Journal of Occupational Rehabilitation, VoL 6, No. 1, 1996
Factors Influencing the Cost of Chronic Low Back Injt/ries: An Analysis of Data from Independent Medical Examinations William A. Earman, 1 Gunnar B. Andersson, 2 Frank Leavitt, 3,4 Thomas W. McNeill, 2 I. Harun Durudogan, and Joan Reagan 1
Cost factors were examined in 157 patients with work-related spine injuries who were referred to a second opinion program between 1985 and 1991. The independent medical examination (IME) included a history, physical examination, and review of imaging and other studies. Data on demographic variables, litigation, work, injury history, physical examination, and imaging studies were recorded. A standard measure of psychological status (Low Back Pain Symptom Check List) was filled out. The instrument uses pain language as a clinical marker of psychological disturbance linked to a range of conflictual issues such as suppressed anger, burdensome feeling of inferiority, damaged self-esteem, role confusion, abnormal mentation, fear of responsibility or intimacy, gender issues, sexual concerns, disturbing arousal, and the like. Since it relies exclusively on pain language for diagnosis, it does not identify the specific nature o f the psychological conflict. Data on treatment, final resolution, and cost were obtained from computerized files of the insurance company. The total cost incurred was $6,551,139. This averaged to $41,727per case. More expensive cases were associated with a surgical intervention, psychological disturbance, litigation, motor weakness, and positive radiographs. These five variables accounted for 48% of the cost variance. Surgery accounted for 19.9% of the variance and contained the most expensive cases ($68,310 vs. $31,423). Psychological disturbance was detected in 27% of the sample and accounted for 10.5% of the cost variance. Litigation was present in 72% of the cases and accounted for 9.1% of the cost variance. Motor strength and radiographs taken together accounted for 8. 4% of the variance. The usefulness of this information was explored from an actuarial and medical perspective. KEY WORDS: low back pain; worker compensation; back injury; second opinion; cost; psychological
disturbance; litigation. ]Department of Orthopedics, Chicago College of Osteopathic Medicine, Olympia Fields, Illinois 60461. 2Department of Orthopedic Surgery, Rush Medical College, Chicago, Illinois 60612. 3Department of Psychology and Social Sciences, Rush Medical College, Chicago, Illinois 60612. 4Correspondence should be directed to Frank Leavitt, Department of Psychology and Social Sciences, Rush Presbyterian St. Luke's Medical Center, 1653 West Congress Parkway, Chicago, Illinois 60612-3828.
1053-0487/96/0300-0005509.50/0O 1996PlenumPublishingCorporation
6
Earman et al.
INTRODUCTION Low back pain is extremely prevalent in all societies (1). In the U.S. alone, about 2% of the working population have a back injury each year, amounting to almost 500,000 injuries and a cost of more than $50 billion per annum. While the incidence and prevalence of back pain as such appears not to have increased over the past two decades, the rate of disability has (2). Independent Medical Examinations (IME) have become increasingly common for industrially injured patients. This trend is motivated by third-party payers' concern about controlling costs, and individual need for advice and reassurance concerning the appropriateness of diagnostic tests and treatment. Second opinions may in some measure be looked upon as admissions to wide variability in medical judgment and experience, to important differences in medical opinion, to iatrogenic effects and that costly errors can be avoided (3). Implicit in this view is the notion that benefits accrue from second opinions resulting in better patient care and containment of costs. Independent Medical Examinations do not ensure against high cost cases partly because very little is known about the economic values of different variables typically collected as part of the IME. The purpose of this research is to identify variables that impact on cost and are measurable at the time of the IME. Empirical identification of cost drivers can potentially serve as a base for development of different approaches to cost containment. Considering the rapidly increasing costs, guch approaches are obviously important. The present study delineates the relative importance of background, medical, and psychological variables in accounting for variations in cost associated with work-related back injuries studied prospectively. Surgical treatment was also included as a cost variable, because of its known relationship to high-cost cases (4).
MATERIALS AND METHODS Participants
Of 255 patients from one insurance company seen for a second opinion between 1985 and 1991, 157 met the two criteria for inclusion. They were (1) low back pain was present for a minimum of 8 weeks, and (2) patient's case had been settled. These patients had been referred by prior agreement to be seen by a single examiner. A disability period of at least 8 weeks triggered a referral. During this period, all patients insured by this particular company, living in an insured district and absent for a work-related spine injury for at least that period were referred for examination. The examination included a patient history, physical examination, and review of radiographs and/or other images available. A diagnosis, treatment plan, and prognosis were provided, but treatment was not assumed. The data were recorded on a special "encounter form" and a letter was directed to the insurance company, but was shared with the treating physician.
Cost of Chronic Low Back Injuries
7
Patients ranged in age from 20 to 65 with a mean of 38.0 (SD = 11.6). They were 26 females (16.6%) and 131 males. Ninety-four (59.9%) were married, 36 single. One hundred and fifteen (73.2%) were Caucasian, 22 Black, and 17 Hispanic. Mean educational level was 11.4 years (SD = 2.3), 74% were high school graduates. Work was divided into five classes based on physical requirements as defined by the U.S. Department of Labor and recorded by employer and in the occupational nurse report to the insurance company. No systematic analysis of the actual work performance was performed. The information from both sources agreed in all cases. Because of low numbers, light and sedentary were later combined. Fifty-four (34.4%) were considered to be in very heavy jobs, 62 (39.5%) in heavy jobs, 31 (19.7%) in medium heavy jobs, and 7 (4.4%) had light or sedentary jobs. Data Collection
From the records, 21 factors were selected for analyses. Twelve factors were based on the history and physical examination, four on imaging studies, and five concerned the injury, length of disability, work, and treatment. The physical examination was a standard orthopedic exam with inspection, palpation, measurements of range of motion, tests of root tension, reflexes, and motor and sensory functions of the lower extremity. No special equipment was used. Radiographs and other imagery were considered abnormal when obvious degenerative changes, deformity, herniations, and stenosis were present. The surgeon evaluation was compared to the radiologist's for correlation, and where different (11 cases), an independent surgeon made the final determination. Evaluation whether the study was normal or abnormal simplified the decision, but of course means that a finer analysis of individual changes cannot be made. Information on psychological status was derived from a 103 word Low Back Pain Symptom Check List (5) filled out by patients on the first visit. The pain checklist was developed as a clinical marker of psychological disturbance by recording and categorizing pain language of groups with documentable organic disease of the back and of groups without organic disease and positive psychological findings. Its other advantage is that it detects pre-existing psychological disturbance in patients with low back pain rather than psychological symptoms that develop in reaction to refractory pain (6). The use of pain language as a clinical marker of psychological disturbance has been validated in a number of studies (7-9). Data on treatment, final resolution, and cost were obtained from the computerized files of the insurance company. Cost factors included all medical and surgical payments, expenses for drugs, equipment and rehabilitation services, compensation for time off work, legal fees, and final payment at closing of case. Statistical Methods
Simple descriptive statistics were used to summarize various attributes of the sample. Cost data were analyzed using two-tailed t-tests where differences were determined between two groups. One-way analysis of variance was used when cost
8
Earman et al.
data involved more than two groups. Scheffe's test was used for a posteriori contrasts. Categorical data were analyzed using the chi-square test. Data were analyzed for cases according to the categories listed in each table. A stepwise multiple regression analysis was employed to identify variables tapping independent sources of cost. Variables with significant univariate effects (p < .05) were entered into the regression analysis. The predictor variables were age, sex, marital status, height, site of pain, CT scan, radiograph, motor strength, reflexes, litigation, psychological status, previous surgery, and subsequent surgery. Because cost data are skewed, logarithmic transformations were used to normalize the distribution.
RESULTS The total cost incurred by the 157 patients was $6,551,139, i.e., a mean of $41,727 per case. Expenditures ranged from $1400 to $160,000. Costs exceeded $10,000 in 84% of the cases (see Table I). The 15 (10%) most expensive cases contributed $1,919,427 (29.4%) to the total costs. Three demographic variables were associated with cost: age, sex, and marital status (Table II). Cases involving older workers were appreciably more expensive than those involving younger workers. In the older than 50 year group, dollar amounts were on the average $25,000 more per case than in the next highest cost age group. Patients were predominantly male. Females made up only 16.6% (26/157). Female cases averaged $19,000 less than cases involving males. Cases involving married workers cost more than cases involving unmarried workers. The difference was $15,000 per case.
Table I. Distribution of Costs Incurred by 157 Patients Injured on the Joba Total settlement Less than $10,000 10,000-19,999 20,000-29,999 30,000-39,999 40,000-49,999 50,000-59,999 60,000-69,999 70,000-79,999 80,000-89,999 90,000 and above
Percentage 16 21 15 10 7 6 7 4 3 11
Cumulative percentage 16 37 52 62 69 75 82 86 89 100
aSeven unclosed cases with follow-up data of more than 2 years included.
Cost of Chronic Low Back Injuries Table II. Individual Factors Associated with Total Cost
N
Mean total cost
Age 20-30 yr 31--40 yr 41-50 yr older than 50
46 49 36 25
Educationa Less than 12 High school More than 12
F
P-value
31,830 40,461 39,662 65,323
4.87
0.002
24 53 14
52,838 40,593 56,462
1.36
NS
Sex Male Female
131 26
44,648 25,799
2.85
0.01
Race White Black Hispanic
115 22 17
44,643 31,402 31,953
1.83
NS
Marital Single Married
36 94
31,930 47,296
4.25
0.04
Height Less than 5'8" 5'8"-5'11" 6'0" and above
35 53 32
29,311 49,886 54,493
4.39
0.01
Weight 90 -160 lb 161-180 Ib 181-199 Ib Above 200 Ib
31 26 21 42
37,067 45,529 50,917 47,893
0.65
NS
aNs vary due to missing data.
The distribution of cases by patient's height revealed a significant difference; however, the findings were largely due to sex differences. Ninety percent of women workers were. less than 5'8" tall and were also the low-cost cases (Table II). Occupational variables involving type of job and mechanism of injury did not produce appreciable differences in costs. There was a trend for medium and light jobs (Table III) to cost less than heavier ones. The two most common mechanisms of injury were falling and lifting. Costs associated with these accident groups were almost identical. Distribution of pain produced significant cost differences (Table III). Pain limited to the back was noticeably less costly than pain involving the back and one, or both extremities. The difference in dollar amounts per case was approximately $18,000.
10
Earman et al.
Table 11I. Total Costs as a Function of Type of Job, Cause of Injury, and Site of Injury Mean
Variables
N
(%)
Cost
F
P-value
Type of job Heavy Medium heavy Medium Light
54 62 31 7
35.1 40.3 20.1 4.5
49,622 41,944 30,829 21,864
2.47
0.06
Site of Pain Back Back--1 leg Back--2 legs
67 64 21
43.2 41.3 13.5
31,095 48,881 49,486
4.51
0.01
Mechanism of injurya Falling Lifting Pulling Bending
68 53 14 10
43.9 33.8 8.9 6.4
41,683 40,949 30,135 41,889
1.83
NS
aFour most commonly mentioned mechanisms of injury.
Table IV. Comparison of Medical Findings and Costs N
Mean to.tal cost
Normal Abnormal
25 40
MRI Normal Abnormal
Variable
T
P-value
35,657 57,768
2.14
0.04
13 31
42,240 51,671
0.85
NS
Myelogram Normal Abnormal
13 10
46,271 57,049
0.69
NS
Radiograph Normal Abnormal
66 34
33,534 53,899
2.57
0.01
C[" scan
Cost d a t a b a s e d on C~, M R I , M y e l o g r a m , and R a d i o g r a p h variables are listed in Table IV. A n a b n o r m a l r a d i o g r a p h was associated with appreciably higher expenditures. Similar t r e n d s were found for positive C T scans. A b n o r m a l M R I Scans a n d M y e l o g r a m s were not associated with significantly higher costs; however, dollar costs were in the $50,000 range for a b n o r m a l findings on any o f the four tests. C o m p a r i s o n of the five clinical p a r a m e t e r s listed in Table V generally showed t r e n d s t o w a r d higher costs w h e n a b n o r m a l findings were present; however, only m o t o r strength and reflex a b n o r m a l i t y r e a c h e d significance at the 0.05 level. R e -
11
Cost of Chronic Low Back Injuries Table
Variables
V. Comparison of Clinical Parameters and Cost Mean N total cost T
P-value
Straight leg raise Normal Abnormal
144 11
41,392 41,744
0.03
NS
strength Normal Abnormal
130 25
38,791 55,070
2.02
0.05
Sensation Normal Abnormal
142 9
40,791 48,513
0.60
NS
Reflex Normal Abnormal
141 14
39,565 60,055
1.99
0.05
Sensory Normal Abnormal
140 9
40,659 48,512
0.61
NS
Motor
duction in motor strength was also the abnormality most frequently (25/155) diagnosed in the sample. A positive straight leg raising test was not associated with higher costs. Length of employment and previous claims were not associated with significantly higher costs (Table VI). The prevalence of litigation in the sample was high (72%). Cases in litigation cost appreciably more. The difference in dollar amount was approximately $28,000. Similar trends were observed in the subpopulation that did not receive surgery. The cost for nonsurgical cases in litigation was $23,668 more per case. Information on psychological status was available on 95 cases, as determined from the Low Back Pain Symptom Check List (5). Psychological disturbance was detected in 26% of this sample, and was significantly associated with cost. The average case cost for workers with psychological disturbance was $29,000 higher. Psychological disturbance was also significantly associated with litigation. 88% (22/25) of workers with psychological disturbance were in litigation involving their cases. By comparison, 58.5% (38/65) of workers without psychological disturbance were in litigation. Six cases could not be classified. Five of these were in litigation. Differences in frequency were significant (Z2 = 7.9; elf = 4; p < .02). Surgical patients contained the most expensive cases (Table VII). Thirteen workers had at least one previous surgical intervention for back pain. This variable was associated with the most expensive cases. In comparison to others in the sample, the difference in dollar amounts for workers with a previous back operation was on average $34,000 per case. Forty-three cases (27%) in the sample received surgical treatm, ent. These cases averaged $37,000 more per case.
Earman et
12 Table
al.
VI. Comparison of EmploymentHistory, Previous Claims, Litigation, and PsychologicalStatus with Costs Mean
Variables
N
total cost
T
P-value
Employment 0-24 mos More than 24
27 32
45,484 41,304
0.42
NS
Previous claim Yes No
11 112
53,469 41,700
0.96
NS
Litigation Total sample Yes No
111 43
49,707 21,081
4.54
0.001
Nonsurgical cases Yes No
72 39
39,727 16,059
4.95
0.001
Psychologicalstatus Normal Abnormal
70 25
35,179 64,276
3.46
0.001
Surgical cases were also significantly associated with litigation. 90.7% (39/43) of the patients treated surgically had legal representation. By comparison, 64.9% (72/111) of the nonoperated cases had lawyers. The 25.8% difference was statistitally significant (Z2 = 10.3; df = 1; p < 0.002). The multiple-correlation coefficient for variables listed in Table VIII was 0.693. This figure indicates that 48.0% (.6932 ) of the variance in total cost was accounted for by the five variables listed in the table. The variables are ranked according to the percentage of cost variance independently accounted for by each variable. Surgery independently accounted for the largest percentage of the variance, 19.9%. Motor strength and radiographs taken together accounted for 8.6% of the variance. The most important nonmedical variable was psychological status which accounted for 10.5% of the variance. Litigation accounted for 9.1%. The demographic variables did not contribute significantly to variance in cost. The significant univariate effects found for age, sex, marital status, and height represent variance shared with other variables.
DISCUSSION The cost of health care for back injured workers who have been off work for at least 8 weeks is high. Third-party payer expenditures averaged $41,727 per case, for a total of $6,551,139. In 84% of the cases, costs exceeded $10,000. Because of the study design (the IME was triggered by a work absence of at least 8 weeks), it is not possible to compare these cases to those of the average
Cost of Chronic Low Back Injuries
13
Table VII. Comparison of Surgery and Costs N
Mean total cost
Previous surgery Yes No
13 142
Present surgery Yes No
43 114
Variables
T
P-value
72,368 38,223
3.30
0.001
68,310 31,423
6.20
0.001
Table VIII. Stepwise Multiple Regression Analysis Using Total Cost as the Criterion Variable Predictor variables Surgery Psychological status Litigation Motor strength Radiographs
Increment in R 2
t
.199 .105 .091 .051 .035
3.29b 4.74b 3.60b 3.51b 2.60a
Cumulative R 2 = .481 b Final F = 18.9 b < .01. P < .001.
worker's compensation back case. These expenditures are less, however, than those reported for a comParable patient sample among 29,421 Tennessee Worker's Compensations claims closed in 1986. Federspiel et al. (10) report an average payout of $55,800 for 7970 back patients. This is $14,000 more per case than recorded for cases in our sample. Costs were directly related to the return-to-work interval. A striking difference in cost was found between cases returning to work soon after injury and those absent for more than 90 days, which constitute the majority of our patients. The results of the regression analysis indicate that five variables are of economic value in understanding 48% of the cost variance. Four of the variables independently impact on cost and are measurable at the time of an Independent Medical Examination (IME). Information gleaned from two nonmedical variables and two medical variables account for approximately 28% of the cost variance. The nonmedical variables contributing to expense are psychological status and litigation. The medical variables associated with high cost cases are radiographs and motor strength. Surgical treatment independently accounts for 19.9% of the cost variance. This is not surprising. Although this variable is not directly measurable at the time of the IME, it is often part of the recommendation that surgery either be or not be considered. Should the independent examiner suggest a surgical procedure, the
14
Earman et al.
insurer must anticipate a higher cost for the case. Also, if differences in opinion exist between the treating and evaluating physician, an additional IME may be costeffective. The individual effects of IME data are of a smaller magnitude in comparison to surgery in determining variation in cost, but their combined effects account for a substantial portion (28.2%) of the cost variance. IME information can be used to identify cases with enhanced risk for high cost early. By way of illustration, cases identified at the time of the IME with motor weakness and positive imaging studies will be more expensive cases independent of whether surgery is involved. Taken together, these two medical variables account for 8.6% of cost variance. The other cost impacting variables measurable at the time of the IME provide a similar basis for predicting cost. To the extent that cases involve psychological disturbance or the presence of an attorney, the probability of higher expense rises with the percentage of variance independently accounted for by each of these variables. IME data of this nature, if replicable in broader and more diverse samples, could form an actuarial basis for treatment and settlement decisions. Controlling costs among workers injured in the workplace depends on better management of variables that impact on costs and that are amenable to manipulation. High cost patients in this study are distinguishable on three variables that meet these criteria. They are an attorney involvement in 72% of the cases, surgical intervention in 27% of the cases, and psychological disturbance in 26% of the cases. A substantial program of cost savings could be developed on the basis of the litigation variable alone. By way of illustration, the cost of care incurred by an injured worker was $28,000 greater, if an attorney was retained. These costs are inflated by surgery, since surgery independently accounts for 19.9% of the cost variance and 39 of the 43 (90.7%) surgical cases in the sample retained legal counsel. The causal linkage between surgery and litigation is not clear. Superficially, it may appear that litigation increases the odds of a surgical procedure. Conversely, surgical cases may involve more incapacitating injuries that require additional medical interventions and raise legitimate concerns about long-term financial security. Closer attention to those cases may result in lowered cost. In the 111 cases not involving surgery, the cost of care incurred by an injured worker with a lawyer was $23,666 greater. Thus, legal participation increased total charges by $1,703,952 ($23,666 • 72 cases). If entry of lawyers in nonsurgical cases (n = 72) was reduced by 20%, then savings of $340,804 (5.2%) are possible. Financial numbers of this magnitude in a small sample suggest that making the system work better will involve taking a closer look at how lawyers enter the clinical situation, and how their entry transforms the care and management of the injured worker (11). If litigation puts pressure on the medical and insurance systems, then questions of why injured workers in such large numbers choose the legal route to address their medical problems may be a good starting point. Grappling with this issue may require tracing developments within the employment and clinical settings that lead to an erosion of trust on the part of the injured worker. A recent survey of injured workers who hire lawyers suggests that many of those entering the legal system do so with a great deal of reluctance (12). In a number of instances, entry of a lawyer was perceived as a hindrance to settlement of the case.
Cost of Chronic Low Back Injuries
15
It is also noteworthy that a large percentage (26%) of the cases injured at work have a psychological dimension, and that those with psychological disturbance seldom see their cases through without hiring a lawyer. In our sample, 88% of the injured workers with identifiable psychological disturbance were in litigation. A linkage between emotional problems and litigation has been previously noted. Talo, Hendler, and Brodie (13) reported 85% of their workers' compensation sample with active litigation, and 89% with completed litigation had an Axis I psychiatric disorder. While the cause and effect relationship between psychological disturbance and litigation can not be established on the basis of clinical linkage, there are a number of possible ways to explain the overrepresentation of cases with psychological disturbance in litigation. One is failure to address or recognize psychological factors in the evaluation process (14) and therefore failure to treat hidden psychological issues. Cases of this nature often produce the paradoxical situation of long-term disability and minimal evidence of organic disease; a predicament rife with potential for fostering a spiral of antagonistic reactions ultimately causing injured workers to los.e faith in the health care system and to turn to legal counsel for help. Controlling costs for this subgroup depends in part on identifying patients with psychological disturbance early in the evaluation process, since this group requires different treatment services. This issue is likely untouched in most medical settings, because of the enormous difficulties in justifying routine psychological consultation for pain that is clearly linked to an accident in the minds of patients. Further, most psychological tests that might provide information on the issue are too cumbersome or objectionable for routine use within the medical setting (15). Tests such as the Low Back Pain Symptom Check List used in this study (5) are acceptable in office practice because they use medical data (in this case pain language) as clinical markers of psychological disturbance. In the era of managed care and capitated reimbursement, providers and insurance carriers may be reluctant to adopt psychological screening measures out of fear that adding a mental claim to a physical claim could conceivably add costs. Adoption of this view may be economically shortsighted if physical symptoms prompting the medical visits are incorrectly attributed to physical disease when medical evidence is absent and hidden psychological disturbance is present. Logically, appropriate treatment enhances chances of recovery, whereas indifference to causal factors likely complicates healing and may well foster the high utilization of medical services and the disproportionate costs accounted for by the small percentage of the low back pain population that is typically nonresponsive to traditional medical care. It is well established (16) that somatizing patients (their physical symptoms are assumed to arise from psychological cause) are "disproportionately higher users of medical services, laboratory investigations and surgical procedures" (p. 1931). Type of work and injury did not significantly influence costs in the present study. This is probably explained by the comparative homogeneity of the group from a work perspective and the fact that work absence was 8 weeks or longer. Several caveats need to be mentioned. First, it is not clear from the present data how representative this sample is of second opinion populations. The patients
16
Earman et al.
constituted a consecutive group but were all from a regional area, all insured by the same company, and had all been out of work for at least 8 weeks. Biases in sample selection might well magnify cost problems unreasonably. Replication in other second opinion populations is required. A second limitation was that the study targeted variables measurable at the time of the IME. Because of a selected focus, some variables likely to impact cost such as duration of disability and duration to case settlement were not considered. Finally, the study data do not bear on the economic value of second opinions. Neither the clinical nor financial usefulness of second opinions was addressed. The study is a first step in understanding the nature of some of the variables that impact on cost in second opinion populations.
REFERENCES 1. White AA, Gordon SL. Synopsis: Workshop on idiopathic low-back pain. Spine 1982; 7: 141-149. 2. Andersson GBJ. The epidemiology of spinal disorders. In: Frymoyer JW, ed. The adult spine: Principles and practice. New York: Raven Press, 1991, pp. 107-146. 3. Rosenfeld I. Second opinion. New York: Simon & Schuster, 1981. 4. Bigos SJ, Spengler DM, Martin NA, Zeh J, Fisher L, Nachemson A, Wang MH. Back injuries in industry: A retrospective study II. Injury factors. Spine 1986; 11:246-256. 5. Leavitt F. Detecting psychological disturbance using verbal pain measurement. The Back Pain Classification Scale. In: Melzack R, ed. Pain measurement and assessment. New York: Raven Press, 1983, pp. 79-84. 6. Leavitt F, Garron DC. The detection of psychological disturbance in patients with low back pain. J Psychosom Res 1979; 23: 149-154. 7. Leavitt F. Use of verbal pain measurement in the detection of hidden psychological morbidity among low back pain patients with objective organic findings. Psychol & Health 1987; 1: 315-326. 8. McNeill TW, Sinkora G, Leavitt F: Psychologic classification of low-back pain patients: A prognostic tool. Spine 1986; 11: 955-959. 9. Sanders SH: Cross-validation of the Back Pain Classification Scale with chronic, intractable pain patients. Pain 1985; 22: 271-277. 10. Federspiel CF, Guy D, Kane D, Spengler D. Expenditure for non-specific back injuries in the work place. J Occup Med 1986; 31: 919-924. 11. Haddad GH. Analysis of 2932 workers' compensation back injury cases. The impact of cost on the system. Spine 1987; 12: 765-769. 12. Tait, RC, Chibnall JT, Richardson WD. Litigation and employment status: Effects on patients with chronic pain. Pain 1990; 43: 37-46. 13. Talo S, Hendler N, Brodie J. Effects of active and completed litigation on treatment results: Workers' compensation patients compared with other litigation patients. 3". Occup Med 1989; 31: 265-269. 14. Waddell G, Kummel EG, Lotto WN, Graham JD, Hall H, McClulloch JA. Failed lumbar disc surgery and repeat surgery following industrial injuries. J Bone Joint Surg 1990; 15: 31-35. 15. Main CA, Wood PLR, Hollis S, Spanswick CC, Waddel G. The distress and risk assessment method. A simple patient classification to identify distress and evaluate the risk of poor outcome. Spine 1992; 17: 42-52. 16. Barsky AJ, Borus JF. Somatization and medicalization in the era of managed care. JAMA 1995; 27: 1931-1934.
Factors Influencing the Cost of Chronic Low Back Injt/ries: An Analysis of Data from Independent Medical Examinations William A. Earman, 1 Gunnar B. Andersson, 2 Frank Leavitt, 3,4 Thomas W. McNeill, 2 I. Harun Durudogan, and Joan Reagan 1
Cost factors were examined in 157 patients with work-related spine injuries who were referred to a second opinion program between 1985 and 1991. The independent medical examination (IME) included a history, physical examination, and review of imaging and other studies. Data on demographic variables, litigation, work, injury history, physical examination, and imaging studies were recorded. A standard measure of psychological status (Low Back Pain Symptom Check List) was filled out. The instrument uses pain language as a clinical marker of psychological disturbance linked to a range of conflictual issues such as suppressed anger, burdensome feeling of inferiority, damaged self-esteem, role confusion, abnormal mentation, fear of responsibility or intimacy, gender issues, sexual concerns, disturbing arousal, and the like. Since it relies exclusively on pain language for diagnosis, it does not identify the specific nature o f the psychological conflict. Data on treatment, final resolution, and cost were obtained from computerized files of the insurance company. The total cost incurred was $6,551,139. This averaged to $41,727per case. More expensive cases were associated with a surgical intervention, psychological disturbance, litigation, motor weakness, and positive radiographs. These five variables accounted for 48% of the cost variance. Surgery accounted for 19.9% of the variance and contained the most expensive cases ($68,310 vs. $31,423). Psychological disturbance was detected in 27% of the sample and accounted for 10.5% of the cost variance. Litigation was present in 72% of the cases and accounted for 9.1% of the cost variance. Motor strength and radiographs taken together accounted for 8. 4% of the variance. The usefulness of this information was explored from an actuarial and medical perspective. KEY WORDS: low back pain; worker compensation; back injury; second opinion; cost; psychological
disturbance; litigation. ]Department of Orthopedics, Chicago College of Osteopathic Medicine, Olympia Fields, Illinois 60461. 2Department of Orthopedic Surgery, Rush Medical College, Chicago, Illinois 60612. 3Department of Psychology and Social Sciences, Rush Medical College, Chicago, Illinois 60612. 4Correspondence should be directed to Frank Leavitt, Department of Psychology and Social Sciences, Rush Presbyterian St. Luke's Medical Center, 1653 West Congress Parkway, Chicago, Illinois 60612-3828.
1053-0487/96/0300-0005509.50/0O 1996PlenumPublishingCorporation
6
Earman et al.
INTRODUCTION Low back pain is extremely prevalent in all societies (1). In the U.S. alone, about 2% of the working population have a back injury each year, amounting to almost 500,000 injuries and a cost of more than $50 billion per annum. While the incidence and prevalence of back pain as such appears not to have increased over the past two decades, the rate of disability has (2). Independent Medical Examinations (IME) have become increasingly common for industrially injured patients. This trend is motivated by third-party payers' concern about controlling costs, and individual need for advice and reassurance concerning the appropriateness of diagnostic tests and treatment. Second opinions may in some measure be looked upon as admissions to wide variability in medical judgment and experience, to important differences in medical opinion, to iatrogenic effects and that costly errors can be avoided (3). Implicit in this view is the notion that benefits accrue from second opinions resulting in better patient care and containment of costs. Independent Medical Examinations do not ensure against high cost cases partly because very little is known about the economic values of different variables typically collected as part of the IME. The purpose of this research is to identify variables that impact on cost and are measurable at the time of the IME. Empirical identification of cost drivers can potentially serve as a base for development of different approaches to cost containment. Considering the rapidly increasing costs, guch approaches are obviously important. The present study delineates the relative importance of background, medical, and psychological variables in accounting for variations in cost associated with work-related back injuries studied prospectively. Surgical treatment was also included as a cost variable, because of its known relationship to high-cost cases (4).
MATERIALS AND METHODS Participants
Of 255 patients from one insurance company seen for a second opinion between 1985 and 1991, 157 met the two criteria for inclusion. They were (1) low back pain was present for a minimum of 8 weeks, and (2) patient's case had been settled. These patients had been referred by prior agreement to be seen by a single examiner. A disability period of at least 8 weeks triggered a referral. During this period, all patients insured by this particular company, living in an insured district and absent for a work-related spine injury for at least that period were referred for examination. The examination included a patient history, physical examination, and review of radiographs and/or other images available. A diagnosis, treatment plan, and prognosis were provided, but treatment was not assumed. The data were recorded on a special "encounter form" and a letter was directed to the insurance company, but was shared with the treating physician.
Cost of Chronic Low Back Injuries
7
Patients ranged in age from 20 to 65 with a mean of 38.0 (SD = 11.6). They were 26 females (16.6%) and 131 males. Ninety-four (59.9%) were married, 36 single. One hundred and fifteen (73.2%) were Caucasian, 22 Black, and 17 Hispanic. Mean educational level was 11.4 years (SD = 2.3), 74% were high school graduates. Work was divided into five classes based on physical requirements as defined by the U.S. Department of Labor and recorded by employer and in the occupational nurse report to the insurance company. No systematic analysis of the actual work performance was performed. The information from both sources agreed in all cases. Because of low numbers, light and sedentary were later combined. Fifty-four (34.4%) were considered to be in very heavy jobs, 62 (39.5%) in heavy jobs, 31 (19.7%) in medium heavy jobs, and 7 (4.4%) had light or sedentary jobs. Data Collection
From the records, 21 factors were selected for analyses. Twelve factors were based on the history and physical examination, four on imaging studies, and five concerned the injury, length of disability, work, and treatment. The physical examination was a standard orthopedic exam with inspection, palpation, measurements of range of motion, tests of root tension, reflexes, and motor and sensory functions of the lower extremity. No special equipment was used. Radiographs and other imagery were considered abnormal when obvious degenerative changes, deformity, herniations, and stenosis were present. The surgeon evaluation was compared to the radiologist's for correlation, and where different (11 cases), an independent surgeon made the final determination. Evaluation whether the study was normal or abnormal simplified the decision, but of course means that a finer analysis of individual changes cannot be made. Information on psychological status was derived from a 103 word Low Back Pain Symptom Check List (5) filled out by patients on the first visit. The pain checklist was developed as a clinical marker of psychological disturbance by recording and categorizing pain language of groups with documentable organic disease of the back and of groups without organic disease and positive psychological findings. Its other advantage is that it detects pre-existing psychological disturbance in patients with low back pain rather than psychological symptoms that develop in reaction to refractory pain (6). The use of pain language as a clinical marker of psychological disturbance has been validated in a number of studies (7-9). Data on treatment, final resolution, and cost were obtained from the computerized files of the insurance company. Cost factors included all medical and surgical payments, expenses for drugs, equipment and rehabilitation services, compensation for time off work, legal fees, and final payment at closing of case. Statistical Methods
Simple descriptive statistics were used to summarize various attributes of the sample. Cost data were analyzed using two-tailed t-tests where differences were determined between two groups. One-way analysis of variance was used when cost
8
Earman et al.
data involved more than two groups. Scheffe's test was used for a posteriori contrasts. Categorical data were analyzed using the chi-square test. Data were analyzed for cases according to the categories listed in each table. A stepwise multiple regression analysis was employed to identify variables tapping independent sources of cost. Variables with significant univariate effects (p < .05) were entered into the regression analysis. The predictor variables were age, sex, marital status, height, site of pain, CT scan, radiograph, motor strength, reflexes, litigation, psychological status, previous surgery, and subsequent surgery. Because cost data are skewed, logarithmic transformations were used to normalize the distribution.
RESULTS The total cost incurred by the 157 patients was $6,551,139, i.e., a mean of $41,727 per case. Expenditures ranged from $1400 to $160,000. Costs exceeded $10,000 in 84% of the cases (see Table I). The 15 (10%) most expensive cases contributed $1,919,427 (29.4%) to the total costs. Three demographic variables were associated with cost: age, sex, and marital status (Table II). Cases involving older workers were appreciably more expensive than those involving younger workers. In the older than 50 year group, dollar amounts were on the average $25,000 more per case than in the next highest cost age group. Patients were predominantly male. Females made up only 16.6% (26/157). Female cases averaged $19,000 less than cases involving males. Cases involving married workers cost more than cases involving unmarried workers. The difference was $15,000 per case.
Table I. Distribution of Costs Incurred by 157 Patients Injured on the Joba Total settlement Less than $10,000 10,000-19,999 20,000-29,999 30,000-39,999 40,000-49,999 50,000-59,999 60,000-69,999 70,000-79,999 80,000-89,999 90,000 and above
Percentage 16 21 15 10 7 6 7 4 3 11
Cumulative percentage 16 37 52 62 69 75 82 86 89 100
aSeven unclosed cases with follow-up data of more than 2 years included.
Cost of Chronic Low Back Injuries Table II. Individual Factors Associated with Total Cost
N
Mean total cost
Age 20-30 yr 31--40 yr 41-50 yr older than 50
46 49 36 25
Educationa Less than 12 High school More than 12
F
P-value
31,830 40,461 39,662 65,323
4.87
0.002
24 53 14
52,838 40,593 56,462
1.36
NS
Sex Male Female
131 26
44,648 25,799
2.85
0.01
Race White Black Hispanic
115 22 17
44,643 31,402 31,953
1.83
NS
Marital Single Married
36 94
31,930 47,296
4.25
0.04
Height Less than 5'8" 5'8"-5'11" 6'0" and above
35 53 32
29,311 49,886 54,493
4.39
0.01
Weight 90 -160 lb 161-180 Ib 181-199 Ib Above 200 Ib
31 26 21 42
37,067 45,529 50,917 47,893
0.65
NS
aNs vary due to missing data.
The distribution of cases by patient's height revealed a significant difference; however, the findings were largely due to sex differences. Ninety percent of women workers were. less than 5'8" tall and were also the low-cost cases (Table II). Occupational variables involving type of job and mechanism of injury did not produce appreciable differences in costs. There was a trend for medium and light jobs (Table III) to cost less than heavier ones. The two most common mechanisms of injury were falling and lifting. Costs associated with these accident groups were almost identical. Distribution of pain produced significant cost differences (Table III). Pain limited to the back was noticeably less costly than pain involving the back and one, or both extremities. The difference in dollar amounts per case was approximately $18,000.
10
Earman et al.
Table 11I. Total Costs as a Function of Type of Job, Cause of Injury, and Site of Injury Mean
Variables
N
(%)
Cost
F
P-value
Type of job Heavy Medium heavy Medium Light
54 62 31 7
35.1 40.3 20.1 4.5
49,622 41,944 30,829 21,864
2.47
0.06
Site of Pain Back Back--1 leg Back--2 legs
67 64 21
43.2 41.3 13.5
31,095 48,881 49,486
4.51
0.01
Mechanism of injurya Falling Lifting Pulling Bending
68 53 14 10
43.9 33.8 8.9 6.4
41,683 40,949 30,135 41,889
1.83
NS
aFour most commonly mentioned mechanisms of injury.
Table IV. Comparison of Medical Findings and Costs N
Mean to.tal cost
Normal Abnormal
25 40
MRI Normal Abnormal
Variable
T
P-value
35,657 57,768
2.14
0.04
13 31
42,240 51,671
0.85
NS
Myelogram Normal Abnormal
13 10
46,271 57,049
0.69
NS
Radiograph Normal Abnormal
66 34
33,534 53,899
2.57
0.01
C[" scan
Cost d a t a b a s e d on C~, M R I , M y e l o g r a m , and R a d i o g r a p h variables are listed in Table IV. A n a b n o r m a l r a d i o g r a p h was associated with appreciably higher expenditures. Similar t r e n d s were found for positive C T scans. A b n o r m a l M R I Scans a n d M y e l o g r a m s were not associated with significantly higher costs; however, dollar costs were in the $50,000 range for a b n o r m a l findings on any o f the four tests. C o m p a r i s o n of the five clinical p a r a m e t e r s listed in Table V generally showed t r e n d s t o w a r d higher costs w h e n a b n o r m a l findings were present; however, only m o t o r strength and reflex a b n o r m a l i t y r e a c h e d significance at the 0.05 level. R e -
11
Cost of Chronic Low Back Injuries Table
Variables
V. Comparison of Clinical Parameters and Cost Mean N total cost T
P-value
Straight leg raise Normal Abnormal
144 11
41,392 41,744
0.03
NS
strength Normal Abnormal
130 25
38,791 55,070
2.02
0.05
Sensation Normal Abnormal
142 9
40,791 48,513
0.60
NS
Reflex Normal Abnormal
141 14
39,565 60,055
1.99
0.05
Sensory Normal Abnormal
140 9
40,659 48,512
0.61
NS
Motor
duction in motor strength was also the abnormality most frequently (25/155) diagnosed in the sample. A positive straight leg raising test was not associated with higher costs. Length of employment and previous claims were not associated with significantly higher costs (Table VI). The prevalence of litigation in the sample was high (72%). Cases in litigation cost appreciably more. The difference in dollar amount was approximately $28,000. Similar trends were observed in the subpopulation that did not receive surgery. The cost for nonsurgical cases in litigation was $23,668 more per case. Information on psychological status was available on 95 cases, as determined from the Low Back Pain Symptom Check List (5). Psychological disturbance was detected in 26% of this sample, and was significantly associated with cost. The average case cost for workers with psychological disturbance was $29,000 higher. Psychological disturbance was also significantly associated with litigation. 88% (22/25) of workers with psychological disturbance were in litigation involving their cases. By comparison, 58.5% (38/65) of workers without psychological disturbance were in litigation. Six cases could not be classified. Five of these were in litigation. Differences in frequency were significant (Z2 = 7.9; elf = 4; p < .02). Surgical patients contained the most expensive cases (Table VII). Thirteen workers had at least one previous surgical intervention for back pain. This variable was associated with the most expensive cases. In comparison to others in the sample, the difference in dollar amounts for workers with a previous back operation was on average $34,000 per case. Forty-three cases (27%) in the sample received surgical treatm, ent. These cases averaged $37,000 more per case.
Earman et
12 Table
al.
VI. Comparison of EmploymentHistory, Previous Claims, Litigation, and PsychologicalStatus with Costs Mean
Variables
N
total cost
T
P-value
Employment 0-24 mos More than 24
27 32
45,484 41,304
0.42
NS
Previous claim Yes No
11 112
53,469 41,700
0.96
NS
Litigation Total sample Yes No
111 43
49,707 21,081
4.54
0.001
Nonsurgical cases Yes No
72 39
39,727 16,059
4.95
0.001
Psychologicalstatus Normal Abnormal
70 25
35,179 64,276
3.46
0.001
Surgical cases were also significantly associated with litigation. 90.7% (39/43) of the patients treated surgically had legal representation. By comparison, 64.9% (72/111) of the nonoperated cases had lawyers. The 25.8% difference was statistitally significant (Z2 = 10.3; df = 1; p < 0.002). The multiple-correlation coefficient for variables listed in Table VIII was 0.693. This figure indicates that 48.0% (.6932 ) of the variance in total cost was accounted for by the five variables listed in the table. The variables are ranked according to the percentage of cost variance independently accounted for by each variable. Surgery independently accounted for the largest percentage of the variance, 19.9%. Motor strength and radiographs taken together accounted for 8.6% of the variance. The most important nonmedical variable was psychological status which accounted for 10.5% of the variance. Litigation accounted for 9.1%. The demographic variables did not contribute significantly to variance in cost. The significant univariate effects found for age, sex, marital status, and height represent variance shared with other variables.
DISCUSSION The cost of health care for back injured workers who have been off work for at least 8 weeks is high. Third-party payer expenditures averaged $41,727 per case, for a total of $6,551,139. In 84% of the cases, costs exceeded $10,000. Because of the study design (the IME was triggered by a work absence of at least 8 weeks), it is not possible to compare these cases to those of the average
Cost of Chronic Low Back Injuries
13
Table VII. Comparison of Surgery and Costs N
Mean total cost
Previous surgery Yes No
13 142
Present surgery Yes No
43 114
Variables
T
P-value
72,368 38,223
3.30
0.001
68,310 31,423
6.20
0.001
Table VIII. Stepwise Multiple Regression Analysis Using Total Cost as the Criterion Variable Predictor variables Surgery Psychological status Litigation Motor strength Radiographs
Increment in R 2
t
.199 .105 .091 .051 .035
3.29b 4.74b 3.60b 3.51b 2.60a
Cumulative R 2 = .481 b Final F = 18.9 b < .01. P < .001.
worker's compensation back case. These expenditures are less, however, than those reported for a comParable patient sample among 29,421 Tennessee Worker's Compensations claims closed in 1986. Federspiel et al. (10) report an average payout of $55,800 for 7970 back patients. This is $14,000 more per case than recorded for cases in our sample. Costs were directly related to the return-to-work interval. A striking difference in cost was found between cases returning to work soon after injury and those absent for more than 90 days, which constitute the majority of our patients. The results of the regression analysis indicate that five variables are of economic value in understanding 48% of the cost variance. Four of the variables independently impact on cost and are measurable at the time of an Independent Medical Examination (IME). Information gleaned from two nonmedical variables and two medical variables account for approximately 28% of the cost variance. The nonmedical variables contributing to expense are psychological status and litigation. The medical variables associated with high cost cases are radiographs and motor strength. Surgical treatment independently accounts for 19.9% of the cost variance. This is not surprising. Although this variable is not directly measurable at the time of the IME, it is often part of the recommendation that surgery either be or not be considered. Should the independent examiner suggest a surgical procedure, the
14
Earman et al.
insurer must anticipate a higher cost for the case. Also, if differences in opinion exist between the treating and evaluating physician, an additional IME may be costeffective. The individual effects of IME data are of a smaller magnitude in comparison to surgery in determining variation in cost, but their combined effects account for a substantial portion (28.2%) of the cost variance. IME information can be used to identify cases with enhanced risk for high cost early. By way of illustration, cases identified at the time of the IME with motor weakness and positive imaging studies will be more expensive cases independent of whether surgery is involved. Taken together, these two medical variables account for 8.6% of cost variance. The other cost impacting variables measurable at the time of the IME provide a similar basis for predicting cost. To the extent that cases involve psychological disturbance or the presence of an attorney, the probability of higher expense rises with the percentage of variance independently accounted for by each of these variables. IME data of this nature, if replicable in broader and more diverse samples, could form an actuarial basis for treatment and settlement decisions. Controlling costs among workers injured in the workplace depends on better management of variables that impact on costs and that are amenable to manipulation. High cost patients in this study are distinguishable on three variables that meet these criteria. They are an attorney involvement in 72% of the cases, surgical intervention in 27% of the cases, and psychological disturbance in 26% of the cases. A substantial program of cost savings could be developed on the basis of the litigation variable alone. By way of illustration, the cost of care incurred by an injured worker was $28,000 greater, if an attorney was retained. These costs are inflated by surgery, since surgery independently accounts for 19.9% of the cost variance and 39 of the 43 (90.7%) surgical cases in the sample retained legal counsel. The causal linkage between surgery and litigation is not clear. Superficially, it may appear that litigation increases the odds of a surgical procedure. Conversely, surgical cases may involve more incapacitating injuries that require additional medical interventions and raise legitimate concerns about long-term financial security. Closer attention to those cases may result in lowered cost. In the 111 cases not involving surgery, the cost of care incurred by an injured worker with a lawyer was $23,666 greater. Thus, legal participation increased total charges by $1,703,952 ($23,666 • 72 cases). If entry of lawyers in nonsurgical cases (n = 72) was reduced by 20%, then savings of $340,804 (5.2%) are possible. Financial numbers of this magnitude in a small sample suggest that making the system work better will involve taking a closer look at how lawyers enter the clinical situation, and how their entry transforms the care and management of the injured worker (11). If litigation puts pressure on the medical and insurance systems, then questions of why injured workers in such large numbers choose the legal route to address their medical problems may be a good starting point. Grappling with this issue may require tracing developments within the employment and clinical settings that lead to an erosion of trust on the part of the injured worker. A recent survey of injured workers who hire lawyers suggests that many of those entering the legal system do so with a great deal of reluctance (12). In a number of instances, entry of a lawyer was perceived as a hindrance to settlement of the case.
Cost of Chronic Low Back Injuries
15
It is also noteworthy that a large percentage (26%) of the cases injured at work have a psychological dimension, and that those with psychological disturbance seldom see their cases through without hiring a lawyer. In our sample, 88% of the injured workers with identifiable psychological disturbance were in litigation. A linkage between emotional problems and litigation has been previously noted. Talo, Hendler, and Brodie (13) reported 85% of their workers' compensation sample with active litigation, and 89% with completed litigation had an Axis I psychiatric disorder. While the cause and effect relationship between psychological disturbance and litigation can not be established on the basis of clinical linkage, there are a number of possible ways to explain the overrepresentation of cases with psychological disturbance in litigation. One is failure to address or recognize psychological factors in the evaluation process (14) and therefore failure to treat hidden psychological issues. Cases of this nature often produce the paradoxical situation of long-term disability and minimal evidence of organic disease; a predicament rife with potential for fostering a spiral of antagonistic reactions ultimately causing injured workers to los.e faith in the health care system and to turn to legal counsel for help. Controlling costs for this subgroup depends in part on identifying patients with psychological disturbance early in the evaluation process, since this group requires different treatment services. This issue is likely untouched in most medical settings, because of the enormous difficulties in justifying routine psychological consultation for pain that is clearly linked to an accident in the minds of patients. Further, most psychological tests that might provide information on the issue are too cumbersome or objectionable for routine use within the medical setting (15). Tests such as the Low Back Pain Symptom Check List used in this study (5) are acceptable in office practice because they use medical data (in this case pain language) as clinical markers of psychological disturbance. In the era of managed care and capitated reimbursement, providers and insurance carriers may be reluctant to adopt psychological screening measures out of fear that adding a mental claim to a physical claim could conceivably add costs. Adoption of this view may be economically shortsighted if physical symptoms prompting the medical visits are incorrectly attributed to physical disease when medical evidence is absent and hidden psychological disturbance is present. Logically, appropriate treatment enhances chances of recovery, whereas indifference to causal factors likely complicates healing and may well foster the high utilization of medical services and the disproportionate costs accounted for by the small percentage of the low back pain population that is typically nonresponsive to traditional medical care. It is well established (16) that somatizing patients (their physical symptoms are assumed to arise from psychological cause) are "disproportionately higher users of medical services, laboratory investigations and surgical procedures" (p. 1931). Type of work and injury did not significantly influence costs in the present study. This is probably explained by the comparative homogeneity of the group from a work perspective and the fact that work absence was 8 weeks or longer. Several caveats need to be mentioned. First, it is not clear from the present data how representative this sample is of second opinion populations. The patients
16
Earman et al.
constituted a consecutive group but were all from a regional area, all insured by the same company, and had all been out of work for at least 8 weeks. Biases in sample selection might well magnify cost problems unreasonably. Replication in other second opinion populations is required. A second limitation was that the study targeted variables measurable at the time of the IME. Because of a selected focus, some variables likely to impact cost such as duration of disability and duration to case settlement were not considered. Finally, the study data do not bear on the economic value of second opinions. Neither the clinical nor financial usefulness of second opinions was addressed. The study is a first step in understanding the nature of some of the variables that impact on cost in second opinion populations.
REFERENCES 1. White AA, Gordon SL. Synopsis: Workshop on idiopathic low-back pain. Spine 1982; 7: 141-149. 2. Andersson GBJ. The epidemiology of spinal disorders. In: Frymoyer JW, ed. The adult spine: Principles and practice. New York: Raven Press, 1991, pp. 107-146. 3. Rosenfeld I. Second opinion. New York: Simon & Schuster, 1981. 4. Bigos SJ, Spengler DM, Martin NA, Zeh J, Fisher L, Nachemson A, Wang MH. Back injuries in industry: A retrospective study II. Injury factors. Spine 1986; 11:246-256. 5. Leavitt F. Detecting psychological disturbance using verbal pain measurement. The Back Pain Classification Scale. In: Melzack R, ed. Pain measurement and assessment. New York: Raven Press, 1983, pp. 79-84. 6. Leavitt F, Garron DC. The detection of psychological disturbance in patients with low back pain. J Psychosom Res 1979; 23: 149-154. 7. Leavitt F. Use of verbal pain measurement in the detection of hidden psychological morbidity among low back pain patients with objective organic findings. Psychol & Health 1987; 1: 315-326. 8. McNeill TW, Sinkora G, Leavitt F: Psychologic classification of low-back pain patients: A prognostic tool. Spine 1986; 11: 955-959. 9. Sanders SH: Cross-validation of the Back Pain Classification Scale with chronic, intractable pain patients. Pain 1985; 22: 271-277. 10. Federspiel CF, Guy D, Kane D, Spengler D. Expenditure for non-specific back injuries in the work place. J Occup Med 1986; 31: 919-924. 11. Haddad GH. Analysis of 2932 workers' compensation back injury cases. The impact of cost on the system. Spine 1987; 12: 765-769. 12. Tait, RC, Chibnall JT, Richardson WD. Litigation and employment status: Effects on patients with chronic pain. Pain 1990; 43: 37-46. 13. Talo S, Hendler N, Brodie J. Effects of active and completed litigation on treatment results: Workers' compensation patients compared with other litigation patients. 3". Occup Med 1989; 31: 265-269. 14. Waddell G, Kummel EG, Lotto WN, Graham JD, Hall H, McClulloch JA. Failed lumbar disc surgery and repeat surgery following industrial injuries. J Bone Joint Surg 1990; 15: 31-35. 15. Main CA, Wood PLR, Hollis S, Spanswick CC, Waddel G. The distress and risk assessment method. A simple patient classification to identify distress and evaluate the risk of poor outcome. Spine 1992; 17: 42-52. 16. Barsky AJ, Borus JF. Somatization and medicalization in the era of managed care. JAMA 1995; 27: 1931-1934.
