H e a l t h C a r e Po l i c y a n d Q u a l i t y • O r i g i n a l R e s e a r c h Amrhein et al. Physician MRI Equipment Ownership and Use of Shoulder MRI Health Care Policy and Quality Original Research

Downloaded from www.ajronline.org by 37.44.207.103 on 01/26/17 from IP address 37.44.207.103. Copyright ARRS. For personal use only; all rights reserved

JOURNAL CLUB:

JOURNA L

CLUB

Timothy J. Amrhein1 Matthew P. Lungren2 Ben E. Paxton2 Ramesh Srinivasan 3 Sin-Ho Jung 4 Miao Yu5 James D. Eastwood2 Ramsey K. Kilani2 Amrhein TJ, Lungren MP, Paxton BE, et al.

Keywords: imaging utilization, medical economics, MRI, self-referral

Shoulder MRI Utilization: Relationship of Physician MRI Equipment Ownership to Negative Study Frequency OBJECTIVE. The purpose of this article is to determine whether ownership of MRI equipment by ordering physicians influences the frequency of negative shoulder MRI scans. MATERIALS AND METHODS. A retrospective review was performed of 1140 consecutive shoulder MRI scans ordered by two separate referring physician groups serving the same geographic community. The first group (financially incentivized) owned the scanners used and received technical fees for their use. The second group (non–financially incentivized) did not own the scanners used and had no direct financial interest. All examinations were performed with identical protocols and were interpreted by a single radiologist group without financial interest in the imaging equipment used. The frequency of negative examinations and the number of abnormalities in each positive study was tabulated for each group. RESULTS. A total of 1140 shoulder MRI scans met inclusion criteria; 255 were negative (142 for the financially incentivized group and 113 for the non–financially incentivized group). There were 25.6% more negative scans in the financially incentivized group (p = 0.047). There was no statistically significant difference in the average number of lesions per positive scan (1.67 for the financially incentivized group and 1.71 for the non–financially incentivized group; p = 0.34). No statistically significant difference was found in the frequency of 19 of 20 examined lesions. CONCLUSION. Shoulder MRI examinations referred by physicians with a financial interest in the imaging equipment used were significantly more likely to be negative. Positive examinations exhibited no statistically significant difference in the number of lesions per scan or in the frequency of 19 of 20 lesion subtypes. This finding suggests a highly similar distribution and severity of disease among the two patient groups.

DOI:10.2214/AJR.12.9977 Received September 17, 2012; accepted after revision November 14, 2012. 1 Department of Radiology and Radiological Science, Medical University of South Carolina, MSC 322, 169 Ashley Ave, Charleston, SC 29425. Address correspondence to T. J. Amrhein ([email protected]). 2 Department of Radiology, Duke University Medical Center, Durham, NC. 3 Department of Orthopedics, University of Michigan, Ann Arbor, MI. 4 Department of Biostatistics and Bioinformatics, Duke University, Durham, NC. 5

Center for Human Genetics, Department of Medicine, Duke University Medical Center, Durham, NC.

AJR 2013; 201:605–610 0361–803X/13/2013–605 © American Roentgen Ray Society

D

iagnostic imaging is a large component of annual health care expenditure. There have been a variety of efforts to rein in imaging-related costs, including the Deficit Reduction Act of 2007. However, even before the advent of the Deficit Reduction Act, the growth of noninvasive imaging had slowed considerably, from a 4.1% compound annual growth rate between 1998 and 2005 to a 1.4% compound annual growth rate between 2005 and 2008. Although overall medical imaging utilization growth has slowed among both radiologists and nonradiologist physicians, the growth rate of the latter group has remained twice that of radiologists [1]. Physicians referring their own patients for imaging to facilities in which they or their partners have a financial interest has been referred to as “self-referred imaging” [2–4]. Self-referral of imaging studies has been sin-

gled out as a potentially important factor contributing to rising health care costs because, critics assert, financial interest can lead to overutilization. Previously published studies have found that the practice of imaging selfreferral is increasing and that physicians who own diagnostic imaging equipment or facilities are more likely to refer their patients for imaging at facilities in which they have a financial interest [5–15]. Despite these observations of increased utilization, the assertion that self-referral leads to inappropriate overutilization, however, remains controversial, in part because data about rates and severity of illness, as well as details on various financial arrangements of imaging equipment ownership, have not been available. Many previous studies that have attempted to link imaging self-referral to overutilization have been criticized mainly for their inability to assess the appropriateness of medi-

AJR:201, September 2013 605

Downloaded from www.ajronline.org by 37.44.207.103 on 01/26/17 from IP address 37.44.207.103. Copyright ARRS. For personal use only; all rights reserved

Amrhein et al. cal imaging utilization. Determining the appropriateness of medical imaging utilization is a complex and challenging task confounded by multiple factors, including clinical setting (hospital vs outpatient practice), disease prevalence (young vs elderly patients), referral biases (e.g., orthopedic vs neurologist), disease diagnosis inhomogeneity (“grading” of lesions), and terminology differences between interpreting radiologists. Much of the previous work investigating driving factors in imaging self-referral has focused on the volume of patients referred for imaging, on the basis of International Classification of Diseases, Ninth Revision, code, between groups of referring physicians to assess for differences in imaging utilization characteristics. This approach may be of limited value because the proportion of normal examinations and final imaging diagnoses were not evaluated. However, information about the rate of normal examinations (“negative studies”), as well as the prevalence and severity of disease within the subject population, are likely important factors when attempting to measure utilization appropriateness. In addition, other important information regarding interpreting physicians, imaging equipment, imaging protocols, and financial arrangements would be needed to avoid significant bias. To our knowledge, one prior study in the literature has documented negative study rates between two referring physician groups to assess imaging appropriateness [15]. However, that study evaluated findings on lumbar spine MRI scans only; thus, there is a need to determine whether these results can be validated across different types of studies, body regions, and geographic practice locations. With regard to advanced imaging use in the clinical workup of shoulder abnormalities, data exist to support that most patients do not require MRI and instead support the assertion that the clinical examination is sufficient to decide further therapy. For example, when evaluating a patient for a superior labrum anterior to posterior (SLAP) lesion of the shoulder, traditional provocative physical examination maneuvers have been found to be more sensitive for the diagnosis than MRI and MR arthrography [16]. A comparison of the proportion of negative shoulder MRI examinations, after controlling for potential confounding variables, may allow an accurate assessment of the differences in imaging referral patterns between two physician groups (financially incentivized and non–financially incentivized); not only would this

606

validate previously published studies on imaging appropriateness, but it also would add to the existing body of literature addressing the issue of self-referral. Furthermore, an analysis of the prevalence and severity of imaging-confirmed injuries between these groups may serve as a surrogate for general disease prevalence within the two different patient populations, because groups with equal disease prevalence should manifest an equivalent number of positive MRI findings. The purpose of our study was to determine whether the presence of financial interest in an ordering physician group affects utilization of shoulder MRI. We set out to test the null hypothesis that no difference in utilization exists between physician groups with self-referred imaging practices and physician groups that do not engage in self-referred imaging. To do this, we tested two subordinate hypotheses: there is no difference in the rate of negative MRI scans between self-referred and non–self-referred groups, and there is no difference, among positive scans, in individual lesion prevalence in the two groups.

tions with the following findings were considered positive: rotator cuff tear, subscapularis tendon tear, rim rent tear, labral tear (including subtypes SLAP, anterior labroligamentous periosteal sleeve avulsion, Bankart, and bony Bankart), biceps tendon tear, biceps tenosynovitis, humeral avulsion of the glenohumeral ligament, acute Hill-Sachs lesion or fracture, and neurogenic myopathies (Parsonage-Turner syndrome and quadrilateral space syndrome). A separate category was established to include lesions that would typically require surgical intervention but did not fit into one of the aforementioned categories (i.e., fractures of humerus and scapula, pectoralis major tendon tears, metastatic disease, avascular necrosis of the humeral head, suprascapular notch, and spinoglenoid notch cysts). Patient age was recorded. The percentage of negative scans was tabulated for each group. The relative frequency of each diagnosis was calculated among the positive studies in each group. In two-sample comparisons, Fisher exact test was used for the binomial variables, and Wilcoxon rank sum test was used for a continuous variable or ordered categoric variables. All p values reported here are two sided.

Materials and Methods

Results Negative Scans A total of 1140 shoulder MRI examinations (570 in the financially incentivized group and 570 in the non–financially incentivized group) met inclusion criteria. The mean patient age was 50 years (range, 13–92 years) for the financially incentivized group and 50 years (range, 11–86 years) for the non–financially incentivized group, which was not statistically significantly different (p = 0.65). Among the 1140 scans, 255 were negative (142 in the financially incentivized group and 113 in the non–financially incentivized group). There were 25.6% more negative scans in the financially incentivized group, which was statistically significant (p = 0.047) (Fig. 1).

This retrospective HIPAA-compliant study was approved by the medical center institutional review board, and a waiver of informed consent was granted. Chronologically consecutive shoulder MRI reports were reviewed from January to September of 2009 in one academic musculoskeletal imaging practice consisting of four attending radiologists. MRI scans that met inclusion criteria were analyzed from each of two groups in tandem such that each cohort would, by design, have the same number of studies. The first group was ordered by community orthopedists and had imaging performed on MRI equipment owned by that referral group (financially incentivized). The second group of scans was ordered by a different group of orthopedists in the same community who did not own or have other financial interest in the MRI equipment used (non–financially incentivized). The source of the database, all physician groups, and their locations are purposely kept anonymous. All shoulder MRI examinations from both referring physician groups (financially incentivized and non–financially incentivized) were performed at 1.5-T field strength using identical protocols and were interpreted by the same subspecialty musculoskeletal radiology practice. The interpreting radiology practice had no financial interest in the imaging equipment used for either patient group. Patients with prior shoulder surgery or prior ipsilateral shoulder MRI examinations were excluded. Final reports were reviewed, and examina-

Positive Scans Among the positive scans, there was no statistically significant difference in the average total number of lesions per scan (1.67 for the financially incentivized group and 1.71 for the non–financially incentivized group; p = 0.34) (Figs. 2 and 3). No statistically significant difference was found in the frequency of the following lesions between the two groups: full rotator cuff tear (p = 0.1367), full subscapularis tendon tear (p = 0.3446), partial subscapularis tendon tear (p = 0.9238), rim rent tear (definite) (p = 0.1004), rim rent tear (possible) (p = 0.1652), cuff tendon in-

AJR:201, September 2013

Negative Scans (%)

25

20

15

Financially Incentivized

Non−Financially Incentivized

Physician Referral Group p = 0.047

2.0 1.5 1.0 0.5 0.0

Financially Incentivized

Non−Financially Incentivized

Physician Referral Group p = 0.34

Fig. 1—Percentage of negative scans per physician group.

Fig. 2—Average total number of lesions per positive scan.

jury present (p = 0.9385), labral tear (definite) (p = 0.8177), labral tear (partial) (p = 0.3090), SLAP injury (definite) (p = 0.3429), SLAP injury (possible) (p = 0.3585), anterior labroligamentous periosteal sleeve avulsion (p = 0.625), Bankart (p = 0.6289), bony Bankart (p = 0.7743), labral injury present (p = 0.634), biceps tendon tear (full) (p = 0.8801), biceps tendon tear (partial) (p = 0.0786), biceps tenosynovitis (p = 0.1632), biceps tendon abnormality present (p = 0.9316), humeral avulsion of the glenohumeral ligament (p = 0.3274), Hill-Sachs lesion (p = 0.3148), Parsonage-Turner syndrome or quadrilateral space syndrome (p = 0.290), and miscellaneous (p = 0.0962). A significantly different frequency was found with partial rotator cuff tears (p = 0.0019) (Table 1).

reporting requirements of Medicare providers would be necessary [17]. This study used a unique approach to address the utilization patterns of referring clinicians with financial interest in medical imaging equipment—namely, rather than auditing financial and billing data, final imaging diagnoses on imaging studies were analyzed to control for severity of disease between referring physician groups (financially incentivized vs non–financially incentivized). Because this approach analyzed the interpretations by one small academic radiology group reading examinations for both referring physician groups, we were able to control for a major confounder: radiologist interreader variation. Furthermore, because this radiology group did not own any of the equipment used, no inherent confounding financial interest was present. Despite the inevitable presence of some minor variation in the clinical presentations of individual patients, we think that similar

Discussion The purpose of this study was to determine whether the utilization patterns of shoulder MRI differed between physicians with a financial interest in the MRI equipment used compared with physicians without a financial interest in the MRI equipment used. We report a 25.6% (p = 0.047) increase in the frequency of negative shoulder MRI examinations referred by the financially incentivized physician group. These data suggest that there may be a lower threshold for shoulder MRI referral in the financially incentivized group compared with physicians with no financial incentive in MRI equipment. Exploring the topic of incentivized selfreferral imaging arrangements has been, and continues to be, a significant challenge. There are a multitude of factors in play, and to comprehensively analyze medical imaging services, including a specific financial analysis of self-referral arrangements, changes in

shoulder injuries show similar clinical findings, on average. The significantly higher rate of negative scans in the financially incentivized group identified in this study is in conflict with our null hypothesis. Our data found a statistically significant increase in the number of negative MRI scans in the financially incentivized group compared with the non–financially incentivized group. We also found that there was no statistically significant difference in the rates of nearly all of the individual injury types between the two groups and no particular predominance of one type of injury in either group. This similar pattern of individual shoulder injury types suggests that the abnormalities and prevalence within the two different patient populations are highly similar. In other words, presuming that abnormalities found on a positive study acts as a surrogate for the patient population, the two groups appear to significantly differ solely in their rates of negative shoulder MRI studies. We suspect that this difference may be secondary to differences in physician decision making regarding patients with similar shoulder abnormalities. Although this study does suggest a bias toward increased utilization of shoulder MRI by physicians with a financial interest in the MRI equipment used, by itself, this study does not prove intended overutilization for profit. It is possible that the practice pattern of a given physician group may be particularly imaging intensive, whereas that of a different group might be less so. The age of the ordering physicians may well be a key factor to consider. Young recently trained physicians may be more familiar and comfortable with MRI examinations, whereas older phy-

60

Financially Incentivized Non−Financially Incentivized

50 Number of Positive Studies

Downloaded from www.ajronline.org by 37.44.207.103 on 01/26/17 from IP address 37.44.207.103. Copyright ARRS. For personal use only; all rights reserved

30

Average Number of Abnormalities

Physician MRI Equipment Ownership and Use of Shoulder MRI

Fig. 3—Frequency of number of lesions per positive scan (as percentage of total number of positive scans in each group).

40

30

20

10

0

1

2

3 4 5 Number of Abnormalities

6

AJR:201, September 2013 607

Amrhein et al. TABLE 1: Comparison of the Frequency of Each Lesion Type (Positive Finding) Between Positive MRI Scans in Financially Incentivized and Non–Financially Incentivized Physician Groups Financially Incentivized Group

Non–Financially Incentivized Group

Rotator cuff tear, full

31.5

36.3

Rotator cuff tear, partial

25.9

17.3

Subscapularis tear, full

2.6

3.7

Downloaded from www.ajronline.org by 37.44.207.103 on 01/26/17 from IP address 37.44.207.103. Copyright ARRS. For personal use only; all rights reserved

Lesion Type

Subscapularis tear, partial

14.5

14.2

Partial-thickness articular surface tendon avulsion, definite

14.0

18.2

Partial-thickness articular surface tendon avulsion, possible

1.4

0.4

Cuff tendon injury present

74.8

74.4

Labral tear full, definite

26.2

25.4

Labral tear possible, partial

11.2

13.6

Superior labrum anterior to posterior tear full, definite

2.6

1.5

Superior labrum anterior to posterior tear partial, possible

0.7

0.2

Anterior labroligamentous periosteal sleeve avulsion

0.2

0.7

Bankart

1.6

2.2

Bony Bankart

1.2

1.5

41.6

43.3

Biceps tendon tear, full

5.4

5.0

Biceps tendon tear, partial

6.1

9.4

Biceps tenosynovitis

7.5

5.0

Labral injury present

18.7

19.0

Humeral avulsion of the glenohumeral ligament

Biceps tendon abnormality present

1.4

0.7

Hill-Sachs lesion

4.9

6.6

Quadrilateral space syndrome, Parsonage-Turner syndrome, or myopathy

6.1

4.4

Other

3.0

5.5

Note—Data are percentage of positive scans containing each lesion subtype. No significant difference was observed between the two groups for any lesion, with the exception of partial tears of the rotator cuff.

sicians who trained and practiced before MRI became widely available may prefer other diagnostic methods. Our study has several limitations. One is the potential for an unapparent selection or referral bias that may have increased the total lesion rate in the non–financially incentivized group. However, the combination of closely matched patient demographics and closely matched frequencies of injury subtypes among the positive studies between the two groups makes this source of bias less likely. Another limitation is that our study analyzed the frequency of 20 specific lesions to categorize a positive examination, and it is possible that higher proportions of negative scans in the non–financially incentivized group may have occurred if alternative lesions were classified as positive;

608

nonetheless, the lesions used in our analysis are common causes of shoulder pain and are clinically important indications for more-invasive interventions, such as surgery, for which imaging is commonly performed. Despite the limitations, we think that the inherent conflict of interest present when physicians both order and then perform and collect money on MRI scans is an important factor to consider whenever the matter of health care cost is raised as an issue. Our findings suggest the likelihood of a bias toward ordering more shoulder MRIs among physicians who own the MRI equipment being used. We caution that our findings require validation across geographic regions and with a larger number of practices. These findings also require validation across different types of imaging stud-

ies. We think that medical societies and government agencies are the ideal sponsors of validating work in this area. In conclusion, we observed an excess negative shoulder MRI rate among patients seen by physicians who own MRI equipment compared with the rate found in patients seen by physicians without a financial interest in performing MRI. Further study of this topic among a larger sample of practices is needed to determine the full scope of the issue. References 1. Levin DC, Rao VM, Parker L, Frangos AJ, Sunshine JH. Bending the curve: the recent marked slowdown in growth of noninvasive diagnostic imaging. AJR 2011; 196:W25–W29 2. U.S. Government Accountability Office. Report to Congressional requesters: Medicare Part B imaging services: rapid spending growth and shift to physician offices indicate need for CMS to consider additional management practices. Washington, DC: Government Accountability Office, 2008 3. Armstrong D. MRI and CT centers offer doctors way to profit on scans. The Wall Street Journal web site. Published May 2, 2005. Accessed March 6, 2012 4. Thompson DF. Understanding financial conflicts of interest. N Engl J Med 1993; 329:573–576 5. Levin DC, Rao VM. Turf wars in radiology: updated evidence on the relationship between selfreferral and the overutilization of imaging. J Am Coll Radiol 2008 5:806–810 6. Litt AW, Ryan DR, Batista D, Perry KN, Lewis RS, Sunshine JH. Relative procedure intensity with self-referral and radiologist referral: extremity radiography. Radiology 2005; 235:142–147 7. 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 8. 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 9. Gazelle GS, Halpern EF, Ryan HS, Tramontano AC. Utilization of diagnostic medical imaging: comparison of radiologist referral versus samespecialty referral. Radiology 2007; 245:517–522 10. Aronovitz LG. Referrals to physician-owned imaging facilities warrant HCFA’s scrutiny: General Accounting Office report to the U.S. House of Representatives. Washington, DC: General Accounting Office, 1994 11. Radecki SE, Steele JP. Effect of on-site facilities

AJR:201, September 2013

Downloaded from www.ajronline.org by 37.44.207.103 on 01/26/17 from IP address 37.44.207.103. Copyright ARRS. For personal use only; all rights reserved

Physician MRI Equipment Ownership and Use of Shoulder MRI on use of diagnostic radiology by non-radiologists. Invest Radiol 1990; 25:190–193 12. Kouri BE, Parsons RG, Alpert HR. Physician self-referral for diagnostic imaging: review of the empiric literature. AJR 2002; 179:843–850 13. Levin DC, Rao VM, Parker L, Frangos AJ, Sunshine JH. Ownership or leasing of MRI facilities by nonradiologist physicians is a rapidly growing trend. J Am Coll Radiol 2008; 5:105–109

14. [No authors listed]. A data book: healthcare spending and the medicare program. Washington, DC: Medicare Payment Advisory Commission, 2008 15. Paxton BE, Lungren MP, Srinivasan RC, et al. Physician self-referral of lumbar spine MRI with comparative analysis of negative study rates as a marker of utilization appropriateness. AJR 2012; 198:1375–1379

16. Pandya NK, Colton A, Webner D, Sennett B, Huffman GR. Physical examination and magnetic resonance imaging in the diagnosis of superior labrum anterior-posterior lesions of the shoulder: a sensitivity analysis. Arthroscopy 2008; 24:311–317 17. Kilani RK, Paxton BE, Stinnett SS, Barnhart HX, Bindal V, Lungren MP. Self-referral in medical imaging: a meta-analysis of the literature. J Am Coll Radiol 2011; 8:469–476

F O R YO U R I N F O R M AT I O N

This article has been selected for AJR Journal Club activity. The accompanying Journal Club study guide can be found on the following page.

AJR:201, September 2013 609

APPENDIX 1: AJR Journal Club

Downloaded from www.ajronline.org by 37.44.207.103 on 01/26/17 from IP address 37.44.207.103. Copyright ARRS. For personal use only; all rights reserved

Study Guide

Shoulder MRI Utilization: Relationship of Physician MRI Equipment Ownership to Negative Study Frequency Alan Mautz, Margaret Mulligan, Joseph J. Budovec* Medical College of Wisconsin, Milwaukee, WI [email protected], [email protected], [email protected] Introduction 1. What is the question being asked? Is this question relevant and timely? Will answering the question impact the practice of medicine? 2. What is self-referral? Why is addressing self-referral important? Methods 3. What type of study was this? What was the study design? 4. What were the inclusion and exclusion criteria for the shoulder MRI studies included in this study? What data were included for analysis? 5. Was the blinding used by the interpreting radiologists adequate for the purposes of this study? 6. What were the limitations of this study? Are these limitations adequately discussed? 7. What statistical analyses were used in the data obtained? Results 8. Was the research question answered? 9. Was the study designed well enough and was the sample size large enough to draw appropriate conclusions about the pretest probability of a negative shoulder MRI based on possible self-referral and its impact on the cost of medical imaging? Socioeconomics 10. What is the Stark Law? What are the exemptions? How may the Stark Law exemptions impact practice patterns? Discussion 11. How did the study overcome limitations cited in previous literature regarding imaging self-referral? 12. Are the data obtained conclusive enough to alter health care policy regarding self-referral? What additional studies should be undertaken with regard to the possibility of overutilization? 13. The authors cite limitations to their study including the need for validation across geographic regions and additional practices. Are there circumstances at your institution that would allow a similar analysis of imaging utilization? How would you go about designing a study to evaluate similar questions at your institution? Background Reading 1. Levin DC, Rao VM. Turf wars in radiology: updated evidence on the relationship between self-referral and the overutilization of imaging. J Am Coll Radiol 2008; 5: 806–810 2. Gazelle FS, Halpern EF, Ryan HS, Tramontano AC. Utilization of diagnostic medical imaging: comparison of radiologist referral versus same-specialty referral. Radiology 2007; 245:517–522

*Please note that the authors of the Study Guide are distinct from those of the companion article.

610

AJR:201, September 2013