Utilization Rates of Neuroradiology across Neuroscience Specialties in the Private Office Setting: Who Owns or Leases the Scanners on Which Studies Are Performed?

Discussion

In the private office setting, the number of head and spine CT and MR studies performed in the Medicare patient population has more than doubled between 1998 and 2008. However, not all of the studies were growing at the same pace. MR grew faster than CT, and spine MR grew faster than head MR.

During 1998–2008, radiologists, neurosurgeons, multi/IDTFs, and other specialists each saw growth in their total global/technical fees for neuro CT and MR, while neurologists did so only in head and spine MR. Between 1998 and 2008, neurologists' head CT utilization rate decreased and spine CT utilization rate remained unchanged. For radiologists, neurologists, neurosurgeons, and other specialists, MR utilization rate has increased the most; however, neurologists and neurosurgeons saw slightly faster growth in head MR and radiologists and other specialists in spine MR. Despite seeing the greatest growth in head MR, most of the neurosurgeons' private office neuroradiology remains spine MR.

In 1998, radiologists performed 83.6% of total private office neuroradiology, compared with 56.6% in 2008. The lost market share was captured by multi/IDTFs and other specialists, and marginally by neurologists and neurosurgeons.

Our observations suggest that during 1998–2008, nonradiologist equipment owners/lessees have been increasing their market penetration in private office neuroradiology. Furthermore, nonradiologists have been preferentially gaining ground in the highest RVU imaging modalities, growing faster in MR than in CT.

In this study, we analyzed billing records for global and technical charges, and did not look at professional charges, which are reimbursed for interpretation of studies. We did not explore the extent to which nonradiologist equipment owners/lessees entered the neuroradiology interpretation market for professional fees or the extent to which nonradiologists contract radiologists for interpreting the neuroradiology studies for which they charge the technical fees.

As owner/lessees, nonradiologists are able to capture technical fees for performed procedures, which are higher than professional fees reimbursed for the interpretation of studies. Technical fee to professional fee ratios range from 3.95 to 4.21 for neuroradiologic CT and from 4.23 to 5.86 for neuroradiologic MR, according to the Highmark Medicare online fee calculator.¹⁶

Our results do not come as a surprise, as there are numerous other reports documenting the recent growth in high RVU imaging utilization in the private office setting among nonradiologists at a pace greater than that of radiologists.^5,10,11,13 During 1998–2002, in the private office setting, radiologists saw growth in radionuclide myocardial perfusion imaging utilization of 16%, while cardiologists increased their volumes by 101%.⁵ Between 2000 and 2005, all of radiologists' private office MR imaging examinations (non-neuroradiology and neuroradiology) increased by 83%, compared with a 254% increase experienced by nonradiologists,¹⁰ and during 2001–2006, all of radiologists' private office CT studies (non-neuroradiology and neuroradiology) increased by 85% in contrast to a 263% growth seen among nonradiologists.¹¹ Also, from 2002 through 2007, radiologists saw growth in their private office PET studies by 259%, while, during the same time period, nonradiologists increased their PET business by 737%.¹³

Between 2000 and 2005, the growth in Medicare MR imaging examinations by nonradiologists was driven by 10 specialties, arranged according to their relative share of the total MR imaging, from greatest to least: orthopedic surgery (volume increase of 408%); neurology (171%); primary care (242%); internal medicine subspecialties, excluding cardiology (1397%); neurosurgery (309%); pathology/rehabilitation medicine (550%); radiation oncology (253%); cardiology (762%); anesthesiology (1972%); and other surgeons (101%).¹⁰ Similarly, the 2001–2006 growth in Medicare CT examinations by nonradiologists, again arranged based on their relative share of the total CT imaging, from greatest to least, was as follows: primary care (volume increase of 215%); internal medicine subspecialties, excluding cardiology, medical oncology, and nephrology (395%); urology (1631%); cardiology (2825%); medical oncology (289%); other surgeons (355%); radiation oncology (191%); neurology (55%); nonphysician practitioners (810%); and vascular surgeons (3107%).¹¹ These and other studies^6,8 suggest that, in large part, the steep utilization increases seen among nonradiologist imaging equipment owners/lessees are the result of self-referral practices.

The practice of self-referral is a consequence of a conflict of interest that exists when a physician refers a patient for an imaging study that is performed on equipment that the physician has a financial stake in. Federal regulations like Stark laws limit financially motivated imaging but, at the same time, legalize the “in-office ancillary services exemption,” which, presumably for patient convenience and speediness of care, allows physicians to install and own imaging equipment in their offices. However, recent analyses suggest that the in-office ancillary services exemption is a loophole for self-referral, with doubtful benefits to the patients.^17–19 Sunshine and Phargavan examined the 2007 Medicare Research Identifiable Files to identify self-referring physicians and found that although 74% of self-referred conventional radiographs were performed on the spot, only 15% of CT and MR were done on the same day.¹⁷ Hughes et al compared duration of illness, imaging episode cost, and total episode of care cost between physicians who self-refer for imaging and physicians who refer for imaging to radiology facilities for 10 nonchronic conditions matched with commonly utilized diagnostic imaging studies.¹⁸ They found that self-referral was associated with shorter duration of illness in only 3 of 20 condition-diagnostic imaging pairs (radiographs for chest pain, respiratory illness, and sinusitis), and that it was also associated with higher episode imaging cost in 15 of 20, and higher total episode of care cost in 14 of 20, condition-diagnostic imaging pairs.¹⁸ Similarly, Baker showed a 2%-per-episode cost increase for orthopedists and a 6% increase for neurologists when they looked at outpatient practice patterns before and after MR imaging equipment was purchased for in-office use.¹⁹

For all specialties, most of the growth in neuroradiology utilization rates occurred between 1998 and 2005. Since 2005, the number of performed studies has flattened or even, in some cases, decreased (Figs 1–3). This may in part be due to the 2005 DRA (signed into law in February 2006 and took effect on January 1, 2007), which resulted in reduction of the technical fees for radiographic studies performed in the private office setting.^20,21 Our results, similar to prior published data, show that the change in the reimbursement schedule disproportionally affected radiologists compared with nonradiologists.²² While both neurologists and neurosurgeons saw continued growth of neuroradiology, albeit at a slower pace compared with the pre-2006 values (Table 2), radiologists saw a decrease in the number of the total neuroradiology studies performed. There are 2 possible explanations behind this observation: new nonradiologists entered the business of imaging services or existing nonradiologist equipment owners increased their individual volumes through self-referral (an option not available to radiologists who do not have comparable access to patients).²²

The purpose behind our paper and presentation of our data is not to sound the alarm that “the sky is falling” but rather to document the extent of nonradiologists' penetration of the private office neuroradiology, as well as to serve as a baseline for comparison of trends over time, given the current climate of cost control, self-referral concerns, and regulatory changes. Starting with the 2005 DRA, radiology became less profitable per case, a trend that will likely continue. Driven by sensitivity to cost, the governments (federal and local) and private insurers are taking steps to scrutinize the appropriateness of imaging examinations and indirectly limit self-referral. For example, the 2010 Affordable Care Act mandated disclosure of physicians' imaging equipment ownership interests.²¹ On January 21, 2011, Maryland's Court of Appeals upheld the state's 1993 anti-self-referral law that prohibits nonradiologists from referring patients for CT and MR imaging studies within their offices²³ (though it is likely that the upheld law will be challenged legislatively). Private insurers started requiring all imaging sites to obtain external accreditation from organizations like the ACR or Intersocietal Accreditation Commission and also started requiring outpatient centers to provide multimodality imaging to qualify for interpretation reimbursements. Therefore, physicians who only have a CT or MR scanner are getting eliminated from potential reimbursement by many of the insurers. This factor is a trend in favor of lessening nonradiology ownership unless nonradiologists are willing to also support mammography, sonography, and plain films—the lower RVU modalities. Finally, major payers deployed radiology benefits management programs, which create “gatekeepers” for imaging studies (typically based on a modified version of the ACR Appropriateness Criteria), leading to limited growth in advanced imaging utilization.^24–25

Because our study looks specifically at trends in neuro CT and MR, it can also serve as a baseline reference in future evaluations of the effects of the recently proposed neuroimaging fellowships for neurologists.^26,27 As more neurologists complete their neuroimaging fellowships, one could conceive of more neurologists becoming imaging equipment owners/lessees. To date, only cardiologists have offered formal training in performing and interpreting of subspecialty imaging examinations, such as cardiac nuclear imaging, and as a result, at many locales, cardiologists have marginalized radiologists' role.²⁸

Our data, inclusive of 2008, suggest variable effectiveness of the imaging utilization mechanisms described herein; however, at least for the time being, the overall trend appears to be against inappropriate self-referral. Many additional strategies and approaches addressing the challenges related to imaging utilization facing radiology today have been suggested.^21,28–30 In our opinion, any proposed solution has to revolve around the principle of performing imaging examinations only for the benefit of the patient rather than financial gain. Referring physicians could be assisted by an objective computerized order-entry support system incorporating evidence-based imaging algorithms.³¹ The ACR has developed Appropriateness Criteria for imaging; however, not many clinicians are familiar with those,³² and thus an automated, easily accessible system would be helpful. The current laws allowing in-office self-referral imaging are based on the assumption that patient care is made more convenient and prompt. However, given the recent evidence,^17,18 restriction of the in-office ancillary services exception to conventional radiography and sonography, and prohibition of CT, MR, and PET, makes sense. Finally, today's outpatient healthcare is largely based on a fee-for-service model, which incentivizes utilization. An alternative approach would involve bundled payments per episode of illness, or per member annual capitation, with quality of care benchmarks to prevent undertreatment.

Our study has a number of limitations. Although our data show an increased number of neuroradiology studies being billed for global/technical fees by nonradiologists, our study design is unable to distinguish between a demand-driven growth in private office neuroradiology (ie natural growth based on clinical need) and a supply-driven growth (ie, overutilization). Further research is needed to clarify whether nonradiologists are simply meeting the demand of markets not well served by radiologists or whether nonradiologists are artificially raising the need for private office neuroradiology by altering their own practice profile. All of our analyses are based on the PSPSMFs, which contain data for the nationwide Part B datasets for beneficiaries in the Medicare fee-for-service program (approximately 85% of the total Medicare population). Although, at many private practices, Medicare constitutes the largest single payer, Medicare reimbursements account for a portion of the total imaging volume, and, additionally, some practices exclude (ie do not accept) any Medicare patients. Therefore, our results may not accurately reflect utilization trends in every practice setting and may have limited generalizability. Also, it is important to note that some of the impressive growth rates that we calculated occurred among provider specialties (or in imaging categories), which had small, absolute volumes or market shares, and should therefore be interpreted with caution (ie addition of each new MR or CT scanner results in lower incremental growth, and, thus, small industry participants can grow at faster rates than large ones).

In our study, we focused on MR and CT imaging of the brain, head and neck, and spine but did not look at any other imaging studies. Given that we were interested in the neuroradiology trends among radiologists, neurologists, and neurosurgeons, we did not fully explore utilization trends within other relevant provider specialties like orthopedic surgery (spine imaging) or primary care (all neuroradiology). In our analyses, we treated the multi/IDTF specialist category as separate from radiology, though some radiologist equipment owners/lessees likely belonged to that category. In so doing, we hoped to present a “cleaner” set of results, acknowledging that we may be underestimating the actual market trends among all radiologists, including IDTF owners. Finally, we had no way to control or account for the effects of the recent changes in the reimbursements rates for outpatient imaging (DRA of 2005), which may explain some of the providers' growth propensity.