HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Maitino, A. J. Articles by Sunshine, J. H. Search for Related Content PubMed PubMed Citation Articles by Maitino, A. J. Articles by Sunshine, J. H.

Practice Patterns of Radiologists and Nonradiologists in Utilization of Noninvasive Diagnostic Imaging among the Medicare Population 1993–1999¹

¹ From the Department of Radiology, Thomas Jefferson University, Suite 3390, Gibbon Bldg, 111 S 11th St, Philadelphia, PA 19107 (A.J.M., D.C.L., L.P., V.M.R.); and American College of Radiology, Reston, Va (J.H.S.). From the 2001 RSNA scientific assembly. Received October 2, 2002; revision requested December 12; revision received December 30; accepted January 27, 2003. Address correspondence to A.J.M. (e-mail: andrea.maitino@jefferson.edu).

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
PURPOSE: To compare nationwide trends in noninvasive diagnostic imaging (NDI) practice patterns of radiologists and of nonradiologists among the Medicare population during the 6 years from 1993 to 1999.

MATERIALS AND METHODS: Medicare Part B claims files from 1993, 1996, and 1999 were analyzed for all procedure codes related to NDI. NDI codes were classified into 22 diagnostic categories within seven imaging modality groups. For each NDI code, physicians performing the services were classified as radiologists or nonradiologists by using the provider specialty code designated in claims in the files. The data were analyzed to determine the overall utilization rates and relative value unit (RVU) rate changes between 1993 and 1999 among radiologists and nonradiologists.

RESULTS: In 1993, the overall NDI utilization rate per 100,000 Medicare fee-for-service beneficiaries was 215,652 for radiologists and 79,942 for nonradiologists. In 1999, the rate was 207,270 for radiologists and 100,059 for nonradiologists, which is a 3.9% decrease among radiologists and a 25.2% increase among nonradiologists. In the 6-year interval from 1993 to 1999, the overall RVU rate increased 6.9% among radiologists and 32.4% among nonradiologists. The percentage of NDI performed by radiologists decreased from 73.0% in 1993 to 67.4% in 1999.

CONCLUSION: Overall, the utilization rate of advanced, high-technology imaging is increasing among both radiologists and nonradiologists. However, it is increasing at a considerably more rapid rate among nonradiologists.

© RSNA, 2003

Index terms: Diagnostic radiology • Economics, medical • Radiology and radiologists, departmental management • Radiology and radiologists, socioeconomic issues

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
It has been previously shown that in the years between 1993 and 1999, the utilization of noninvasive diagnostic imaging (NDI) in the Medicare fee-for-service population increased 3.8% and the work associated with the imaging, as determined with relative value unit (RVU) rates per 100,000 Medicare fee-for-service beneficiaries, increased 14.6% (1). The increase in RVU rates was associated with the increased use of more complex, high-technology imaging examinations. Health policy planners have been concerned with the increase in utilization of diagnostic imaging and high-technology medicine, but the earlier study to which we referred suggested that the increase in diagnostic radiology work in a 6-year period was moderate (1).

Exploring these increases in relation to the relative shares of procedures that are performed by radiologists and nonradiologists is of interest because of the issues of self-referral and image quality. In two important studies by Hillman et al (2,3), a considerably higher utilization of imaging examinations was noted among physicians who had the opportunity to self-refer than among physicians who referred their patients to radiologists. The same phenomenon was identified throughout the Florida Medicare population, as noted in a report by the United States General Accounting Office (4). It has also been reported that there are apparent quality problems among the images obtained by nonradiologists. For example, in a pilot project in which 98 chest radiographs were reviewed at Hershey Medical Center, inadequate image quality was found in 3% of images obtained by radiologists but in 28% of images obtained by nonradiologists (5). The purpose of our study was to compare practice patterns between radiologists and nonradiologists performing NDI examinations among the Medicare population during the 6 years from 1993 to 1999.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
The methods used in this study have previously been described (1). The nationwide Centers for Medicare and Medicaid Service Part B Physician/Supplier Procedure Summary Master Files (the Medicare Part B databases) for 1993, 1996, and 1999 were evaluated for all procedure codes related to NDI. The Medicare Part B databases are anonymous public files and are therefore exempt from review by institutional review board. For each NDI procedure code in the Medicare Part B databases, we classified the physicians performing the services as radiologists or nonradiologists by using the provider specialty code designated in claims in the files. Those NDI examinations in which the specialty of the provider was classified as a multispecialty group (without further indicating the specialty of the physician performing the examination) were excluded from analysis, although the overall volumes and percentages of examinations were tabulated.

The overall utilization rates and Medicare total professional component RVU rates per 100,000 Medicare fee-for-service beneficiaries were calculated for 1993, 1996, and 1999 among radiologists and nonradiologists. NDI examinations were classified into seven imaging modalities, which were then further classified into 22 diagnostic categories. The seven imaging modalities and their component diagnostic categories were as follows: (a) radiography (chest, skeletal, abdominal, gastrointestinal fluoroscopy); (b) mammography; (c) ultrasonography (US) (general, vascular, breast, echocardiography, obstetric); (d) computed tomography (CT) (body, cranial, spinal, musculoskeletal); (e) nuclear imaging (general, cardiovascular); (f) magnetic resonance (MR) imaging (cranial, spinal, musculoskeletal, body, cardiovascular); and (g) bone densitometry.

Four of the investigators (A.J.M., D.C.L., L.P., V.M.R.) reviewed each of the CPT-4 codes in the 70000 series and the echocardiography and vascular US codes in the 90000 series. Each code that represented an NDI procedure was assigned to the appropriate category and modality. SAS 8e Release 2 for Windows (SAS Institute, Cary, NC) was then used to tabulate the number of procedures in each of the categories and modalities. The changes in percentage of NDI utilization and RVU rates among radiologists and nonradiologists during the 6-year period between 1993 and 1999 were calculated for each NDI modality and category. Since cardiovascular imaging is often self-referred (6), we then separately analyzed these trends for all cardiovascular imaging and compared them with those for all noncardiovascular imaging. For this purpose, cardiovascular imaging included four categories: echocardiography, vascular US, cardiovascular nuclear imaging, and cardiovascular MR imaging. The 1999 RVU scale was used to calculate the RVU rates for 1993 and 1996 rather than the RVU scale for those particular years. We used these "synthetic" RVU rates because the assigned RVUs for some codes change from year to year, and we believed it was important to use a single, consistent scale in order to ascertain changes in the relative amount of work. Since the Medicare Part B database contains information on a complete population, no inferential statistical analysis is required, which would have been the case if we had been trying to infer population statistics from sample data.

The RVU rates represent a proxy for the relative amount of work associated with each procedure. On the basis of an earlier recommendation by the American College of Radiology, we made the assumption that screening mammography should carry 80% of the professional component RVUs of diagnostic mammography.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Table 1 shows the overall volume of NDI performed by radiologists, nonradiologists, and multispecialty groups in 1993, 1996, and 1999. In 1993, a total of 104,231,053 NDI examinations were performed among the Medicare fee-for-service population. Radiologists performed 71,812,070 (68.9%) examinations and nonradiologists performed 26,620,918 (25.5%) examinations. In 1999, a total of 109,029,422 NDI examinations were performed. Radiologists performed 69,539,392 (63.8%) examinations, which is a 5.1% reduction in their share, while nonradiologists performed 33,569,932 (30.8%) examinations, which is a 5.3% increase in their share. The share of NDI examinations performed by multispecialty groups remained relatively constant in the 6-year period: 5,798,065 (5.6%) in 1993, 6,063,549 (5.9%) in 1996, and 5,920,098 (5.4%) in 1999.

The RVU rates per 100,000 Medicare fee-for-service beneficiaries for 1993, 1996, and 1999 are shown in Table 3. In 1993, the overall RVU rate was 139,526 for radiologists and 57,836 for nonradiologists. In 1999, this rate increased to 149,208 (an increase of 6.9%) for radiologists and 76,551 (an increase of 32.4%) for nonradiologists. Table 4 lists the percentage of NDI examinations performed by radiologists and nonradiologists by category and modality in 1993 and 1999. The overall percentage of NDI performed by radiologists decreased from 73.0% in 1993 to 67.4% in 1999. Their rate of participation decreased in 15 of the 22 imaging categories. Substantial reductions in percentage of imaging performed by radiologists were observed for general US (from 75.1% in 1993 to 65.3% in 1999), cardiovascular nuclear imaging (from 58.8% to 34.6%), and musculoskeletal MR imaging (from 97.0% to 89.0%). The percentages of NDI RVU rates among radiologists and nonradiologists are listed by category and modality in Table 5. The radiologists’ overall share of RVUs decreased from 70.7% in 1993 to 66.1% in 1999, which is a decrease of 4.6%. Radiologists’ share of RVUs decreased in 16 of the 22 categories; the largest reductions were in cardiovascular nuclear imaging (from 58.3% in 1993 to 36.1% in 1999), obstetric US (from 53.7% to 40.9%), and musculoskeletal MR imaging (from 97.0% to 89.1%).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Between 1993 and 1999, the overall utilization rates of NDI decreased 3.9% among radiologists but increased 25.2% among nonradiologists. During the same period, the overall RVU rates increased 6.9% among radiologists and 32.4% among nonradiologists. Results of a previous study showed that overall, the utilization of NDI increased only 3.8% between 1993 and 1999 (1). Our data reveal the somewhat surprising fact that the utilization actually decreased among radiologists but sharply increased among nonradiologists. This indicates that almost all growth in the utilization rates of NDI is attributable to imaging performed by nonradiologists.

Cardiovascular imaging procedures (echocardiography, vascular US, cardiovascular nuclear imaging, and cardiovascular MR imaging) constitute a considerable proportion of the increase in utilization of NDI by nonradiologists. Levin et al (6) showed that in 1998, 83.3% of all cardiovascular NDI was performed by nonradiologists and that 61.5% of this was performed by cardiologists. Most of this disparity was related to cardiologists’ strong domination of echocardiography, in which they performed 79.8% of imaging. In our analyses, the rate of utilization of cardiovascular imaging by nonradiologists is so substantial that when other categories of NDI are analyzed separately from it, there is little change in the utilization of NDI by nonradiologists. Utilization of noncardiovascular NDI between 1993 and 1999 increased only 0.3% by nonradiologists and decreased 5.4% by radiologists. Examining the trends in RVU rates shows a substantially greater increase in the RVU rates among nonradiologists than among radiologists for cardiovascular NDI (42.2% among nonradiologists, 28.9% among radiologists) and noncardiovascular NDI (13.7% among nonradiologists, 5.5% among radiologists). Among nonradiologists, however, the rates of both cardiovascular and noncardiovascular NDI utilization and RVU remained relatively stable during the 6-year period.

When we focused on noncardiovascular advanced imaging (US, CT, MR imaging, nuclear imaging, and bone densitometry), we found that radiologists perform far more examinations per 100,000 beneficiaries than do nonradiologists, but nonradiologists are increasing their utilization more rapidly. In 1993, the utilization rate of such examinations was 4,453 per 100,000 among nonradiologists and 41,577 among radiologists; radiologists performed more than nine times as many examinations. By 1999, the utilization rate had increased 110.0% to 9,351 among nonradiologists and had increased 22.7% to 51,003 among radiologists; radiologists performed more than five times as many examinations.

Aside from cardiovascular imaging, there are four other imaging categories in which nonradiologist participation is substantial. These are bone densitometry, skeletal radiography, chest radiography, and general US. Bone densitometry was a newly reimbursable technology in 1993 that gained popularity because of a new focus on osteoporosis screening. Consistent with this growth in popularity, large increases in utilization rates of bone densitometry were noted for both radiologists and nonradiologists, which indicates that the trends were not related to different practice patterns. Utilization of skeletal and chest radiography are declining at about the same rate among radiologists and nonradiologists, which again suggests similar practice patterns. General US, by contrast, is a category in which utilization is decreasing among radiologists and increasing among nonradiologists. Given the relative absence of the cost and safety barriers that limit the use of other technologies, this increase in US utilization by nonradiologists most likely results from self-referral (2,3), although wider dispersion of technology and the establishment of new diagnostic uses for US could also help explain the increase. These latter two explanations seem unlikely as major causes. Such divergence in utilization trends between radiologists and nonradiologists suggests the need for further exploration.

There were limitations in this study. This was a secondary analysis of an administrative Medicare data set. We made the assumption that the specialty of the physician who billed for the service is the same as that of the physician who actually performed the service, but this may not always be the case. The data set contained only pure utilization information that applied to a restricted set of categories. There were no diagnosis codes in this data set, and we were unable to study individual patients during episodes of care. We cannot determine appropriateness with data sets like this, and, further, we cannot explore other areas of interest like quality or outcomes. Another limitation is that our database covers only the Medicare fee-for-service population and not those enrolled in the Medicare managed care plans. Some researchers believe that the healthier segment of the population has a greater tendency to migrate to managed care plans and that utilization rates for NDI performed on this population might therefore differ from those presented in this study. However, there is no firm evidence that this is the case.

In conclusion, although radiologists performed more than two-thirds of all NDI examinations in 1999, the utilization of NDI by radiologists has decreased 3.9% between 1993 and 1999, while utilization by nonradiologists has increased by 25.2%. Thus, virtually all increase in utilization rates of NDI was attributable to an increase in utilization by nonradiologists. Most of the discrepancy is due to the considerable increase in the number of cardiovascular imaging examinations performed by nonradiologists. Utilization of cardiovascular imaging increased 66.8% among nonradiologists and 47.3% among radiologists during the interval of this study. If cardiovascular imaging is excluded from the analysis, radiologists performed more than three-quarters of NDI in 1999; utilization by radiologists decreased 5.4% during the 6 years and utilization by nonradiologists increased by 0.3%. Overall, the rate of utilization of advanced, high-technology imaging is increasing among both radiologists and nonradiologists; it is increasing at a considerably more rapid rate among nonradiologists. Further research will be needed to assess the appropriateness of the accelerating use of NDI by nonradiologists and the image quality that results.

	FOOTNOTES

 
Abbreviations: NDI = noninvasive diagnostic imaging, RVU = relative value unit

Author contributions: Guarantors of integrity of entire study, all authors; study concepts, all authors; study design, D.C.L., L.P., V.M.R.; literature research, A.J.M.; data acquisition, J.H.S.; data analysis/interpretation, A.J.M., D.C.L., L.P.; statistical analysis, L.P.; manuscript preparation, A.J.M.; manuscript definition of intellectual content, all authors; manuscript editing, A.J.M., D.C.L., L.P.; manuscript revision/review and final version approval, all authors

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 

1. Maitino AJ, Levin DC, Parker L, Rao VM, Sunshine JH. Nationwide trends in rates of utilization of noninvasive diagnostic imaging among the Medicare population between 1993 and 1999. Radiology 2003; 227:113-117.[Abstract/Free Full Text]
2. Hillman BJ, Joseph CA, Mabry MR, Sunshine JH, Kennedy SD, Noether M. Frequency and costs of diagnostic imaging in office practice: a comparison of self-referring and radiologist-referring physicians. N Engl J Med 1990; 323:1604-1608.[Abstract]
3. Hillman BJ, Olson GT, Griffith PE, et al. Physicians’ utilization and charges for outpatient diagnostic imaging in a Medicare population. JAMA 1992; 268:2050-2054.[Abstract]
4. Subcommittee on Health. Committee on Ways and Means. House of Representatives. Referrals to physician-owned imaging facilities warrant HCFA’s scrutiny: report to the chairman. GAO/HEHS-95-2 Washington, DC: U.S. General Accounting Office, 1994; 1-61.
5. Hopper KD, Rosetti GF, Edminston RB, et al. Diagnostic radiology peer review: a method inclusive of all interpreters of radiographic examinations regardless of specialty. Radiology 1990; 177:335-339.[Abstract/Free Full Text]
6. Levin DC, Parker L, Sunshine JH, Pentecost MJ. Cardiovascular imaging: who does it and how important is it to the practice of radiology? AJR Am J Roentgenol 2002; 178:303-306.[Abstract/Free Full Text]

This article has been cited by other articles:

				G. W. L. Boland Enhancing CT Productivity: Strategies for Increasing Capacity Am. J. Roentgenol., July 1, 2008; 191(1): 3 - 10. [Abstract] [Full Text] [PDF]

				G. S. Gazelle, E. F. Halpern, H. S. Ryan, and A. C. Tramontano Utilization of Diagnostic Medical Imaging: Comparison of Radiologist Referral versus Same-Specialty Referral Radiology, November 1, 2007; 245(2): 517 - 522. [Abstract] [Full Text] [PDF]

				A. S. Pearlman, T. Ryan, M. H. Picard, and P. S. Douglas Evolving Trends in the Use of Echocardiography: A Study of Medicare Beneficiaries J. Am. Coll. Cardiol., June 12, 2007; 49(23): 2283 - 2291. [Abstract] [Full Text] [PDF]

				J. M. Boone Multidetector CT: Opportunities, Challenges, and Concerns Associated with Scanners with 64 or More Detector Rows. Radiology, November 1, 2006; 241(2): 334 - 337. [Full Text] [PDF]

				G W L Boland Government reform of the National Health Service: implications for radiologists and diagnostic services Br. J. Radiol., November 1, 2006; 79(947): 861 - 865. [Abstract] [Full Text] [PDF]

				M. E. Detsky, D. R. McDonald, M. O. Baerlocher, G. A. Tomlinson, D. C. McCrory, and C. M. Booth Does this patient with headache have a migraine or need neuroimaging? JAMA, September 13, 2006; 296(10): 1274 - 1283. [Abstract] [Full Text] [PDF]

				D. D. Saket, C. C. Nwanze, C. D. Maynard, J. H. Sunshine, and H. P. Forman Update on the Diagnostic Radiologist Employment Market: Findings Through 2004 Am. J. Roentgenol., December 1, 2005; 185(6): 1408 - 1415. [Abstract] [Full Text] [PDF]

				L. Berlin Errors of Omission Am. J. Roentgenol., December 1, 2005; 185(6): 1416 - 1421. [Full Text] [PDF]

				M. Bhargavan and J. H. Sunshine Workload of Radiologists in the United States in 2002-2003 and Trends Since 1991-1992 Radiology, September 1, 2005; 236(3): 920 - 931. [Abstract] [Full Text] [PDF]

				M. T. Beinfeld and G. S. Gazelle Diagnostic Imaging Costs: Are They Driving Up the Costs of Hospital Care? Radiology, June 1, 2005; 235(3): 934 - 939. [Abstract] [Full Text] [PDF]

				A. W. Litt, D. R. Ryan, D. Batista, K. N. Perry, R. S. Lewis, and J. H. Sunshine Relative Procedure Intensity with Self-Referral and Radiologist Referral: Extremity Radiography Radiology, April 1, 2005; 235(1): 142 - 147. [Abstract] [Full Text] [PDF]

				L. Berlin and J. W. Berlin Leasing Imaging Facilities to Referring Physicians: Fee Shifting or Fee Splitting? Radiology, January 1, 2005; 234(1): 44 - 48. [Full Text] [PDF]

				J. H. Sunshine, C. D. Maynard, J. Paros, and H. P. Forman Update on the Diagnostic Radiologist Shortage Am. J. Roentgenol., February 1, 2004; 182(2): 301 - 305. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Maitino, A. J. Articles by Sunshine, J. H. Search for Related Content PubMed PubMed Citation Articles by Maitino, A. J. Articles by Sunshine, J. H.