Recent Shifts in Place of Service for Noninvasive Diagnostic Imaging: Have Hospitals Missed an Opportunity?

Article Outline

• Abstract
• Materials and Methods
• Results
• Discussion
• References
• Copyright

Purpose

The aim of this study was to examine recent shifts in place of service for noninvasive diagnostic imaging (NDI) and determine whether hospitals have lost business to private outpatient imaging facilities.

Method and Materials

The nationwide Medicare Part B databases for 1996 through 2006 were used, and all Current Procedural Terminology^®, fourth edition, codes for NDI were studied. Utilization rates per 1,000 Medicare beneficiaries were calculated. Medicare uses place-of-service codes to differentiate examinations performed in hospital inpatients, hospital outpatients, and hospital emergency departments from those performed in private office settings. Changes in utilization rates in these locations were compared over the course of the decade, with particular emphasis on possible outpatient NDI shifts between hospital outpatient departments and private offices or imaging centers. Also, Medicare physician specialty codes were used to determine whether radiologists or other specialists were more responsible for growth.

Results

Between 1996 and 2006, Medicare NDI utilization rates per 1,000 -hospital inpatients increased from 1,056.5 to 1,211.8 (+15%). Emergency department rates increased from 222.1 to 392.2 (+77%). Hospital outpatient rates increased from 793.4 to 993.2 (+25%), while private office rates went from 883.3 to 1,442.2 (+63%). Total outpatient imaging rates (both hospital and office) went from 1,676.7 to 2,435.4 (+45%). As a result of the more rapid growth in private office imaging, hospitals' share of this market dropped from 47% in 1996 to 41% in 2006. Private office imaging utilization rates between 1996 and 2006 grew by 71% among nonradiologist physicians, compared with 44% among radiologists.

Conclusion

Medicare NDI utilization rates increased in all places of service between 1996 and 2006. Growth in hospital outpatient imaging was slower than that in private imaging facilities. Because NDI can be a profitable business, it seems that hospitals have lost an important opportunity. Much of this loss of business was to nonradiologist physicians, whose private office imaging utilization rate grew considerably more rapidly than that of radiologists.

Key Words: Noninvasive diagnostic imaging, utilization of imaging, hospital finances, medical economics, radiology and radiologists, socioeconomic issues

Imaging has been shown in recent years to be the most rapidly growing of all physician services [1, 2, 3]. Although this is of concern to policymakers and payers, it has created economic opportunities for nonradiologist physicians, hospitals, and radiologists. Despite recent cuts in reimbursement, such as the Deficit Reduction Act of 2005, it still seems to be possible to make some profit on the technical-component reimbursement for outpatient imaging if facilities can operate efficiently. In recent years, some progressive hospitals and radiology groups have expanded their outpatient imaging capability. Hospitals have found that imaging is one major area in which revenues generally exceed expenses and that can have a positive contribution margin. At the same time, other trends seem to be in evidence. Some nonradiologist physicians have installed advanced imaging equipment in their offices, taking advantage of the in-office ancillary services exception of the Stark laws to do so [1, 4, 5]. Emergency department (ED) imaging has been on the increase, as EDs are seeing greater numbers of patients who are uninsured and as ED physicians try to cope with the added volume. Inpatient imaging also seems to be increasing as hospitalized patients have higher levels of illness acuity. Publicly owned companies are building or acquiring imaging centers that often compete directly with hospitals. In this mix of developments, it is of interest to study trends in utilization rates of noninvasive diagnostic imaging (NDI) in the 4 primary locations in which such imaging is conducted: hospital inpatient settings, hospital outpatient facilities, private offices or imaging centers, and EDs. In this study, we assessed these trends during the period from 1996 to 2006, using a nationwide patient database, and we address their implications for hospitals and the US health care system.

Back to Article Outline

Materials and Methods 

The data were taken from the Centers for Medicare & Medicaid Services Part B Physician/Supplier Procedure Summary Master Files for 1996 through 2006. These files contain information on all medical services provided to Medicare fee-for-service beneficiaries (83% of the total Medicare population in 2006). For each code in Current Procedural Terminology, fourth edition (CPT-4), the files provide data on annual procedure volume, the specialty of the physician provider, and the location (or place of service) where the medical service was provided. Our study covered the NDI codes in the 70000 CPT-4 series, as well as those in the 90000 series pertaining to echocardiography and vascular ultrasound. We excluded CPT-4 codes relating to interventional procedures, ophthalmic ultrasound, radiation therapy planning, and radioimmunoassays.

The NDI utilization rates per 1,000 Medicare fee-for-service beneficiaries were calculated, and trends between 1996 and 2006 were studied for the 4 primary location or place-of-service codes where imaging is performed: hospital inpatient settings, hospital outpatient facilities, private offices or imaging centers, and EDs. Using Medicare's physician specialty codes, we categorized physician providers as radiologists, all other physicians as a group, and independent diagnostic testing facilities (IDTFs). Trends were determined for these specialty categories. One of the peculiarities of this Medicare data set is that it considers IDTFs to be a medical “specialty” rather than places of service. Nevertheless, we were able to capture IDTF data because they have the separate specialty code. The exact specialty of any physician owner of an IDTF cannot be determined. In studying the trends among the specialty categories, we excluded a small number of claims with the specialty code for multispecialty groups because the actual provider specialties cannot be determined from those claims. Procedure volume and utilization rates were determined by tabulating global and professional-component claims. Technical-component-only claims were not included in the tabulation because that would have led to double counting of examinations.

Back to Article Outline

Results 

Table 1 shows the changes in utilization rates per 1,000 Medicare fee-for-service beneficiaries between 1996 and 2006 for the 4 primary locations or places of service where imaging is conducted: hospital inpatient settings, hospital outpatient facilities, private offices or imaging centers, and EDs. Rates per 1,000 Medicare beneficiaries increased in all 4 places of service. Among hospital inpatients, the rate grew from 1,056.5 to 1,211.8 (+15%) over the study period, the lowest increase of the 4 primary places of service. In hospital outpatient facilities, a 25% increase was seen (793.4 to 993.2). This compared with the 63% increase that was seen in private offices or imaging centers (883.3 to 1,442.2). The utilization rates in EDs were considerably lower than in the other 3 locations, but growth was the fastest in EDs, at 77% (222.1 to 392.2). Total outpatient imaging rates (the sum of hospital outpatient facility and private office rates) increased from 1,676.7 in 1996 to 2,435.4 in 2006 (+45%). Hospitals' share of this market dropped from 47% in 1996 to 41% in 2006.

Table 1. Medicare noninvasive diagnostic imaging rates per 1,000 beneficiaries in all places of service for all specialties, 1996 through 2006

Note: Hosp inpt = studies done on inpatients in hospitals; hosp outpt = studies done in hospital outpatient facilities; private offices include freestanding imaging centers; ED = emergency department. A small number of examinations were performed at places other than these 4 primary locations (eg, nursing homes).

Figure 1 illustrates the trends on a year-to-year basis. Note that the hospital inpatient and hospital outpatient facility trend lines closely parallel each other. The much more rapid growth in private offices and imaging centers is apparent. In 1996, the private office utilization rate was only marginally higher than the rate in hospital outpatient facilities, but by 2006, it was approximately 45% higher. We did not graph the studies done in “other locations” because their rates were far lower than those in the 4 primary locations.

• 
  • View Large Image
  • Download to PowerPoint

• Fig 1. 

  Shifts in place of service between 1996 and 2006 for Medicare noninvasive diagnostic imaging. The vertical axis shows examinations per 1,000 Medicare fee-for-service beneficiaries. Office = examinations performed in private offices or imaging centers; hosp inpt = examinations performed on hospital inpatients; hosp outpt = examinations performed in hospital outpatient facilities; ED = examinations performed in emergency departments.

Table 2 shows the utilization rate changes in hospital outpatient facilities between 1996 and 2006 among radiologists and other physicians. In that setting, radiologists did by far the larger number of examinations, although the growth rate was higher among other physicians. Radiologists' rate increased from 683.4 in 1996 to 852.3 in 2006 (+25%). The rate among other physicians went from 78.8 to 140.3 (+78%).

Table 2. Medicare noninvasive diagnostic imaging utilization rates per 1,000 beneficiaries in hospital outpatient facilities, 1996 through 2006

The situation was quite different in private offices and imaging centers, as shown in Table 3. There, the rate increases were as follows: among radiologists, from 275.9 in 1996 to 398.5 in 2006 (+44%), and among other physicians, from 571.8 to 977.0 (+71%). Thus, nonradiologist physicians performed a considerably larger proportion of imaging in offices, and their rate of growth was considerably more rapid than among radiologists. In 2006, radiologists' share of the private office or imaging center market was 28%, other physicians' share was 68%, and IDTFs' share was 4%. These numbers are based on raw utilization rates and do not take into account the complexity or relative costs of the examinations.

Table 3. Medicare noninvasive diagnostic imaging utilization rates per 1,000 beneficiaries in private offices or imaging centers, 1996 through 2006

Note: IDTF = independent diagnostic testing facility. In 1996, there was no code for IDTFs, so no data were recorded for them.

Back to Article Outline

Discussion 

The trends revealed in the tables and figure provide some interesting insight into the NDI market in recent years. Figure 1 demonstrates that in the late 1990s, NDI utilization rates remained relatively flat. However, between 2000 and 2005, there was a progressive increase in all 4 major places of service. In 3 of them—hospital inpatient settings, hospital outpatient facilities, and EDs—the slopes of the lines were relatively similar, indicating that the accrual of new examinations was relatively similar. On the other hand, the slope of the private office line was much steeper, indicating that growth in that place of service was considerably more rapid. Between 2005 and 2006, the rates declined slightly for hospital inpatients and hospital outpatient facilities but continued to increase in private offices. It remains to be seen whether these trends will continue in subsequent years. Although the imaging utilization rates in EDs were the lowest among the 4 primary imaging locations, the growth rate in EDs was the highest. There are several reasons for this, which will be the subject of a future study.

The trends suggest that outpatient imaging is gradually shifting from hospitals to the private office setting. There are several likely explanations. First, it is possible that hospitals have been slow to add outpatient imaging capacity and that physicians and entrepreneurs have filled a gap in supply. Second, it may be that physicians and entrepreneurs have been more aggressive in adding new equipment in their offices and have driven demand up by a combination of marketing and self-referral. However, the fact that the private office NDI utilization rate among radiologists increased by 44% between 1996 and 2006 would seem to militate against this, because radiologists are not in a position to drive demand.

A third factor fueling imaging growth in private offices may be patient preferences [6]. As noted above, many hospitals have been slow to convert their outpatient imaging practices to a truly outpatient setting. In an effort to save equipment and space costs, they continue to do outpatient imaging in the hospital radiology department itself, where outpatient and inpatient studies are intermingled. This means that outpatients often have to put up with unpleasant circumstances, such as seeing sick inpatients on stretchers, or having their examinations delayed because of ED or inpatient emergencies. Physicians and entrepreneurs have responded by building attractive private office facilities, where patients feel more comfortable and where other amenities such as free parking are offered. Not surprisingly, outpatients have opted for these more pleasant and convenient office settings.

Regardless of the cause, what does seem clear is that the NDI utilization rate among outpatients is increasing. The implication for hospitals is that if they want to continue having an important share of this market, they need to increase their capacity and market it actively to referring physicians, patients, and health care insurance carriers [6]. In past years, many hospitals have relied solely on their capital budgets to pay for imaging equipment, and this has limited their ability to expand (Frank Kyle, Outpatient Imaging Affiliates LLC, personal communication, 2002). In contrast, physicians and entrepreneurs have more frequently relied on lease arrangements to finance their equipment acquisitions and updates. It is likely that hospitals have to adopt this policy as well.

Nonradiologists have been very enterprising in developing imaging capabilities in their private offices, as shown in Table 3. From 1996 to 2006, their office NDI utilization rate increased by 405.2 studies per 1,000 Medicare beneficiaries (from 571.8 in 1996 to 977.0 in 2006), a 71% increase. The private office utilization rate among radiologists during that same time interval increased by 122.6 studies per 1,000, a 44% increase. Thus, nonradiologists accrued new studies (most of which were self-referred) much more rapidly in their offices than did radiologists. The situation is quite different in hospital outpatient facilities, which only perform examinations referred by other physicians. There, radiologists strongly predominate, and the growth rate has been more modest.

Our study had certain limitations. Because it focused on utilization rates, we tabulated only global and professional-component claims, to avoid double counting examinations. This methodology by necessity somewhat overestimates the participation of radiologists while underestimating the participation of those nonradiologists in the “other physicians” category. This is because in some instances (particularly the advanced imaging modalities), nonradiologist physicians own equipment in their offices, perform studies on their patients, bill technical components only, and contract with radiologists, who perform the interpretations and bill separate professional components [4, 5]. In instances such as this, our methodology would capture the examinations as “radiologist” examinations, even though nonradiologist physicians were initiating the examinations and collecting the technical-component fees. This limitation does not apply to cardiologists, who make up a large component of the “other physicians” category, because they virtually always read their own studies, rather than contracting with radiologists to do so. Another limitation, as noted above, is that the complexity of examinations was not taken into account; this was purely a study of utilization rates per 1,000 Medicare beneficiaries. We plan to address this in a future study of relative value unit rate growth.

In summary, growth in the private office NDI utilization rate has been rapid in recent years, particularly among nonradiologist physicians. Hospitals are disadvantaged by this trend because imaging is generally felt to be profitable for most institutions. They may want to consider expanding their outpatient imaging capabilities more actively by creating environments deemed more desirable by patients. This study also has certain other health policy implications. The Deficit Reduction Act substantially reduced reimbursement for Medicare private office NDI beginning in 2007, especially for the advanced imaging modalities. Once 2007 and later data become available, it will be important to see what effect this has on utilization rates and patient access.

Back to Article Outline

References 

1. Iglehart JK. The new era of medical imaging—progress and pitfalls. N Engl J Med. 2006;354:2822–2828
   • View In Article
   • CrossRef
2. Centers for Medicare and Medicaid Services. Payment for imaging services under the Medicare physician fee schedule, Herb Kuhn, director, CMM: House Subcommittee on Health of the Committee on Energy and Commerce. http://www.cms.gov/apps/media/press/testimony.asp?counter=1903Accessed May 29, 2007
   • View In Article
3. Blue Cross Blue Shield Association. 2004 medical cost reference guide. http://athensbenefits.com/pdf/2004_12_13_blue_cross_blue_shield_medical_cost_reference_guide.pdfAccessed May 29, 2007
   • View In Article
4. 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
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (113 KB)
   • CrossRef
5. Mitchell JM. The prevalence of physician self-referral arrangements after Stark II: Evidence from advanced diagnostic imaging. Health Aff (Millwood). 2007;26:W415–W423Available at: http://content.healthaffairs.org/cgi/content/abstract/hlthaff.26.3.w415. Accessed December 28, 2007.
   • View In Article
   • CrossRef
6. Boland GWL. Hospital-owned and operated outpatient imaging centers: strategies for success. J Am Coll Radiol. 2008;5:900–906
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (97 KB)
   • CrossRef