Pain, 59 (1994) 175-187 17.

5
0 1994 Elsevier Science B.V. All rights reserved 0304-3959/94/$07.00

PAIN 2586

Brief group cognitive-behavioral intervention

for te~poro~an~ibular disorders

Judith A. Turner b7d,Leanne Wilson a, Donna Massoth a, Coralyn Whitney aYc,

Samuel F. Dworkin a9d9*,
Kimberly H. Huggins a, Jeffrey Burgess a, Earl Sommers a and Edmond Truelove a
Departments of a Oral deficit (X-631, b Rehabi~tation ~ed~c~e and ’ Dental Public Health Sciences, School of medicine,
University of W~h~ngton, Seattie, WA 98195 (USA], and
d Department of Psychiatry and Behavioral Sciences, University of Washington, Seattle, WA 98195 (USA)

(Received 13 August 1993, revision received received 20 February 1994, accepted 3 March 1994)

Summary Temporomandibular disorders (TMD) are currently viewed as an interrelated set of clinical
conditions presenting with signs and symptoms in masticatory and related muscles of the head and neck, and the
soft tissue and bony components of the temporomandibular joint. Epidemiologic and clinical studies of TMD
confirm its status as a chronic pain problem. In this report we present results from a randomized clinical trial which
compared, at 3- and 1Zmonth follow-ups, the effects of usual TMD treatment on TMD pain and related physical
and psychological variables with the effects of a co~itive-behavioral (CB) intervention delivered to small groups of
patients before usual TMD treatment began. The purpose of this study was to determine whether a minimal CB
intervention followed by dental TMD treatment enhanced the effects of usual clinical dental treatment. A second
purpose of the study was to determine whether patients classified as high in somatization and psychosocial
dysfunction would respond less favorably to this minimal intervention than would those low in somatization and
dysfunction. Patients who participated in the CB inte~ention followed by usual treatment showed greater long-term
decreases in reported pain level and pain interference in daily activities than did patients who received only usual
treatment. The benefits of CB intervention were not seen when the CB and UT groups were compared at 3-month
follow-up. During the 3-12-month follow-up interval, however, the UT group maintained essentially the same level
of improvement in characteristic pain while the CB group continued to improve, as hypothesized. During this same
follow-up interval, the CB group also showed a strong trend toward continued improvement in pain interference.
Such effects were not observed for depression, somatization, or clinical measures of jaw range of motion.
Additionally, as hypothesized, dysfunctional chronic pain patients did not appear to benefit from the brief CB
intervention. Intent to treat analyses were also performed to assess generalizability of the results.

Key words: Temporomandibular disorder; Chronic pain; Cognitive-behavioral; Group; Dysfunction; Somatization;
Intent to treat

Introduction tissue and bony components of the temporomandibular

joint. Although a common condition, with a &month
Temporomandibular disorders (TMD) are currently prevalence of appro~mately 12% in the population we
viewed as an interrelated set of clinical conditions studied Won Korff et al. 1988), the etiology of TMD is
presenting with signs and symptoms in masticator-y and poorly understood and its diagnosis is acknowledged to
related muscles of the head and neck, and the soft be complex (Bell 1986; Fricton et al. 1987).
Epidemiologic and clinical studies of TMD confirm
its fundamental status as a pain problem, more specifi-
* Co~es~~~~g author: Samuel F. Dworkin, Department of Oral cally a chronic pain problem (Bell 1986; Fricton et al.
Medicine (K-631, School of Medicine, University of Washington, 1987; Dworkin et al. 1992b). About 95% of TMD
Seattle, WA 98195, USA. Tel.: 206-543-5912; FAX: 206-685-8412. patients seek treatment to relieve pain in the region of

SSDI 0304-3959(94)00062-J
the ear, the temporomandibular joint (TMJ) and/or patient follow-up, such as brief telephone counseling.
the muscles of mastication (Dworkin et al. 199Oa). These minimal interventions involve a smaller number
Comparison of TMD with other common chronic pain of sessions with trained mental health professionals
conditions, such as headache and back pain, for (e.g., clinical psychologists) than the 8-16 treatment
chronicity, intensity, psychosocial profile, and use of programs that are typical for CB interventions for
health care resources confirms that in all these major chronic pain. They are also often characterized as
respects TMD is essentially a chronic pain condition. occurring independent of any biomedically based treat-
Empirical support for this view comes from large-scale ments. Most relevant to TMD, a series of studies has
longitudinal population-based studies we have con- shown minimal CB interventions for headache to be
ducted (Dworkin et al. 1990a; Von Korff et al. 19931, effective (Jurish et al. 1983; Richardson and McGrath
and from the extensive work of Turk, Rudy and col- 1989; Nash and Holroyd 1992). Nash and Holroyd
leagues (Turk and Rudy 1988; Rudy et al. 1990) in a (1992) indicate that CB treatments for headache can be
pain clinic setting, as well as psychosocial assessment made less costly and thereby more widely available to
of TMD patients by numerous workers (Marbach et al. the degree that self-management skills can be learned
1983; Keefe and Dolan 1986; Schnurr et al. 1990; with minimal assistance from a therapist.
McCreary et al. 1991). Although TMD patients are exposed to many forms
of treatment, from physically based modalities such as
Cognitive-behavioral treatments and chronic pain occlusal adjustment (Clark and Adler 1985) and TMJ
Cognitive-behavioral (CB) treatment methods have surgery (Benson and Keith 1985) to behavioral and
been incorporated into the overall management of the other psychologically based modalities, such as
most common chronic pain conditions and their use is biofeedback, relaxation and diverse forms of psy-
especially widespread in pain clinics. CB chronic pain chotherapy (Moulton 1966; Pomp 19741, the manage-
programs typically involve multiple components, in- ment of TMD is not associated with the same
cluding: (1) information to increase knowledge and widespread use of CB approaches reported for other
awareness of factors influencing chronic pain prob- common chronic pain conditions. When biobehavioral
lems; (2) cognitive and behavioral therapies aimed at methods are used, they tend to be limited to biofeed-
increasing physical and functional activities, and adap- back and, to a lesser extent, relaxation therapies (Oke-
tive responses to pain; and (3) skills training such as son et al. 1983; Burdette and Gale 1988). These biobe-
the use of relaxation, biofeedback, hypnosis and other havioral treatments are delivered on an individual basis
self-control strategies to modify perception of pain and and the efficacy of group interventions for TMD has
related bodily sensations (Fordyce 1976; Turk et al. not been evaluated in controlled studies.
1983; Keefe and Gil 1986; Turner and Roman0 1990). With notable exceptions (Stam et al. 1984; Funch
The importance of addressing psychological and behav- and Gale 1986), the tendency to view psychologically
ioral aspects of chronic pain problems is now widely based therapies as treatments of last resort seems to
recognized because physical findings are often not con- prevail in the clinical TMD literature. For example,
sistent with observed pain behaviors and disability. Clark and colleagues reported research (Clark 1986)
Numerous studies have examined the efficacy of CB which provided TMD patients with a package of CB
methods for chronic pain problems. Comprehensive interventions (e.g., cognitive restructuring, coping
reviews by Turner and Roman0 (Turner 1982; Turner methods, relaxation skills training). However, patients
and Roman0 1990) and Turk et al. (1983) support the receiving this intervention were those who had not
use of multi-component CB treatments, including the succeeded at biologically based TMD therapies. Keefe
use of group approaches. Some researchers have also has advocated early introduction of CB methods into
found that patients who participate in CB treatment chronic pain management and points, specifically, to
show continued improvement on longer-term follow-up the absence of CB programs for the management of
past the end of such treatment (Keel 1982; Turner et TMD (Keefe and Dolan 1986). Scott and Gregg (1980)
al. 1990). reviewed psychological aspects of TMD treatment and
also suggested early intervention with biobehavioral
Minimal interventions methods to minimize the likelihood of prolonged TMD
A recent innovation based on CB principles, and pain problems. Because TMD tends to be a chronic
referred to as minimal intervention or minimal therapy recurrent pain condition (Dworkin et al. 1992a), many
(Glynn et al. 1990; Glasgow et al. 1991), also seems patients seek repeated bouts of treatment in response
promising as a pragmatic approach to management of to the cyclic nature of the pain condition. Dichotomiz-
chronic pain problems. Minimal interventions empha- ing TMD treatments into either physical or behavioral
size use of information and education in the form of approaches relegates behavioral methods only to ‘re-
self-care materials coupled with brief professional sistant’ cases which may deny many TMD patients
guidance at critical points and low-cost methods for opportunities to learn more efficacious methods of
 177

long-term self-management of their problem earlier in assessment of pain clinic patients, have demonstrated,
the course of the condition. for example, that TMD patients characterized as dys-
In this report we present results from a randomized functional show greater depression and report more
clinical trial which compared, at 3- and 12month fol- physical symptoms than TMD patients categorized as
low-ups, the effects of usual TMD treatment on pain ‘adaptive capers’, although dysfunctional TMD pa-
and related physical and psychological variables with tients and adaptive capers do not differ in commonly
the effects of a CB intervention delivered to small assessed physical signs of TMD.
groups of patients before usual TMD treatment began. Using an alternative classification system (Van Korff
By introducing this intervention prior to the onset of et al. 1992), dysfunctional chronic pain patients are
usual treatment, we hoped to test a clinical model that defined as having high pain intensity and pain-related
would make apparent the early introduction of biobe- interference with daily activities. Patients classified as
havioral methods, along with the more biologically dysfunctional were found to be more depressed, to
based usual treatments that dentists provide. have higher somatization scores, and to use more med-
The purpose of this study was to determine whether ications and health care than non-dysfunctional pa-
a minimal CB intervention followed by conventional tients.
TMD treatment was more beneficial for TMD pain
and related limitations in mandibular function than Study hypotheses
clinical dental treatment alone. A second purpose of The present study was designed to introduce a 2-ses-
the study was to determine whether patients classified sion CB component prior to the onset of the patient’s
as high in somatization and psychosocial dysfunction clinical treatment. Because the intervention was de-
would respond less favorably to this minimal interven- signed to be placed before conventional TMD treat-
tion than would those low in somatization and dysfunc- ment began, no attempt was made to distinguish among
tion. the clinical subtypes of TMD when recruiting study
subjects. We hypothesized that the CB intervention
followed by usual TMD treatment, compared to usual
Increasing attention has been called to the role that treatment alone, would be associated with greater de-
reporting of multiple non-specific physical symptoms creases in pain and pain interference, and with greater
may have in determining health care behavior (Bridges improvement in mandibular function and psychological
and Goldberg 1985; Katon 1985). We have shown that distress. We also hypothesized that patients would
somatization, defined for present purposes as the re- show greater improvement at 1Zmonth than at 3-month
porting of non-pain-related physical s~ptoms such as follow-up in the CB inte~ention group but not in the
tremors, heart palpitations, etc., is associated with de- usual treatment group. Of secondary interest, we hy-
pression in chronic pain patients (Dworkin et al. 1990e) pothesized that somatization scores and grades of dys-
and with increased number of muscle sites tender to functional chronic pain would not be lowered by the
palpation in TMD patients (Wilson et al. 1991). Soma- CB intervention, which was not designed in content or
tization has also been shown to be associated with length to address such complex problems.
increased likelih~d of seeking treatment from multi-
ple providers (Katon et al. 1986). For example, using
norms for populations from which our study samples
Methods
are drawn, we found that approximately 50% of TMD
patients scoring in the top quartile for somatization
@CL-90-R age/ sex-adjusted scale scores) reported a Subjects
Subjects ~tentially available to participate in the study included
history of seeing 5 or more TMD providers, compared
395 patients experiencing pain and related symptoms of TMD re-
to approximately 18% of TMD patients whose somati- cruited from the TMJ Clinic of Group Health Cooperative of Puget
zation scores were in the lowest quartile. McCreary et Sound (GHC) or Orofacial Pain and Dysfunction Clinic at the
al. (1992), in a study relating psychological factors to University of Washington School of Dentistry (VW). Criteria for
TMD treatment outcome, concluded that unless soma- study inclusion were referral for treatment of TMD with a self-report
tization issues are addressed with TMD patients, suc- of facial ache or pain in the muscles of mastication, the TM joint, the
region in front of the ear or inside the ear, other than infection.
cessful treatment outcome is threatened. Exclusion criteria included pain attributable to confirmed migraine
or head pain condition other than tension headache; acute infection
Psychosocial dysfunction or other significant disease of the teeth, ear, eye, nose or throat; or
The assessment of psychosocial functioning in history of significant or debilitating chronic physical or mental ill-
chronic pain patients, including TMD patients, has ness. Patients requiring emergency TMD treatment were also ex-
cluded from the study. All participants were recruited into the study
received appreciable attention (Osterweis et ai. 1987; prior to their initial examination with the TMD dental specialist, by
Social Security Administration 1987; Turk and Rudy clinical field examiners who were not involved with usual treatment
1988). Rudy et al. (1989), using methods derived from or with the CB intervention.
I 7s

Of the 395 patients who met eligibility criteria, 185 agreed to I?sycho.wciul r,ariabies. These variables were assessed by rtems
participate. Using a block randomization schedule (Pocock 1983). 95 from the Symptom Checklist 90.Revised (SCL-90-R) (Derogates IOS3)
were assigned to the CB intervention and 90 were assigned to the Depression and Somatization scales. Age- and sex-adjusted stand-
usual treatment (UT) group. Extensive interview and clinical exami- ardized scale scores were computed for each subject using popula-
nation data were collected at (pre-treatment) baseline and at 3- and tion norms (Van Korff et al. lY88; Dworkin et al. 199Oe) derived
12-month follow-ups. Of those randomized. 14X(80%: CB = 69 and from a random sample survey of the community p[)puiatiol~ from
UT = 7’1) coInpleted the 3month follow-up. Outcome data for this which approximately 60% of the present clinical study sample was
report come from the sample of I39 patients (75%; CB = 66 and drawn.
UT = 73) who completed the entire study through 12-month follow- Dysfwtctiona~ chronzc pain. We used a O-IV scale developed hy
up. All study participants provided signed, informed consent prior to Von Korff et al. (1992) to grade patients as functional or dysfunc-
randomization. Overall, approximately 85% of patients were female, tional. The scale incorporates characteristic pain level, degree of
81% completed more than high school education and 96% were interference due to pain and number of days of activity lost due to
Caucasian. The average age of all participants was 371f-IO.3 years. pain. The reliability and validity of this graded chronic pain scale
More than two-thirds reported experiencing TMD pain for greater have been established with separate samples of headache. back pain
than 1 year. Approximately 58% of both the CB and UT groups were and TMD patients and the scale has been shown to be useful for
composed of GHC patients and 42% of each group were IJW clinic relating pain-related dysfunction to psychological, psychosocial and
patients. CB and UT groups did not differ significantly in age, clinical variables (Graff-Radford et al. 1991; Von Korff et al. 1992).
gender, level of education. race, or pain-related and clinical vari- For present purposes, functional TMD patients were defined as
ables. as summarized in Table 1. those whose characteristic pain was less than SOon a O-100 scale and
whose combined pain interference and activity limitation scores were
311or below on O-100 scale. Dysfunctional patients were those who
Clinical meuswes. Mandibular range of motion measures were scored above these cut-off points (see Von Korff et al. 1992 for
obtained to assess the patient’s ability to open the jaw without pain detailed description of methods to compute scores for determining
and the extent to which the jaw can be assisted open by the attending functional and dysfunctional chronic pain status). The quantitative
examiner. Measures were recorded in millimeters (mm) of: (a) unas- criteria used allow functional TMD patients to be reliably defined as
sisted mandibular opening without pain and (b) maximum assisted minimally impacted by their pain condition, reporting low-moderate
mandibular opening. pain which is not highly persistent (i.e., present on many days), and
S~~f-re~~f measures. TMD History Questionnaire inquired into not associated with activity limitation. The composite measure of
sociodemographic variables, including age, gender, race, income, graded chronic pain is included in the study analyses even though it
education and marital status as well as TMD treatment history. incorporates measures of characteristic pain and pain interference,
Puin measures. Visual analog scales (VAS) were used to assess which are also analyzed separately as major dependent variables in
present pain intensity, average and worst pain intensity in the past 2 the study because it extends the measurement of chronic pain to a
months as well as pain interference with daily activities. Two pain-re- multi-dimensional assessment which directly quantifies the impact of
lated measures were analyzed from data gathered. TMD pain on important behaviors in daily living. We were interested
Churzz~[er~~~ic pain. The measure characteristic pain represents in observing the potential for the minimal intervention to differen-
the average of VAS scores for average, present and worst pain. This tially affect pain-related variables of interest when they were ana-
has been shown to be somewhat more reliable than a single measure lyzed unidimensionally and when incorporated into a multidimen-
of average pain (Dworkin et al. 1990d). sional measurement.
t’uin interference. Pain interference with daily activities in the Self-rating of change in TMD condition and response to treatment.
previous 2 months was measured by a O-10 point scale anchored by This is a single item assessed at 3- and Z-month follow-up whether
0= no interference and lO=as unable to carry on any activities. patients viewed their TMD condition as improved, stabiiized or
Pain interference has been shown to be a usefui measure of the worsened. A series of items inquired into evaluation of TMD treat-
impact of chronic pain on ability to perform usual daily activities ment received, knowledge regarding factors thought to exacerbate
(Van Korff et al. 1992). TMD and methods used for self-management of the condition.

TABLE I
BASELINE CHARACTERISTICS OF CB AND UT GROUPS: DEMOGRAPHIC AND DEPENDENT MEASURES

UT
(“n”=66) (n = 73)
Demographics
Age i + SD) 38.4 ( * 11.30 35.9 (k9.21)
Gender (% females) 83 86
In Pain > 1 year (%I 66.7 72.6
Completed High School (%) 97.0 98.6

Dependent measures !? (*SD) ?? (_+SD)

Characteristic Pain (O-10) 5.2 (+ 1.97) 4.5 ( + 1.90)
Pain Interference (O-10) 3.4 (i2.37) 3.1 (k2.72)
Somatization (age/sex standardized) 0.6 ti 1.55) 0.75 ( + 1.45)
Depression (age/sex standardized) 0.19 (9 1.08) 0.26 f& i.rs>
Unassisted Opening (mm) 35.1 (* 10.49) 36.4 ( f 11.95)
Maximum Assisted Opening (mm) 43.4 (+ 9.38) 46.5 ( f 8.27)

AI1 group differences n.s.

 179

Finally, CB participants were asked to evaluate their satisfaction to standardized protocols and calibrated to acceptable levels of
with the CB intervention and its perceived usefulness. reliability for assessing the variables covered by the examination. The
training protocol, calibration and reliability of the dental hygienist
TMD clinical examiners have been described previously (Dworkin et
Procedures al. 1988, 1990b).
CB interuention. The CB intervention was delivered in a small All subjects who dropped out from the study prior to completion
group format of from 2 to 7 (mode = 4) TMD patients. The CB of the 12-month follow-up were asked to complete an abbreviated
groups met for two 2-h sessions spaced 1 week apart. For most questionnaire inquiring into the status of their pain and jaw function
patients, group sessions began before onset of dental treatment. The in order to allow intent to treat analyses of all subjects (Turk and
groups were team-led by a study dentist and study psychologist. The Rudy 1990a; Lee et al. 1991; Peter et al. 1992).
teams were drawn from a panel of 4 dentist-specialists and 4 clinical
psychologists, all experienced in the treatment of chronic pain pa-
tients. Such teams used a detailed manual and set of materials to
provide information concerning the nature and typical course of Results
TMD; biomedical and biobehavioral management of TMD; the rela-
tionships among jaw muscle fatigue, muscle tension, and the psy-
Baseline comparisons
chophysiologic aspects of stress; the basics of pain physiology with an
emphasis on chronic pain; how to self-monitor TMD signs and GHC vs. UW patients. Approximately 60% of sub-
symptoms; and an introduction to cognitive and behavioral pain and jects were drawn from the TMJ Clinic of GHC and
stress-coping strategies. Patients learned and had an opportunity to 40% from the Orofacial Pain and Dysfunction Clinic,
briefly practice a progressive relaxation method and a simple physio- UW. Analyses of baseline clinical and demographic
therapy exercise for jaw muscles. Patients also developed a daily
personal plan for adherence to these pain and stress reduction and
data revealed no significant differences (t tests for
physiotherapy exercises. Each patient was provided a personalized independent means, (Y< 0.05) between the GHC and
notebook containing study materials and forms, a relaxation (audio) UW groups (e.g., characteristic pain was 4.8 vs. 4.9,
tape, reminder cards containing exercise schedules and brief descrip- respectively, for the GHC and UW groups; unassisted
tions, and an annotated list of relevant articles and books. The
vertical opening was 36.3 vs. 34.1 mm; SCL-90-R-de-
psychologist called participants between the CB sessions to clarify
and/or reinforce the group discussion content and called 1 month
pression score was 0.2 vs. 0.4; and age was 37.5 vs. 37.1
after the second session to discuss the patient’s progress in imple- years).
menting the daily personal plan. CB patients also received usual Non-participants and study drop-outs vs. study com-
treatment by their dentist TMD-specialist following the 2-session pleters. Of those refusing to participate in the study,
intervention.
77% gave either the time or the location of sessions as
Psychologists and dentists were provided with a detailed thera-
pist’s manual with scripts for each of the 2 sessions. Psychologists
their reason for non-participation (sessions were sched-
and dentists were trained together in the use of study materials and uled in the evening at sites other than UW and GHC
methods, practicing first in groups among themselves and then with clinics. Of these refusals, 58% agreed to answer 3 key
non-patient clinic personnel and finally with pilot testing of study questions regarding their pain condition over the past 2
conditions using several groups of TMD patients. Psychologists and
weeks: (1) number of days of facial pain in last 2
dentists rotated in their team composition to insure consistency of
presentation and regular discussions among all clinician-researchers
months, (2) number of days of limited activity in last 2
insured acceptable consistency among those conducting the 2-session months, and (3) average pain intensity in last 2 months
interventions. Since several combinations of dentist-psychologist were (VAS).
employed, inadequate sample sizes precluded formal statistical anal- There was no significant difference between groups
yses of outcomes by intervention team.
for numbers of days of facial pain, with 52% of subjects
UT condition. Dental ‘treatment-as-usual’ was delivered by one
of the study dentist/TMD specialists at the TMJ Clinic of Group
in both groups reporting the maximum of 60 days in
Health Cooperative or the Orofacial Pain and Dysfunction Clinic of pain. More refusals (84%) reported no activity-limited
the Department of Oral Medicine, University of Washington. Usual days compared to 72% of subjects participating, a
treatment was conservative and typically included use of flat-plane statistically significant difference. In contrast to this
occlusal splints, non-steroidal anti-inflammatory medications, passive
finding, refusals showed a trend to have higher pain
and active range of jaw motion exercises, modification of parafunc-
tional and/or dietary habits and regular use of cold and heat packs. intensity levels (mean = 5.29) compared to those partic-
No attempt was made to influence patient dental treatment. ipating (mean = 4.80, P = 0.07) in the study.
Data collection procedures. A previously developed and stand- Baseline analyses were performed comparing those
ardized interview and clinical examination (Dworkin et al. 1988) was subjects who completed the study (through 12-month
used in this study to gather extensive data on the most common
clinical signs associated with TMD as well as self-report data related
follow-up, n = 139) with those who dropped out after
to pain and dysfunction, psychosocial variables, treatment history, baseline assessment (n = 46) on 50 clinical and demo-
self-management and coping behaviors and satisfaction with treat- graphic variables. Analyses (t tests for independent
ment. A subset of the most commonly assessed clinical and psychoso- means for continuous variables, chi-square analyses for
cial variables associated with TMD and relevant to the experimental categorical variables) yielded some significant differ-
hypotheses was selected for analysis. All clinical and self-report data
were gathered at baseline and at 3- and 12-month follow-up by
ences. Study dropouts reported significantly lower in-
dental hygienist examiners blind to the subject’s original random come (P = 0.003) and had more recent onset of pain,
assignment to the CB or UT study conditions. Dental hygienists who with a median of 3.5 years pain duration in the com-
served as TMD field examiners/interviewers were trained according pleted study group vs. 1.4 years in the drop outs
 Characteristic Pain Score (O-10) Pain Interference Score (O-10)
6 r---------_ 6 pi- --- ~-~

* Group x Time Interaction: p-O.015

Qroup x Time Interaction: P.O.066
OA -- ..J . . . _-
BASELINE S-MONTHS 12-MONTHS SASELINE 3-MONTHS 12-MONTHS
STUDY PERIOD STUDY PERIOD
Fig. 1. A: characteristic pain (O-10): group CB vs. UT. Baseline, 3- and 12-month follow-ups (*SEMI, respectively, for UT 4.54 (0.22), 3.14
(0.28), 3.03 (0.27), and 5.17 (0.2.Q 3.73 (0.31), 2.74 (0.32). B: pain interference score (o-10): group CB vs. UT. Baseline, 3- and 12-month
follow-ups (f SEM), respectively, for UT 3.07 (0.321, 2.16 (0.291, 1.75 (0.31), and 3.38 (0.29), 2.11 (0.30, 1.09 (0.22).

(P = 0.05). We also observed that 47% of those failing observed. For example, in the present study character-
to complete the study were assigned to the UT vs. 64% istic pain was correlated 0.05 with maximum assisted
to the CB group (P = 0.05). jaw opening and -0.13 with unassisted jaw opening.
CB ZIS.UT groups ~~t~~ycompleter~~.Baseline com- Similarly, somatization was correlated - 0.01 with max-
parisons of the CB and UT groups on relevant demo- imum assisted opening and - 0.13 with unassisted
graphic and dependent variables revealed no signifi- opening. The same pattern was observed for the rela-
cant differences between the groups, as summarized in tion between pain interference scores and clinical signs.
Table I. Similarly, preliminary analyses did not reveal Thus, multiple univariate ANOVAs were performed
systematic differences between those subjects who re- rather than an overall multivariate analysis of variance
ceived both sessions before usual treatment began, and (MANOVA) followed by post-hoc univariate analyses,
those who (for logistic reasons) may have completed because there was no theoretical interest in analyzing
the second group session shortly after an initial clinical the changes on all the dependent variables as a group
treatment visit. (Huberly and Morris 1989).
C~a~acte~tic pain. As depicted in Fig. la, a signifi-
Follow-up corn~a~o~ of CB and UT cant Group X Time interaction was observed for char-
The initial set of major analyses compared the UT acteristic pain (F = 4.23, df = 2,272, P = 0.015). The
and CB groups at 3- and Z-month follow-up for differ- CB group continued to decrease in characteristic pain
ences from baseline in characteristic pain level, vertical between the 3- and 1Zmonth follow-ups at a signifi-
range of jaw motion, depression, and pain interference, cantly greater rate than did the UT group; the latter
using repeated-measures analyses of variance group’s mean characteristic pain level remained essen-
(ANOVA) with 2 Groups (CB, UT) X3 Time Points tially constant from the 3- to the 12-month follow-up.
(baseline, 3-month, 12-month). As in previous research Thus, while no significant main effects were observed
(Rudy et al. 1989), weak relationships among pain or - that is, differences in characteristic pain level be-
psychological measures and clinical signs of TMD were tween the groups were not significant at baseline, 3- or

TABLE II
MEAN VALUES OF UT (n = 73) AND CB (n = 66) GROUPS FGR CLINICAL AND PSYCHOLOGICAL VARIABLES
Repeated-measures ANOVA for Group effects (CB vs. UT); Time effects (baseline, 3- and 12-month follow-ups); Group X Time interaction.

Dependent variables Baseline 3 Months 12 Months P

UT CB UT CB UT CB Group Time Group x Time

Maximum assisted mandibular opening 46.5 43.1 46.9 45.2 46.9 45.6 n.s. 0.05 n.s.
Unassisted mandibular opening 36.4 35.1 38.9 38.1 38.5 39.2 n.s. 0.001 us.
Depression score * 0.26 0.12 - 0.06 0.04 - 0.20 - 0.06 n.s. 0.001 Il.%
Somatization score * 0.75 0.86 0.54 0.59 0.44 0.44 n.s, 0.001 n.s.

* SCL-90-R scores age/sex-adjusted and standardized to population norms.

 181

1Zmonths - the significant interaction indicates those ble for the CB and UT groups. Baseline, 3- and 12-
receiving the minimal CB intervention preceding usual month mean values for these variables are summarized
TMD treatment showed a ~ntinuing benefit in pain in Table II. Thus, the CB inte~ention did not enhance
reduction past the 3-month follow-up that was not seen the effect of usual treatment on these physical and
with those receiving only usual treatment. psychological parameters.
Pain interference. Results were somewhat compara- TMD pain dysfunction. In addition to the analyses
ble, as seen in Fig. lb, for the measure of pain interfer- just described, a secondary set of repeated-measures
ence. There was a strong trend for the CB group only ANOVA included functional vs. dys~nctional graded
to continue to decrease in pain interference through chronic pain status as an additional variable because it
the 1Zmonth follow-up period. However, the Group was of interest to examine if functional vs. dysfunc-
(CB, UT) X Time interaction just failed to reach statis- tional chronic pain status influenced responses to the
tical significance at the 0.05 level (F = 3.04, df = 2,272, CB intervention. Results of these analyses of 2 groups
P = 0.066). (CB, UT) X 3 time periods (baseline, 3-month, 12”
Clinicaland ~~~~010~~~~~e~~~re~.Clinical physical month follow-ups) x 2 levels of chronic pain grade
measures of vertical range of jaw opening (unassisted (functional, dysfunctional) are depicted in Fig. 2a-c,
opening and maximum assisted opening) and measures Characteristic pain (see Fig. 2a) was analyzed as a
of psychological status (depression and somatization) dependent variable only to demonstrate that this vari-
all showed improvement over time that was compara- able, which enters into the criteria for defining func-
Characteristic Pain Score (O-10) SCL-90-R Somatization Scale Score
I

6
1.6 *-..
--._
u.
. .._ --w_
1.2 *-----__-__
_.._ ------------a
4
_.
CI----___
0.8 1 -------------

I
2 t -.

”
/A I “1
t :---?“----------,
B

BASELINE 3-MONTHS IP-MONTHS BASELINE 3-MONTHS 12-MONTHS

STUDY PERIOD STUDY PERIOD
I
- CBfFune. N.31 + Usual TxfFunc. N-42 - CBJFunc. N-31
--*-CBJDy4. N-34 --r--Usual TxJOya. N.29 - +- CBlOya. N.34

Maximum Assisted Jaw Opening (mm)

--___ ---___

BASELINE 3-MONTHS It?-MONTHS

STUDY PERIOD

+ Uw4, TxJFunc. N.43 -r CBJFuno. N-26

--- Ulual TxJDy& N-26 -+I-- GBJDye. N=33 /

Fig. 2. a: Characteristic pain (O-10): group CB vs. UT by functional and dysfunctional patients (k SEMI. Baseline, 3- and la-month follow-ups,
respectively, for UT Functional (3.42 (0.201, 2.07 (0.291, 2.26 (0.33)); CB Functional (3.83 (0.321, 2.67 (0.381, 2.02 (0.40)); UT Dysfunctional (6.24
(0.24), 4.76 (OAO), 4.19 (0.38)); and CB Dysfunctional (6.40 (0.20), 4.71 (0.411, 3.39 (0.46)). b: Somatization score for group CB vs. UT by
functional and dysfunctional patients (* SEM). Baseline, 3- and 12-month follow-ups, respectively, for UT Functional (0.17 (0.151, 0.08 (0.14,
-0.01 P (0.13)); CB Functional (0.31 (0.X8), 0.24 (0.201, 0.07 (0.15)); UT Dysfunctional (1.60 (0.29), 1.21 (0.28), 1.08 (0.26)X and CB
Dysfunctional (1.35 (0.28), 0.92 (0.26), 0.77 (0.33)). c: Maximum assisted jaw opening for group CB VS. UT by functional and dysfunctional
patients (& SEM). Baseline, 3- and lbmonth foI~ow-ups, respectively, for UT Functional (46.19 (1.271, 46.74 (1.241, 47.05 (1.21)); CB Functional
(43.00 (2.11), 44.88 (1.78), 45.88 (1.72)); UT Dysfunctional (46.93 (1.57), 47.14 (1.361, 46.68 (1.39)); and CB Dysfunctional (43.24 (1.471, 45.48
(1.25), 45.30 (1.25)).
tional vs. dysfunctional chronic pain level, does indeed Percent of Subjects
separate patients into functional and dysfunctional 100

sub-groups. Mean baseline, 3- and I2-month follow-up

levels of characteristic pain are significantly different (t
tests for independent means, baseline t = 8.02, df =
295, P < 0.01; 3-months. t = 7.12, P < 0.01; 12-months,
t = 4.71. P < 0.011, when functional and dysfunctiona
patients are compared at each time point.
Somatization scores were examined as a dependent
variable because it was of interest to observe how these
scores were related to chronic pain grade and to ob-
serve how they changed for the CB and UT groups. Cognitive-Behavioral Treatment-as-Usual
Somatization score does not enter into the determina-
tion of functional vs. dysfunctional status, yet Fig. 2b m Improved Not improved
reveals a pattern of significant differences in somatiza- Chi-Square-4.88, p q0.03
tion scores between functional and dysfunctional pa-
Fig. 3. Self-reported improvement in TMD. Conditiun: baseline to
tients at each time point (baseline: t = 5.82, dJ”= 223,
I?-month follow-up (some/much improved vs. no change/worse).
P < 0.01; 3 months: c = 4.28, #= 223, P < 0.01; 12
months: t = 4.19. df = 223, P < 0.01). By contrast, max-
imum assisted jaw opening. selected for these analyses
because it is one of the few clinically relevant and the assigned treatment. Analysis by intent to treat
objective physical signs associated with TMD that is protects against possible bias introduced by differential
not dependent on self-report, shows no significant dif- drop-out rates in CB vs. UT groups. Analyses by intent
ferences at any time point, for either the functional or to treat, although commonly reported for biomedical
dysfunctional TMD patients, as shown in Fig. 2c (base- randomized trials (Lee et al. 19911, are rarely reported
line: t = 0.03, n.s.; 3 months: t = 0.14, n.s.; 12 months: in randomized clinical trials of biobehaaiorai interven-
t = 0.47. n.s.). For example, while somatization re- tions.
turned to the population mean for functional patients In the present study, randomization was originally
at 12-months, this decline was from only slightty ele- performed on 185 TMD patients and resulted in com-
vated basefine levels. For dysfunctional patients, while parable numbers of patients assigned to the CB and
somatization did lower from baseline levels. at 12- UT groups. Results of the data analyses presented so
month follow-up somatization levels still remained high far were for 139 (CB = 66; UT = 73) patients who
(i.e., 75th percentile, see Fig. 2b), according to norma- completed the 12-month follow-up. Of the remaining
tive values derived from our studies of the general (i.e., 46 (185 - 139) patients, baseline and l%-month follow-
non-TMD clinic) population with demographic charac- up data on the major dependent variables were avail-
teristics (Van Korff et al. 1988; Dworkin et al. 1990e) abie for 25 patients. Repeated-measures 2 Groups
comparable to the clinic sample studied here. The (CB, UT) X 2 Time periods (baseline, 1Zmonth follow-
maximum assisted jaw opening measure remained rela- up) ANOVA were performed on the resulting sample
tively constant from baseline through Q-month follow- of 164 (139 + 25) patients who were originally random-
up (see Fig. 2~). ized to CB (n = 83) and UT (n = 81). Results revealed
Additional analyses for between-group differences a significant Group x Time interaction for characteris-
at baseline, 3- and 12 month follow-up also compared tic pain (F = 5.79, df = 1,161, P = 0.0171, as was found
functional and dysfunctional patients for changes over with the sample of 12-month study completers. The
time. We observed no significant interaction effects Group X Time interaction was not significant for any
between chronic pain grade (functional vs. dysfunc- other dependent variable (pain interference, somatiza-
tional) and group assignment (CB vs. UT) for somatiza- tion or depression and the clinical measures of
tion or for maximum assisted opening. Although the mandibular opening).
sample sizes in all these secondary analyses were small, Assessment of leaf-repotted im~~oL~erne~t. At 12-
and power to detect significant differences were low, month follow-up, as shown in Fig. 3, significantly more
these exploratory analyses are provocative and, at the patients in the CB group (86.4%) than in the UT group
least, consistent with our hypothesis that the minimal (70.1%) reported improvement in their TMD condition
CB intervention would not influence somatization. compared to their baseline status. Conversely, more
Analyses by intent to treat. These analyses allow than twice as many UT subjects (29.9%) as CB patients
determination of the effectiveness of the CB interven- (13.6%) reported no improvement or worsening of
tion for all patients randomized to the CB and UT their condition at IZmonth foIlow-up (x2 = 4.88, df =
conditions, regardless of whether they actually received I, P = 0.027).
 183

TABLE III
KNOWLEDGE OF POSITIVE AND NEGATIVE FACTORS INFLUENCING TMD
Comparison of CB and UT groups for ‘agree’, ‘disagree’, ‘don’t know’ (%I. At Q-month follow-up.

TMD item Patients reporting (%) P

Agree Disagree Don’t know

CB UT CB UT CB UT
Oral habits (nail biting, clenching or grinding teeth)
are often significant in the development of TMD.
(agree = correct response) 94 74 5 4 1 22 0.001

All individuals with clicking joints should have treatment.

(disagree = correct reponse) 20 15 56 38 24 47 0.019

Relaxation treatments are an effective treatment for TMD.

(agree = correct response) 89 79 6 1 5 19 0.013

Bite adjustment is an essential treatment for TMD.

(disagree = correct response) 26 35 47 21 27 44 0.005

Crooked or missing teeth, or a poor fit of the upper teeth

to the lower teeth is a primary cause of TMD.
(disagree = correct response) 24 32 50 21 26 47 0.001

The normal relaxed resting position for the mouth

is to have the upper and lower teeth touching.
(disagree = correct response) 11 17 87 54 2 29 0.001

Eualuatians of TMD treatment received and knowl- was most striking at the positive and negative extremes.
edge concerning TMD and self-management. All pa- Approximately two-thirds of CB patients, compared to
tients responded to questionnaire items inquiring into only one-third of UT patients, evaluated their TMD
evaluation of TMD treatment received. For CB pa- treatment as ‘very/ extremely helpful’; conversely,
tients, this included the combination of the CB inter- about twice as many UT patients (21%) as CB patients
vention plus usual treatment; for UT patients, re- (11%) viewed their overall TMD treatment as ‘not at
sponses reflected evaluation of only TMD usual treat- all/ minimally helpful’ (x2 = 10.5, df = 2, P < 0.005).
ment provided by the dentist-specialist. For the item, When 12month responses were analyzed for knowl-
‘how helpful has treatment been for your face and jaw edge of factors thought to negatively influence TMD
pain’, using a O-10 scale anchored with 0 = not at a11 and/ or positive self-management strategies thought to
helpful and 10 = extremely helpful, data were analyzed be helpful in ameliorating pain and discomfort, CB
by comparing CB and UT groups at 1Zmonth follow- patients were significantly better informed about the
up. Responses were collapsed into 3 categories: (a) TMD condition and how patients might help them-
O-2 = not at all to minimally helpful; (b) 3-7 = selves. Table III summarizes these data and indicates
moderately helpful; and cc> 8-10 = very to extremely that for each item, CB patients were significantly (chi-
helpful. Significantly more favorable responses were square analyses) better informed than UT patients
given by the CB patients. The disparity in evaluations about factors such as oral habits that were likely to
of overall TMD treatment between UT and CB groups exacerbate their condition, as well as about useful

TABLE IV
POST-TREA~ENT SATISFA~ON WITH THE CB TREATME~ PROGRAM

Percentage subjects (n = 70)

Usefulness 0E Not useful Somewhat useful Very useful
Developing personal plan 0 12.9 87.1
Stress management methods 2.9 8.6 88.5
Jaw exercises 4.3 20.0 75.7
Relaxation methods 1.4 12.9 85.7
Confident about ability to apply program methods for: Not confident Somewhat confident Very confident
Self-monitoring of TMD 4.3 21.4 74.3
Maintaining correct jaw posture 0 12.9 87.1
Muscle relaxation 7.1 14.3 78.6
183

self-management strategies and preferred customary Although the overall rate of self-reported improvc-
jaw position to reduce risk of increased TMD pain and ment was high, which is typical for studies reporting on
discomfort. the efficacy of a wide variety of TMD treatments
Patient satisfaction with the CB program. Given the (Greene and Marbach 19821, significantly more CB
positive implications of the data concerning CB vs. UT than UT patients reported an overall improvement in
group evaluations of TMD treatment and knowledge their TMD condition 1 year after baseline. It was also
about self-management, we would expect overall satis- encouraging to note the high rate of acceptance of the
faction with the CB program. Data gathered by post- CB component among those attending the groups. The
program questionnaires, self-administered immediately dentist introduced biobehavioral concepts in the den-
after the second session, did indicate very high satisfac- tist-psychologist conjointIy led groups, perhaps enhanc-
tion with the program. Table IV summarizes patient ing patient acceptance of the applicability of these
global assessment of the program as well as assessment non-biomedical approaches to their TMD condition. A
to selected components. In addition, overall, 94% were central feature of the CB program was the emphasis
extremely satisfied with the program, 6% were some- placed on developing a personal plan for managing
what satisfied and none were dissatisfied. About 86% TMD, which each patient was required to prepare in
indicated they would be ‘very likely’ to recommend the the second group session. At l-year follow-up. 72% of
program to a friend and about 14% indicated a ‘fair’ CB participants reported that they were still following
likelihood to do so; again, no one indicated they would their personal plan while about 28% indicated they
‘not recommend’ the CB program. used it rarely or not at all. Similarly, 81% of the CB
patients evaluated the personal plan as ‘somewhat’ to
‘very impor~ant~ in the management of their TMD
Discussion condition (19% said it was of little to no importance).
Interestingly, only 65% of patients in this CB followed
A 2-session CB intervention was designed to in- by IJT group thought that dental treatment was some-
crease knowledge of etiology and treatment of TMD as what to very important and 35% thought dental treat-
well as provide skills in self-monitoring the condition ment was of little or no importance to the management
and in use of behavioral strategies to manage chronic of their TMD condition at l-year follow-up. Also, at 1
TMD pain. This intervention was introduced before year, we observed that CB patients demonstrated more
usual clinical treatment for TMD began. Our hypothe- knowledge than did UT patients about the nature and
sis that patients who participated in the CB interven- self-management of their condition. Moreover, they
tion followed by usual treatment would show greater also seemed more positive than UT patients about the
long-term decreases in reported pain level and pain rest of their usual TMD treatment, as provided by the
interference in daily activities than would patients who 2 specialized clinics engaged in this study. Taken to-
received only usual treatment was confirmed. The ben- gether, these data support the view that the CB pro-
efits of CB intervention were not seen when the CB gram was not only well received by the majority of
and UT groups were compared at 3-month follow-up. patients but was also experienced as beneficial in help-
At this time, both groups showed comparable and ing them improve their TMD condition.
fairly steep improvement from baseline. During the Despite these positive outcomes, results from the
3-1Zmonth follow-up interval, however, the UT group present study must be interpreted with important
maintained essentially the same level of improvement reservations. The effects of CB vs. UT, although pre-
in characteristic pain while the CB group continued to sent after a reasonable follow-up, are modest in size.
improve, as h~othesized. During this same follow-up For pain interference, while a strong trend is noted,
interval, the CB group also showed a strong trend clear statistical significance was not demonstrated. in
toward continued reduction in pain interference. addition, a longer follow-up would have allowed us to
A similar pattern was not observed for relevant determine if the trend towards continued improvement
clinical variables involving range of mandibular motion, shown only in the CB group after 3 months reflected
or for the psychological variables of depression and an enduring pattern that extended beyond the 1 year
somatization. The improvement in all these variables follow-up to which the present study was limited. It is
was significant over time, independent of CB or UT well known that clinical trials of this type are difficult
group assignment. Thus, our hypothesis that patients in to conduct and many practical as well as experimental
the CB group would show significantly greater im- design problems have been described (Lee et al. 1991).
provement in vertical range of motion than would the Some of these involve subjects dropping from the study
UT group was not confirmed. Our expectation that the at differential rates for experimental and control groups
modest CB inte~ent~on would not have a significant (Turk and Rudy 1990a), the design of appropriate
impact on somatization and dysfunctional chronic pain control groups (Whitney and Von Korff 19921, and
grade was borne out. issues in outcomes assessment (Dworkin et al. 1990bl.
 185

These issues have received increased recent attention to our intervention and to elucidate possible mecha-
in the chronic pain literature (Turk and Rudy 1990b; nisms of action to account for the patterns of response
Peter et al, 1992) and our experience confined that observed. For example, we observed that dys~nctional
logistic problems encountered in conducting clinical TMD patients showed significantly higher levels of
research with groups of patients could be formidable. somatization at baseline than did functional patients.
Analyses by intent to treat are not commonly reported For this group, somatization did decrease somewhat
in biobehavioral trials. However, they have been advo- over the 12-month follow-up of this study, but re-
cated by clinical trials methodologists and were de- mained at levels in the top quartile for somatization
scribed in this paper to enhance interpretation of the using norms for the population from which the present
generalizability of these findings. study are largely drawn. By contrast, somatization
Another methodologic strength of the present study scores for functional patients returned to the mean
is its relatively large sample size (139 subjects across values for the population (age/ sex-adjusted population
both groups with complete data available for analyses mean for somatization is equal to zero). Our data have
after 1Zmonth follow-up), yielding adequate statistical an important limitation with regard to somatization, in
power to conduct the planned statistical comparisons. that they are limited to self-report of non-specific
For our treatment group sizes, for example, there is physical symptoms on the SCL-90-R and behavioral
84% power to detect significant between-group differ- data with regard to health care utilization are lacking.
ences of at least 0.5 SD at the a! = 0.05 level. Thus, we Nevertheless, our data supports the concern of Mc-
feel confident that the effect sizes we observed are Creary et al. (1992) that somatization may have an
what can reasonably be expected with inte~entions of important negative influence on outcomes of treatment
this type, but that clinical trials seeking to demonstrate for chronic pain. The observation that self-report mea-
such effects will require comparably sized experimental sures of pain and somatization seem to change over
and control groups. time (Fig. 2a,b) while an objective physical finding
With regard to issues of diagnosis, subjects were (e.g., maximum assisted jaw opening), which does not
randomly assigned to CB and UT conditions without involve self-report does not show either time- or
consideration of their clinical TMD (muscle, internal group-related changes (Fig. 2c), may also have imphca-
derangement or degenerative joint disease) diagnostic tions for the kinds of changes one can expect chronic
status. We have been intensely interested in the prob- pain patients to accomplish. In a similar vein, the
lems of diagnosis (Dworkin et al. 1990~) and have present study supports the notion that dysfunctional
recently (Dworkin and LeResche 1992) contributed to chronic pain patients might not respond readily to
making available empirically derived and operationally modest CB inte~entions which do not address the
defined research diagnostic criteria for TMD (RDC/ more complex aspects of their pain dysfunction, such
TMD). However, these criteria were not available to as somatization. These initial conclusions with regard
the study’s TMD dentist-clinicians when the present to: (1) the role of chronic pain dysfunction and its
study was designed and undertaken and reflect a limi- resistance to usual TMD treatment with and without
tation of our present analyses. Since a number of minimal CB interventions, and (2) the potential for
dentists participated in this clinical trial and they did somatization to influence treatment outcomes, require
not have available an agreed upon set of diagnostic more extensive investigation to validate their applica-
criteria and standardized examination procedures, it bility to TMD and to determine whether patterns ob-
was deemed most advisable not to include for analyses served in chronic TMD pain patients are generalizable
TMD diagnoses based on clinical data gathered in to other pain conditions.
non-standardized fashion. Our subsequent clinical In summary, the present study supports the utility of
TMD research includes these research diagnostic crite- a brief group CB intervention, placed before conven-
ria for classifying TMD subjects. tional clinical treatment for TMD began, to ameliorate
Dysfunctional chronic pain - associated with self- the report of TMD pain. The effects observed from
reports of more intense and persistent pain and mani- such a biobehavioral intervention seem long-lasting,
festations of depression and maladaptive coping behav- albeit modest in size. Further research is needed to
iors, but poorly correlated with physical pathology - explicate which components of CB inte~entions such
has been documented as present in appreciable num- as those used here are most powerful, e.g., the use of
bers of chronic pain patients seen in pain clinics (Turk small groups, placement of the CB components before
and Rudy 1988; Rudy et al. 1989) and identified in usual clinical dental treatment and the efficacy of
population-based studies Non Korff et al. 1990, 1991). having the dentist identified with biobehavioral meth-
The CB intervention was not targeted towards chang- ods as we11 as biomedical treatments. It also appears
ing levels of dysfunctional chronic pain (or of somatiza- that biobehavioral treatments, like biomedical treat-
tion). It was, however, of additional interest in this ments, are not equally effective across the spectrum of
study to observe the presence of differential responses chronic pain patients. Present evidence indicates that
somatization tends to correspond with dysfunctional and psychiatric disturbance: An epidemiologic investigation, Arch.
chronic pain status. Findings presented with regard to Gen. Psychiatry, 47 (199Oe) 239-244.
Dworkin, S.F., LeResche, L.. Van Korff, M., Dicker. B. and Som-
somatization as a potentially critical variable need to
met’s, E.. Constant, remitted and cyclic pain patterns in TMD:
be extended to determine if the readiness to report Three year follow-up, J. Dent. Res., (1992a) 441.(Abstract)
multiple non-specific physical symptoms is a predictor, Dworkin, SF., Von Korff, M. and LeReschc, I.., Epidemiologic
or ‘marker’ variable, capable of identifying individuals studies of chronic pain: A dynamic-ecologic perspective. Ann.
who resist biobehavioralIy based methods for coping Behav. Med., I4 (199263 ?I- 1t
Dworkin, S.F. and LeResche. I... Research diagnostic criteria for
with pain while engaging in excessive health care uti-
Temporomandibular Disorders, J. Craniomandib. Disord. Facial
lization. Oral Pain, 6 (1992) 301-355.
Fordyce, WE., Behavioral Methods in Chronic Pain and Illness.
C.V. Mosby, St. Louis, 1976.
Fricton, J.R., Kroening. R.J. and Hathaway, K.M., TMJ and Cranio-
Acknowledgements
facial Pain: Diagnosis and Management, lshiyaku EuroAmerica,
St. Louis, 1987.
The authors wish to acknowledge the invaluable Funch, D.P. and Gale, EN., Predicting treatment completion in a
contributions of Pamela Achziger, BS, RDH, Peggy behavioral therapy program for chronic temporomandibular pain,
Biddle, RDH, Melinda Lane, BS, RDH, Christine J. Psychosom. Res.. 30 (1986) 57-62.
Glasgow, R.E., Hollis, J.F., McRae, S.G., Lando, HA. and LaChance,
Tweedy, BS, RDH and Sandra Ulm, BS, RDH, as
P.. Providing an integrated program of low intensity tobacco
clinical examiners and Robert Harrison, BS, for data cessation services in a health maintenance organization. Health
management and analysis. Research supported by Educ. Res., 6 (1991) X7-99.
NIDR Program Project grant 1 POl-DE-08773. Glynn, T.J., Boyd, G.M. and Gruman, J.C.. Essential elements of
self-help/minimal iI~te~ention strategies for smoking cessation,
Health Educ. Q., 17 (1990) 329-345.
Graff-Radford, S.B., Reeves, J.L. and Janetta, P.J., Evaluation and
References management of the difficult facial pain patient. 10th Annual
Scientific Meeting, American Pain Society, New Orleans. 1991.
Bell, WE., Temporomandibular disorders: classification, diagnosis, Greene, C.L. and Marbach. J.J., Epidemiologic studies of mandibu-
management, Year Book Medical Publishers, Chicago, Illinois, lar dysfunction: a critical view, J. Prosthet. Dent.. 48 (1982)
1986, 184- 190.
Benson, B.J. and Keith, D.A., Patient response to surgical and Huberly, C.J. and Morris, J.D., Multivariate analysis versus multiple
nonsurgical treatment for internal derangement of the temporo- univariate analysis, Psychol. Bull., 105 (1989) 302-308.
mandibular joint, J. Oral Maxillofac. Surg., 430 (1985) 770-777. Jurish, SE., Blanchard. E.B.. Andrasik, F., Teders, S.J., Neff. D.F.
Bridges, K.W. and Goldberg, D.P., Somatic presentation of DSM III and Arena, J.G., Home-versus clinic-based treatment of vascular
psychiatric disorders in primary care, J. Psychosom. Res., 29 headache, J. Consult. Clin. Psychol., 51 (1983) 743-751.
(1985) 563-569. Katon, W., Somatization in primary care, J. Fam. Pratt., Zl(4) (1985)
Burdette, B.H. and Gale, E.N.. The effects of treatment on mastica- 257-258.
tory muscle activity and mandibular posture in myofascial pain- Katon, W., Berg, A.. Robins, A. and Risse, S., Depression: Pattern
dysfunction patients, J. Dent. Res., 8 (1988) 1126-l 130. of medical utilization and somatization in primary care. In: S.
Clark, G.T., Treatment outcomes for TMJ disorders. Sixth Annual McHugh and T.M. Vallis (Eds.), Illness behavior: a multidisci-
Session, Washington D.C, American Pain Society, 15 (1986) (Ab- plinary model, Plenum Press, New York, 1986, pp. 355-364.
stract) Keefe, F.J. and Dolan, E., Pain behavior and pain coping strategies
Clark, G.T. and Adler, R.A., A critical evaluation of occlusal ther- in low back pain and myofascial pain dysfunction syndrome
apy: Occlusal adjustment procedures, J. Am. Dent. Assoc., 110 patients, Pain, 24 (1986) 49-56.
(1985) 743-750. Keefe, F.J. and Gil, K.M., Behavioral concepts in the analysis of
Derogatis, L.R., SCL-90-R: Administration, Scoring and Procedures chronic pain syndromes, J. Consult. Clin. Psychol.. 54 (1986)
Manual - II for the Revised Version, Clinical Psychometric Re- 776-7133.
search, Towson MD, 1983, Keel, P.J., Group therapy for chronic pain patients: Concept and
Dworkin, SF., LeResche, L. and DeRouen, T., Reliability of clinical first experiences. In: R. Roy and B. Tunks (Eds.), Chronic pain.
measurement in temporomandibular disorders, Clin. J. Pain, 4 psychological factors in rehabilitation, Williams and Williams,
(1988) 89-99. Baltimore. 1982,
Dworkin, SF., Huggins, K.H., LeResche, L., Von Korff, M., Howard, Lee. Y.J., Ellenberg, J.H., Hirtz, D.G. and Nelson, K.B., Analysis of
J., Truelove, E. and Sommers, E., Epidemiology of signs and clinical trials by treatment actually received: Is it really an op-
symptoms in temporomandibular disorders: clinical signs in cases tion? Stat. Med., 10 (1991) 1595-1605.
and controls, J. Am. Dent. Assoc., 120 (199Oa) 273-281. Marbach, J.J., Richlin, D.M. and Lipton, J.A., Illness behavior,
Dworkin, SF., LeResche, L., DeRouen, T. and Von Korff, M., depression, and anhedonia in myofascial fact and back pain
Assessing clinical signs of temporomandibular disorders: reliabil- patients, Psychother. Psychosom., 39 (1983) 47-54.
ity of clinical examiners, J. Prosthet. Dent., 63 (199Ob) 574-580. McCreary, C.P., Clark, G.T., Merril, V. and Oakley, MA., Psycho-
Dworkin, SF., LeResche, L. and Von Korff, M.R., Diagnostic stud- logical distress and diagnostic subgroups of temporomandibular
ies of temporomandibular disorders: challenges from an epidemi- patients, Pain, 44 (1991) 29-34.
ologic perspective, Anesth. Prog., 37 (1990~) 147-154. McCreary, C.P., Clark, G.T., Oakley, M.E. and Flack, V., Predicting
Dworkin, SF., Van Korff, M., Whitney, C.W., LeResche, L., Dicker, response to treatment for tem~~romandjbular disorders. J. Cran-
B.C. and Barlow, W., Measurement of characteristic pain in field iomandib. Disord. Facial Oral Pain, h(3) (1992) 161-169.
research, Pain, 55 (19%) 290. Moulton, R.E., Emotional factors in non-organic tem~romandjbu-
Dworkin, S.F., Von Korff, M.R. and LeResche, L., Multiple pains tar joint pain, Dent. Clin. North Am., Nov (1966) 609-624.
 187

Nash, J.M. and Holroyd, K.A., Home-based behavioral treatment for Medicine. A Cognitive Behavioral Perspective, Guilford Press,
recurrent headache: A cost-effective alternative? Am. Pain Soci- New York, 1983,
ety J., 2 (1992) l-6. Turk, DC. and Rudy, T.E., Toward an empirically derived taxonomy
Okeson, J., Kemper, J., Moody, P.M. and Haley, J.V., Evaluation of of chronic pain patients: Integration of psychological assessment
occlusal splint therapy and relaxation procedures in patients with data, J. Consult. Clin. Psychol., 56 (1988) 233-238.
temporomandibular disorders, J. Am. Dent. Assoc., 107 (1983) Turk, D.C. and Rudy, T.E., Neglected factors in chronic pain treat-
418-424. ment outcome studies--referral patterns, failure to enter treat-
Ostenveis, M., Kleinman, A. and Mechanic, D., Pain and disability: ment, and attrition, Pain, 43 (1990a) 7-25.
Clinical, behavioral and public policy perspectives, National Turk, DC. and Rudy, T.E., Neglected topics in the treatment of
Academy Press, Washington, D.C., 1987, chronic pain patients--relapse, noncompliance, and adherence
Peter, J., Large, R.G. and Elkind, G., Followup results from a enhancement, Pain, 43 (1990b) 7-26.
randomized controlled trial evaluating in- and out-patient pain Turner, J.A., Comparison of group progressive-relaxation training
management programs, Pain, 50 (1992) 41-50. and cognitive-behavioral group therapy for chronic low back pain,
Pocock, S.J., Clinical trials: A practical approach, Wiley, New York, J. Consult. Clin. Psychol., 50 (1982) 757-765.
1983, Turner, J.A., Clancy, S., McQuade, K.J. and Cardenas, D.D., Effec-
Pomp, A.M., Psychotherapy for the myofacial pain-dysfunction syn- tiveness of behavioral therapy for chronic low back pain: a
drome: A study of factors coinciding with symptom remission, J. component analysis, J. Consult. Clin. Psychol., 58 (1990) 573-579.
Am. Dent. Assoc., 89 (1974) 629-632. Turner, J.A. and Romano, J.M., Cognitive-behavioral therapy. In:
Richardson, G.M. and McGrath, P.J., Cognitive-behavioral therapy J.J. Bonica (Ed.), The Management of Pain, Lea and Febiger,
for migraine headaches: a minimal-therapist-contact approach Philadelphia, 1990, pp. 1711-1721.
versus a clinic-based approach, Headache, 29 (1989) 352-357. Von Korff, M., Dworkin, S.F., LeResche, L. and Kruger, A., An
Rudy, T.E., Turk, D.C., Zaki, H.S. and Curtin, H.D., An empirical epidemiologic comparison of pain complaints, Pain, 32 (1988)
taxometric alternative to traditional classification of temporo- 173-183.
mandibular disorders, Pain, 36 (1989) 311-320. Von Korff, M., Dworkin, S.F. and LeResche, L., Graded chronic
Rudy, T.E., Turk, D.C., Brena, SF. and Brody, M.C., Quantification pain status: An epidemiologic evaluation, Pain, 40 (1990) 279-291.
of biomedical findings of chronic pain patients: development of Von Korff, M., LeResche, L., Wilson, L. and Dworkin, SF., Depres-
an index of pathology, Pain, 42 (1990) 167-182. sion and risk of onset of chronic pain: A 3 year follow-up of a
Schnurr, R.F., Brooke, RI. and Rollman, G.B., Psychosocial corre- pain-free population sample, Society for Behavioral Medicine,
lates of temporomandibular joint pain and dysfunction, Pain, 42 12th Meeting #2B (1991) 17.
(1990) 153-166. Von Korff, M., Ormel, J., Keefe, F.J. and Dworkin, S.F., Grading the
Scott, D.S. and Gregg, J.M., Myofascial pain of the temporo- severity of chronic pain, Pain, 50 (1992) 133-149.
mandibular joint: A review of the behavioral-relaxation therapies, Von Korff, M., Deyo, R.A., Cherkin, D.C. and Barlow, W., Back
Pain, 9 (1980) 231-241. pain in primary care: Outcomes at one year, Spine, In Press
Social Security Administration,, Report of the commission on the (1993)
evaluation of pain, Social Security Bulletin, 50 (1987) 13-44. Whitney, C.W. and Von Korff, M., Regression to the mean in
Stam, H.J., McGrath, P.A. and Brooke, RI., The effects of a treated versus untreated chronic pain, Pain, 50 (1992) 281-285.
cognitive-behavioral treatment program on temporomandibular Wilson, L., Dworkin, S.F., LeResche, L., Whitney, C.W. and Dicker,
pain and dysfunction syndrome, Psychosom. Med., 46 (1984) B.G., Somatization and diffuseness of clinical pain symptoms, J.
534-545. Behav. Med., 12(Sp. 1~s.) #2C (1991) 17-18.
Turk, D.C., Meichenbaum, D. and Genest, M., Pain and Behavioral