Our Human Research Committee approved a prospective protocol for the
comparison of psychological factors in patients with discrete and idiopathic
pain complaints. Following a single initial routine office evaluation in the
practice of one of three hand surgeons (D.R., S.-G.P.L., or J.B.J.), patients
were classified into one of the three groups on the basis of the clinical
impression. The first group (the discrete pain cohort) included patients who
had a single, discrete, identifiable cause of upper limb pain for which all of
the symptoms and signs were consistent, with the treating physician being very
confident about the diagnosis. The second group (the idiopathic pain cohort)
included patients who had vague, diffuse complaints and inconsistent findings
on examination, with the physician being very uncertain about the cause of the
pain. The third group (the intermediate group) included patients who could not
be assigned to either of the other two groups. Patients in the discrete pain
and idiopathic pain cohorts rated the severity of their pain and completed the
same battery of five psychological questionnaires. Patients in the
intermediate group were excluded. Patients with previous operative treatment
were also excluded. One hundred and seven patients (fifty-six with discrete
pain and fifty-one with idiopathic pain) were enrolled in the study.
The present investigation focused on the initial impression of the treating
physician. The distinction of patients into these three groups was based on
the initial office visit alone, without regard to subsequent visits or
diagnostic tests. Patients in the discrete pain group were those in whom the
findings were specific enough for the physician to be extremely confident in
the diagnosis on the basis of the interview and examination alone. Patients in
the idiopathic pain group were those in whom the pain was so diffuse, vague,
and/or extreme that there was no single pathological or anatomical process
that could account for all of the symptoms or their severity. Patients with
idiopathic pain were thought to have such noncharacteristic findings that no
diagnostic tests would be able to explain the entire clinical presentation.
Any patient who did not fit into either of these two extreme categories was
considered intermediate and was excluded.
The diagnoses of the patients in the discrete pain cohort included
trapeziometacarpal osteoarthritis (eleven patients), de Quervain tenosynovitis
(thirteen), a single trigger digit (twenty-four), carpal tunnel syndrome
(four), and lateral epicondylitis (four).
The discrete pain cohort included twenty-two men (39%) and thirty-four
women, and the idiopathic pain cohort included fourteen men (27%) and
thirty-seven women (p = 0.20). The average age (and standard deviation) was 55
± 15 years (range, twenty-eight to eighty-two years) in the discrete
pain cohort and 41 ± 15 years (range, twenty-two to eighty-six years)
in the idiopathic pain cohort (p < 0.001).
In the discrete pain cohort, thirty-one patients worked full-time, five
worked part-time, three were homemakers, fourteen were retired, one was
unemployed but able to work, one was unemployed and unable to work, and one
reported his work status as "other." In the idiopathic pain
cohort, thirty patients worked full-time, seven worked part-time, two were
homemakers, five were retired, two were unemployed but able to work, two were
unemployed and unable to work, and three reported their work status as
"other." There were no significant differences between the two
groups with regard to employment status.
The idiopathic pain cohort had significantly more single individuals
(nineteen compared with nine, p = 0.02), whereas the discrete pain cohort had
significantly more married individuals (thirty-six compared with nineteen, p =
0.01). There was no significant difference between the discrete and idiopathic
pain cohorts with regard to the number of patients living with a partner
(three compared with six), the number of patients who were separated or
divorced (six compared with four), or the number of patients who were widowed
(two compared with three). The cohorts were comparable with respect to race.
The discrete pain cohort included fifty individuals who described their race
as white (not of Hispanic origin), two who described it as African-American,
three who described it as Asian or Pacific Islander, and one who described it
as American Indian or Alaskan Native. The idiopathic cohort included
forty-seven individuals who described their race as white, two who described
it as African-American, one who described it as Asian or Pacific Islander, and
one who described it as Hispanic. The discrete and idiopathic cohorts also
reported having similar levels of education (16.0 ± 3.0 years compared
with 15.7 ± 2.9 years).
Patients in the idiopathic pain cohort had seen an average of 1.8 ±
1.7 doctors before presenting to one of us, whereas the discrete pain cohort
had seen an average of 0.8 ± 0.7 doctors (p < 0.001). Nineteen
patients had been offered disputable diagnoses, including carpal tunnel
syndrome that was not verified by electrophysiological testing (seven
patients), repetitive strain injury (five), tendonitis (four), fibromyalgia
(three), and thoracic outlet syndrome (one), before presentation to us.
Psychiatric comorbidity (defined as a history of a psychiatric disorder or
current use of psychiatric medication) was present in seven (12.5%) of the
fifty-six patients in the discrete pain cohort and in fourteen (27.5%) of the
fifty-one patients in the idiopathic pain cohort (p = 0.10).
Measures
Pain Scale
Pain was assessed with use of 10-point Likert scales, with 1 point
representing no pain and 10 points representing the worst pain ever
experienced. Patients rated pain in several categories, including pain in
general, worst pain, pain at rest, pain when lifting a heavy object, pain with
repeated movements, and pain at night.
Pain Anxiety Symptoms Scale
The Pain Anxiety Symptoms Scale is a validated, forty-question inventory
designed to assess anxiety about
pain24. Its four
ten-item subscales are reliable measures of differentiable aspects of pain
anxiety: (1) cognitive anxiety, (2) fear of pain, (3) escape and avoidance,
and (4) physiological anxiety. The cognitive-anxiety subscale evaluates
symptoms such as the inability to concentrate and the frequency of unwanted
thoughts when the patient is in pain. The fear-of-pain subscale measures the
frequency of thoughts provoking fear and a profound dread of negative
consequences when the patient is in pain. The escape-and-avoidance subscale
rates the frequency of behaviors aimed at minimizing the severity and duration
of pain. Finally, the physiological-anxiety subscale measures the patient's
physical responses to pain, such as sweating or feeling dizzy. With use of a
6-point rating system (with 0 corresponding to "never" and 5
representing "always"), the possible score on each subscale ranges
from 0 to 50 points. In addition, a combined total Pain Anxiety Symptoms Scale
score (possible range, 0 to 200 points) has proved useful for the evaluation
of generalized pain
anxiety24. Based on
a series of 180 patients with pain who were seen in a multidisciplinary
setting, McCracken et al. reported a mean value (and standard deviation) of
26.02 ± 12.47 for the cognitive-anxiety subscale, 26.88 ± 9.88
for the escape-and-avoidance subscale, 18.70 ± 12.37 for the
physiological-anxiety subscale, and 94.24 ± 39.16 for the total Pain
Anxiety Symptoms
Scale24.
Pain Catastrophizing Scale
The Pain Catastrophizing Scale was used to determine the extent to which
the patients coped with their pain by catastrophizing (i.e., performing
cognitive activities that exacerbate the fearful aspects of the pain
experience)25. The
Pain Catastrophizing Scale is a validated scale that measures three components
of pain catastrophizing: rumination, magnification, and
helplessness25,26.
Rumination, the tendency to regurgitate thoughts about pain and to relive
painful experiences, is measured with a four-question subscale. Magnification,
the inclination to "make a mountain out of a molehill" or to
believe that pain represents an undiagnosed serious or worsening problem, is
measured with a three-question subscale. Finally, helplessness, an inability
to mitigate pain symptoms or a submission to pain, is assessed with use of a
six-question subscale. Questions are scored on a 5-point system, with 0 points
corresponding to "not at all" and 4 points corresponding to
"all of the time." The possible scores range from 0 to 16 points
on the rumination subscale, from 0 to 12 points on the magnification subscale,
and from 0 to 24 points on the helplessness subscale. In addition, a combined
total Pain Catastrophizing Scale score (possible range, 0 to 52) has proved to
be effective for the assessment of generalized coping
strategies25. Data
have been published for two populations: a community sample of 215
individuals, and a pain clinic outpatient sample of sixty
individuals26. The
mean values (and standard deviations) for the community sample were 5.89
± 4.15 points for rumination, 3.2 ± 2.45 points for
magnification, 4.78 ± 4.67 points for helplessness, and 13.87 ±
10.11 points for the total Pain Catastrophizing Scale. The mean values for the
pain clinic sample were appreciably higher: 9.02 ± 3.34 points for
rumination, 4.58 ± 2.54 points for magnification, 8.65 ± 5.2
points for helplessness, and 22.25 ± 10.16 points for the total Pain
Catastrophizing Scale.
Wahler Physical Symptom Inventory
The Wahler Physical Symptom Inventory is widely used to measure the
frequency of somatic complaints and has been shown to have acceptable internal
consistency and
validity27. For the
purposes of the present study, it was useful because its "score is based
exclusively on somatic complaints without an admixture of emotional and mood
symptoms such as worry, anxiety and
depression."27
To evaluate the occurrence and intensity of patient complaints of general
physical symptoms, the Wahler Physical Symptom Inventory uses a
forty-two-question format and a 6-point rating system, with 0 indicating that
the patient experiences the symptom "almost never" and 5
indicating that the patient experiences the symptom "nearly every
day." The final Wahler Physical Symptom Inventory score, ranging from
0.0 to 5.0, is an average of individual responses. Wahler reported norms based
on 111 male college students (0.56 ± 0.17), 134 female college students
(0.66 ± 0.15), 236 male psychiatric outpatients (0.99 ± 0.56),
and 317 female psychiatric outpatients (1.49 ±
0.69)27.
Body Consciousness Questionnaire
The Body Consciousness Questionnaire, a validated instrument demonstrating
satisfactory test-retest reliability, has three
subscales28. Of
special interest to us was the 5-question Private Body Consciousness subscale,
which assesses attention to internal physical sensations (e.g., dry mouth,
hunger, and body temperature). Questions are rated on a 6-point scale, with 0
representing an "extremely uncharacteristic" quality and 5
representing an "extremely characteristic" quality. The Private
Body Consciousness scores range from 0 to 25. Miller and Buss reported
gender-specific societal norms based on 275 men (11.7 ± 3.0) and 353
women (12 ±
3.3)28.
Multidimensional Health Locus of Control Scale
Locus of control was measured with the Multidimensional Health Locus of
Control Scale. This eighteen-item measure assesses the perception that the
forces determining one's health are (1) primarily internal, (2) under the
control of powerful others, or (3) a matter of
chance29. Patients
who score high on the "Internal Control" subscale demonstrate
personal responsibility for their health. Those who score high on the
"Powerful Others" subscale are likely to rely on others (e.g.,
their doctor) to control their health. Finally, those who score high on the
"Chance" subscale are not likely to rely on their own actions or
the actions of a doctor to cure their pain because they believe that their
health is a matter of chance. The Multidimensional Health Locus of Control
Scale is scored on a 6-point scale, with 1 representing "strongly
disagree" and 6 representing "strongly agree." Each
dimension consists of six questions with scores ranging from 6 to 36 points.
On the basis of data gathered from a normal population of individuals waiting
at a metropolitan airport, Wallston et al. reported mean values (and standard
deviations) of 25.1 ± 4.9 for the "Internal Control"
subscale, 20.0 ± 5.2 for the "Powerful Others" subscale,
and 15.6 ± 5.8 for the "Chance"
subscale29.
Statistical Analysis
A power analysis was performed on the basis of the Private Body
Consciousness subscale of the Body Consciousness Questionnaire. Previous
studies have suggested that the mean and the standard deviation of the Private
Body Consciousness subscale for the general population are approximately 12
and 3 points,
respectively28. To
detect a cohort difference of one-fourth of a standard deviation—judged
to be the smallest clinically important difference—at an alpha of 0.05
and a power of 80%, we estimated that a minimum of twenty-four patients would
be needed in each cohort.
Commercial software was used for analysis (Statistical Package for the
Social Sciences, Chicago, Illinois). Fisher's exact test was used to compare
dichotomous variables. The Student t test and the Mann-Whitney test were used
to compare continuous variables. A multiple stepwise logistic regression model
using the backward elimination algorithm was used to account for potential
confounding among the various psychological measures. Factors associated with
differences between cohorts that were associated with a significance level of
p = 0.075 during univariate analysis were included as covariates in the
model.
Univariate Analyses
Patients with idiopathic pain had more extreme complaints about pain
than patients with discrete pain did (Table
I). Patients with idiopathic pain had significantly more pain with
repeated movements (mean score [and standard deviation], 6.8 ± 2.6
compared with 5.2 ± 2.8; p = 0.01) and pain at rest (4.1 ± 2.6
compared with 3.0 ± 2.4; p = 0.02) than patients with discrete pain
did. The differences between the two groups with regard to pain at night (4.9
± 2.8 compared with 3.9 ± 2.8; p = 0.06) and pain in general
(5.3 ± 2.4 compared with 4.4 ± 2.5; p = 0.10) showed trends
toward significance. Patients in both groups had similar scores for the
severity of pain when lifting a heavy object (5.9 ± 2.8 compared with
5.4 ± 3.2; p = 0.34) and for the severity of the worst pain (7.3
± 2.1 compared with 6.3 ± 2.7; p = 0.48).
Patients with idiopathic pain had greater anxiety about their pain than did
those with discrete pain (Table
II). According to the Pain Anxiety Symptoms Scale, patients with
idiopathic pain had significantly greater cognitive anxiety (21.7 ±
10.1 compared with 16.8 ± 8.7; p = 0.008) and had nearly significantly
greater fear of pain (14.4 ± 8.1 compared with 11.8 ± 6.8; p =
0.07) and overall pain anxiety (69.2 ± 29.0 compared with 59.3 ±
25.9; p = 0.07) than did those with discrete pain. The differences between the
groups with regard to physiological anxiety (11.1 ± 8.4 compared with
9.8 ± 8.2, p = 0.43) and pain-escaping behavior (22.2 ± 7.5
compared with 21.0 ± 8.7; p = 0.47) were not significant.
Patients with idiopathic pain had less-adaptive personality traits than did
those with discrete pain (Table
III). Patients with idiopathic pain demonstrated significantly
greater helplessness than did patients with discrete pain (8.2 ± 5.5
compared with 5.1 ± 4.6; p = 0.002). Patients with idiopathic pain also
had significantly higher scores on the pain-magnification subscale (4.5
± 3.3 compared with 2.8 ± 2.7; p = 0.007) and the total combined
Pain Catastrophizing Scale (20.4 ± 11.7 compared with 14.0 ±
11.3, p = 0.005). The difference between the groups with regard to the
rumination subscale score was nearly significant (7.7 ± 4.5 compared
with 6.1 ± 4.9, p = 0.08). There was nearly a significant difference
between the patients with idiopathic pain and those with discrete pain in
terms of somatic complaining as measured with the Wahler Physical Symptom
Inventory (1.11 ± 0.56 compared with 0.93 ± 0.49; p = 0.07).
There was no significant difference between the idiopathic pain cohort and
the discrete pain cohort with regard to the average score on the Private Body
Consciousness subscale (14.1 ± 5.0 compared with 14.7 ± 5.8; p =
0.55). With regard to the Multidimensional Health Locus of Control Scale, the
idiopathic pain cohort scored higher on the "Internal Control"
subscale (25.1 ± 4.8 compared with 23.8 ± 4.5; p = 0.17) and the
"Chance" subscale (16.4 ± 4.9 compared with 14.9 ±
4.8, p = 0.11), whereas the discrete pain cohort scored higher on the
"Powerful Others" subscale (17.9 ± 5.4 compared with 16.0
± 5.8; p = 0.08). All of these differences failed to reach
significance.
Multivariate Analysis
Eight covariates met the criteria for inclusion in the logistic regression
model: the overall Pain Anxiety Symptoms Scale score, the cognitive-anxiety
subscale score, the fear-of-pain subscale score, the combined total Pain
Catastrophizing Scale score, the Pain Catastrophizing Scale magnification
score, the Pain Catastrophizing Scale rumination score, the Pain
Catastrophizing Scale helplessness score, and the Wahler Physical Symptom
Inventory score. The regression model identified the total Pain
Catastrophizing Scale score as the sole predictor of the cohort to which a
patient would belong (p = 0.008). This model was used to determine a patient's
probability of belonging to each cohort. A total Pain Catastrophizing Scale
score of 0 corresponded with a 25% probability that the patient would be in
the idiopathic pain cohort, a score of 10 corresponded with a 40% probability,
a score of 20 corresponded with a 50% probability, a score of 30 corresponded
with a 65% probability, and a score of 40 corresponded with a 75%
probability.