37
Pain, 43 (1990) 37-46 Elsevier
PAIN 01646
Litigation and employment status: effects on patients with chronic pain Raymond
C. Tait, John T. Chibnali
and William D. Richardson
Division of Behavioral Medicine, St. Louis University School of Medicine, St. Louis, MO 63104 (U.S.A.) (Received
4 October
1989, revision
received 28 March 1990, accepted
6 April 1990)
summary
In order to study the effects of compensation and litigation, 201 chronic pain patients were selected from a sample of 464: 99 were working, 15 were working and litigating, 53 were receiving Worker’s Compensation, and 34 were receiving Worker’s Compensation and litigating. Employment (working vs. Worker’s Compensation) and litigation status (litigating vs. not litigating) were analyzed in a 2 x 2 factorial design with measures of pain, disability, psycholo~c~ distress, and selected derno~ap~~ as dependent variables. Compared to Worker’s Compensation patients, working patients reported significantly less disability (down-time, days spent in bed, interference of pain in daily activities) and pain of a longer duration. Compared to litigating patients, non-litigating patients reported less pain (on the McGill Pain Questionnaire) and less disability (stopping activity, interference of pain in daily activities). On two measures of psychological distress (depression, anxiety), there were significant interactions: Worker’s Compensation patients who were litigating reported less distress than non-litigants, while working patients who were litigating reported more distress than non-litigants. The results indicate clear differences in self-reports of disability associated with both employment and litigation status. They also suggest that litigation may function as a coping response for patients who are distressed by the adversarial nature of the Workers’ Compensation system. Limitations of the study as well as suggestions for further research also are discussed.
Key words.
Chronic
pain;
Litigation;
Employment
status
Introduction
Financial incentives have long been thought to have significant and negative effects on both symptom complaints [5,20] and response to treatment [4,9,10,31]. As early as 1945 the term “compensation neurosis” was coined to describe the psychological effects thought to be associated with compensation [14]: “A state of mind, born out of
Correspondence to: Dr. R.C. Tait, Division of Behavioral Medicine, St. Louis University School of Medicine, 1221 S. Grand Blvd., St. Louis, MO 63104, U.S.A. 03~-3959/~/$03.50
@ 1990 Elsevier Science Publishers
fear, kept alive by avarice, stimulated by lawyers, and cured by a verdict.” While the belief in a compensation (or accident or litigation) neurosis is widespread, there is increasing evidence, especially in regard to chronic pain patients, that its effects may be overstated or, at least, over-simplified [21]. Several studies have found no significant differences in pain severity or psycholo~cal distress related to compensation status [15,18,19,28], while others have revealed no changes in clinical status once compensation had been resolved [2,13]. It has been suggested that instead of clarifying the effects of compensation, the confusing state of the literature probably serves
B.V. (Biomedical
Division)
3x
only to enhance the biases of those who read it 122.291. Several problems have been identified in the pain literature that may have contributed to the divergent findings. One involves studies where employment and compensation effects may have been confounded. Recent research comparing compensation and non-compensation patients found that effects of compensation status on treatment outcomes were mediated by employment status; when differences in employment status were controlled, there were no differences in outcomes between patients receiving Worker’s Compensation and those not receiving it [Xl. Another problem involves failure to adequately differentiate patient groups. For example, Jamison et al. [ll] observed that compensation effects vary, depending on whether compensation is timelimited or not. Similarly, it has been noted that the compensation system is heterogeneous, including such diverse groups as veterans, Social Security Disability recipients, Worker’s Compensation benefits recipients, and litigants [2,9,23]. Failure to adequately specify group composition may have added unwanted variability to research effects [28]. Finally, it has been observed that much of the research has addressed only pain-related and psychological variables. Broadening the focus to include other variables has been suggested in order to facilitate the development of a more comprehensive understanding of the effects of compensation systems on patients involved in them [28,32]. The present study attempted to investigate compensation and litigation effects in ways that address some of these methodolo~c problems. To broaden the scope of the study and to maximize its construct validity, dependent variables were included that covered several dimensions previously found useful in studies of chronic pain patients [26]: pain characteristics, pain history, psychological distress, and pain-related disability. The study also used a factorial design in order to separate the effects of compensation and litigation status. Patients were categorized based on both their current employment and litigation status. Thus, patients were either working at their usual job, or were not working and collecting Workers Compensation benefits. Also, they were
either actively involved in litigation seeking compensation for their pain, or were not involved in pain-related legal proceedings. The compensation patients in this study (i.e.. those who were not working) were covered by Worker’s Compensation health insurance, a United States government program which insures workers against injuries sustained on the job. While this system is mandated by the federal government. each state is free to establish guidelines that govern the actual administration of the compensation program. Although there are variations in state laws, certain benefits typically apply: (1) injured workers are entitled to payments equivalent to a percentage of their wage (percentages and time periods vary across states); (2) they are entitled to medical evaluation and treatment for on-the-job injuries; (3) if their injuries prevent them from returning to the job they held when injured, workers are entitled to vocational rehabilitations (4) if the injured worker remains disabled, he/she can apply for Social Security Disability benefits. In practice, several difficulties can arise from this process. Most notably, many injured workers whose primary complaint is pain are unable to regain function in a “timely” way; others report symptoms that exceed objectively verifiable physical evidence. Such events may open the door to an adversarial relationship between the injured worker (who wants benefits or continued benefits) and the insurance carrier (which demands verifiable, objective evidence of injury). The result may be the retention of attorneys and the beginning of litigation. Thus, the compensation patients in this study were not working, but were receiving benefits from the Worker’s Compensation system. Some of the compensation patients had retained attorneys (compensation-litigation group), while others had not retained attorneys and were not in litigation (compensation-only group). The other two study groups were composed of working patients. These patients were currently working full time at some occupation, and drawing a salary from that activity. Of the patients in this group, those who were not litigating either had never been involved in litigation, or had resolved all legal activity prior to participating in
39
the study. Those who were working and litigating were a small, uncommon group. These patients, although functional enough to carry out their occupational duties every day, also had initiated litigation against some party, such as an employer, an insurance company, or an individual (e.g., someone involved in a traffic accident may sue the other driver). The purpose of the present study, then, was to look for differences in pain characteristics, psychological distress, pain history, and pain-related disability among these four groups. Based on recent research [15,18,19,28] neither litigation nor employment status was expected to be associated with pain or psychological distress, although both were expected to have effects on patient reports of disability [6]. No interactions were expected.
Methods Subjects Subjects were 201 chronic pain patients who presented for evaluation at a multi-disciplinary treatment center located in a large university medical center. These subjects were selected from a total of 444 consecutive patients seen over a 30-month period (January, 1985 to June, 1987). The 201 patients were chosen because they fit into categories in the 2 x 2 factorial design. Of the 201, 15 were working and litigating, 99 were working but not litigating, 34 were receiving Worker’s Compensation and litigating, and 53 were receiving Worker’s Compensation but were not litigating. Of the 243 patients not in the study, 53 were receiving Social Security Disability Benefits, 37 were retired and receiving Social Security payments, 41 were not working or receiving pain-related benefits but were actively litigating, and 101 were not working, receiving benefits, or litigating (missing data: n = 11). The subjects in the study reported a mean age of 39.8 years (S.D. = 9.85) and a mean level of education of 12.1 years (SD. = 2.55). One-hundred twenty-one (60.8%) were males and 78 (39.2%) were females; 74.4% were married, 11.8% were single, 13.4% were separated or divorced, and 0.5% were widowed. With regard to circumstances of
pain onset, 66.2% reported that a work-related accident led to the pain condition, 11.8% indicated that some other type of accident (e.g., auto) caused their pain, and for 4.6% pain began following an illness or surgery. For nearly 15% of the sample (14.9%) there was no clear precipitant for the pain (i.e., it “just began”) and 2.6% reported that “other unspecified events” led to the pain problem. On a 5-point Likert-type scale (0 = mild, 5 = excruciating), patient ratings of usual pain level had a mean of 2.9 (S.D. = 0.74) and duration of the pain problem averaged 46.9 months (S.D. = 93.0). Pain distribution for the study subjects was somewhat varied: 5 (2.6%) reported ventral pain; 12 (6.3%) reported extremity pain (upper or lower); 6 (3.2%) reported upper quadrant pain (head, neck, shoulder); 10 (5.3%) reported low back pain; 20 (10.5%) reported upper quadrant and extremity pain; 65 (34.2%) reported low back and extremity pain; 5 (2.6%) reported upper quadrant and low back pain; 67 (35.3%) reported upper quadrant, low back, and extremity pain (missing data = 11). The majority of patients were diagnosed as having pain of a musculoskeletal nature, although there were some cases in which other factors also were diagnosed (e.g., neurovascular compression syndrome, post-traumatic headache). Comparisons were made between the 201 study patients and the other 243 patients on several variables. T tests revealed that the study patients were younger (M = 39.8 vs. M = 46.3) t (422) = 5.1, P < 0.001; were more educated (M = 12.1 vs. M = 11.6) t (435) = - 2.0, P < 0.05; reported lower usual pain intensity (M = 2.9 vs. M = 3.2), t (399) = 3.3, P < 0.01; and reported pain of a shorter duration (M = 46.9 vs. M = 68.7) t (433) = 3.8, P < 0.01. As can be seen in Table I, chisquare tests also revealed that study patients were more likely to be male, less likely to be widowed, and more likely to be married than non-study patients. Circumstances of pain onset also differed, largely because the study patients were more likely to report pain related to a work injury. Demographic differences such as those reported above are not surprising in light of the composition of the study group. The vast majority of these patients were of working age (only 1.5% were 65 or older). Of the non-study patients, how-
40
TABLE
I
DIFFERENCES PATIENTS
BETWEEN
Variable
STUDY
AND
NON-STUDY
Study group
Non-study
group
n
48
n
%
Gender a Male Female
121 78
60.8 39.2
102 141
42.0 58.0
Marital status b Single Married Separated Divorced Widowed
23 145 5 21 1
11.8 74.4 2.6 10.8 0.S
31 153 9 22 13
13.6 67.1 3.9 9.6 5.7
Circumstances of onset ’ Work accident Other accident Illness/surgery ‘Just began’ Other events
129 23 9 29 5
66.2 11.8 4.6 14.9 2.6
63 58 46 50 19
26.7 24.6 19.5 21.2 8.1
a XL (1) =15.5, P i 0.001. h x2 (4) = 10.3, P < 0.05. ’ x2 (4) = 73.2, P i 0.001.
ever, a substantial number were 65 or older (16.5%). Thus, it was not surprising for this group to have pain of a longer duration and to have a higher percentage of females and widows. Also, some of these patients had been declared legally disabled, while many were neither working nor receiving benefits for their pain. Materials All subjects completed a pre-evaluation assessment battery containing the following measures: Pain Disability Index (PDI) [25]; Quality of Life Scale (QOLS) [7]; McGill Pain Questionnaire (MPQ) [17], from which one score was derived the sum of the psychophysically scaled adjectives; Beck Depression Inventory, short form (BDI) [3]; Psychosomatic Symptom Checklist (PSC) [l]; Spielberger Trait Anxiety Inventory (STAI) [24]; and a pain drawing which yielded a measure of percentage of painful body surface (PBS) [16]. Subjects also rated their worst (WORST), least (LEAST), and usual (USUAL) pain levels on a
5-point Likert scale (1 = mild; 2 = discomforting; 3 = severe; 4 = horrible; 5 = excruciating), and the frequency with which they stopped activity (STOP), lay down (DOWN), and stayed in bed (IN BED) due to pain. The activity ratings were made on 5-point scales that ranged from 1 (16 or more times/day) to 5 (0 times/day) for STOP and DOWN, and from 1 (7 days/week) to 5 (0 days/week) for IN BED. In a demographic section. subjects provided other relevant information (e.g., education, age, surgical history) as well as the number of months they had been in pain (DURATION). Design The design of the study was a between-subjects 2 x 2 factorial, with employment status as one independent variable (working vs. Worker’s Compensation) and litigation status as the other independent variable (litigating vs. not litigating). Fourteen dependent variables were analyzed separately using 2-way univariate analyses of variance. The univariate approach was necessary due to missing data and small cell sizes. The regression approach was used to calculate sums of squares for the analyses in order to accommodate unequal and disproportional cell sizes. To correct for violations of homogeneity of variance 2 variables were transformed: DURATION values were transformed to base 10 logarithms to remove positive skewness, and WORST pain values were exponentially transformed to remove negative skewness.
Results Comparisons on demographic variables Table II presents demographic data for patients broken down by employment and litigation status. No effects for age were found. Educational level, however, differed significantly for the working and compensation groups, with working patients reporting higher levels of education, F (1, 196) = 9.9, P < 0.01. A similar main effect was found for pain duration, with working patients reporting pain of a longer duration than those receiving worker’s compensation, F (1, 192) = 4.09, P ( 0.05.
41 TABLE
II
EMPLOYMENT
AND LITIGATION
GROUPS:
DEMOGRAPHICS Workers Compensation
Working
Group
Variable
M
SD.
N
M
SD.
N
14 96
39.3 38.7
11.54 9.41
32 52
Age
Litigating Non-litigating
36.9 41.1
5.87 9.88
Education ’
Litigating Non-litigating
12.1 12.4
1.36 2.62
15 98
11.2 11.2
2.68 2.08
34 53
Pain duration
Litigating Non-litigating
35.1 68.4
33.86 125.89
14 96
30.0 21.8
38.35 23.90
33 53
Pain duration (log 10) b
Litigating Non-litigating
1.3 1.4
0.46 0.57
14 96
1.3 1.2
0.40 0.41
33 53
’ Employment effect, P < 0.01. b Employment effect, P i 0.05 (log transformation required).
There was also a marginally significant difference for gender distribution among the experimental groups (x2 (3) = 6.86, P < 0.10). The largest difference was associated with the Worker’s Compensation/non-litigating group, which was composed of more males (75%) and fewer females (25X) than the other groups (working/litigating: male = 50%, female = 50%; working/non-litigating: male = 558, female = 45%; Workers Compensation/litigating: male = 59% female = 41%). Finally, the groups were compared on circumstances of pain onset. This comparison also functioned as a check on the composition of the study groups. Table III shows that circumstances
TABLE
to expectations: (1) of pain onset conformed working/litigating patients reported pain onset following an accident either at work or elsewhere; (2) working/non-litigating patients reported very mixed circumstances of onset; (3) all worker’s compensation patients reported pain following a work-related accident. While the distributions precluded a &i-square analysis, the differences are obvious. Comparisons on pain characteristics variables Table IV shows the effects of employment and litigation status on pain characteristics. The table shows very few differences on any of the variables
III
CIRCUMSTANCES Group
OF PAIN ONSET BY WORK STATUS AND LITIGATION
STATUS
Circumstances of onset Accident at work
Other accident
Surgery or illness
Workinglitigating
11 (73.3%)
4 (26.7%)
0
Workingnon-litigating
34 (35.4%)
19 (19.8%)
9 (9.4%)
Worker’s Comp. litigating
::OO.OS)
Worker’s Comp. non-litigating
51 (100.0%)
No known cause 0
f390.28)
‘ Other’ events 0
5 (5.2%)
0
0
0
0
0
0
0
0
42 TABLE
IV
EMPLOYMENT Variable
AND
LITIGATION
STATUS:
Group
EFFECTS
ON PAIN CHARACTERISTICS
Working
Worker’s
M
S.D.
N
M
Compensation
..__~~~
S.D.
N
WORST
Litigating Non-litigating
4.60 4.37
0.63 0.90
15 89
4.73 4.64
0.45 0 52
LEAST
Litigating Non-litigating
2.13 1.80
0.83 0.73
15 X8
2.24 2.0X
0.79 O.XX
USUAL
Litigating Non-litigating
3.13 2.16
0.64 0.71
15 89
3.12 3.04
0.x2 0.73
PBS
Litigating Non-litigating
26.08 23.50
19.42 17.16
15 90
30.60 24.38
16.52 18.43
MPQ a
Litigating Non-litigating
38.36 32.87
16.00 13.22
14 91
38.26 33.31
1I .92
a Litigation effect, P < 0.05. Nnte: Values for WORST were transformed
exponentially
for analysis.
measuring pain intensity. The one univariate main effect that did emerge involved the MPQ: litigating patients used more intense adjectives to describe pain than non-litigating patients, F (1, 181) = 4.2, P -c 0.05. In light of the fact that only 1 of the 5 univariate analyses showed between-group differences, the differences that did emerge on the MPQ should be viewed with appropriate caution.
TABLE
The untransformed
15.57
values are shown
here.
Comparisons on disability variables
In contrast to the lack of differences in pain characteristics associated with employment or litigation status, a number of main effects were found with disability variables. These are shown in Table V. Significant litigation effects were found for the PDI, F (1, 185) = 7.7, P -c 0.01, and for STOP, F (1, 185) = 6.5, P < 0.05, both variables reflecting
V
EMPLOYMENT
AND
LITIGATION
STATUS:
EFFECTS
ON DISABILITY Worker’s
Compensation
Variable
Group
Working M
SD.
N
M
S.D.
N
pDI
Litigating Non-litigating
44.20 36.94
14.66 13.24
15 91
51.48 46.72
10.28 10.16
33 50
Litigating Non-litigating
2.14 2.93
1.35 1.27
14 94
2.09 2.44
1.oo
1.20
32 50
hb
Litigating Non-litigating
3.87 3.91
1.00 1.08
15 93
3.29 3.38
1.13 1.13
31 53
IN BED b
Litigating Non-litigating
4.20 4.23
0.86 1.16
15 91
3.63 3.92
0.93 0.98
30 53
ra.bbb
STOP a
DOWN
_.____
a Litigation effect, P < 0.05. aa Litigation effect, P < 0.01. ’ Employment effect, P < 0.05. ” Employment effect, P < 0.01. bbb Employment effect, P -C0.001. Noret For STOP, DOWN, and IN BED, lower scores reflect greater
disability.
43
pain-related interference with activity. Interestingly, no differences emerged in measures of down-time or days spent in bed. Also shown in Table V, the effects of employment status cut across 3 of the 4 disability variables: PDI, DOWN, and IN BED. Surprisingly, working patients did not report more stoppage of activity than compensation patients. Nonetheless, the differences that did emerge suggested that the compensation patients were more disabled than their working counterparts, with greater interference of pain in daily activities, F (1,185) = 15.5, P < 0.001, more down-time, F (1, 188) = 8.8, P < 0.01, and more days spent in bed, F (1, 185) = 5.3, P < 0.05.
\\ 11/‘>
NonLitigating
----
Litigating
\\
\\
‘\
\\
10
91
Vv’orking
Worker’s Compensation
Fig. 1. Employment and litigation status: effects on depression. Scores on the BDI can range from 0 to 39.
Comparisons on psychological distress variables As can be seen in Table VI, there were no main effects for litigation or employment status on the variables measuring psychological distress. There were, however, several significant employment by litigation interactions involving variables designed to measure depression and anxiety. The interactions were broken down further into simple effects. As can be seen in Fig. 1, the interaction is in cross-over form for the BDI. Patients who were working and litigating were more depressed than patients who were working but not litigating, F (1, 179) = 3.94, P < 0.05. Among patients receiving Worker’s Compensa-
TABLE
-
‘\ /
tion, however, there was a trend for those who were litigating to be less depressed than those who were not, F (1, 179) = 2.94, P < 0.10. A similar pattern was obtained for the STAI (see Fig. 2). Among patients who were working, those who were litigating reported more anxiety than those who were not, F (1, 156) = 4.15, P < 0.05. Among patients who were receiving Worker’s Compensation, however, litigating patients reported less anxiety than non-litigating patients, F (1, 156) = 3.88, P < 0.05. As above, however, this comparison produced a somewhat weaker effect.
VI
EMPLOYMENT
AND LITIGATION
STATUS:
EFFECTS
ON PSYCHOLOGICAL
Working
DISTRESS
Variable
Group
Worker’s N
M
Beck a
Litigating Non-litigating
12.45 10.03
6.03 5.50
15 92
10.10 12.41
1.63 7.09
30 46
Litigating Non-litigating
53.38 46.95
7.24 10.27
13 77
45.89 50.69
9.91 11.29
28 42
QOLS
Litigating Non-litigating
25.29 27.24
6.00 8.54
14 90
23.15 24.10
9.38 9.02
33 49
PSC
Litigating Non-litigating
66.19 59.12
40.29 32.26
14 88
71.47 67.88
39.08 33.21
30 49
M
Spielberger
b
a Interaction: F (1, 179) = 5.1, P < 0.05. b Interaction: F (1,156) = 8.0, P < 0.01. Noret For the QGLS, lower scores reflect lower quality
S.D.
of life.
Compensation S.D.
N
-
NonLitigoting
----
Litigating
AA 1
Working
Worker’s Compensation
Fig. 2. Employment and litigation status: effects on anxiety. Scores on the STAI can range from 20 to 80.
Discussion Results relevant to pain characteristics and disability levels were generally consistent with prior research. There were few differences between groups on pain characteristics, a common finding [15,18,19,28]. The between-group differences on disability measures, involving both the compensation and litigation variables, also is generally consistent with prior research [5,23]. The absence of main effects for the psychological distress variables also has been reported previously [18,19,28]. The cross-over interactions that emerged for measures of depression and anxiety, however, have not been reported previously (probably because they have not been examined in previous studies). These interactions showed the effects of litigation to be mediated by work status; litigating Worker’s Compensation patients reported less distress than non-litigants, while distress levels were higher among litigating, working patients than among non-litigants. While any interpretation of this pattern is speculative, it is interesting to speculate on factors that might have occasioned the pattern of findings that emerged. Among Worker’s Compensation patients, the decision to retain an attorney clearly was associated with reports of less distress, suggesting that the decision to hire an attorney was an effective coping response to the stresses that
they faced [12,18]. This contrasts with the distress experienced by working patients who retained attorneys; levels of distress among this group were higher than among the non-litigants. Why might litigation be an effective response for the latter patients but not for the former? Both working and Worker’s Compensation patients face stresses associated with ongoing pain and dysfunction. Worker’s Compensation patients, however, face the additional stress of involvement in an adversarial medical-legal system 1181; the decision to hire an attorney is likely to be an effective means of addressing the adversarial stresses of that situation [23], although the longterm impact of such a decision may be deleterious to rehabilitation [21]. By contrast, the working patient faces primarily the stresses of pain and dysfunction. If these patients experience significant pain-related disability (as the data in this study suggest), they are likely to be facing the stress of meeting high functional demands (e.g., in the workplace) with impaired physical capabilities. Hiring an attorney does not resolve this stressor: in fact, it may be an act driven by the realization that work demands will exceed physical capabilities for the foreseeable future. Thus, the decision to litigate does not resolve immediate stresses, although it may provide long-term help. In order to explore whether there was any clinical evidence to support the above speculation. several patients involved in Worker’s Compensation proceedings were interviewed regarding their decisions to retain attorneys. Each patient indicated that the Worker’s Compensation proceedings had been adversarial. For each, the decision to hire an attorney was prompted by a concern that the adversarial proceedings would work against an appropriate treatment and/or settlement. These interviews clearly were consistent with the data. Parenthetically, we should note that the majority of patients interviewed also indicated some regret that attorneys had got involved: the attorneys were perceived as actually delaying the settlement of their cases. While it is tempting to stop with the above speculations, we should note that other explanations can be supplied. For example, patients who were working and litigating may have presented
45
themselves as more distressed than non-litigating patients in order to dramatize their misery. Similarly, among Worker’s Compensation patients, those involved in litigation may have downplayed their emotional distress in order to avoid being seen as symptom magnifiers, while the non-litigants might not have faced that contingency. Clearly, further research is needed to address this unexpected finding regarding psychological distress. There also are methodologic shortcomings to the study. First, it relied on self-report data; these should be complemented by medical and/or behavioral data to get a more adequate assessment of patients with chronic pain [21]. Second, this was a between-groups design. This design presumes that litigation and/or work status is static when, in actuality, the decision to litigate or not probably is much more fluid. To obtain information more specifically relevant to the effects of litigation on patients, a within-subjects design would be preferable (although not without its own shortcomings), because each patient then would serve as his own control. The study also was limited by the small number of working/litigating patients, a consequence of the very small percentage (3%) of such patients who presented for treatment during the two years of data collection. Thus, these results should be viewed with caution until further research on this group can be conducted. Also, considerations of sample size and missing data forced the use of univariate analyses of variance. Such analyses can yield results tending to over-estimate group differences. Similarly, we should note that this is a correlational study where strict experimental procedures were impossible, so cause-and-effect conclusions cannot be drawn. Thus, there may be rival explanations of group differences other than employment and litigation status. For example, significant differences in gender distribution, circumstances of pain onset, and years of education were found among the groups. These differences could have mediated the results. Similarly, other unassessed variables also could have contributed to the group differences that were found (e.g., socioeconomic status, type of occupation).
Finally, we should note that the patients in this study were drawn from a sample of patients who were evaluated at a comprehensive pain center. They had not responded well to treatment provided by their primary physicians. Previous research has suggested that patients not managed successfully by their physicians may exhibit more depression and illness behavior than those patients who have been successfully managed [6]. Extrapolating from this study to patients not seen in comprehensive pain centers should be done with caution, and future research should examine these successfully managed patients. Despite its limitations, the study does have value. First, it points out the usefulness of multidimensional assessment in comparisons of chronic pain patients. Second, the data suggest that the effects of litigation and compensation are not so straightforward as might be expected. While there were main effects relevant to disability, the measures of psychological distress followed a more complex pattern, suggesting that the decision to litigate may have been a coping response to the stresses of Worker’s Compensation proceedings. Further research investigating this phenomenon is needed in order to better understand factors that might be involved in these proceedings that militate against chronic pain patients who are served by it. References Attanasio, V., Andrasik, F., Blanchard, E.B. and Arena, J.G., Psychometric properties of the SUNYA revision of the Psychosomatic Symptom Checklist, J. Behav. Med., 1 (1984) 247-257. 2 Balla, J.I. and Moraitis, S., Knights in armour: a follow-up study of injuries after legal settlement, Med. J. Aust., 2 (1970) 355-361. 3 Beck, A.T. and Beck, R.W., Screening depressed patients in family practice: rapid technique, Postgrad. Med., 52 (1972) 1
81-85.
Block, A.R., Kremer, E. and Gaylor, M., Behavioral treatment of chronic pain: variables affecting treatment efficacy, Pain, 8 (1980) 367-375. 5 Brena, S.F., Chapman, S.L., Stegall, P.G. and Chyatte, S.B., Chronic pain states: their relationship to impairment and disability, Arch. Phys. Med. Rehab., 60 (1979) 387-389. 6 Chapman, C.R., Sola, A.E. and Bonica, J.J., Illness behavior and depression compared in pain center and private
4
practice
patients,
Pain, 6 (1979) l-7.
46 7 Chibnall,
20 Miller,
8
21
9
10
11
12
13
14 I5
16
J.T. and Tait, R.C.. The Quality of Life Scale: a preliminary study, Submitted for review. Dworkin, R.H., Handlin, D.S., Richlin. D.M., Brand, L. and Vannucci, C., Unraveling the effects of compensation, litigation, and employment on treatment response in chronic pain, Pain, 23 (1985) 49-59. Fowler. D.R. and Mayfield, D.G., Effect of disability compensation: disability symptoms and motivation for treatment, Arch. Environ. Hlth, 19 (1969) 719-725. Hammonds, W., Brena, SF. and Unikel. I.P., Compensation for work-related injuries and rehabilitation of patients with chronic pain. Sth. Med. J. (Bgham. AL). 71 (1978) 664-666. Jamison, R.N., Matt, D.A. and Parris, W.C.V., Effects of time-limited vs. unlimited compensation on pain behavior and treatment outcome in low back pain patients, J. Psychosom. Res., 32 (1988) 277-283. Keefe, F.J., Caldwell, D.S., Queen, K.T.. Gil, K.M., Martinez, S., Crisson, J.E., Ogden, W. and Nunley. J.. Pain coping strategies in osteoarthritis patients. J. Cons. Clin. Psychoi., 55 (1987) 208232. Kelly, R. and Smith, B.N., Post-traumatic syndrome: another myth discredited, J. Roy. Sot. Med., 74 (1981) 275-277. Kennedy. F.. The mind of the injured worker: ita effect on disability periods, Compens. Med., 1 (1946) 19-24. Leavitt, F., Garron, D.C., McNeill. T.W. and Whisler, W.W., Organic status, psychological disturbance, and pain report characteristics in low-back-pain patients on compensation, Spine, 7 (1982) 398-402. Margolis, R.B., Tait, R.C. and Krause. S.J., A rating system for use with patient pain drawings, Pain, 24 (1986) 57-65. Melzack, R., The McGill Pain Questionnaire: major properties and scoring methods. Pain, 1 (1975) 277-299. Meizack, R., Katz, J. and Jeans. M.E.. The role of compensation in chronic pain: analysis using a new’ method of scoring the McGill Pain Questionnaire. Pain. 23 (1985) 101-112. Mendelson, G.. Compensation, pain complaints. and psychological disturbance, Pain, 20 (1984) 169- 177.
‘2 23
24
H.. Accident neurosis I. Br. Med. J., 1 (1961) 919-925. Osterweis, M., Kleinman, A. and Mechanic. D. (Eds.), Pain dnd Disability: Clinical, Behavioral, and Public Policy Prospectives, National Academy Press, Washington, DC. 7987. Parker. N., Accident neurosis. Med. J. Amt.. 2 (1970) 362-365. Peck. C.J.. Fordyce, W.E. and Black, R.G., The effect of the pendency of claims for compensation upon behavior indicative of pain, Wash. Law Rev., 53 (1978) 251 278. Spielberger, CD.. Goruch, R.L. and Lushene. R.E., STAI; Manual for the State-Trait Anxiety Inventory, Consulting Psychologists Press, Palo Alto. CA, 1970. Tait, R.C., Pollard, C.A.. Margolis, R.B., Duckro, P.N. and Krause, S.J., The Pain Disability Index: psychometric and validity data. Arch. Phys. Med. Rehab., 68 (1987) 438441. ‘Tait, R.C., Chibnall, J.. Duckro. P.N. and Deshields, T.I.... Dimensions of chronic pain: a factor analytic study. C‘lin. J. Pain. 5 (1989) 3233328. fait, R.C.. Duckro. P.N., Margoiis, R.B. and Wiener, R.. Quality of life following treatment: a preliminary study of outpatients and inpatients writh chronic pain, Int. J. Psychiat. Med.. 19 (1988) 271 -2X1. Tait, R.C.. Margolis, R.B., Krause, S. and Liebowitz. E.. Compensation status and symptoms reported by patients with chrome pain. Arch. Phys. Med. Rehah., 69 (1988) 1027”- 1029. Thompson, G.N., Post-traumatic psychoneurosis a statisttcal survey. Am. J. Psychiat., 121 (1965) 1043-1048. Travell, J.G. and Simon& D.G.. Myofascial Pain and Dy5function: The Trigger Point Manual, Williams and Williams, Baltimore, MD, 1983. Warfield, C.A. and Crews. D.A., Work status and response to epidural steroid injection. J. Occup. Med.. 29 (1987) 315-316. We&bill, V.E., ‘Compensation neurosis’: a review of the Irterature, .I. Psychosom. Res.. 27 (1983) 97- 104
Pain, 43 (1990) 37-46 Elsevier
PAIN 01646
Litigation and employment status: effects on patients with chronic pain Raymond
C. Tait, John T. Chibnali
and William D. Richardson
Division of Behavioral Medicine, St. Louis University School of Medicine, St. Louis, MO 63104 (U.S.A.) (Received
4 October
1989, revision
received 28 March 1990, accepted
6 April 1990)
summary
In order to study the effects of compensation and litigation, 201 chronic pain patients were selected from a sample of 464: 99 were working, 15 were working and litigating, 53 were receiving Worker’s Compensation, and 34 were receiving Worker’s Compensation and litigating. Employment (working vs. Worker’s Compensation) and litigation status (litigating vs. not litigating) were analyzed in a 2 x 2 factorial design with measures of pain, disability, psycholo~c~ distress, and selected derno~ap~~ as dependent variables. Compared to Worker’s Compensation patients, working patients reported significantly less disability (down-time, days spent in bed, interference of pain in daily activities) and pain of a longer duration. Compared to litigating patients, non-litigating patients reported less pain (on the McGill Pain Questionnaire) and less disability (stopping activity, interference of pain in daily activities). On two measures of psychological distress (depression, anxiety), there were significant interactions: Worker’s Compensation patients who were litigating reported less distress than non-litigants, while working patients who were litigating reported more distress than non-litigants. The results indicate clear differences in self-reports of disability associated with both employment and litigation status. They also suggest that litigation may function as a coping response for patients who are distressed by the adversarial nature of the Workers’ Compensation system. Limitations of the study as well as suggestions for further research also are discussed.
Key words.
Chronic
pain;
Litigation;
Employment
status
Introduction
Financial incentives have long been thought to have significant and negative effects on both symptom complaints [5,20] and response to treatment [4,9,10,31]. As early as 1945 the term “compensation neurosis” was coined to describe the psychological effects thought to be associated with compensation [14]: “A state of mind, born out of
Correspondence to: Dr. R.C. Tait, Division of Behavioral Medicine, St. Louis University School of Medicine, 1221 S. Grand Blvd., St. Louis, MO 63104, U.S.A. 03~-3959/~/$03.50
@ 1990 Elsevier Science Publishers
fear, kept alive by avarice, stimulated by lawyers, and cured by a verdict.” While the belief in a compensation (or accident or litigation) neurosis is widespread, there is increasing evidence, especially in regard to chronic pain patients, that its effects may be overstated or, at least, over-simplified [21]. Several studies have found no significant differences in pain severity or psycholo~cal distress related to compensation status [15,18,19,28], while others have revealed no changes in clinical status once compensation had been resolved [2,13]. It has been suggested that instead of clarifying the effects of compensation, the confusing state of the literature probably serves
B.V. (Biomedical
Division)
3x
only to enhance the biases of those who read it 122.291. Several problems have been identified in the pain literature that may have contributed to the divergent findings. One involves studies where employment and compensation effects may have been confounded. Recent research comparing compensation and non-compensation patients found that effects of compensation status on treatment outcomes were mediated by employment status; when differences in employment status were controlled, there were no differences in outcomes between patients receiving Worker’s Compensation and those not receiving it [Xl. Another problem involves failure to adequately differentiate patient groups. For example, Jamison et al. [ll] observed that compensation effects vary, depending on whether compensation is timelimited or not. Similarly, it has been noted that the compensation system is heterogeneous, including such diverse groups as veterans, Social Security Disability recipients, Worker’s Compensation benefits recipients, and litigants [2,9,23]. Failure to adequately specify group composition may have added unwanted variability to research effects [28]. Finally, it has been observed that much of the research has addressed only pain-related and psychological variables. Broadening the focus to include other variables has been suggested in order to facilitate the development of a more comprehensive understanding of the effects of compensation systems on patients involved in them [28,32]. The present study attempted to investigate compensation and litigation effects in ways that address some of these methodolo~c problems. To broaden the scope of the study and to maximize its construct validity, dependent variables were included that covered several dimensions previously found useful in studies of chronic pain patients [26]: pain characteristics, pain history, psychological distress, and pain-related disability. The study also used a factorial design in order to separate the effects of compensation and litigation status. Patients were categorized based on both their current employment and litigation status. Thus, patients were either working at their usual job, or were not working and collecting Workers Compensation benefits. Also, they were
either actively involved in litigation seeking compensation for their pain, or were not involved in pain-related legal proceedings. The compensation patients in this study (i.e.. those who were not working) were covered by Worker’s Compensation health insurance, a United States government program which insures workers against injuries sustained on the job. While this system is mandated by the federal government. each state is free to establish guidelines that govern the actual administration of the compensation program. Although there are variations in state laws, certain benefits typically apply: (1) injured workers are entitled to payments equivalent to a percentage of their wage (percentages and time periods vary across states); (2) they are entitled to medical evaluation and treatment for on-the-job injuries; (3) if their injuries prevent them from returning to the job they held when injured, workers are entitled to vocational rehabilitations (4) if the injured worker remains disabled, he/she can apply for Social Security Disability benefits. In practice, several difficulties can arise from this process. Most notably, many injured workers whose primary complaint is pain are unable to regain function in a “timely” way; others report symptoms that exceed objectively verifiable physical evidence. Such events may open the door to an adversarial relationship between the injured worker (who wants benefits or continued benefits) and the insurance carrier (which demands verifiable, objective evidence of injury). The result may be the retention of attorneys and the beginning of litigation. Thus, the compensation patients in this study were not working, but were receiving benefits from the Worker’s Compensation system. Some of the compensation patients had retained attorneys (compensation-litigation group), while others had not retained attorneys and were not in litigation (compensation-only group). The other two study groups were composed of working patients. These patients were currently working full time at some occupation, and drawing a salary from that activity. Of the patients in this group, those who were not litigating either had never been involved in litigation, or had resolved all legal activity prior to participating in
39
the study. Those who were working and litigating were a small, uncommon group. These patients, although functional enough to carry out their occupational duties every day, also had initiated litigation against some party, such as an employer, an insurance company, or an individual (e.g., someone involved in a traffic accident may sue the other driver). The purpose of the present study, then, was to look for differences in pain characteristics, psychological distress, pain history, and pain-related disability among these four groups. Based on recent research [15,18,19,28] neither litigation nor employment status was expected to be associated with pain or psychological distress, although both were expected to have effects on patient reports of disability [6]. No interactions were expected.
Methods Subjects Subjects were 201 chronic pain patients who presented for evaluation at a multi-disciplinary treatment center located in a large university medical center. These subjects were selected from a total of 444 consecutive patients seen over a 30-month period (January, 1985 to June, 1987). The 201 patients were chosen because they fit into categories in the 2 x 2 factorial design. Of the 201, 15 were working and litigating, 99 were working but not litigating, 34 were receiving Worker’s Compensation and litigating, and 53 were receiving Worker’s Compensation but were not litigating. Of the 243 patients not in the study, 53 were receiving Social Security Disability Benefits, 37 were retired and receiving Social Security payments, 41 were not working or receiving pain-related benefits but were actively litigating, and 101 were not working, receiving benefits, or litigating (missing data: n = 11). The subjects in the study reported a mean age of 39.8 years (S.D. = 9.85) and a mean level of education of 12.1 years (SD. = 2.55). One-hundred twenty-one (60.8%) were males and 78 (39.2%) were females; 74.4% were married, 11.8% were single, 13.4% were separated or divorced, and 0.5% were widowed. With regard to circumstances of
pain onset, 66.2% reported that a work-related accident led to the pain condition, 11.8% indicated that some other type of accident (e.g., auto) caused their pain, and for 4.6% pain began following an illness or surgery. For nearly 15% of the sample (14.9%) there was no clear precipitant for the pain (i.e., it “just began”) and 2.6% reported that “other unspecified events” led to the pain problem. On a 5-point Likert-type scale (0 = mild, 5 = excruciating), patient ratings of usual pain level had a mean of 2.9 (S.D. = 0.74) and duration of the pain problem averaged 46.9 months (S.D. = 93.0). Pain distribution for the study subjects was somewhat varied: 5 (2.6%) reported ventral pain; 12 (6.3%) reported extremity pain (upper or lower); 6 (3.2%) reported upper quadrant pain (head, neck, shoulder); 10 (5.3%) reported low back pain; 20 (10.5%) reported upper quadrant and extremity pain; 65 (34.2%) reported low back and extremity pain; 5 (2.6%) reported upper quadrant and low back pain; 67 (35.3%) reported upper quadrant, low back, and extremity pain (missing data = 11). The majority of patients were diagnosed as having pain of a musculoskeletal nature, although there were some cases in which other factors also were diagnosed (e.g., neurovascular compression syndrome, post-traumatic headache). Comparisons were made between the 201 study patients and the other 243 patients on several variables. T tests revealed that the study patients were younger (M = 39.8 vs. M = 46.3) t (422) = 5.1, P < 0.001; were more educated (M = 12.1 vs. M = 11.6) t (435) = - 2.0, P < 0.05; reported lower usual pain intensity (M = 2.9 vs. M = 3.2), t (399) = 3.3, P < 0.01; and reported pain of a shorter duration (M = 46.9 vs. M = 68.7) t (433) = 3.8, P < 0.01. As can be seen in Table I, chisquare tests also revealed that study patients were more likely to be male, less likely to be widowed, and more likely to be married than non-study patients. Circumstances of pain onset also differed, largely because the study patients were more likely to report pain related to a work injury. Demographic differences such as those reported above are not surprising in light of the composition of the study group. The vast majority of these patients were of working age (only 1.5% were 65 or older). Of the non-study patients, how-
40
TABLE
I
DIFFERENCES PATIENTS
BETWEEN
Variable
STUDY
AND
NON-STUDY
Study group
Non-study
group
n
48
n
%
Gender a Male Female
121 78
60.8 39.2
102 141
42.0 58.0
Marital status b Single Married Separated Divorced Widowed
23 145 5 21 1
11.8 74.4 2.6 10.8 0.S
31 153 9 22 13
13.6 67.1 3.9 9.6 5.7
Circumstances of onset ’ Work accident Other accident Illness/surgery ‘Just began’ Other events
129 23 9 29 5
66.2 11.8 4.6 14.9 2.6
63 58 46 50 19
26.7 24.6 19.5 21.2 8.1
a XL (1) =15.5, P i 0.001. h x2 (4) = 10.3, P < 0.05. ’ x2 (4) = 73.2, P i 0.001.
ever, a substantial number were 65 or older (16.5%). Thus, it was not surprising for this group to have pain of a longer duration and to have a higher percentage of females and widows. Also, some of these patients had been declared legally disabled, while many were neither working nor receiving benefits for their pain. Materials All subjects completed a pre-evaluation assessment battery containing the following measures: Pain Disability Index (PDI) [25]; Quality of Life Scale (QOLS) [7]; McGill Pain Questionnaire (MPQ) [17], from which one score was derived the sum of the psychophysically scaled adjectives; Beck Depression Inventory, short form (BDI) [3]; Psychosomatic Symptom Checklist (PSC) [l]; Spielberger Trait Anxiety Inventory (STAI) [24]; and a pain drawing which yielded a measure of percentage of painful body surface (PBS) [16]. Subjects also rated their worst (WORST), least (LEAST), and usual (USUAL) pain levels on a
5-point Likert scale (1 = mild; 2 = discomforting; 3 = severe; 4 = horrible; 5 = excruciating), and the frequency with which they stopped activity (STOP), lay down (DOWN), and stayed in bed (IN BED) due to pain. The activity ratings were made on 5-point scales that ranged from 1 (16 or more times/day) to 5 (0 times/day) for STOP and DOWN, and from 1 (7 days/week) to 5 (0 days/week) for IN BED. In a demographic section. subjects provided other relevant information (e.g., education, age, surgical history) as well as the number of months they had been in pain (DURATION). Design The design of the study was a between-subjects 2 x 2 factorial, with employment status as one independent variable (working vs. Worker’s Compensation) and litigation status as the other independent variable (litigating vs. not litigating). Fourteen dependent variables were analyzed separately using 2-way univariate analyses of variance. The univariate approach was necessary due to missing data and small cell sizes. The regression approach was used to calculate sums of squares for the analyses in order to accommodate unequal and disproportional cell sizes. To correct for violations of homogeneity of variance 2 variables were transformed: DURATION values were transformed to base 10 logarithms to remove positive skewness, and WORST pain values were exponentially transformed to remove negative skewness.
Results Comparisons on demographic variables Table II presents demographic data for patients broken down by employment and litigation status. No effects for age were found. Educational level, however, differed significantly for the working and compensation groups, with working patients reporting higher levels of education, F (1, 196) = 9.9, P < 0.01. A similar main effect was found for pain duration, with working patients reporting pain of a longer duration than those receiving worker’s compensation, F (1, 192) = 4.09, P ( 0.05.
41 TABLE
II
EMPLOYMENT
AND LITIGATION
GROUPS:
DEMOGRAPHICS Workers Compensation
Working
Group
Variable
M
SD.
N
M
SD.
N
14 96
39.3 38.7
11.54 9.41
32 52
Age
Litigating Non-litigating
36.9 41.1
5.87 9.88
Education ’
Litigating Non-litigating
12.1 12.4
1.36 2.62
15 98
11.2 11.2
2.68 2.08
34 53
Pain duration
Litigating Non-litigating
35.1 68.4
33.86 125.89
14 96
30.0 21.8
38.35 23.90
33 53
Pain duration (log 10) b
Litigating Non-litigating
1.3 1.4
0.46 0.57
14 96
1.3 1.2
0.40 0.41
33 53
’ Employment effect, P < 0.01. b Employment effect, P i 0.05 (log transformation required).
There was also a marginally significant difference for gender distribution among the experimental groups (x2 (3) = 6.86, P < 0.10). The largest difference was associated with the Worker’s Compensation/non-litigating group, which was composed of more males (75%) and fewer females (25X) than the other groups (working/litigating: male = 50%, female = 50%; working/non-litigating: male = 558, female = 45%; Workers Compensation/litigating: male = 59% female = 41%). Finally, the groups were compared on circumstances of pain onset. This comparison also functioned as a check on the composition of the study groups. Table III shows that circumstances
TABLE
to expectations: (1) of pain onset conformed working/litigating patients reported pain onset following an accident either at work or elsewhere; (2) working/non-litigating patients reported very mixed circumstances of onset; (3) all worker’s compensation patients reported pain following a work-related accident. While the distributions precluded a &i-square analysis, the differences are obvious. Comparisons on pain characteristics variables Table IV shows the effects of employment and litigation status on pain characteristics. The table shows very few differences on any of the variables
III
CIRCUMSTANCES Group
OF PAIN ONSET BY WORK STATUS AND LITIGATION
STATUS
Circumstances of onset Accident at work
Other accident
Surgery or illness
Workinglitigating
11 (73.3%)
4 (26.7%)
0
Workingnon-litigating
34 (35.4%)
19 (19.8%)
9 (9.4%)
Worker’s Comp. litigating
::OO.OS)
Worker’s Comp. non-litigating
51 (100.0%)
No known cause 0
f390.28)
‘ Other’ events 0
5 (5.2%)
0
0
0
0
0
0
0
0
42 TABLE
IV
EMPLOYMENT Variable
AND
LITIGATION
STATUS:
Group
EFFECTS
ON PAIN CHARACTERISTICS
Working
Worker’s
M
S.D.
N
M
Compensation
..__~~~
S.D.
N
WORST
Litigating Non-litigating
4.60 4.37
0.63 0.90
15 89
4.73 4.64
0.45 0 52
LEAST
Litigating Non-litigating
2.13 1.80
0.83 0.73
15 X8
2.24 2.0X
0.79 O.XX
USUAL
Litigating Non-litigating
3.13 2.16
0.64 0.71
15 89
3.12 3.04
0.x2 0.73
PBS
Litigating Non-litigating
26.08 23.50
19.42 17.16
15 90
30.60 24.38
16.52 18.43
MPQ a
Litigating Non-litigating
38.36 32.87
16.00 13.22
14 91
38.26 33.31
1I .92
a Litigation effect, P < 0.05. Nnte: Values for WORST were transformed
exponentially
for analysis.
measuring pain intensity. The one univariate main effect that did emerge involved the MPQ: litigating patients used more intense adjectives to describe pain than non-litigating patients, F (1, 181) = 4.2, P -c 0.05. In light of the fact that only 1 of the 5 univariate analyses showed between-group differences, the differences that did emerge on the MPQ should be viewed with appropriate caution.
TABLE
The untransformed
15.57
values are shown
here.
Comparisons on disability variables
In contrast to the lack of differences in pain characteristics associated with employment or litigation status, a number of main effects were found with disability variables. These are shown in Table V. Significant litigation effects were found for the PDI, F (1, 185) = 7.7, P -c 0.01, and for STOP, F (1, 185) = 6.5, P < 0.05, both variables reflecting
V
EMPLOYMENT
AND
LITIGATION
STATUS:
EFFECTS
ON DISABILITY Worker’s
Compensation
Variable
Group
Working M
SD.
N
M
S.D.
N
pDI
Litigating Non-litigating
44.20 36.94
14.66 13.24
15 91
51.48 46.72
10.28 10.16
33 50
Litigating Non-litigating
2.14 2.93
1.35 1.27
14 94
2.09 2.44
1.oo
1.20
32 50
hb
Litigating Non-litigating
3.87 3.91
1.00 1.08
15 93
3.29 3.38
1.13 1.13
31 53
IN BED b
Litigating Non-litigating
4.20 4.23
0.86 1.16
15 91
3.63 3.92
0.93 0.98
30 53
ra.bbb
STOP a
DOWN
_.____
a Litigation effect, P < 0.05. aa Litigation effect, P < 0.01. ’ Employment effect, P < 0.05. ” Employment effect, P < 0.01. bbb Employment effect, P -C0.001. Noret For STOP, DOWN, and IN BED, lower scores reflect greater
disability.
43
pain-related interference with activity. Interestingly, no differences emerged in measures of down-time or days spent in bed. Also shown in Table V, the effects of employment status cut across 3 of the 4 disability variables: PDI, DOWN, and IN BED. Surprisingly, working patients did not report more stoppage of activity than compensation patients. Nonetheless, the differences that did emerge suggested that the compensation patients were more disabled than their working counterparts, with greater interference of pain in daily activities, F (1,185) = 15.5, P < 0.001, more down-time, F (1, 188) = 8.8, P < 0.01, and more days spent in bed, F (1, 185) = 5.3, P < 0.05.
\\ 11/‘>
NonLitigating
----
Litigating
\\
\\
‘\
\\
10
91
Vv’orking
Worker’s Compensation
Fig. 1. Employment and litigation status: effects on depression. Scores on the BDI can range from 0 to 39.
Comparisons on psychological distress variables As can be seen in Table VI, there were no main effects for litigation or employment status on the variables measuring psychological distress. There were, however, several significant employment by litigation interactions involving variables designed to measure depression and anxiety. The interactions were broken down further into simple effects. As can be seen in Fig. 1, the interaction is in cross-over form for the BDI. Patients who were working and litigating were more depressed than patients who were working but not litigating, F (1, 179) = 3.94, P < 0.05. Among patients receiving Worker’s Compensa-
TABLE
-
‘\ /
tion, however, there was a trend for those who were litigating to be less depressed than those who were not, F (1, 179) = 2.94, P < 0.10. A similar pattern was obtained for the STAI (see Fig. 2). Among patients who were working, those who were litigating reported more anxiety than those who were not, F (1, 156) = 4.15, P < 0.05. Among patients who were receiving Worker’s Compensation, however, litigating patients reported less anxiety than non-litigating patients, F (1, 156) = 3.88, P < 0.05. As above, however, this comparison produced a somewhat weaker effect.
VI
EMPLOYMENT
AND LITIGATION
STATUS:
EFFECTS
ON PSYCHOLOGICAL
Working
DISTRESS
Variable
Group
Worker’s N
M
Beck a
Litigating Non-litigating
12.45 10.03
6.03 5.50
15 92
10.10 12.41
1.63 7.09
30 46
Litigating Non-litigating
53.38 46.95
7.24 10.27
13 77
45.89 50.69
9.91 11.29
28 42
QOLS
Litigating Non-litigating
25.29 27.24
6.00 8.54
14 90
23.15 24.10
9.38 9.02
33 49
PSC
Litigating Non-litigating
66.19 59.12
40.29 32.26
14 88
71.47 67.88
39.08 33.21
30 49
M
Spielberger
b
a Interaction: F (1, 179) = 5.1, P < 0.05. b Interaction: F (1,156) = 8.0, P < 0.01. Noret For the QGLS, lower scores reflect lower quality
S.D.
of life.
Compensation S.D.
N
-
NonLitigoting
----
Litigating
AA 1
Working
Worker’s Compensation
Fig. 2. Employment and litigation status: effects on anxiety. Scores on the STAI can range from 20 to 80.
Discussion Results relevant to pain characteristics and disability levels were generally consistent with prior research. There were few differences between groups on pain characteristics, a common finding [15,18,19,28]. The between-group differences on disability measures, involving both the compensation and litigation variables, also is generally consistent with prior research [5,23]. The absence of main effects for the psychological distress variables also has been reported previously [18,19,28]. The cross-over interactions that emerged for measures of depression and anxiety, however, have not been reported previously (probably because they have not been examined in previous studies). These interactions showed the effects of litigation to be mediated by work status; litigating Worker’s Compensation patients reported less distress than non-litigants, while distress levels were higher among litigating, working patients than among non-litigants. While any interpretation of this pattern is speculative, it is interesting to speculate on factors that might have occasioned the pattern of findings that emerged. Among Worker’s Compensation patients, the decision to retain an attorney clearly was associated with reports of less distress, suggesting that the decision to hire an attorney was an effective coping response to the stresses that
they faced [12,18]. This contrasts with the distress experienced by working patients who retained attorneys; levels of distress among this group were higher than among the non-litigants. Why might litigation be an effective response for the latter patients but not for the former? Both working and Worker’s Compensation patients face stresses associated with ongoing pain and dysfunction. Worker’s Compensation patients, however, face the additional stress of involvement in an adversarial medical-legal system 1181; the decision to hire an attorney is likely to be an effective means of addressing the adversarial stresses of that situation [23], although the longterm impact of such a decision may be deleterious to rehabilitation [21]. By contrast, the working patient faces primarily the stresses of pain and dysfunction. If these patients experience significant pain-related disability (as the data in this study suggest), they are likely to be facing the stress of meeting high functional demands (e.g., in the workplace) with impaired physical capabilities. Hiring an attorney does not resolve this stressor: in fact, it may be an act driven by the realization that work demands will exceed physical capabilities for the foreseeable future. Thus, the decision to litigate does not resolve immediate stresses, although it may provide long-term help. In order to explore whether there was any clinical evidence to support the above speculation. several patients involved in Worker’s Compensation proceedings were interviewed regarding their decisions to retain attorneys. Each patient indicated that the Worker’s Compensation proceedings had been adversarial. For each, the decision to hire an attorney was prompted by a concern that the adversarial proceedings would work against an appropriate treatment and/or settlement. These interviews clearly were consistent with the data. Parenthetically, we should note that the majority of patients interviewed also indicated some regret that attorneys had got involved: the attorneys were perceived as actually delaying the settlement of their cases. While it is tempting to stop with the above speculations, we should note that other explanations can be supplied. For example, patients who were working and litigating may have presented
45
themselves as more distressed than non-litigating patients in order to dramatize their misery. Similarly, among Worker’s Compensation patients, those involved in litigation may have downplayed their emotional distress in order to avoid being seen as symptom magnifiers, while the non-litigants might not have faced that contingency. Clearly, further research is needed to address this unexpected finding regarding psychological distress. There also are methodologic shortcomings to the study. First, it relied on self-report data; these should be complemented by medical and/or behavioral data to get a more adequate assessment of patients with chronic pain [21]. Second, this was a between-groups design. This design presumes that litigation and/or work status is static when, in actuality, the decision to litigate or not probably is much more fluid. To obtain information more specifically relevant to the effects of litigation on patients, a within-subjects design would be preferable (although not without its own shortcomings), because each patient then would serve as his own control. The study also was limited by the small number of working/litigating patients, a consequence of the very small percentage (3%) of such patients who presented for treatment during the two years of data collection. Thus, these results should be viewed with caution until further research on this group can be conducted. Also, considerations of sample size and missing data forced the use of univariate analyses of variance. Such analyses can yield results tending to over-estimate group differences. Similarly, we should note that this is a correlational study where strict experimental procedures were impossible, so cause-and-effect conclusions cannot be drawn. Thus, there may be rival explanations of group differences other than employment and litigation status. For example, significant differences in gender distribution, circumstances of pain onset, and years of education were found among the groups. These differences could have mediated the results. Similarly, other unassessed variables also could have contributed to the group differences that were found (e.g., socioeconomic status, type of occupation).
Finally, we should note that the patients in this study were drawn from a sample of patients who were evaluated at a comprehensive pain center. They had not responded well to treatment provided by their primary physicians. Previous research has suggested that patients not managed successfully by their physicians may exhibit more depression and illness behavior than those patients who have been successfully managed [6]. Extrapolating from this study to patients not seen in comprehensive pain centers should be done with caution, and future research should examine these successfully managed patients. Despite its limitations, the study does have value. First, it points out the usefulness of multidimensional assessment in comparisons of chronic pain patients. Second, the data suggest that the effects of litigation and compensation are not so straightforward as might be expected. While there were main effects relevant to disability, the measures of psychological distress followed a more complex pattern, suggesting that the decision to litigate may have been a coping response to the stresses of Worker’s Compensation proceedings. Further research investigating this phenomenon is needed in order to better understand factors that might be involved in these proceedings that militate against chronic pain patients who are served by it. References Attanasio, V., Andrasik, F., Blanchard, E.B. and Arena, J.G., Psychometric properties of the SUNYA revision of the Psychosomatic Symptom Checklist, J. Behav. Med., 1 (1984) 247-257. 2 Balla, J.I. and Moraitis, S., Knights in armour: a follow-up study of injuries after legal settlement, Med. J. Aust., 2 (1970) 355-361. 3 Beck, A.T. and Beck, R.W., Screening depressed patients in family practice: rapid technique, Postgrad. Med., 52 (1972) 1
81-85.
Block, A.R., Kremer, E. and Gaylor, M., Behavioral treatment of chronic pain: variables affecting treatment efficacy, Pain, 8 (1980) 367-375. 5 Brena, S.F., Chapman, S.L., Stegall, P.G. and Chyatte, S.B., Chronic pain states: their relationship to impairment and disability, Arch. Phys. Med. Rehab., 60 (1979) 387-389. 6 Chapman, C.R., Sola, A.E. and Bonica, J.J., Illness behavior and depression compared in pain center and private
4
practice
patients,
Pain, 6 (1979) l-7.
46 7 Chibnall,
20 Miller,
8
21
9
10
11
12
13
14 I5
16
J.T. and Tait, R.C.. The Quality of Life Scale: a preliminary study, Submitted for review. Dworkin, R.H., Handlin, D.S., Richlin. D.M., Brand, L. and Vannucci, C., Unraveling the effects of compensation, litigation, and employment on treatment response in chronic pain, Pain, 23 (1985) 49-59. Fowler. D.R. and Mayfield, D.G., Effect of disability compensation: disability symptoms and motivation for treatment, Arch. Environ. Hlth, 19 (1969) 719-725. Hammonds, W., Brena, SF. and Unikel. I.P., Compensation for work-related injuries and rehabilitation of patients with chronic pain. Sth. Med. J. (Bgham. AL). 71 (1978) 664-666. Jamison, R.N., Matt, D.A. and Parris, W.C.V., Effects of time-limited vs. unlimited compensation on pain behavior and treatment outcome in low back pain patients, J. Psychosom. Res., 32 (1988) 277-283. Keefe, F.J., Caldwell, D.S., Queen, K.T.. Gil, K.M., Martinez, S., Crisson, J.E., Ogden, W. and Nunley. J.. Pain coping strategies in osteoarthritis patients. J. Cons. Clin. Psychoi., 55 (1987) 208232. Kelly, R. and Smith, B.N., Post-traumatic syndrome: another myth discredited, J. Roy. Sot. Med., 74 (1981) 275-277. Kennedy. F.. The mind of the injured worker: ita effect on disability periods, Compens. Med., 1 (1946) 19-24. Leavitt, F., Garron, D.C., McNeill. T.W. and Whisler, W.W., Organic status, psychological disturbance, and pain report characteristics in low-back-pain patients on compensation, Spine, 7 (1982) 398-402. Margolis, R.B., Tait, R.C. and Krause. S.J., A rating system for use with patient pain drawings, Pain, 24 (1986) 57-65. Melzack, R., The McGill Pain Questionnaire: major properties and scoring methods. Pain, 1 (1975) 277-299. Meizack, R., Katz, J. and Jeans. M.E.. The role of compensation in chronic pain: analysis using a new’ method of scoring the McGill Pain Questionnaire. Pain. 23 (1985) 101-112. Mendelson, G.. Compensation, pain complaints. and psychological disturbance, Pain, 20 (1984) 169- 177.
‘2 23
24
H.. Accident neurosis I. Br. Med. J., 1 (1961) 919-925. Osterweis, M., Kleinman, A. and Mechanic. D. (Eds.), Pain dnd Disability: Clinical, Behavioral, and Public Policy Prospectives, National Academy Press, Washington, DC. 7987. Parker. N., Accident neurosis. Med. J. Amt.. 2 (1970) 362-365. Peck. C.J.. Fordyce, W.E. and Black, R.G., The effect of the pendency of claims for compensation upon behavior indicative of pain, Wash. Law Rev., 53 (1978) 251 278. Spielberger, CD.. Goruch, R.L. and Lushene. R.E., STAI; Manual for the State-Trait Anxiety Inventory, Consulting Psychologists Press, Palo Alto. CA, 1970. Tait, R.C., Pollard, C.A.. Margolis, R.B., Duckro, P.N. and Krause, S.J., The Pain Disability Index: psychometric and validity data. Arch. Phys. Med. Rehab., 68 (1987) 438441. ‘Tait, R.C., Chibnall, J.. Duckro. P.N. and Deshields, T.I.... Dimensions of chronic pain: a factor analytic study. C‘lin. J. Pain. 5 (1989) 3233328. fait, R.C.. Duckro. P.N., Margoiis, R.B. and Wiener, R.. Quality of life following treatment: a preliminary study of outpatients and inpatients writh chronic pain, Int. J. Psychiat. Med.. 19 (1988) 271 -2X1. Tait, R.C.. Margolis, R.B., Krause, S. and Liebowitz. E.. Compensation status and symptoms reported by patients with chrome pain. Arch. Phys. Med. Rehah., 69 (1988) 1027”- 1029. Thompson, G.N., Post-traumatic psychoneurosis a statisttcal survey. Am. J. Psychiat., 121 (1965) 1043-1048. Travell, J.G. and Simon& D.G.. Myofascial Pain and Dy5function: The Trigger Point Manual, Williams and Williams, Baltimore, MD, 1983. Warfield, C.A. and Crews. D.A., Work status and response to epidural steroid injection. J. Occup. Med.. 29 (1987) 315-316. We&bill, V.E., ‘Compensation neurosis’: a review of the Irterature, .I. Psychosom. Res.. 27 (1983) 97- 104
