August 1998

Grant Results

SUMMARY

From 1994 to 1996, researcher teams at the University of Virginia and the Primary Care Outcomes Research Institute (PCORI) at the New England Medical Center conducted two distinct but parallel studies of:

An additional cost/utilization study by one of the research teams sought to determine if physicians who own their own diagnostic imaging equipment and charge for both performing and interpreting these tests are more likely to order more imaging tests per episode of care than physicians who refer patients to radiologists.

The research team operated under contract with the United Mine Workers of America Health and Retirement Funds.

Key Findings
Key findings of the two studies and the additional study are as follows:

  • The cost and utilization study suggests that the imaging reimbursement policy had no consistent impact on utilization or costs.
  • The health outcomes and patient satisfaction study suggests that the imaging reimbursement policy had little if any effect on the utilization practices of self-referring physicians.
  • The additional study of diagnostic imaging utilization trends among physicians found that after the new payment policy, there were increased numbers of claims and greater expenditures.
  • Whether one or the other study tells a more accurate story of utilization changes is not evident; however, neither supports the expected decrease in costs and utilization in response to the decrease in reimbursement.
  • The results suggest that even a 30 percent fee reduction may not affect utilization.

Funding
The Robert Wood Johnson Foundation (RWJF) supported this project through a grant of $200,000.

 See Grant Detail & Contact Information
 Back to the Table of Contents


THE PROBLEM

Diagnostic imaging studies accounted for 5 percent of the nation's $800 billion in health care costs in 1992. One controversial aspect of diagnostic imaging has been physician self-referral, i.e., physicians who refer patients for imaging procedures in their own offices or to freestanding centers in which they have a financial interest. With respect to self-referral in office practice, three studies (Levin et al., 1990; Hillman et al., 1990; and, Hillman et al., 1992) found that self-referring physicians charge significantly more than radiologists for comparable imaging examinations. Using the claims database of the United Mineworkers of America Health and Retirement Funds (Funds), the 1992 study found that self-referring physicians had higher utilization for all 10 clinical presentations that were tracked, and, adjusting for complexity of the cases, their charges were higher for 7 of the 10 clinical presentations. As a result of these findings, the Funds implemented a policy in January 1993 to limit the incentives for self-referral for imaging services among nonradiologist physicians. While still paying physicians the standard fee for taking the images, the new policy eliminated payments to self-referring, nonradiologists for reading the images, effectively reducing payments to these self-referring nonradiologists by about 30 percent. The Funds assumed that its costs would be reduced without reducing the quality of care for its beneficiaries.

 Back to the Table of Contents


THE PROJECT

The Funds contributed $258,864 to this project to assess whether utilization and cost of diagnostic imaging would decline without affecting quality. Under subcontracts to the Funds, researchers at the University of Virginia and the Primary Care Outcomes Research Institute (PCORI) at the New England Medical Center tested three hypotheses underlying this change in policy:

  • Self-referring physicians would evidence less use of diagnostic imaging in the period following the payment policy change relative to their utilization of imaging in the prior period.
  • Cost savings to the Funds associated with the decrease in utilization and reduced payment allowance for self-referred imaging would not be offset by increases in radiologist-referred imaging or increased utilization of other services.
  • Neither patient satisfaction nor complaint-specific or general health outcomes (adjusted for the underlying chronic disease burden of Funds' beneficiaries) would be diminished for Funds' beneficiaries relative to other patients of the same physicians.

Methodology

The project was designed as two distinct but parallel studies:

  1. an evaluation of changes in the Funds' costs and utilization of diagnostic imaging under the principal direction of Bruce J. Hillman, M.D., in the department of radiology, University of Virginia;
  2. an assessment of the effects of the policy on patient satisfaction and health outcomes under the principal direction of Ira B. Wilson, M.D., and Sheldon Greenfield, M.D., at PICORI, the New England Medical Center.

FIRST HEALTH Strategies, the Funds' health benefit claims administrator and managed care vendor at the time the grant was awarded, provided computer programming and other logistical support services for both studies. The Funds' field offices provided logistical support for the recruitment of physicians and patients into the patient satisfaction and health outcomes study.

Study 1. The first study was designed as a pre/post comparison of the periods prior to and following implementation of the new diagnostic imaging payment policy in January 1993 using methods, which replicated, to the extent possible, the methods developed and applied by Hillman et al. in their 1992 study of self-referral within the Funds' population. The Funds' claims-history database was used to generate data files on episodes of care for individuals with specific medical conditions, i.e., low back pain, upper respiratory complaints, a complex of urinary tract symptoms, and knee complaints. The pre-policy period covered the twelve months ending December 31, 1992, and coincided with implementation of the policy on January 1, 1993. The post-policy period covered the twelve months from July 1, 1993 to June 30, 1994. A six-month period, through June 30, 1993, was allowed for physicians to become familiar with the new policy.

Study 2. In the second study, a prospective approach for primary data collection was used to assess the effects of the new payment policy on patients' health outcomes and satisfaction, employing a combination of existing and newly developed survey instruments. This study was designed to compare differences in outcomes and satisfaction of Funds' beneficiaries treated after initiation of the new policy by a sample of self-referring physicians and radiologist-referring physicians. Other Medicare patients treated by these physicians who were not Funds' beneficiaries were included in the study to serve as a control population.

For physician recruitment, 420 generalist physicians who each treat more than 30 of the Funds' beneficiaries were asked to participate. Physicians were informed that the study was focusing on reimbursement policy in general, not the imaging policy in particular. To achieve a desired level of statistical significance, it was determined that 60 physicians were needed, each seeing 60 patients (30 Funds' and 30 Medicare beneficiaries). Four weeks after the office visit of a participating patient, a survey was mailed to obtain their evaluation of that episode of care. Of the 2,559 patients who initially agreed to participate in the study, 1,432 (56 percent) returned completed surveys (37 percent back, 20 percent knee, and 43 percent respiratory). Respondents were not different from nonrespondents in gender, age, insurance type, or medical condition.

An additional study. Hillman et al. also conducted a retrospective economic evaluation of claims and expenditures for diagnostic imaging examinations filed by physicians practicing in the 20 US counties that had the greatest number of Funds' beneficiaries, both in 1992 and 1993, after the new policy went into effect. Forty-five percent of all Funds' beneficiaries reside in these 20 counties. This analysis was not contemplated at the time of the grant but was undertaken when monitoring reports developed by FIRST HEALTH and the Funds revealed the existence of dramatic trend differences in these counties.

 Back to the Table of Contents


FINDINGS

The findings of the two studies and the additional study are as follows:

  • The cost and utilization study (1996) suggests that the imaging reimbursement policy had no consistent impact on utilization or costs. While, in general, imaging frequency declined slightly for all indications (back pain, knee pain, and respiratory symptoms) except for urinary tract symptoms, and for both self- and radiologist-referral, the magnitude of this trend was small, and frequently offset by changes in the number of exams performed per episode.
  • The health outcomes and patient satisfaction study (1996) suggests that the imaging reimbursement policy had little if any effect on the utilization practices of self-referring physicians. Self-referring physicians appear to perform imaging with the same frequency for both their Funds' and non-Funds' patients. Consistent with the absence of changes in imaging frequency, there were no apparent differences in outcomes or satisfaction comparing Funds' and non-Funds' patients and comparing self- and radiologist-referring physicians. The comparable levels of outcome and satisfaction suggest that the additional imaging performed by self-referring physicians ultimately makes little difference to the patient. An abstract published in the Journal of General Internal Medicine, "The Relationship of Patient Demand for Radiology Procedures to Utilization Rate," utilizes study data to suggest that patients are important and active participants in radiology-utilization decisions. Efforts to reduce radiology-utilization rates that do not involve patients may be unsuccessful or result in patient dissatisfaction.
  • The additional study by Dr. Hillman of diagnostic imaging utilization trends among physicians in the 20 counties with the highest concentrations of Funds' beneficiaries found that after the new payment policy, there were increased numbers of claims and greater expenditures. The study's results were published in "Responses to a Payment Policy Denying Professional Charges for Diagnostic Imaging by Nonradiologist Physicians," which appeared in the September 20, 1995 issue of the Journal of the American Medical Association (JAMA). Despite $811,466 in claims disallowed by the new policy, the Funds paid 12 percent more for diagnostic imaging performed in the 20 counties in 1993 than it did during 1992. The Funds also reimbursed 41 percent more claims per beneficiary for diagnostic imaging in 1993 than in 1992. These findings contradict the findings of Study 1.
  • "Whether one or the other study tells a more accurate story of utilization changes is not evident. However, neither supports the expected decrease in costs and utilization in response to the decrease in reimbursement," wrote Ralph Horn, assistant director of provider relations at the Funds in the final report to RWJF. He also notes that the 1996 cost and utilization study is more rigorous in its efforts to distinguish the behavior of self- and radiologist-referring physicians.
  • The results suggest that even a 30 percent fee reduction may not affect utilization.

The change to the Funds' imaging reimbursement policy implemented in 1993 was rescinded November 1, 1996.

Limitations

The study team from PCORI encountered more trouble than anticipated in recruiting physicians. After disappointing results in soliciting physicians by mail and having assistants make telephone calls, Drs. Wilson and Greenfield personally contacted each physician candidate by phone. Only 83 of 420 invited physicians agreed to participate, and of these, only 57 actually enrolled patients in the study. The 2,559 patients who agreed to participate were below the 3,000–4,000 originally targeted. Of those patients who agreed to participate, only 1,433 (56 percent) returned surveys, a response rate that was 15 to 20 percent lower than expected. The PCORI group still had confidence that the sample was sufficient to detect clinically important differences in health and satisfaction of Funds' and non-Funds' patients. The reduced numbers did affect the ability to perform analyses comparing subgroups. Explanations for the low-response rate included low-literacy levels and poor health in the elderly population.

The federal Medicare anti-self-referral law (Stark-II amendment) took effect on January 1, 1995, and imposed extremely strict penalties on self-referral to freestanding facilities for patients in Medicare, Medicaid, and some other health plans, but it exempted in-office settings with which the referring physician had a relationship. Since in-office referral is, by far, the predominant form of self-referral, it is not clear what effect, if any, the implementation of this law had on responses to the patient outcomes and satisfactions surveys, 70 percent of which were conducted after January 1, 1995. The pre- and post-reimbursement policy study periods for the cost and utilization study were well before the implementation date for the law; therefore, that study was not affected.

Communications

The results of the additional study were published in JAMA in September 1995. Several articles on the results of the two main studies were in progress at the end of the grant period but have not been published. An abstract appeared in the Journal of General Internal Medicine.

 Back to the Table of Contents


LESSONS LEARNED

  1. Studies that involve surveys of physicians may require extra time to be completed because physicians may be unresponsive to mailed surveys, necessitating personal contact. The researchers for the second study had to make personal phone calls to recruit physicians, and even then got a far smaller response than they had initially anticipated. This smaller sample affected their ability to conduct analyses of subgroups.
  2. Patient literacy needs to be taken into account when designing studies involving surveys of Medicare patients. The response rate of patients in the second study was also less than expected, which the researchers attributed in part to low literacy among the patients being surveyed.

 Back to the Table of Contents


GRANT DETAILS & CONTACT INFORMATION

Project

Impact of Curtailing Payment for Imaging in Referring Doctor's Facility

Grantee

United Mine Workers of America Combined Benefit Fund (Washington,  DC)

  • Amount: $ 200,000
    Dates: June 1994 to June 1996
    ID#:  022799

Contact

Ralph Horn
(814) 262-1006

 Back to the Table of Contents


BIBLIOGRAPHY

(Current as of date of this report; as provided by grantee organization; not verified by RWJF; items not available from RWJF.)

Articles

Hillman BJ, Olson GT, Colbert RW, and Bernhardt LB. "Responses to a Payment Policy Denying Professional Charges for Diagnostic Imaging by Nonradiologist Physicians," Journal of the American Medical Association, 274(11): 885–887, 1995. Abstract available online.

Wilson IB, Dukes K, Greenfield S, Kaplan S, and Hillman B. "Patients' Role in the Use of Radiology Testing for Common Office Practice Complaints." Archives of Internal Medicine, 161(2): 256–263, 2001. Abstract available online.

Wilson IB, Kaplan SH, Dukes K, Tripp T, Greenfield S. "The Relationship of Patient Demand for Radiology Procedures to Radiology Utilization." (abstract) Journal of General Internal Medicine, 11(Supplement): 93, 1996.

Wilson IB, Greenfield S, Kaplan SH, et al. "Variations in Imaging Rates Among Self- and Radiologist-Referring Physicians: Relationships to Health Outcomes and Satisfaction." In progress.

Presentations and Testimony

Ira B Wilson and Sheldon Greenfield, "The Relationship of Patient Demand for Radiology Procedures to Utilization Rate," Abstract read to the 1996 National Meeting of the Society of General Internal Medicine.

 Back to the Table of Contents


Report prepared by: Todd Shapera
Reviewed by: Timothy F. Murray
Reviewed by: Molly McKaughan
Program Officer: Lewis G. Sandy

Most Requested