Special Article
Increased Costs and Rates of Use in the California Workers' Compensation System as a Result of Self-Referral by Physicians
N Engl J Med 1992; 327:1502-1506November 19, 1992
- Abstract
- Abstract
Background.
There is widespread concern that ownership by physicians of testing or treatment facilities to which they refer patients leads to overuse of such facilities. We determined the patterns of use of three services — physical therapy, psychiatric evaluation, and magnetic resonance imaging (MRI) — among physicians treating patients whose care was covered under workers' compensation. We then compared the rates of use among physicians who referred patients to facilities of which they were owners (self-referral group) with the rates among physicians who referred patients to independent facilities (independent-referral group).
Methods.
We used a large data base to analyze claims under workers' compensation in California from October 1, 1990, through June 30, 1991, to determine the frequency and cost of these three selected services and determined whether the referring physicians were practicing self-referral or independent referral. We evaluated the cost per case for all three services, measured the frequency with which physical therapy was initiated, and evaluated the medical appropriateness of MRI.
Results.
We found that physical therapy was initiated 2.3 times more often by the physicians in the self-referral group (68 percent) than by those in the independent-referral group (30 percent; P<0.01). The mean cost per case for physical therapy was significantly lower in the self-referral group ($404±102) than in the independent-referral group ($440±167; P<0.01).
The mean cost of psychiatric evaluation services was significantly higher in the self-referral group than in the independent-referral group (psychometric testing, $1,165±728 vs. $870±482; P<0.01; psychiatric evaluation reports, $2,056±1,063 vs. $1,680±578; P<0.01). The total cost per case of psychiatric evaluation services was 26.3 percent higher in the self-referral group ($3,222± 1,451) than in the independent-referral group ($2,550±742; P<0.01).
Of all the MRI scans requested by the self-referring physicians, 38 percent were found to be medically inappropriate, as compared with 28 percent of those requested by physicians in the independent-referral group (P<0.05). There was no significant difference in the cost per case between the two groups.
Conclusions.
This study demonstrates that self-referral increases the cost of medical care covered by workers' compensation for each of the three types of service studied. (N Engl J Med 1992;327:1502–6.)
Media in This Article
Table 1
Distribution of AVGs and Mean Cost per Case in the Self-Referral and Independent-Referral Groups.*Table 2
Frequency of Use of Services and Cost per Case in the Self-Referral and Independent-Referral Groups.Article Activity
- Article
THERE is growing concern about conflict of interest in medicine in the United States.1 2 3 4 5 6 Recent studies have focused on whether physicians' ownership of testing or treatment centers increases the number of tests and services performed.7 8 9 10 Research in Florida indicates that physician-owned facilities generate significantly higher rates of use and costs than independently owned facilities.7 , 8 Studies of physician ownership in California have found that the higher concentration of physician-owned magnetic resonance imaging (MRI) facilities in California has increased rates of use between 34 percent and 56 percent above the rates for the rest of the country.9 The study by Hillman et al. of diagnostic imaging demonstrated that physicians who referred patients to facilities of which they were owners (those who practiced self-referral) charged 4.4 to 7.5 times more per episode of care than other physicians.10 In response to these findings, the states of Florida, Michigan, and New Jersey have enacted legislation that restricts self-referral by physicians.
The American Medical Association (AMA) Council on Ethical and Judicial Affairs stated in December 1991: "In general, physicians should not refer patients to a health care facility outside their office practice at which they do not directly provide care or services when they have an investment interest in the facility."11 In June 1992, however, the AMA's House of Delegates adopted a new policy that allows doctors to make such referrals if patients are informed of the doctor's financial interest in the facility and of any available alternatives.12 This reversal on the part of the AMA reflects the lack of consensus within organized medicine about physicians' ownership of medical facilities. There have also been two recent efforts by the federal government to limit self-referral on the part of physicians. Since January 1992, physicians have been prohibited from referring patients to clinical laboratories in which they have an ownership interest. In addition, the "safe harbor" regulations published in the Federal Register defined more clearly the investment, ownership, and reimbursement arrangements in which physicians may participate without fear of violating anti-kickback provisions of Medicare and Medicaid.13
To our knowledge, the effects of physician self-referral within the workers' compensation system have not been systematically analyzed. To investigate this issue, we evaluated a total of 6581 California workers' compensation cases for which claims were filed with a large workers' compensation insurance company during a nine-month period in 1990 and 1991. We analyzed the effect of physicians' self-referral on three high-cost medical services covered under workers' compensation: physical therapy, psychiatric evaluation, and MRI. We evaluated the cost per case for all three services, measured the frequency with which physical therapy was initiated, and prospectively analyzed the medical appropriateness of MRI.
Methods
This study was designed to compare the patterns of use of three services — physical therapy, psychiatric evaluation, and MRI —among physicians who refer patients to facilities of which they are owners (self-referral group) and physicians who refer patients to independent facilities (independent-referral group).
Since differences in case mix between physicians in the self-referral and independent-referral groups might account for differences in rates of use or cost, we classified all cases according to the Ambulatory Visit Groups (AVG) classification scheme,14 which we have modified for workers' compensation cases.15 The AVG system is analogous to the system of diagnosis-related groups currently used by Medicare to reimburse acute care hospitals. On the basis of the patient's diagnosis (the diagnostic code from the International Classification of Diseases, 9th Revision, Clinical Modification) and the medical-procedure codes of the California Relative Value Studies and Current Procedural Terminology for outpatient services in each case, the AVG system is used to assign that case to 1 (and only 1) of 571 groups.
Type of Referral
Throughout the study, self-referral was defined as a referral for a medical service made by a physician or clinic to an entity owned entirely or in part by the referring physician or clinic. Self-referral was defined by either of the following two patterns: referral services were provided under the same tax identification number as the primary service, or referral services were provided under a different tax identification number from the primary service, but one or more owners were common to both entities.
When services were delivered under different tax identification numbers, we searched commercially available data bases that list officers, stockholders, and partners of facilities (the California Fictitious Business Name Listing, the Executive Business Listing, and other state and national data bases on corporations and limited partnerships available from Information America, Atlanta). If this search failed to identify common ownership, we then directly telephoned the referring physician's office and inquired about common ownership.
Physical-Therapy and Psychiatric-Evaluation Services
We used one of California's largest data bases on workers' compensation claims (that of the Industrial Indemnity Co., San Francisco) to analyze the frequency and cost of physical-therapy and psychiatric-evaluation services provided to injured workers. The data base was selected because it was complete and contained information on a large number of patients distributed throughout California. Information about each case was stored longitudinally; thus, the data base contained claims information for all services provided to the injured worker during the entire nine-month period of the study.
Data on all patients covered by workers' compensation in California who received physical-therapy or psychiatric-evaluation services from October 1, 1990, through June 30, 1991, were analyzed. Our analysis compared the rates of use and costs of physical-therapy and psychiatric-evaluation services for physicians in the self-referral and independent-referral groups.
Since musculoskeletal injuries make up the majority of all workers' compensation medical cases, we were able to limit our evaluation of physical therapy to providers with substantial experience in treating industrial musculoskeletal injuries. We defined this degree of experience as the treatment of 10 or more cases of musculoskeletal injury during the study period. There were 76 providers who met this criterion; they treated 1257 cases of musculoskeletal injury. Using the method described above, we were able to determine in all instances whether the referring providers were in the self-referral group or the independent-referral group.
In California, patients covered by workers' compensation are most frequently referred for psychiatric-evaluation services to document a claim of "stress." This evaluation virtually always includes both psychometric testing and a psychiatric-evaluation report. (We documented this fact in a preliminary analysis of our data base.) We therefore limited our analysis of psychiatric-evaluation services to cases in which there was both psychometric testing and a psychiatric-evaluation report. Altogether, 1751 (39 percent) of the cases within the data base met this criterion. A random sample of 220 cases (13 percent) was selected for analysis of ownership. We were able to determine ownership and self-referral or independent-referral status in each of these cases.
MRI
We also compared the cost and appropriateness of MRI scans in the self-referral and independent-referral groups. Appropriateness of referral for an MRI scan was determined under a prospective precertification program. All physicians' requests for MRI scans (regardless of the body part to be examined) were referred by Industrial Indemnity to a national, independent utilization-review firm for precertification review of medical appropriateness. The firm's criteria for appropriateness were established by a panel of board-certified specialists in orthopedics, industrial medicine, and radiology. After initial development by an independent board-certified radiologist and the medical directors of the utilization-review firm and its parent (one of the three largest companies managing health maintenance organizations in the United States), the criteria were reviewed and revised by a panel of independent, practicing experts in managed care who were all board-certified in orthopedics, neurology, neurosurgery, or radiology.
On the basis of medical documentation of the patient's injuries and conversations with the physician who requested the MRI, the review firm gave an opinion on the medical appropriateness of the procedure before it was performed. The reviewers were blinded to the physician's relation with the MRI center.
The classification of a procedure as medically inappropriate could be appealed. To be certain that the reviewer's decision did not merely defer an appropriate scan to another date, cases in which the MRI was categorized as medically inappropriate were followed for an additional six months. In all cases in which a scan was approved within six months after the original request, the MRI was considered to be medically appropriate.
All 864 requests for MRI scans from January 1, 1991, through June 30, 1991, were evaluated. We were able to determine whether the physician had an ownership interest in the facility in 502 (58 percent) of these cases.
Statistical Analysis
Continuous variables are presented as means ±SD and were compared by two-tailed t-tests. The proportion of cases in each group was assessed by the chi-square test. For all analyses, a P value of less than 0.05 was considered to indicate statistical significance. Results were analyzed with use of the Crunch4 Statistical Package (Oakland, Calif.).
Results
Physical Therapy
Table 1Table 1
Distribution of AVGs and Mean Cost per Case in the Self-Referral and Independent-Referral Groups.* shows the 1257 cases of musculoskeletal injury (whether or not the patients received physical therapy) according to AVG and type of provider (whether the provider practiced self-referral or independent referral). Four AVGs account for 92 percent of all cases; there was no significant difference in the distribution of AVGs between the self-referral and independent-referral groups.As shown in Table 2Table 2
Frequency of Use of Services and Cost per Case in the Self-Referral and Independent-Referral Groups., physical therapy was initiated more than twice as often by physicians in the self-referral group (in 68 percent of the cases) as by those in the independent-referral group (30 percent; P<0.01). The mean cost per case for physical therapy in the self-referral group ($404±102) was significantly lower than that in the independent-referral group ($440±167; P<0.01).Psychiatric-Evaluation Services
Table 1 classifies the random sample of 220 cases in which patients received both psychometric testing and psychiatric-evaluation services, according to AVG and provider type. There was no significant difference in the distribution of AVGs between the two types of providers. As Table 2 shows, the mean cost per case for psychiatric-evaluation services was 26 percent higher in the self-referral group ($3,222±1,451) than in the independent-referral group ($2,550± 742; P<0.01). This difference was due to the higher cost of psychometric testing ($1,165±728 vs. $870±482; P<0.01) and the greater number of tests per case and to the higher costs of psychiatric-evaluation reports ($2,056±1,063 vs. $1,680±578; P<0.01) (since psychometric tests are reimbursed according to the California Official Medical Fee Schedule, which pays the same amount for each psychometric test regardless of the test, the cost per case for these reports is directly proportional to the number of tests performed).
MRI Scans
Tables 1 and 2 show the results of our study of the medical appropriateness of MRI scans. A total of 502 requests for precertification were received from imaging centers in which ownership could be identified. In Table 1, these cases are classified according to AVG and provider type. There was no significant difference in the distribution of cases between the self-referral and independent-referral groups.
As shown in Table 2, 38 percent of the scans requested by physicians in the self-referral group were found to be medically inappropriate, as compared with 28 percent of those requested by physicians in the independent-referral group (P<0.05). There was no significant difference in cost per MRI procedure between the two groups.
Discussion
This study demonstrates that self-referral increases the cost of medical care under workers' compensation for each of the three types of service studied, but by a different mechanism in each instance: by substantially increasing the percentage of injured workers who receive physical therapy (which more than offsets the slight decrease in cost per case); by increasing the number of psychometric tests and the cost of psychiatric-evaluation reports; and by increasing the frequency of requests for clinically inappropriate MRI scans. These higher rates of use and higher costs have important implications for workers' compensation expenditures, since self-referral is the predominant form of referral for these services.
Physical Therapy
According to the California Workers' Compensation Institute (CWCI) 1990 Medical Fee Survey of 39 private and public insurers, physical therapy represents 56 percent of all outpatient procedures and 34 percent of all outpatient costs for the treatment of injured workers in California.16 This represents an increase of 31 percent in the volume of services in relation to other outpatient procedures since the CWCI's 1988 study.16
Injured workers usually receive a prescription for treatment from a physician (an orthopedic specialist or physician at an industrial medical or multispecialty clinic) to the physical therapist for specific treatment. Over the years, many physicians and clinics that treat patients covered by workers' compensation have established physical-therapy departments within their general operations or have established separate physical-therapy facilities that they own but that are operated as distinct financial entities.
In Florida, Mitchel and Scott recently found that 40 percent of physical-therapy facilities were owned by physicians.7 Our study focused on California physicians who treat large numbers of musculoskeletal injuries and found that 91 percent of all physical therapy was performed by providers who engage in self-referral (Table 2), and the frequency with which physical therapy was initiated was 2.3 times greater in the self-referral group than the independent-referral group. The cost per case of physical therapy, however, was about 10 percent higher in the independent-referral group.
In this study, there was no significant difference in case mix between the self-referral and independent-referral groups (Table 1). In the absence of measures of severity of illness among outpatients, it is therefore impossible to determine whether the lower cost per case in the self-referral group reflects more efficient care or the provision of physical therapy to patients with less severe injuries, since self-referring practitioners initiate physical therapy at more than twice the rate of independent providers.
Regardless of which hypothesis is correct, this small difference in cost per case is more than offset by the dramatically greater frequency with which self-referring providers initiate physical therapy. As Table 3Table 3
Additional Cost Incurred by the California Workers' Compensation System for Each 1000 Injuries Treated at Self-Referral Rather Than Independent-Referral Rates. shows, for every 1000 workers with musculoskeletal injuries, the costs incurred by the California workers' compensation system would be $143,672 (110 percent) higher if these injured workers were evaluated by self-referring rather than independently referring practitioners.Psychiatric-Evaluation Services
The CWCI estimates that approximately 6 percent of the total medical payments under workers' compensation were for psychiatric services in 1991.16 California state law defines a valid claim of work-related stress as one in which the work environment contributes 10 percent or more to a worker's total stress level. Some argue that this definition of compensable workplace stress has created a referral environment that encourages excessive evaluation and testing.
We found that 70 percent of all psychiatric-evaluation services were requested by providers who had an ownership interest in the entity that provided both psychometric testing and psychiatric-evaluation reports (Table 2). Furthermore, evaluation costs were 26 percent higher when this ownership relation existed.
As indicated above, a referral for evaluation virtually always results in charges for two services: psychometric testing and a psychiatric-evaluation report that synthesizes the findings of the psychometric tests with the findings from the psychiatric history and examination. Therefore, if a provider with an economic interest in a facility were motivated more by monetary incentives than one without such an economic interest, we would expect this to be reflected in greater use and higher costs of psychometric testing, as well as a more extensive and therefore more costly evaluation report, which would be required to integrate the results of more extensive testing. As shown in Table 2, the cost of each psychiatric service and the mean cost per case were significantly higher in the self-referral group than the independent-referral group; the differences in cost were as follows: psychometric testing, 34 percent; psychiatric evaluation reports, 22 percent; and total evaluation, 26 percent.
As Table 3 shows, for every 1000 workers receiving psychiatric-evaluation services, the costs incurred by the California workers' compensation system would be $672,000 (26 percent) higher if these workers were treated by physicians in the self-referral group rather than the independent-referral group.
MRI Scans
MRI has gained prominence as the diagnostic imaging tool of choice in the assessment and documentation of specific types of injuries. California currently has approximately 400 MRI machines (Mitchel J: personal communication). Recent studies have shown that such a concentration of imaging centers is associated with higher rates of use. After adjustment for the characteristics of the population, Californians undergo 51 percent more MRI procedures than the national average.9 Leape et al. similarly concluded that an increased concentration of providers increases rates of use.17 In their study, regions with a high rate of carotid endarterectomy had twice as many surgeons performing the operation as regions where the rate was low.
We found MRI scans to be medically inappropriate 38 percent more often when ordered by self-referring physicians, suggesting increased rates of use in this group. The higher rate of inappropriateness in the self-referral group may help explain the Florida study's finding that rates of use in these physician-owned facilities were 14 to 65 percent higher than in a control area.7
Table 3 illustrates the effects of these requests for medically inappropriate scans. For every 1000 requests for MRI scans, the costs incurred by the California workers' compensation system would be $89,456 (31 percent) higher if these requests were made by self-referring physicians rather than by physicians in the independent-referral group.
We are indebted to Mr. Will B. Murphey for his support and encouragement in this project.
Source Information
From the National Medical Audit unit, William Mercer, Inc., San Francisco (A.S., G.J., A.M.); and Current Health Concepts, Sausalito, Calif. (N.S.). Address reprint requests to Dr. Johnson at William Mercer, Inc., 3 Embarcadero Ctr., Suite 1250, San Francisco, CA 94111.
- References
References
1
Relman
Full Text | Web of Science | Medline2
Iglehart
Full Text | Web of Science | Medline3
Financial arrangements between physicians and health care businesses: report to Congress. Bethesda, Md.: Department of Health and Human Services, Office of Inspector General, 1989.
4
Morreim
Web of Science | Medline5
Relman 6
Waldholz 7
Mitchell JM, Scott E. Joint ventures among health care providers in Florida. Vol. 2. Contract report for Florida Health Care Cost Containment Board, September 1991. Tallahassee: State of Florida, 1991.
8
Dyckman Z. Impact of physician joint venture activity on medical care costs in Florida. Columbia, Md.: Center for Health Policy Studies, 1992.
9
Idem. Impact of physician joint venture activity on costs of diagnostic imaging services in California. Columbia, Md.: Center for Health Policy Studies, 1992.
10
Hillman
Full Text | Web of Science | Medline11
Council on Ethical and Judicial Affairs, American Medical Association
CrossRef | Web of Science12
A.M.A. eases policy on self-referrals . New York Times. June 24, 1992: A12.
13
Department of Health and Human Services
Medline14
Schneider K, Lichtenstein JL, Fetter RB, Freeman JL, Newbold RC. The new ICD-9-CM ambulatory visit groups classification scheme definitions manual. New Haven, Conn.: Yale University Press, 1986.
15
Smithline
Medline16
Jones M. Medical fees, 1990. San Francisco: California Workers' Compensation Institute, 1991 (executive memo).
17
Leape
Full Text | Web of Science | Medline
- Citing Articles (28)
Citing Articles
1
(2011) Reply by Authors. The Journal of Urology 185:5, 1993-1994
CrossRef2
Mitchell S. Anscher, Barbara M. Anscher, Cathy J. Bradley. (2010) The Negative Impact of Stark Law Exemptions on Graduate Medical Education and Health Care Costs: The Example of Radiation Oncology. International Journal of Radiation Oncology*Biology*Physics 76:5, 1289-1294
CrossRef3
Edward J. Bernacki, Xuguang Tao, Larry Yuspeh. (2010) The Impact of Cost Intensive Physicians on Workersʼ Compensation. Journal of Occupational and Environmental Medicine 52:1, 22-28
CrossRef4
Morey J. Kolber. (2006) Stark Regulation: A Historical and Current Review of the Self-Referral Laws. HEC Forum 18:1, 61-84
CrossRef5
D BECKER, D KESSLER, M MCCLELLAN. (2005) Detecting Medicare abuse. Journal of Health Economics 24:1, 189-210
CrossRef6
John D. Goodson, Arlene S. Bierman, Oliver Fein, Kimberly Rask, Eugene C. Rich, Harry P. Selker. (2001) The Future of Capitation. The Physician Role in Managing Change in Practice. Journal of General Internal Medicine 16:4, 250-256
CrossRef7
Deyo, Richard A., Weinstein, James N., . (2001) Low Back Pain. New England Journal of Medicine 344:5, 363-370
Full Text8
Kenneth H Wathne, Jan B. Heide. (2000) Opportunism in Interfirm Relationships: Forms, Outcomes, and Solutions. Journal of Marketing 64:4, 36-51
CrossRef9
Richard A. Deyo, Marie Schall, Donald M. Berwick, Tom Nolan, Penny Carver. (2000) Continuous Quality Improvement for Patients with Back Pain. Journal of General Internal Medicine 15:9, 647-655
CrossRef10
Richard A. Deyo. (2000) Editorial: Quality and Its Dimensions. Spine 25:6, 661
CrossRef11
Arthur Oleinick, Jeremy V. Gluck, Kenneth E. Guire. (1998) Diagnostic and Management Procedures for Compensable Back Injuries Without Serious Associated Injuries. Spine 23:1, 93-110
CrossRef12
Alvin I. Mushlin, Cathleen Mooney, Robert G. Holloway, Allan S. Detsky, David H. Mattson, Charles E. Phelps. (1997) The Cost-Effectiveness of Magnetic Resonance Imaging for Patients with Equivocal Neurological Symptoms. International Journal of Technology Assessment in Health Care 13:01, 21
CrossRef13
Bruce J Hillman. (1997) Medical imaging in the 21st century. The Lancet 350:9079, 731-733
CrossRef14
Pablo Lázaro, Kathryn Fitch. (1996) Economic Incentives and the Distribution of Extracorporeal Shock wave Lithotripters and Linear Accelerators in Spain. International Journal of Technology Assessment in Health Care 12:04, 735
CrossRef15
Rollin M. Gallagher. (1996) Referral Delay in Back Pain Patients on Worker's Compensation. Psychosomatics 37:3, 270-284
CrossRef16
Woolhandler, Steffie, Himmelstein, David U., . (1995) Extreme Risk — The New Corporate Proposition for Physicians. New England Journal of Medicine 333:25, 1706-1708
Full Text17
David M. Schuster. (1995) Hospital 2000: The Interface of Radiology within a Combined "Complementary-Allopathic" Medicine Framework. The Journal of Alternative and Complementary Medicine 1:4, 333-337
CrossRef18
Laurence C. Baker, Alan B. Krueger. (1995) Medical costs in workers' compensation insurance. Journal of Health Economics 14:5, 531-549
CrossRef19
Jean M. Mitchell, Tim R. Sass. (1995) Physician ownership of ancillary services: Indirect demand inducement or quality assurance?. Journal of Health Economics 14:3, 263-289
CrossRef20
Dan R. Berlowitz, Amy K. Rosen, Mark A. Moskowitz. (1995) Ambulatory care casemix measures. Journal of General Internal Medicine 10:3, 162-170
CrossRef21
Leonard N. Matheson, Richard G. Brophy, Karel Douglas Vaughan, Christine Nunez, Katherine A. Saccoman. (1995) Workers' compensation managed care: Preliminary findings. Journal of Occupational Rehabilitation 5:1, 27-36
CrossRef22
Steven C. Hackett. (1995) Heterogeneity and Share Contracting in Medical Group Practices. International Journal of the Economics of Business 2:1, 103-121
CrossRef23
DEBI PRASADMISHRA. 1995. Signaling and monitoring strategies of service firms: Interdisciplinary perspectives. , 249-288.
CrossRef24
Thomas J. Smith, Christopher E. Desch, Bruce E. Hither. (1994) Ways to Reduce the Cost of Oncology Care Without Compromising the Quality. Cancer Investigation 12:2, 257-265
CrossRef25
(1993) Self-Referral by Physicians. New England Journal of Medicine 328:17, 1274-1279
Full Text26
Kassirer, Jerome P., . (1993) Medicine at Center Stage. New England Journal of Medicine 328:17, 1268-1269
Full Text27
(1992) Editorial comment. Journal of Occupational Rehabilitation 2:4, 171-173
CrossRef28
Relman, Arnold S.. (1992) Self-Referral — What's at Stake?. New England Journal of Medicine 327:21, 1522-1524
Full Text
