these practices, collective-
ly termed "consciousness disciplines," are based on assumptions about hu-
man nature which differ in fundamental ways from the paradigms upon which
Western psychology and behavior science rest (Walsh, 1980). These assump-
tions include (1) that "our usual state of consciousness is severely suboptimal"
and (2) that "through intensive mental training it is possible to attain states
of consciousness and psychological well-being beyond those currently
described by traditional Western psychologies, as well as profound insight
into the nature of mental processes, consciousness, and reality" (Walsh, 1980).
Commenting on methods of psychological transformation, C. G. Jung once
remarked that the "methods and philosophical doctrines [that] have been de-
veloped [in the East] simply put all Western attempts along these lines into
the shade" (Jung, 1969). If such views have any substance, they suggest that
the meditative traditions may have important and unique viewpoints and
methods to offer behavioral science in general and clinical behavioral medi-
cine in particular (Deikman, 1982). It is also plausible that the relaxation
exercises and cognitive and behavioral therapies developed within the Western
Meditation for Self-Regulation of Pain 165

psychological paradigm might be further developed and deepened via ex-

posure to rigorous meditation practice and a systematic study of its empiri-
cal effects (Burns, 1973). There is no doubt that the Eastern traditions can
also benefit from the psychological sophistication of the West (see, e.g., But-
ler, 1983) and that both paradigms can be enriched by cross-fertilization.
It was with these notions in mind that we chose to create a behavioral medi-
cine clinic based on training in meditation.
This paper describes the clinical use of relatively intensive training in
the consciousness discipline known generically as mindfulness or awareness
meditation in a hospital outpatient stress reduction program and the out-
come for 90 patients referred to it for chronic pain conditions in its first 2
years of operation. Preliminary results have been reported (Kabat-Zinn and
Burney, 1981; Kabat-Zinn, 1982). Initial observations suggested that train-
ing in meditation is acceptable to a broad spectrum of medical outpatients
and can be effective in reducing pain and pain-related behaviors for a range
of chronic pain conditions.
Mindfulness meditation has roots within Theravada Buddhism, where
it is known as satipatana vipassana or insight meditation (Nyanaponika,
1962), in Mahayana Buddhism in Soto Zen practices (Suzuki, 1970), and in
the Yogic traditions as expressed in the contemporary writings of J. Krish-
namurti (1979), Vimila Thakar (1977), and Nisargadatta Maharaj (1973). This
form of meditation is a highly developed, coherent, systematic, and mul-
timodal utilization of attention. One of its primary goals is the development
of "insight" into the actuality of phenomena, achieved by the cultivation of
what the Buddhists refer to as "bare attention" or "detached observation"
(Nyanaponika, 1962). This is a moment-to-moment effort to perceive a
phenomenon and to allow it to register with full awareness, as it is, without
gross distortion of the bare percept from associated and second-order mean-
ings to the ego of the observer (see Naranjo and Ornstein, 1971). The medi-
tation instructions themselves are an active support in minimizing distortion
of this kind.
In the case of pain perception, the cultivation of detached observation
of the pain experience may be achieved by paying careful attention to and
distinguishing as separate events the actual primary sensations as they occur
from moment to moment and any accompanying thoughts about pain. The
rationale for the choice of this form of meditation and a description of its
use in the stress reduction program have been presented elsewhere (Kabat-
Zinn, 1982).
In this outcome study, we sought to address the specific questions list-
ed below. For clarity, after each question the indices which measure the rele-
vant parameters in this study are listed in parentheses (see Methods).
166 Kabat-Zinn, Lipworth, and Burney

Can mindfulness meditation training in the context of stress reduction

effectively
(1) reduce pain levels over an extended period of time (10 weeks)?
(PRI, BPM);
(2) lead to improvement in body image and reduced somaticizing?
(BPPA, SOM scale of SCL-90-R);
(3) help in coping with persistent pain so that it will interfere less with
the performance of routine activities of normal living? (TLI);
(4) reduce the characteristically elevated negative affective states in
chronic pain patients, in particular depression, hostility, low self-
esteem, and anxiety? (POMS, SCL-90-R);
(5) compare favorably in outcome with more traditional and more
expensive medical treatments for the same pain conditions? (com-
parison with nonmeditating Pain Clinic patients);
(6) produce positive long-term improvements in pain, coping be-
haviors, and affect? (follow-up questionnaire); and
(7) lead to a continued, voluntary practice of the meditation follow-
ing training? (follow-up questionnaire).

METHODS

Program Design

The meditation training took place in a 10-week Stress Reduction and

Relaxation Program (SR&RP). The SR&RP is a clinical service of the Divi-
sion of Preventive and Behavioral Medicine in the Department of Medicine
at the University of Massachusetts Medical Center. Chronic pain is one rea-
son for referral to this program. Approximately 60% of the patients are
referred for stress-related medical problems having nothing to do with pain.
This report concerns only those patients referred in the first 2 years of the
program with a diagnosed pain condition of greater than 6 months' dura-
tion, well substantiated by medical history, who had not improved with tradi-
tional medical care. All patients were physician referred.
Each individual was seen initially in an evaluation interview which in-
cluded a detailed description of the program. The description emphasized
that the program was educational in nature and that a high degree of dis-
cipline on the patient's part was required. It was explained that the SR&RP
was based on intensive, daily practice of meditation and on the practical
application of meditation for coping with stress and pain. The program was
explicitly differentiated from behavior modification programs and from group
therapy. If the patient chose to enroll, a battery of interviewer-administered
Meditation for Self-Regulation of Pain 167

and self-report questionnaires (see below) was given. The information from
these instruments constituted the pre-meditation-training data base (pre).
The SR&RP courses are conducted in cycles three times a year. Each
cycle consists of ten 2-hr classes, one per week, in which a variety of forms of
fulness meditation are taught and practiced [for details see Discussion and
Kabat-Zinn (1982)].5 All subjects in this study were required to meditate for-
mally for a minimum of 45 min per day, 6 days per week, for homework,
using an audiocassette tape in the beginning weeks for guidance. Instruction
and practice of Hatha Yoga were included as a form of meditative exercise
for those who could do it. It functioned primarily to improve musculoskele-
tal strength and flexibility and reduce disuse atrophy. The Yoga was taught
emphasizing mindfulness (Kabat-Zinn, 1982).
Each SR&RP course was conducted by an instructor on the SR&RP
staff. The instructors have practiced mindfulness meditation regularly for
many years and continue to engage in periodic retreats for intensive training
and practice.
Following the course, each patient was seen individually in a second
evaluation interview, during which post-meditation-training data (post) were
obtained.

Patient Characteristics

The subjects in this study were trained in meditation in five consecu-

tive 10-week cycles of the SR&RP in 1980 and 1981. Referrals were from
four major sources within the hospital: the Pain Control Center (Pain Clin-
ic; PC) [low back, neck, shoulder, arm, leg, and facial pain and multiple-
site pain (chronic pain syndrome) (N -- 21); the Orthopedic Clinic (similar
to the PC profile)] (N = 18); the Neurology Clinic (headaches, including
migraine and tension; low back pain; and peripheral nerve problems) (N =
8); and the Adult Primary Care Clinic (headaches and chest pain) (N = 23).
The remaining 20 subjects in this study were referred for pain problems from
medical subspecialists such as gastroenterologists and cardiologists, from psy-
chiatrists, or from physicians outside the hospital.
All patients were enrolled if they met the entry criteria for the SR&RP
and agreed to make the necessary commitment of time and effort. Over 90%
of the patients contacted after referral came for an initial evaluation inter-
view, and 80 to 90% of those enrolled in the program. Of the patients begin-
ning the program, 80-90% completed it. These percentages varied within these
limits for different cycles of the program. Data from all the cycles have been

5The SR&RP was recentlychanged to an 8-week course including an additional 8-hr intensive
"retreat" session.
168 Kabat-Zinn, Lipworth, and Burney

pooled and averaged in reporting the results except for the follow-up study,
in which each cycle is plotted separately. The patient characteristics are shown
in Table IA. The majority had long histories of medical treatment for their
conditions, with little or no improvement in either pain status or affective
and" cognitive/behavioral status prior to enrolling in the SR&RP.
Subsequent Intervention. The SR&RP offers as a sequel an "advanced"
course to deepen the process begun in the initial mediation training. This
graduate SR&RP is an 8-week course with a format similar to that the of
basic SR&RP. The periods of meditation are longer and less guided. Some
of the patients in this study had taken one or more graduate courses at the
time some of the follow-up data were obtained (see Results).

Follow-Up

Follow-up data were solicited from all patients who completed the
SR&PR by periodic mailing of questionnaires at approximately 2.5, 4.5, 7,
12, and 15 months after completion of the program. In addition to follow-
up information on pain and psychological status, detailed information was
obtained about whether and how much individuals were meditating and about
the techniques they had found the most useful.

Comparison Group

One of our objectives was to compare the outcome of this nontradi-

tional approach with that of a more traditional medical, pain-specific approach
with similar patients. To this end we compared outcomes between pain pa-
tients trained in meditation and pain patients undergoing treatment but who
were not trained in any form of self-regulation. This was possible because
the same battery of data instruments given to the meditators pre and post
was also employed with all outpatients in the Pain Clinic at the initial visit
and at a clinic visit 10 weeks later. A cohort of 21 consecutive patients who
were being treated by the traditional methods of the PC (which include nerve
blocks, TENS, physical therapy, analgesics, antidepressants) and who had
not been referred to the SR&RP at the time of the comparison was moni-
tored over a 10-week period (PC comparison group). The outcome for these
individuals was compared with that for those patients (N = 21) referred to
the SR&RP from the Pain Clinic (PC referrals). During the meditation train-
ing some, but not all, of the PC referrals received periodic treatments at the
Pain Clinic and continued to take prescribed medication. The remainder had
completed their treatmentcourse in the pain Clinic. In all cases, however,
individuals were referred to the SR&RP because they had continued to have
pain without improvement at the time of referral.
Meditation for Self-Regulation of Pain 169

It is important to note that this is not a prospective randomized study

but a descriptive comparison of two functioning hospital clinics. In both clin-
ics, referral includes elaborate placebo concomitants such as enthusiastic refer-
ral to a special program and high expectation of pain relief. It was
hypothesized that both cohorts in the comparison would show positive
changes in pain status associated with these placebo elements to a similar
extent and that quantitative differences due to the specific interventions might
be distinguishable in the comparison. However, the fact that the patients were
not distributed randomly to the two interventions to be compared means that
conclusions based on the observations must be limited (see Discussion). The
relevant differences between the comparison groups are cited in Table IB.

Pain Indices and Psychological/Behavioral Measures

A number of different self-report indices was used to assess the multi-

ple aspects of pain and certain pain-related behaviors of interest to us in ad-
dressing the questions posed in the Introduction. The McGill-Melzack Pain
Rating Index (PRI) (Melzack, 1975) measured present-moment pain; the Body
Parts Problem Assessment (BPPA) Scale (Kabat-Zinn, 1983) measured how
problematic the patient viewed various body parts; the Table of Levels of
Interference (TLI) measured how pain affected activities of normal living
(Kabat-Zinn, 1982); and the three-color Body Pain Map (BPM) 8 assessed
changes in pain distribution, intensity, and frequency (Kabat-Zinn, 1982).
In addition, the total number of symptoms reported in the preceding month
was monitored using a medically oriented symptom Checklist (MSCL) (Kabat-
Zinn, 1982). These indices measure overlapping aspects of pain and are not
completely independent of each other.
Affective status was assessed using the Profile of Mood States (McNair
et al., 1971) and is represented in the results by the summary Total Mood
Disturbance (TMD) score.
Psychological symptomatology was assessed using the revised Hopkins
Symptom Checklist (SCL-90-R). The SCL-90-R is a validated 90-item inven-
tory (Derogatis, 1977) consisting of nine symptom dimensions: Somatiza-
tion (perceptions of bodily dysfunction), Obsessive-Compulsive, Interpersonal
Sensitivity (feelings of personal inadequacy and inferiority, lack of self-
esteem); Depression, Anxiety, Hostility, Phobic Anxiety, Paranoid Ideation,
and Psychoticism. The summary score, termed the General Severity Index
(GSI), combines information on numbers of symptoms and intensity of per-

~Previouslyused with dermatomesand referred to as the DermatomePain Map (DPM).

 Table I. Patient Characteristics
Major diagnostic category
Mean Mean No. o f
No. of No. of age chronicity' surgeries Neck/shoul-
males females (years) (years) for pain LBP Headache der Other
(A) Total population
(N = 90) 30 60 44 8.1 -- 31 24 15 20
(B) PC referrals
(iV = 21) 5 16 48 8.0 1.95 13 2 6
PC comparison group
(iV = 21) 13 8 37 4.6 1.10 9 9 3

e,

ca.
Meditation for Self-Regulation of Pain 171

ceived distress (Derogatis, 1977). The SCL-90-R has been shown to corre-
late well with comparable scales on the M M P I (Derogatis et al., 1976).
The POMS and the SCL-90 (R) were employed together to obtain pro-
files of affective status and psychological symptoms since chronic pain is
known to cause or be accompanied by severe mood disturbance as well as
by depression, loss of self-esteem, irritability, and anxiety. These instruments
have been shown to have independent predictive variance (Haskell et al.,
1969).
A Summary Outcome Questionnaire was used with the SR&RP patients
both post-meditation training and at follow-up. This instrument was designed
to yield a single number representative of the average degree of change in
10 relevant symptom and behavioral parameters since taking the SR&RP.
It consisted of 10 questions pertaining to pain frequency, severity, use of
drugs to control pain, activity levels, attendance at work, energy levels, feel-
ings in general, ability to cope with stress, frequency of physician visits, and
blood pressure. The rating scale was from 1 to 5, where 3 represented "no
change," 5 "great improvement," 4 "some improvement," 2 "worse," and 1
"much worse," with the exact wording of each scale topic appropriate. Rat-
ings for the 10 questions were averaged to give the Summary Outcome Score.
If certain items were not applicable, the patient circled an option to that ef-
fect, and the average was calculated for the number of questions answered.

Data Analysis

Data were analyzed using standard SPSS programs. The matched t test
was used to determine significance for paired pre-post or pre-follow-up
results for the same subjects over time. The unmatched t test was used to
determine significance in the comparison of outcomes for the PC patients
trained in meditation with the patients in the PC comparison group. The
Bonferroni adjustment was then applied to the P values from all t tests as
recommended by Ingelfinger et al. (1983) to reduce the risk of type 1 error
from multiple comparisons. Further analyses to identify possible predictors
of outcome were undertaken using linear regression and discriminatory anal-
ysis (unpublished results).

RESULTS

Pain Outcome

Outcome was first analyzed for the total group. For every pain index,
the mean value was reduced significantly (P < 0.003) between pre- and postin-
172 Kabat-Zinn, Lipworth, and Burney

Table II. Pain Outcomes for the Total Population

(A) Group means~
PRI BPPA TLI
Pre 19.3 41.6 11.8
Post 8.2 29.6 8.2
% change in mean 58* 29* 30*
N 57 87 61

(B) Individual gainsb

APRI ABPPA ABPM
Level of reduction _>33.3~ _>50% _>33.3% _>50% + +/+ + +
Number reaching
this level 41/57 35/57 48/86 36/86 42/87
% of
total patients 72 61 56 42 48
aPRI, Pain Rating Index; BPPA, Body Parts Problem Assessment score; TLI, Ta-
ble of Levels of Interference with daily activities.
bFraction of individuals achieving the indicated level of reduction or improvement
on each index. The TLI is not included. BPM, Body Pain Map; scored by compari-
son of pre and post drawingsof the patient's pain (see Kabat-Zinn, 1982). The fraction
represents the number of individuals in the population who were scored as either
+ + (moderate improvement) or + + + (great improvement). The PR1 values have
been corrected for zero values (Melzack, 1975).
*P < 0.003 in t test adjusted for multiple comparisons using the Bonferroni method
(Ingelfinger et al., 1983).

tervention assessments (Table IIA). Quantitatively similar resuks were obtained

for the m e a n percentage change (Melzack a n d Perry, 1975) in each index
(data n o t shown). The g r o u p m e a n value o f the P R I was reduced 58%, that
of the B P P A was reduced 29o70, a n d that of the T L I was r e d u c e d 30%.
O u t c o m e was also expressed in terms o f i n d i v i d u a l a c h i e v e m e n t s (Ta-
ble liB) following the format of Melzack and Perry (1975). By the end o f
m e d i t a t i o n t r a i n i n g , the large m a j o r i t y (72 a n d 56%, respectively) o f the pa-
tients had achieved levels o f p a i n r e d u c t i o n o n the P R I a n d negative b o d y
image o n the B P P A of greater t h a n or equal to 33%, a n d 61 a n d 4 2 % , respec-
tively, achieved r e d u c t i o n s o f greater t h a n 50%. The A B P M c o l u m n in Ta-
ble IIB shows that 48% achieved m o d e r a t e to great i m p r o v e m e n t
( + + / + + + ) b e t w e e n initial drawings of their p a i n a n d drawings d o n e fol-
lowing m e d i t a t i o n training.

Sym0tom, M o o d , and Psychological Outcome

In parallel with the pain outcome, the mean scores for the n u m b e r
of symptoms reported for the preceding m o n t h (MSCL), m o o d disturbance
Medilation for Self-Regulation of Pain 173

(TMD), and psychological symptomatology (GSI) were reduced by 35, 55,

and 35%, respectively ( P < 0.003) (Table IIIA). The majority (54 and 59%,
respectively) achieved reductions of greater than 33% on the MSCL and the
GSI, 7 and 37 and 39%, respectively, achieved reductions of greater than 50%
(Table IIIB). There were significant mean reductions in all dimensions of
the SCL-90-R; these were largest for anxiety and depression (unpublished
data).

Overall Outcome

The distribution of Summary Outcome Scores for cycles 3, 4, and 5

reflected these improvements. These subjects (N = 59) had filled out the
Summary Outcome Questionnaire (see Methods) at the end of meditation
training as part of the battery of post outcome measures. This measure has
a scale from 1 to 5, where 3 represents no change, 1 represents a large nega-
tive change, and 5 represents a large positive change. The mean score was
3.9: 76~ of the patients (45 of 59) scored 3.5 or above, and 61% (36 of 59)
scored 3.8 or above. The range from 3.8 to 5.0 empirically reflects a moder-
ate to great improvement in pain and in overall health status. One item asked
for changes in medications for pain control. Of the patients in these cycles
who were using drugs to control pain before taking the SR&RP (N = 39),
17 (44%) reported reduced drug dosages, and an additional 11 individuals
(28%) reported rarely or never using medication for pain relief by the end
of the SR&RP.

Analysis by Pain Category

A comparison of outcomes for the three major diagnostic classes of

pain among the patients was performed. These were (1) low back pain with
or without leg pain (N = 31), (2) headache including migraine and tension
headaches, (N = 24), and (3) neck and shoulder pain (N = 15). The results
are presented in Table IVA. As expected, the mean initial levels for headache
patients were consistently lower than those for patients with low back pain
or neck a n d / o r shoulder pain on all indices? Patients in all three diagnostic
categories achieved comparable degrees of improvement based on the Sum-
mary Outcome Scores available for cycles 3, 4, and 5 (N = 59). Mean Sum-
mary Outcome Scores were 4.0 for the low back-pain patients, 3.9 for the

7Due to negative scaling in the low range of the T M D scale, changes in an individual's T M D
cannot be expressed readily as a percentage.
8The one exception was the number of symptoms reported in the previous m o n t h (MSCL pre
mean, 23.3), which exceeded that for the low back-pain patients.
174 Kabat-Zinn, Lipworth, and Burney

Table III. Symptom, Mood, and Psychological Outcomes for the Total Population
(A) Group means ~
MSCL TMD GSI
Pre 22.3 47.8 0.77
Post 14.4 21.5 0.50
% change in mean 35* 55* 35*
N 87 73 74

(B) Individual gains ~

AMSCL AGSI
Level of reduction (%) _>33.3 _>50 _>33.3 _>50
N u m b e r reaching
this level 47/87 32/87 44/74 29/74
~ of
total patients 54 37 59 39
aMSCL, number of symptoms on a Medical Symptom Checklist; TMD, Total Mood
Disturbance score on the Profile of M o o d States (POMS); GSI, General Severity
Index (SCL-90-R).
bThe T M D is excluded because the percentage change could not be calculated due
to negative scaling.
*P < 0.003 in t test adjusted for multiple comparisons using the Bonferroni method
(Ingelfinger et al., 1983).

headache patients, and 3.9 for the patients with neck and shoulder pain. These
differences were not statistically significant. Patients with neck and shoul-
der pain had higher mean pre and post values than the low back-pain pa-
tients on the BPPA, MSCL, TMD, and GSI. Neck a n d / o r shoulder pain
was consistently reported as more severe and more debilitating than low back
pain.

Analysis by Gender

The female-to-male ratio for the population was 2:1 (Table IVB). Males
consistently had higher initial mean levels of mood disturbance (TMD) and
of psychological symptomatology (GSI) than females. They were also less
successful in lowering the mean scores on these indices than the females dur-
ing meditation training. SCL-90-R profiles for the males showed higher lev-
els of Somatization, Depression, Anxiety, Hostility, and Phobic Anxiety than
those for the females both before and after meditation training (unpublished
data).
The mean Summary Outcome Score for the females in cycles 3, 4, and
5 ( N = 41) was 4.0, and that for the males (N = 18) was 3.8 (Table IVB).
This difference was not statistically significant. Forty-four percent of the
males and sixty-eight percent of the females were in the 3.8 to 5.0 range,
reflecting a moderate to great overall improvement.
"Fable IV. Group Mean Values Pre- and Post-Meditation Training: Breakdown by (A) Diagnosis and (B) Sex"

Mean post
PR1 BPPA BPM TLI MSCL TMD GSI Summary
Outcome
Pre Post Pre Post Pre Post Pre Post Pre Post Pre Post Score b
(A)
L B P ( N = 31) 17.6 8.7 41.2 27.2 1.3 14.5 I1.1 20.8 1 3 . 1 51.2 18.8 0.78 0.47 4.0
( N = i6)
Headaches
(N = 24) 12.5 3.4 34.8 33.6 1.3 8.6 4.7 23.3 14.8 39.4 19.8 0.71 0.47 3.9
(N = 13)
Neck/shoulder
(N = 15) 16.9 8.0 57.5 37.6 1.6 12.1 9.7 25.5 20.9 58.7 33.4 1.0 0.74 3.9
(N 14)
(8)
Males
(N = 30) 16.1 7.3 45.8 33.1 1.3 10.9 8.1 19.4 12.5 51.0 33.1 0.84 0.60 3.8
(N = 18)
Females
(N = 60) 16.1 6.4 40.0 27.2 1.6 12.3 8.3 23.8 15.2 45.1 1 8 . 3 0.75 0.45 4.0
(N = 41)
"PRI, Pain Rating Index; BPPA, Body Parts Problem Assessment score; BPM, Body Pain Map. BPM group
outcomes are expressed as the numerical averages of individual change scores. Numerical values were assigned
as follows: - = - 1 ; 0 0; + = +1; + + = +2; + + + = +3. TLI, Table of Levels of lnterference with
daily activities; MSCL, number of symptoms on a Medical Symptom Checklist; TMD, Total Mood Disturbance
score on the Profile of Mood State (POMS); GSI, General Severity Index (SCL-90-R).
1'For patients in cycles 3, 4, and 5 only. N 59, of whom 43 were in the three dominant pain categories.
176 Kabat-Zinn, Lipworth, and Burney

Comparison of Pain Clinic Meditators with Other Pain Clinic Patients

The reductions in pain and pain-related affect and symptoms observed

consistently on the self-report indices the patients filled out before and after
the program were paralleled by clinical improvements in many cases. These
took the form of an increased capacity to sit, stand, or walk and an improve-
ment in appearance and affect. The combined weight of these observations
suggested a clinically important change in overall health status for the majority
of patients trained in the meditation. To address the question of how these
results might compare with the effects of a more traditional medical inter-
vention, we compared as a cohort the mediators who had been referred for
meditation training from the Pain Clinic with a cohort of 21 consecutive pa-
tients receiving the standard treatment course in the Pain Clinic (see Methods
and Table IB).
Table VA shows that the patients undergoing the traditional treatment
course showed little improvement in pain, symptoms, or affect. The medita-
tors had large mean decreases on all the test measures, which reached sig-
nificance in four of six cases. In no case did the change in the mean value
of any index reach statistical significance for the PC comparison group.
The mean percentage change for each index was also calculated to com-
pare outcomes for the two groups. The mean percentage changes in the PRI,
BPPA, MSCL, and GSI were negative and close to zero for the PC compar-
ison group. The mean percentage change in the TLI was positive and ap-
proximated one-third the value of the mean percentage change for the PC
referrals to the SR&RP. The mean percentage changes in the PRI and the
GSI were statistically significantly different between the two comparison
groups in the unpaired t test corrected for multiple testing (P < 0.005 for
the PRI; P < 0.05 for the GSI).
The initial mean PRI for the PC comparison group was 9 points higher
than the mean PRI for the PC referrals to the SR&RP. This was not a sig-
nificant difference. The initial (pre) mean scores on all indices for both groups
were similar in spite of the major differences in group composition (see
Methods).
Comparison of the subdimensions of the GSI showed that the Pain Clin-
ic meditators were reporting major reductions in Anxiety (65%), Depression
(59%), Hostility (57%), and Somatization (30%), while the Pain Clinic com-
parison group reported mean reductions of 29% for Anxiety, 18070 for Depres-
sion, 7070 for Hostility, and 0% for Somatization. The reduction in
interpersonal Sensitivity (lack of self-esteem) was 45% for the meditators
and 34% for the nonmeditators.
Few individuals in the comparison group achieved the 33.3% improve-
ment level (or the + + / + + + level for the BPM) on any of the pain-related
indices (PRI, BPPA, TLI, BPM) (see Table VB). Approximately 25% of
Meditation for Self-Regulation of Pain 1"/7

the patients achieved the greater than 33.3% reduction level in the number
of medical symptoms (MSL) and in overall psychological symptomatology
(GSI). These individuals constituted a much smaller fraction of their cohort
than did those among the PC referrals trained in the meditation (Table VB).
Individuals achieving greater than a 50% reduction level were far less evi-
dent in the traditional-care cohort than among the meditators.

F o l l o w - U p Studies on the S R & R P Patients

Three follow-up questionnaires were sent to these patients. Depending

on when they took the program, the most recent follow-up to date, mea-
sured from the time the meditation training ended, was 15 months (cycle 1),
12 months (cycle 2), 4.5 months (cycle 3), or 2.5 months (cycle 4). Of the
72 questionnaires sent in the most recent survey to all subjects who complet-
ed the SR&RP in cycles 1 through 4, 56 were returned (78%). In three other
cases partial information (the Summary Outcome Score) was obtained via
telephone interviews. For the patients for whom 15 months of follow-up had
elapsed, 80% (8 of 10) responded; at the 12-month follow-up, 69% (11 of
16) responded; at the 4.5-month follow-up, 67% (16 of 24) responded; and
at the 2.5-month follow-up, 95% (21 of 22) responded.
A comparison of responders to nonresponders was undertaken to exa-
mine potential bias in the follow-up results, since one might assume that the
individuals with more successful outcomes would be more likely to respond
to the questionnaires. In the case of cycle 3, the first cycle in which post Sum-
mary Outcome Scores were obtained and also the cycle with the lowest per-
centage of returns (67%), the nonresponders (8 individuals) had a mean post
score of 3.8; the 159 responders had a mean post score of 4.2. The difference
is not statistically significant and both means were in the range empirically
defined as successful (3.8-5.0). This suggests that nonresponders probably
did not differ remarkably from responders in terms of successful outcome
in the SR&RP itself. BPM scores were also compared for the responders and
nonresponders in all cycles and no significant differences found.
Figure 1 plots the group mean values as a function of time for the (A)
Pain Rating Index (PRI), (B) Body Parts Problem Assessment (BPPA) score,
(C) number of symptoms (MSCL), (D) Total Mood Disturbance (TMD) score,
and (E) General Severity Index for each cycle of the SR&RP for which follow-
up data were obtained. The follow-up values are the means for those individu-
als in each cycle responding to that follow-up questionnaire. In all cases,'~
there was a pronounced and statistically highly significant reduction in the

9The post Summary Outcome Score was missing for one responder.
4The one exception was the BPPA for cycle 1.
178 Kabat-Zinn, Lipworth, and Burney

~
25 PRI TMD
(~)Ns
20 40

15 30
xIlOINS O|I6)NS
X (10}NS
I0
(211"~-X"
5 IO

i i I ~ i i [ I I 1
PRE POST 25 I0 12,5 ~ONTHS PRE POST 2.5 5 I0 12.5

50 .8 GSI

40 NS NS .7
(H) {16)~
50 .6
X (1~

20 (O)NS .5

I0 .4
z-Y- (20#t~X" ~ X (7 ) ~
I I I I I I .~
.3
PRE POST 2.5 5 I0 12.5

.2

25 MSCL .I

20
PRE POST 2 5 !5 I 12.5
O~X{II) NS
15

I0 (10)~ (lO)~)f {8}

p<.05
p<.02
~ p<.OI
~ p < .001
I I I I I i
PRE POST 2.5 5 I0 12.5

Fig. 1. Time dependency of outcome measures with follow-up. (A) Pain Rating Index; (B)
Body Parts Problem Assessment score; (C) number of symptoms on a Medical Symptom Check-
list; (D) Total M o o d Disturbance score on the POMS; (E) General Severity Index on the
SCL-90-R. Pre represents the initial m e a n levels for the patients in each cycle; post represents
the mean levels after the 10-week meditation training. Follow-up times are expressed as months
following the completion of the SR&RP. Open squares represent patients in cycle 1 ( N =
10); crosses, patients in cycle 2 (N = 16); filled circles, patients in cycle 3 (N = 24); and open
triangles, patients in cycle 4 (N = 21). The numbers of individuals responding to follow-up
questionnaires are given in parentheses next to the corresponding data points. The follow-up
points represent the mean scores for the respondents. P values for each follow-up point represent
paired t-tests with pre values. Paired t tests grouping 2.5- to 7-month returns together gave
P values of < 0.0001 on all indices except the PRI when comparing pre and follow-up levels
for each responder. Similarly, t tests for the 12- to 15-month returns gave P values of < 0.01
for the B P P A and MSCL. All P values are adjusted values after application of the Bonferro-
ni correction for multiple tests. The PRI values have been corrected for zero values (Melzack,
1975).
Meditation for Self-Regulation of Pain 179

level of each index between the initial assessment (pre-meditation training)

and the immediate post meditation assessment, reflecting the pre/post results
presented in Tables IIA and IIIA. Mean scores for each measure tended to
remain at the postintervention level for periods of up to 15 months follow-
ing completing of the SR&RP, with the notable exception of the PRI. The
responses between 2.5 and 7 months postintervention were grouped together
for each measure and tested for significance in the paired t test, matching
pre values with follow-up values for the responders and adjusting for multi-
ple tests using the Bonferroni method. With the exception of the PRI, all
were highly significant (P < 0.0001). Combining the 12- and 15-month
follow-up results, the means of the MSCL and the B P P A also differed sig-
nificantly from the pre but not from the post means using the same method
(P < 0.01). In the case of the T M D and GSI, the mean values at the 1-year
follow-up could not be shown to reach statistical significance compared to
the mean pre levels. This appeared to be due to the smaller sample size be-
cause no data on these indices were collected in cycle 1. Nevertheless, the
TMD and GSI means for the responders were always lower than the pre lev-
el at 1 year.
The Pain Rating index (PRI) clearly differs from the other indices in
its follow-up profile. In two of three cycles, on follow-up the PRI rose to
levels exceeding the pre level, and in no case did the mean differ significant-
ly from the pre mean value. The difference in behavior between the PRI and
the other indices on follow-up is discussed below.
The results in Fig. 1 were not significantly affected when individuals
who had taken a graduate SR&RP course were excluded from the follow-up
analysis (data not shown). The finding that the mean improvements observed
post-SR&RP are maintained over time (Figs. 1B through E) is thus not ex-
plainable solely by the additional training experienced by a minority of the
responders.

Compliance with the Meditation

The third follow-up questionnaire probed the frequency and duration

of formal meditation practice with a set of three precise questions which made
it difficult to exaggerate compliance without frank dissimulation. This for-
mat reinforced the face validity of this section of the questionnaire. On the
basis of their responses, individuals were grouped in five classes: (A) regular
meditation practice (__ 3 x per week and _> 15 min at a time); (B) sporatic
practice ( < 3 x per week but > 1 x per week and _> 15 min at a time or
_> 3 x per week and < 15 min at a time); (C) infrequent practice (_< 1 x
per week and _< 15 min at a time); (D) no longer meditating; and (E) no
i n f o r m a t i o n - d i d not answer the questions accurately or at all.
Table V. Comparison of Outcomes Between Pain Clinic Meditators (PC Referrals) and Nonmeditators (PC Comparison Group)
(A) Group means a
PRI b BPPA TLI MSCL TMD ~ GSI
PC referrals
(meditators)
Pre 23.7 47.8 15.4 20.0 41.4 0.62
Post 15.2 30.1 11.2 13.6 8.4 0.33
% change in mean 36 (ns) a 37* 27 (ns) 32** 87*** 47***
N 15 21 14 21 15 15
42~t 28 22 23 -- 41 "~

PC comparison group
(nonmeditators)
Pre 32.5 44.4 17.7 22.9 51.1 0.74
Post 32.4 43.4 15.0 20.6 39.3 0.66
o70 change in mean 0 (ns) 2 (ns) 15 (ns) 10 (ns) 22 (ns) 11 (ns)
N 20 21 21 21 20 20 ,'7
Mean % change - 2 - 4 7 - 5 -- - 6
(B) Individual gains ~

2~ PRI b zXBPPA ~BPM f AMSCL AGSI

>_33.3% > - - 5 0 % >__33.3% _>50% ++/+++ >-33.3% >-50% >-33.3% -->50%

PC referrals
(meditators)
Number 10/14 8/14 10/21 10/21 7/19 10/21 6/21 11/15 6/15
o7o 71~;~ 57~ 48 (ns) 48~ 37 ~ 48 (ns) 29 (ns) 73~ 40~
P C comparison group
(nonmeditators)
Number 2/19 0/19 5/21 2/21 1/19 5/21 4/2t / 5/20 1/20
070 11 0 24 10 5 24 19 25 5
~PRI, Pain Rating Index; B P P A , Body Parts Problem Assessment score; TLI, Table of Levels of Interference with daily activi-
ties; M S C L , n u m b e r of s y m p t o m s on a Medical S y m p t o m Checklist; T M D , Total M o o d Disturbance Score (POMS); GSI, e~
General Severity Index (SCL-90-R).
~Corrected for zero values (see Melzack, 1975).
CMean percentage change not calculated.
aNot significant (P > 0.05).
~Fraction of individuals achieving the indicated level of reduction or improvement on each index.
J Body Pain Map. The fraction represents the n u m b e r of individuals in the population who were scored as either + + (moderate
improvement) or + + + (great improvement).
*P < 0.012; significance of p r e / p o s t t tests of group m e a n s corrected for multiple comparisons.
**P < 0.006.
***P < 0.003.
~P < 0.05; significance of unpaired t test comparing the mean percentage change of PC referrals and the PC comparison group,
corrected for multiple-comparisons.
}-}P < 0.005.
~P < 0.05 comparing PC referrals to the PC comparison group in the chi-square test (ldf) or Fisher's Exact Test applied following
t h e guidelines of Cochrane cited by Armitage (1971).
~ p < o.o25.
~P < O.Ol.
~p < o.ool.
182 Kabat-Zinn, Lipworth, and Burney

More than 70% of the respondents in each cycle described themselves as

still meditating (up to 15 months) following the end of the meditation training.
Patients in classes A and B together constituted 54% (30/56) of all respondents
(23 in class A, 7 in class B). Of the regular meditators (class A), 14 of 23 reported
that they were "meditating everyday (almost)." Twenty-nine percent of the
responders (16/56) were in the class of infrequent practice. Those who report-
ed not meditating at all constituted 13% (7/56) of the respondents. The remain-
ing 5% (3/56) either did not answer this section or answered ambiguously.
Seventy percent of the individuals with Summary Outcome Scores be-
low 3.5 also claimed that they were still meditating. Thus, in many cases in
which marginal or no improvement was detectable on the outcome indices,
some factor appears to have motivated these individuals to continue to prac-
tice the meditation. The percentage of individuals in classes A and B among
this cohort was lower (40%) than for those who had Summary Outcome
Scores above 3.8 (53%).
The follow-up questionnaire also measured informal use of the medi-
tation in addition to assessing the level of compliance with the formal prac-
tice, Patients were asked to rate the frequency with which they used awareness
of breathing in daily life and to rate its utility as a coping strategy in stressful
situations. This mindfulness strategy was used by more patients and more fre-
quently than any of the formal techniques. Patients with Summary Outcome
Scores above 3.5 rated awareness of breathing much higher in usefulness and
reported using it more regularly in daily life activities than did individuals with
scores below 3.5 (data not shown).

DISCUSSION

The data suggest that mindfulness meditation training in the context

of a Stress Reduction and Relaxation Program can be highly effective in
reducing self-reports of both pain and pain-related behaviors in the majori-
ty of the patients referred to it for chronic pain. Significant group improve-
ments were recorded over the 10 weeks of the program for present-moment
pain (PRI), negative body image (BPPA), degree of inhibition of everyday
activities by pain (TLI), medical symptoms (MSCL), mood and affect (TMD),
and psychological symptomalolgy (GSI) including somatization, anxiety,
depression, and self-esteem.
The specific outcome measures employed were chosen to assess the phys-
ical and psychological dimensions of chronic pain and appeared to reflect
accurately our clinical impressions of the patients. Each index was reduced
significantly when the results were averaged over the entire population and
high proportions of individuals made major improvements on all indices.
Meditation for Self-Regulation of Pain 183

This combined assessment of group and individual outcomes suggests a pro-

found change in the majority of the patients due to the SR&RP interven-
tion. How much of this improvement is placebo related, i.e., due to the fact
of intervention rather than to the nature of the intervention, can be inves-
tigated only using placebo control groups in a prospective, randomized
study. However, the mean chronicity of this population (8 years) and the
fact that these patients had received extensive medical treatment without at-
taining this level of improvement speaks against a simple placebo effect, as
does the lack of significant improvement in the comparison group of Pain
Clinic patients, who were exposed to strong positive placebos in the form
of enthusiastic referral to a specialty service, high expectation of relief, and
pain-specific medical treatment protocols.
Thus, within the limits inherent in a descriptive outcome study, the
specific questions posed in the Introduction concerning the empirical effec-
tiveness of mindfulness meditation in the context of a stress reduction pro-
gram have been answered in the affirmative both qualitatively and
quantitatively.
A number of observations merit explicit discussion.
(1) The relative degree of improvement was independent of the refer-
ral source, pain severity, diagnosis (Table IVA), and gender (Ta-
ble IVB). Thus the mechanism(s) which underlies the observed
pain reductions is elicitable by individuals with different types of
pain and over a wide range of intensities. This suggests a gener-
alized applicability of the method for pain reduction and coping.
(2) The mean improvements observed on all indices during the inter-
vention were maintained in the period from 2.5 to 7 months
postintervention, with the exception of the PRI, with a high degree
of statistical significance for the responders. Mean improvements
in negative body image (BPPA) and symptoms (MSCL) were
maintained at a significant level for the period of 12 to 15 months
postintervention as well. Improvements in mood (TMD) and psy-
chological symptomatology (GSI) tended to be maintained but
did not reach statistical significance. The responders represented
the majority of the study subjects. Since these indices in combi-
nation measured important aspects of a person's well-being, their
coordinated improvement suggests a fundamental and long-lasting
improvement in health status.
(3) The mean PRI tended to return to the original (pre) level by the
time of follow-up. This finding can be interpreted in two ways:
either (a) the PRI cannot be used effectively by mail in conjunc-
tion with previous PRI scores obtained in interviews (the pre and
post data were obtained in individual interviews) and the rise is
184 Kabat-Zinn, Lipworth, and Burney

an artifact of mixing the modes of administration of the instru-

ment or (2) the rise on follow-up in the PRI above the postinter-
vention level is real and reflects a worsening of the pain status.
If we assume the latter, it is important to note that the increased
PRI levels on follow-up were apparently no longer interpreted in
the original (pre) way, because the mean B P P A score remained
low on follow-up for the same patients. We believe that the sta-
ble reduction in the B P P A score reflects a newly developed abili-
ty to live with and cope with pain and reinterpret its meaning.
Many of the patients trained in the SR&RP reported on follow-
up that their pain is "still there" but that their relationship to it
has changed, i.e., there is less fear of pain, less self-pity, and less
willingness to let pain or fear of pain restrict activity. However,
an analysis of the components of the total PRI (sensory, affec-
rive, evaluative) on follow-up did not yield discrimination of this
kind as might have been expected. It remains for a carefully con-
trolled study to elucidate this observation.
(4) The high compliance with the meditation practice achieved dur-
ing the SR&RP appears to have been maintained over time.
Seventy percent of the respondents reported that they still medi-
tate (in response to the question, "Do you meditate anymore?").
A likely explaination for such a high proportion of individuals
who claim to still be meditating on follow-up is that the medita-
tion practice has an informal dimension in addition to the more
time-consuming formal discipline. Thus subjects can honestly
report that they still meditate in the sense that, from time to time,
they consciously bring attention to the moment-to-moment events
and experiences of their daily lives. On the basis of precise ques-
tions about the frequency, duration, and type of practice, 41%
could be classified as regular meditators (class A; see Results.).
(5) Among the most successful individuals (those with Summary Out-
come Scores above 3.8), there appeared to be two equal classes
of pain outcome: (a) those for whom the pain was greatly reduced
or eliminated and (b) those who reported that the pain was un-
changed but that they were coping with it differently and there-
fore it was not as problematic as before the meditation training.
Patients with diagnoses of headache predominated in the former
class, but some individuals with chronic gastrointestinal, chest,
and facial pain also reported sustained disappearance of pain. Low
back-pain patients predominated in the latter class, but headache
and other classes of pain were also present.
(6) There were considerable differences in the composition of the two
cohorts of Pain Clinic patients compared here, and the potential
Meditation for Self-Regulation of Pain 185

influence of these differences on the results observed is unknown.

For this reason, the comparison suffers obvious shortcomings and
needs to be repeated in a randomized prospective study.
(7) It is important to emphasize that the intervention makes use of
meditation practice within a context of stress reduction and health
promotion. Because of the short period of the training and its
clinical orientation, it cannot be compared facilely with the years
of intensive meditation training common in the more traditional
contexts in which such consciousness disciplines are pursued.
Nevertheless, the fact that even at an introductory level, such train-
ing appears to be enthusiastically received and useful and practi-
cal for patients with long-term pain problems attests to the
potential power and depth of such approaches (see Burns, 1973).
For mindfulness meditation to be considered a practical tool in clinical
behavioral medicine, it must be investigated and conceptualized within the
theoretical and experimental perspective developing from the study of the
psychological interventions currently in widespread use for chronic pain relief.
At present, the most widely used and accepted of these are progressive relax-
ation, biofeedback, operant conditioning, hypnosis, and cognitive-behavioral
therapies (Turner and Chapman, 1982a,b). One element which the above
methods have in common and which may be of central and underestimated
therapeutic importance is attention regulation. Each of these methods neces-
sitates a conscious primary utilization of attention: either to muscle tension
and relaxation (progressive relaxation and biofeedback), to a feedback stimu-
lus (biofeedback), to exclusive expression of nonpain behaviors (operant con-
ditioning), to suggestions of altered proprioception (hypnosis), or to events,
emotions, and thought patterns in relationship to symptom onset (cognitive-
behavioral procedures). Attention is directed to quite different objects in these
different therapies. Yet it may be the regulation and the intensity of one's
attention, and one's belief in a method based on past experience, rather than
the particular object or process attended to, which are of greatest therapeu-
tic value. Indeed, it is well known that any strategy for the self-regulation of
attention, including purposeful distraction, can be used with some effective-
ness in coping with pain both in the laboratory and in the clinic.
In this regard, recent laboratory studies using the cold pressor stimu-
lus have shown that a strategy of attention to proprioception during the trial
results in significantly less distress and higher tolerance than strategies utilizing
distraction or expression of emotions (Ahles et al., 1983) and that attention
to sensation becomes a relatively better coping strategy the longer the trial
(McCaul and Haugtvedt, 1982). In these laboratory studies, the word "pain"
was studiously avoided because it had been observed (Levinthal et al., 1979)
that mention of the word pain negated any positive effects of attending to
sensations. This lability of the attentional strategy to the mention of the word
186 Kabat-Zinn, Lipworth, and Burney

pain suggests that training in detached observation of sensations through

mindfulness meditation techniques could greatly enhance the positive effects
of the attentional strategy used in these experiments since subjects were un-
trained and merely instructed to describe aloud the sensations that they were
experiencing (Ahles et al., 1983). This possibility could be tested by compar-
ing experienced meditators with naive subjects using the cold pressor test and
carefully documenting the precise strategies subjects actually used during the
trials.
In terms of chronic pain, the results of our study suggest that the sys-
tematic cultivation of a flexible attentional capacity for detached observa-
tion of proprioception can enhance whatever the patient's previously (and
often inadequate) coping strategies have been and reduce the level of dis-
tress. Holroyd and Andrasik (1980) have observed, in carefully controlled
studies with tension-headache patients, that pain relief following EMG-
biofeedback training was the result not of self-control of muscle tension but
of learning to recognize the onset of headache symptoms. There is now in-
creasing evidence that the use of biofeedback for pain control offers no ther-
apeutic benefit in most situations beyond that attributable to the relaxation
that is taught in conjunction with it (Zitman, 1983; Turner and Chapman,
1982a) and/or the cognitive and behavioral changes that often arise spon-
taneously within the therapeutic context (Turk et al., 1979). Indeed, each
patient's own private strategies need to be inquired into and cultivated when
appropriate, rather than imposing the therapist's choice of method. When
Holroyd and Andrasik's patients were taught to recognize the onset of symp-
toms, they spontaneously changed the ways in which they were coping even
when no coping skills were taught. These authors concluded that "it "may
be less crucial to provide clients with specific coping responses than to in-
sure that they monitor the insidious onset of symptoms and are capable of
engaging in some sort of cognitive or behavioral response.., this response
need not be relaxation and in certain circumstances where.., inappropriate,
should not be relaxation."
These observations imply that moment-to-moment mindfulness may,
in this and other interventions, itself be the principal, if implicit, coping
mechanism. While there may be a Variety of cognitive and behavioral strate-
gies to enhance this capacity, the clinical results of our study suggest that
the systematic formal practice of mindfulness meditation, which in this con-
text emphasizes attending from moment to moment to proprioception and
to stress reactions, may provide a therapeutic dimension which includes both
physiological relaxation and cognitive-behavioral changes and which goes
deeper than the methods currently in use. We suggest that the evidence is
strong enough to merit a comparative study of mindfulness meditation and
other psychological interventions under rigorously controlled conditions in
both the laboratory and the clinic.
Meditation for Self-Regulation of Pain 187

COMMENT

To be effective, it is likely that any clinical approach in behavioral medi-

cine seeking to relieve suffering and improve the quality of life for patients
with chronic medical problems must require an active participation by the
patient to develop and utilize his or her full range of internal resources, in-
cluding deep relaxation, physical fitness, self-confidence, and even wisdom.
Mindfulness meditation has a number of unique features which recom-
mend it as a clinical method for teaching self-regulation and as a psychological
intervention in chronic pain. (1) It is much less expensive to introduce than
elaborate inpatient behavior modification programs, and training is readily
accomplished in groups of up to 30 individuals. The high compliance we ob-
served during and after training suggests that meditation can be both enjoy-
able and beneficial to large numbers of patients. (2) Its emphasis on
self-observation and on self-responsibility can enhance realizations of self-
worth and help people to perceive conditioned patterns of (illness) behavior
more clearly. (3) Since this form of meditation is a systematic development
of the basic human capacity to attend intentionally to events, percepts, and
cognitions in the field of consciousness, it has a generalized applicability
within a wide range of perceptual, cognitive, and behavioral contexts, which
includes but is not limited to pain relief. (4) As with other meditative prac-
tices, mindfulness meditation can facilitate deep physiological relaxation (Ben-
son, 1975). In contrast to relaxation techniques, however, it has the further
property of enhancing what the Buddhists call "insight" (Nyanaponika, 1962).
Walsh (1980) and Wilber (1980) suggest, as do the classical Buddhist medi-
tation texts (see Nyanaponika, 1962), that mindfulness meditation accesses
the deep structure or "core" of one's being and can potentiate the experience
of what has become known in contemporary psychological circles (see Walsh
and Vaughan, 1980) as "transpersonal" levels of consciousness. Potential
benefit may thus be derived from training in this form of meditation on a mul-
tiplicity of levels, ranging from relaxation and anxiety reduction to profound
personal transformation (Wilber, 1979). Mindfulness meditation has recently
become a subject of exploration by psychiatrists (Burns, 1973; Burns and
Ohayv, 1980; Walsh, 1977, 1978, 1983; Deikman, 1982; Kutz et al., 1985a)
and clinical psychologists (Brown and Engler, 1980; Deatherage, 1975;
Shapiro, 1980). Attempts to integrate meditation practice into psychothera-
py in appropriate instances are currently in progress (Deatherage, 1975; Kutz,
et al., 1985a,b; Shapiro and Giber, 1978). (5) An increasing number of
Westerners are being trained in mindfulness meditation without cultural, reli-
gious, or ideological overtones at meditation centers in the West. It is cer-
tainly no longer merely an esoteric "Eastern" phenomenon. As yet, it is unclear
whether the other r~ajor class of meditative practices, known generically as
concentration meditation, would achieve similar clinical results. Mindfulness
188 Kabat-Zinn, Lipworth, and Burney

is a generic term, encompassing a range of techniques and traditions, all utiliz-

ing attention in a well-defined way which differs substantially from the con-
centration practices but which, nevertheless, requires a foundation in
concentration. It remains for further studies to clarify this point. (6) The
physiology (see Davidson, 1976), psychophysiology (Woolfolk, 1975), and
phenomenology (Maliszewski et al., 1981) of intensive meditation practice
are becoming fields of serious scientific research. While presently in its in-
fancy, in the future this research may provide an important foundation for
understanding the underlying psychobiological mechanisms of meditation
and of the self-regulation of pain and point to new ways to maximize the
subjective and latent dimensions of human consciousness for achieving whole-
ness and well-being, even in the midst of suffering.

NOTE ADDED IN P R O O F

A recent study has demonstrated the reproducibility of the results report-

ed here and has extended the follow-up time to four years post SR&RP
(Kabat-Zinn et al., 1984).

ACKNOWLEDGMENTS

The first author would like to thank Mr. Byran Tucker, Ms. N o r m a
Rosiello, Ms. Debbie Hanna, R.N., Ms. Suzie Pilapel, and Mr. William Sellers,
B.S., for their assistance in assembling the data files; Drs. James E. Dalen,
Judith Ockene, Robert Goldberg, and Basil Barr for their critical readings
of the manuscript; and A m y Singer for the graphics in Fig. 1.
This work was carried out in conformity with N I H guidelines for
research on human subjects.

REFERENCES

Ahles, T. A., Blanchard, E. B., and Levinthal, H. (1983). Cognitivecontrol of pain: Attention
to the sensory aspects of the cold pressor stimulus. Cognit. Ther. Res. 7: 159-178.
Armitage, P. (1971). Statistical Methods in Medical Research, Wiley, New York, 138.
Benson, H. (1975). The Relaxation Response, Morrow, New York.
Brown, D. P., and Engler, J. (1980). A Rorschach study of the stages of mindfulness medita-
tion. J. Transper. Psychol. 12: 143-192.
Burns, D. M. (1973). Buddhist Meditation andDepth Psychology, Wheel, Kandy, Sri Lanka.
Burns, D., and Ohayv, R. J. (1980). Psychologicalchanges in meditating Western monks in
Thailand. J. Transper. Psychol. 12: 11-24.
Butler, K. (1983). Events are the teacher. Co-Evol. Q. Winter: 112-123.
Cassel, E. J. (1982). The nature of suffering and the goals of medicine. N. Engl. J. Med. 306:
639-645.
Meditation for Self-Regulation of Pain 189

Davidson, J. M. (1976). The physiology of meditation and mystical states of consciousness.

Perspect. Biol. Meal. 345-379.
Deatherage, G. (1975). The clinical use of mindfulness meditation techniques in short-term psy-
chotherapy. J. Transper. Psychol. 2: 133-144.
Deikman, A. J. (1982). The Observing Self, Beacon, Boston.
Derogatis, L. R. (1977). SCL-90-R Manual I Johns Hopkins University School of Medicine,
Baltimore, Md.
Derogatis, L. R., Rickels, K., and Rock, A. F. (1976). The SCL-90 and the MMPI: A step
in the validation of a new self-report scale. Br. J. Psychiat. I28: 280-289.
Haskell, D., Pugatch, D., and McNair, D. M. (1969). Time-limited psychotherapy for whom?
Arch. Gen. Psychiat. 21: 546-552.
Holroyd, K. A., and Andrasik, F. (1980). Self-control of tension headache. In McGuigan, F.
J., Sime, W. E., and Wallace, J. M. (eds.), Stress and Tension Control, Plenum, New
York.
Ingelfinger, J. A., Mosteller, F., Thibodeau, L. A., and Ware, J. H. (1983). Biostatistics in
Clinical Medicine, Macmillan, New York, p. 170.
Jung, C. G. (ed.) (1969). Forward to Introduction to Zen Buddhism. In Psychology and Religion,
Princeton University Press, Princeton, N.J., p. 554.
Kabat-Zinn, J. (1982). An outpatient program in behavioral medicine for chronic pain patients
based on the practice of mindfulness meditation: Theoretical considerations and prelimi-
nary results. Gen. Hosp. Psychiat. 4: 33-42.
Kabat-Zinn, J. (1983). The Body Problem Assessment Scale. In Melzack, R. (ed.), Pain Meas-
urement and Assessment. Raven, New York, pp. 227-231.
Kabat-Zinn, J., and Burney, R. (1981). The clinical use of awareness meditation in the self-
regulation of chronic pain. Pain (Suppl.) 1: $273.
Kabat-Zinn, J., Lipworth, L., Sellers, W., Brew, M., and Burney, R. (1984). Reproducibility
and four year follow-up of a training program in mindfulness meditation for the self-
regulation of chronic pain. Pain (Suppl.) 2: $303.
Krishnamurti, J. (1979). The Wholeness o f Life, Harper and Row, New York.
Kutz, I., Borysenko, J. Z., and Benson, H. (1985a). Meditation and psychotherapy: A ration-
ale for the integration of dynamic psychotherapy, the relaxation response, and mindful-
ness meditation. Am. J. Psychiat. 142: 1-8.
Kutz, I., Leserman, J., Morrisson, C. H., Borysenko, J. Z., Dorrington, C., and Benson, H.
(1985b). Meditation as an adjunct to psychotherapy: An outcome study (submitted for
publication).
Leventhal, H., Brown, D., Shachman, S., and Engquist, G. (1979). Effects of preparatory in-
formation about sensations, threat of pain, and attention on cold pressor distress. J.
Personal. Soc. Psychol. 37: 688-714.
Maliszewski, M., Twemlow, S. W., Brown, D. P., and Engler, J. (1981). A phenomenological
typology of intensive meditation. ReVision 4: 3-27.
McCaul, K. D., and Haugtvedt, C. (1982). Attention, distraction and cold-pressor pain. J. Per-
sonal. Soc. Psychol. 43: 154-162.
McCue, J. D. (1982). The effects of stress on physicians and their medical practice. New Engl.
J. Meal. 306: 458-463.
McNair, D. M., Lorr, M., and Droppleman, L. F. (1971). Profile o f Mood States (POMS),
Educational and Industrial Testing Service, San Diego, Calif.
Melzack, R. (1975). The McGill Pain Questionnaire: Major properties and scoring methods.
Pain 1: 277-299.
Melzack, R., and Perry, C. (1975). Self-regulation of pain: The use of alpha feedback and hyp-
notic training for the control of chronic pain. Exp. Neurol. 46: 452-469.
Melzack, R., and Wall, P. D. (t970). Psychophvsiology of pain. Int. Anesthesiol. Clin. 8: 3-34.
Melzack, R. and Wall, 1-'. D. (1983). Tt~e Challenge o f Pain, Basic, New York.
Naranjo, C., and Ornstein, R. E. (1971). On the Psychology o f Meditation, Penguin, New York,
p. 9.
Nisargadatta Maharaj (1973). I A m That, Vols. I and II, Chetana, Bombay.
Nyanaponika, T. (1962). The Heart o f Buddhist Meditation, Samuel Weiser, New York, pp.
30-45.
190 Kabat-Zinn, Lipworth, and Burney

Shapiro, D. H. (1980). Meditation: Self-Regulation Strategy and Altered State o f Conscious-

ness, Aldine, New York.
Shapiro, D. H., and Giber, D. (1978). Meditation and psychotherapeutic effects. Arch. Gen.
Psychiat. 35: 294-302.
Sternback, R. A. (ed.) (1978). The Psychology o f Pain, Raven, New York.
Suzuki, S. (1970). Zen Mind, Beginner's Mind, Weatherall, New York.
Thakar, V. (1977). Life as Yoga, Motilal Banarsidass, Delhi.
Turk, D. C., Meichenbaum, D. H., and Berman, W. H. (1979). Application of biofeedback
for the regulation of pain: A critical review. Psychol. Bull. 86: 1322-1338.
Turk, D. C., Meichenbaum, D., and Genest, M. (1983). Pain and Behavioral Medicine, Guil-
ford, New York.
Turner, J. A., and Chapman, C. R. (1982a). Psychological interventions for chronic pain: A
critical review. I. Relaxation training and biofeedback. Pain 12: 1-21.
Turner, J. A., and Chapman, C. R. (1982b). Psychological interventions of chronic pain: A
critical review. II. Operant conditioning, hypnosis, and cognitive therapy. Pain 12: 22-46.
Walsh, R. N. (1977). Initial meditative experiences I. J. Transper. Psychol. 9: 151-192.
Walsh, R. N. (1978). Initial meditative experiences II. J. Transper. Psychol. 10: 1-28.
Walsh, R. N. (1980). The consciousness disciplines and the behavioral sciences: Questions of
comparison and assessment. Am. J. Psychiat. 137: 663-673.
Walsh, R. N. (1983). Meditation practice and research. J. Hum. Psychol. 23: 18-50.
Walsh, R. N., and Vaughan, F. (1980). (eds.), BeyondEgo: TranspersonalDimensions in Psy-
chology, J. P. Tarcher, Los Angeles.
Wilber, K. (1979). No Boundary, Shambhala, Boulder, Colo.
Wilber, K. (1980). A developmental model of consciousness. In Walsh, R. N., and Vaughan,
F. (eds.), Beyond Ego: Transpersonal Dimensions in Psychology, J. P. Tarcher, Los
Angeles, pp. 99-114.
Woolfolk, R. L. (1975). Psychophysiologicai correlates of meditation. Arch. Gen. Psychiat.
32: 1326-1333.
Zitman, F. G. (1983). Biofeedback and chronic pain. in Bonica, J. J., et al. (eds.), Advances
in Pain Research and Therapy, Vol. 5, Raven, New York, pp. 795-808.