| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Review |
1 University of North Carolina School of Medicine, 212-4 Conner Dr., Chapel
Hill, NC 28714.
2 Department of Radiology, Carolinas Medical Center, Charlotte Radiology, 3080
Latrobe Dr., Charlotte, NC 28236-6937.
3 Vital Science & Health, P. O. Box 600284, Newtonville, MA 02460.
Received January 16, 2002;
accepted after revision March 13, 2002.
H. R. Alpert was supported by the American College of Radiology.
Review
 Top Review IntroductionNonradiologist Referral to Own...Nonradiologist Referral to an...Effects of Financial Incentives...Summary of FindingsResponses to Self-ReferralConclusionsReferences |
|---|
Legislation has substantially curbed referral to outside facilities in which physicians have a financial interest. To assure quality and contain costs, private insurers have begun to limit physicians to imaging in which they have expertise and to require physician and facility credentialing. These insurers report encouraging results [1, 2].
|
TopReviewIntroductionNonradiologist Referral to Own...Nonradiologist Referral to an...Effects of Financial Incentives...Summary of FindingsResponses to Self-ReferralConclusionsReferences |
|---|
Self-referral takes two principal forms. A physician who is not an imaging specialist (or a nonphysician provider, such as a podiatrist or chiropractor) directing patients to his or her own on-site imaging services was the first form of self-referral. Common examples are obstetricians (or their staff) performing sonographic examinations, internists performing and interpreting chest radiography, and orthopedists performing and interpreting musculoskeletal radiography. We include in the concept of self-referral all radiologic imaging done by nonradiologists because the financial incentives are the same regardless of whether a particular type of imaging is predominantly done by nonradiologists, which may affect its quality. The second form of self-referral consists of physicians referring their patients to outside facilities in which the physicians have a financial interest. This second form, which is often called "joint venture" self-referral, has been targeted by federal legislation with respect to services for Medicaid and Medicare patients. Some states, but far from all, have similar legislation applying to non-Medicaid and non-Medicare patients. Because coverage by legislation is incomplete and arrangements exist that attempt to avoid the effects of legislation, we include this form of self-referral in this review.
A host of claims, both positive and negative, have been raised regarding self-referral. The most prominent concern is that inherent conflicts of interest stimulate excessive utilization of health care services and thus generate excessive costs [4, 14]. Quality of patient care, access to services, and patient safety can purportedly be compromised because financial interest might dilute norms of care and quality assurance [13]. Physicians who refer patients to their own services on-site are nonradiologists. Arguments are made that such physicians are generally less qualified because of a lack of relevant training and experience [15]; order more, and more inappropriate, procedures [16]; produce images of poorer technical quality [16]; and are more prone to errors of interpretation [16] than are radiologists. Self-referring physicians' facilities may have poorer quality control of equipment maintenance, calibration, and film handling [17]. Similar concerns are raised regarding referral to an outside facility in which the physician has a financial interest, with particular focus on higher utilization [5, 18], poor access to services by underserved populations [14], and inadequate quality control [19].
Proponents of self-referral to a physician's own on-site services argue that such self-referral is convenient for patients and physicians [20]. Proponents maintain that some problems such as a fracture should receive imaging and treatment at a single office [21]. Proponents also say that quality of care can be better in self-referral settings because the treating physicians are better qualified to interpret diagnostic tests relevant to their specialties and have better insight into the specific patient's problem [21]. Furthermore, physicians may commit diagnostic and related treatment errors as a result of foregoing tests when referral to a radiologist is inconvenient [21]. Access to care for underserved geographic areas might be better with self-referral because of the greater availability of financing for imaging facilities [18, 22]. Finally, economies of scale and the presence of physician owners who are familiar with the patient and his or her financial limitations may help control costs (Winslow W, personal communication).
Considerable empiric evidence regarding these issues is now available, including studies that examine the prevalence of both forms of self-referral; factors associated with them; and the consequences for utilization, costs, quality, and access.
In this article we review the empiric literature concerning both forms of self-referral, addressing first nonradiologist referral to their own on-site imaging services and then referral to an outside facility in which a physician has a financial interest. For each form of self-referral, we address the prevalence of the form, its effect on utilization and costs, and quality issues. We then review the self-referral—related empiric literature on the effects of financial incentives on physician behavior. Finally, we summarize the responses that have been made to self-referral and present suggestions for action.
Nonradiologist Referral to Own On-Site Imaging Services
|
TopReviewIntroductionNonradiologist Referral to Own...Nonradiologist Referral to an...Effects of Financial Incentives...Summary of FindingsResponses to Self-ReferralConclusionsReferences |
|---|
More narrowly focused studies of prevalence report that nationwide Part B Medicare data for 1997 showed emergency physicians performing 0.7% of all sonographic studies in emergency departments and nonradiologists performing about 93% of echocardiography and 20% of vascular sonography in emergency departments [25]. Nationwide Part B Medicare data for 1993 and a seven-state subset of the same data for 1992 each indicated that nonradiologists were interpreting 2% of the CT and MR neuroradiologic imaging examinations in hospital settings (where 91% of CT and MR imaging was performed) and 9% in private office and free-standing imaging centers (where 9% of the total studies was performed) [26, 27].
Utilization and Costs
Health care services utilization and costs are somewhat distinct issues,
but they are closely related and have often been studied together in the
context of self-referral. The earliest studies indicated higher rates of
procedures by self-referring physicians than by others
[28,29,30].
Childs and Hunter [28], in a
1965 study involving 6902 patients who were 65 years old or older and 763
physicians, observed that patients of self-referring nonradiologists were
twice as likely to have an imaging procedure performed as patients of
non—self-referrers. Moreover, patients of self-referring physicians who
had any imaging performed underwent 65% more procedures than their
counterparts. A national random sample of physicians who were surveyed in 1978
revealed that nonradiologists with imaging facilities on-site had 1.2-1.7
times as high rates of utilization, depending on specialty, as those without
such facilities [29]. The
presence of on-site radiology facilities at a community teaching family
medicine center in Ontario was shown in 1987 to be associated with rates of
radiographic procedures among patients presenting with chest-related diagnoses
2.4 times as high as centers with no on-site radiology facilities
[30]. Landmark studies in the
early 1990s by Hillman et al.
[31,32,33]
and by the United States General Accounting Office
[6] sustained and extended this
early evidence. Hillman et al.
[33] first analyzed all health
insurance claims of employees and their dependents of several large American
corporations, focusing on chest and lumbar spine radiography, obstetric
sonography, excretory urography, cystography, and sonography. This research
was unique for its designation of episodes of care as the units of analysis,
its relatively large sample size (60,829 one-physician and 4,688 two-physician
episodes), and its control for case mix, both by using specific clinical
situations as the frame of reference and by analyzing by physician specialty.
Depending on the specialty and the clinical presentation, self-referring
physicians performed imaging 2.4-11.1 times as often as radiologist-referring
physicians and generated 3.0-17.1 times as high imaging costs per episode. The
findings were subsequently confirmed in a study of private insurance claims of
119,000 beneficiaries of United Mine Workers of America health and retirement
funds [31]. That research
analyzed data for a broader range of clinical presentations in a mostly
retired, elderly population and included more advanced imaging technologies.
Self-referring physicians were found to use imaging 1.7-7.7 times as
frequently as did radiologist-referring physicians.
Large-scale analyses of 1990 Medicare claims for imaging services ordered by Florida physicians were conducted by the General Accounting Office using information collected by researchers at Florida State University for the Florida Heath Care Cost Containment Board [6]. The General Accounting Office examined data pertaining to utilization and costs of each of the major imaging modalities for each form of self-referral. Physicians in each specialty who had in-office imaging capabilities performed an average of four to five times more sonography, echocardiography, and diagnostic nuclear medicine; three times more MR imaging; and twice as much radiography and CT as those without this equipment. Imaging performed by internists and general practitioners with in-office echocardiographic equipment exceeded the utilization rate of cardiovascular specialists without the equipment.
A multivariate regression analysis of radiology utilization, including radiography, CT, and MR imaging, for 650 elderly patients with respiratory symptoms who were seen in 52 generalist physicians' offices in eight states in 1994-1995 showed higher rates of studies (odds ratio, 1.81; p = 0.002) by self-referring physicians [34]. The analyses were adjusted for patients' own perceived needs for radiologic tests and for a range of patient case mix and sociodemographic characteristics. Self-referral was not a statistically significant factor in a similar analysis of 522 patients whose symptoms were lower back pain, possibly because the generalists were less comfortable interpreting lumbar spine images than chest images.
In light of the point made by proponents of self-referral that radiologists' actions can give rise to additional imaging beyond that recommended by the referring physician, the previous data already include any imaging resulting from radiologists' recommendations for additional studies. Moreover, Baumgarten and Nelson [35] found that additional CT was recommended for a minority (19.3%) of patients, and only a fraction (30.5%) of that recommended was actually performed. However, the findings of additional imaging studies when recommendations were followed were almost always helpful, either in confirming a diagnosis or in context of a therapeutic procedure.
Observers have noted that utilization trends of certain generally self-referred procedures have increased at the apparent expense of competing alternative imaging procedures. Levin [16] and Karasick et al. [36] highlighted declining rates of barium enema studies from 1985 to 1992 concurrent with increasing utilization of more expensive colonoscopy examinations. The former are generally referred to a radiologist, and the latter are typically self-referred by either gastroenterologists, surgeons, or internists. Gastroenterologists generally claim that colonoscopy is more effective than barium enema studies, and the findings may, at least in part, represent a new technology replacing an older one.
Issues of differences between self-referrers and radiologists in charges and reimbursements for the same services are much less important than in the past because most payers are now paying all physicians according to fee schedules. Studies found no major differences in charges and payments in earlier years but revealed some minor tendencies for self-referrers to be more expensive [17, 24].
Quality
Empiric information on quality of care in referral to physicians' own
on-site services involves accuracy of interpretation, technical quality of the
image, quality and condition of the equipment, records management, and patient
safety.
Accuracy of interpretation.—Unlike studies of utilization, most studies of accuracy of interpretation are from a single institution or otherwise narrow population base, which may partially explain their varied findings. Also, discrepant interpretations may be few among routine cases but more common (in percentage terms) among more seriously ill patients and for images with less obvious findings.
The largest number of published studies compares emergency physicians and radiologists. Misinterpreted radiographs account for most missed diagnoses in the emergency department [37].
In interpretations of 120 selected radiographs evaluated by a consensus panel comprised of radiologists and emergency physicians at a nationally preeminent hospital [38], radiologists' interpretations were more accurate (74% vs 55%), more sensitive (62% vs 38%), and more specific (86% vs 72%) than those of emergency physicians. Radiologists correctly interpreted 82% of cases that were identified by the consensus panel as "critical" compared with 48% correctly interpreted by emergency physicians. A subsequent similar study at the same institution produced the same basic findings, even after the increased awareness of quality issues generated by the first evaluation. Moreover, radiology residents' interpretations were more sensitive than those of emergency medicine physician faculty (66% ± 8% vs 42% ± 9%), as accurate (71% ± 6% vs 59% ± 6%,) and as specific (76% ± 8% vs 77% ± 8%) [39].
Nearly 39% nonconcordance between attending physicians and radiologists was found in the interpretations of 555 consecutive CT scans among patients seen at the emergency department of a county medical facility [40]. About 24% of the total had false-positive or false-negative interpretations by emergency physicians that were potentially clinically significant using radiologists' interpretations as the standard. A convenience sample of interpretations of 15 cranial CT scans by 103 physicians at a university teaching hospital in which clinical findings were preidentified by the authors found 83% overall accuracy of interpretation among both neurologists and radiologists and 67% among emergency physicians [41]. In that study, 52% of radiologists, 40% of neurologists, and 17% of emergency physicians correctly identified hemorrhage wherever it was present.
In a prospective cohort study at a pediatric emergency department in 1995-1996, clinically significant differences in interpretation between emergency physicians and radiologists were found in 1.1% of 2083 radiographs [42]. The most common discrepancy was an emergency physician reporting an infiltrate on a chest radiograph when the radiologist correctly interpreted the findings as normal. In none of these discrepancies did the authors determine, on the basis of the patient's physical examination and emergency department treatment, that the emergency physician was correct and the radiologist was wrong. Eight-tenths of 1% of 15,585 consecutive radiographs interpreted by attending emergency physicians in a Pittsburgh community hospital were not interpreted correctly according to review by attending radiology staff [43]. Fewer than half of these discrepancies were considered clinically significant and requiring a follow-up intervention. Similarly, emergency medical physicians and radiologists interpreted differently 134 (0.95%) of 14,046 radiographs obtained at two teaching hospitals during a 6-month study. Two-tenths of 1% of all cases had discrepancies that were clinically significant [44].
Some studies have examined the relationship between emergency physicians' confidence in their interpretations and accuracy. Emergency physicians were confident in 58% of 16,410 consecutive radiographs in a Minnesota urban teaching hospital and a community nonteaching hospital [45]. Discordant interpretations between the emergency physicians and radiologists were found in 3.1% of 6811 radiographs for which the emergency medical physician was not confident and in 1.2% of those for which the physician was confident. More than 5% of 1872 consecutive radiographic interpretations by emergency physicians were discrepant with the interpretations by two radiologists in an Illinois teaching hospital [46]. Sixty-five percent (1219) of the emergency physicians rated their level of confidence in their interpretations as very confident, 29% as moderately confident, 5% as mildly confident, and 1% as nonconfident. Treatment was potentially altered in 38 cases (39% of discordant radiographs), in 13 of which the emergency physician was very confident in the accuracy of the interpretation.
Chest radiographs have been another focus of studies of interpretation accuracy. One study compared accuracy of three groups of participants (111 board-certified radiologists, 29 radiology residents, and 22 nonradiologist physicians) recruited from private practices throughout North America and Great Britain who reviewed a standardized, clinically confirmed set of 60 posteroanterior chest radiographs [47]. The area under the receiver operating characteristic curve—a measure for which 1.0 is perfect diagnosis and 0.50 represents a purely random result—was 0.86 for the board-certified radiologists, 0.75 for the radiology residents, and 0.66 for the non-radiologists (p < 0.001 for each difference).
In 1989, 10 randomly sampled providers—seven nonradiologists and three radiologists—underwent a peer review of their interpretations of chest radiographs [48]. The peer review was conducted by board-certified radiologists at Pennsylvania State University for Pennsylvania Blue Shield. Ten randomly sampled chest radiographs were examined for each provider. Five clinically serious errors and nine other errors of interpretation were noted among the 68 radiographs interpreted by the nonradiologists. Only two minor errors of interpretation were found among the 30 radiographs interpreted by the three radiologists. The highest volume providers committed no errors, whereas the lowest volume providers committed four of the five serious interpretive errors and four of the six minor errors identified.
A few investigations have examined the accuracy of interpretation by orthopedic surgeons and compared their accuracy with that of radiologists. Fifteen disagreements between the two specialties regarding 524 preoperative radiographic interpretations were reported in a Canadian university teaching hospital, none of which changed the orthopedic management of any patient [49]. However, 61% of 371 radiologists' preoperative reports regarding fractures and implants were not available before the surgery at a general community hospital in Utah [21]. Four members of the hospital's orthopedic staff reviewed the radiographs and written radiologists' reports for studies of fractures with and without implants. They found that the descriptive findings in 15% of 145 studies of fractures without implants were omitted or were in error. Radiologists did not describe fracture alignment in 9% of studies and erred in 6% of fracture displacement descriptions. Descriptions of orthopedic implant type, effect, position, and stability were in error or absent for 32% of the relevant details in the 226 studies with implants.
Other studies have also compared accuracy of interpretation between radiologists and nonradiologists [50,51,52,53,54,55,56,57,58].
Technical quality.—If image quality is inadequate, accurate interpretation may not be possible. Insurers have sponsored or conducted the studies of image quality that have been published. Pennsylvania Blue Shield's study of 98 chest radiographs found five inadequately marked films, 21 incomplete reports, and eight problems in the technical quality of the image among the nonradiologists' radiographs, and no such findings among the radiologists' radiographs [48]. A quality audit by the same insurer assessed more than 1000 chest, foot and ankle, and spine radiographs obtained by 46 physicians and other providers in eight specialties in 1984 and 1985 [59]. Rates of unacceptable chest image quality were highest among studies by family practitioners (55% of 86) and pulmonary disease specialists (53% of 72) and lowest (6% of 36) among radiologists' studies. Unacceptable foot and ankle image rates were 82% of 147 podiatrists', 7% of 91 orthopedic surgeons', and 1% of 66 radiologists' studies. Sixty percent of 63 spine images by chiropractors were unacceptable, compared with 18% of 141 by orthopedic surgeons and 17% of 236 by radiologists.
U. S. Healthcare examined the quality of fetal sonography in 69 freestanding radiology centers, 31 obstetric and gynecology offices, and 27 hospitals in New Jersey and Pennsylvania to which it sends its patients. More than half the examinations performed by self-referring physicians did not meet guidelines established by the American Institute of Ultrasound in Medicine, the American College of Obstetrics and Gynecology, or the American College of Radiology, as reported by Modern Healthcare, a popular magazine [60].
Condition of the equipment and patient safety.—As part of a privileging program, Blue Cross Blue Shield of Massachusetts examined an extensive range of quality-related professional performance and technical aspects of diagnostic imaging conducted in more than 1000 facilities it reimbursed [1]. Significant numbers of providers were using imaging equipment for which they were not trained or were not certified to use. These providers consisted of 55 of 265 chiropractic, 41 of 179 podiatric, one of 130 orthopedic surgical, five of 40 obstetric and gynecologic, one of 84 primary care, one of 33 cardiology, and two of 16 other medical specialty sites. Approximately half of 179 podiatrist-operated and 265 chiropractor-operated facilities failed the program's site inspections, compared with approximately 3-5% of those operated by radiologists or cardiologists. Deficiencies were found in technical quality of images, equipment maintenance, and patient safety. Poor film technique and a variety of image deficiencies were detected. Common and significant safety hazards included failures to perform annual calibration of equipment, to post radiation exposure signs, and to use radiation monitoring devices, and a lack of a shielding plan or review of the existing plan by a medical physicist.
A quality improvement program conducted by a "large northeastern US HMO [health maintenance organization]" identified serious deficiencies among 78% of 92 nonradiology offices inspected, including 10% with equipment that had not been inspected in the previous year, 16% with inadequate image markings, 62% that lacked a formal radiology report, and 9% with uncorrected previously identified deficiencies [2]. No significant deficiencies were identified among the five radiologists' offices inspected.
Nonradiologist Referral to an Outside Facility in Which the Physician
Has a Financial Interest
|
TopReviewIntroductionNonradiologist Referral to Own...Nonradiologist Referral to an...Effects of Financial Incentives...Summary of FindingsResponses to Self-ReferralConclusionsReferences |
|---|
Physician ownership interest in outside health care facilities occurred more often in MR imaging (24%) and radiology (22%) facilities than in clinical laboratories (17%) or in ambulatory surgery (16%), physical therapy (9%), cardiopulmonary testing (2%), and nursing home (2%) facilities, according to the American Medical Association socioeconomic monitoring system 1990 core survey of non-federal patient care physicians [61]. Eight percent of all physicians had financial interests in some outside facility. Larger proportions of these had ownership interests in, and referred their patients to, outside radiology (16%) or MR imaging (15%) facilities than to clinical laboratory (11%), physical therapy (8%), cardiopulmonary testing (2%), ambulatory surgery (12%), or nursing home (<1%) facilities.
According to the American College of Radiology group practice survey, referring physicians had ownership interests in 21% of offices in which radiology groups practiced, which was equivalent to more than 500 offices in the United States [62]. Referring physicians represented a higher proportion having such financial interests than any other outside party (e.g., nonphysician entrepreneur, hospital, other radiologist).
Florida's high concentration of outside physicians who had investment interests in health care facilities made that state the locale of extensive studies [2, 61]. Physicians had ownership interests in nearly all of the 160 diagnostic imaging centers responding to a 1992 survey. A 1990 survey in the state indicated that 41% of physicians who were indirect owners through parent corporations chose to invest in diagnostic imaging centers more than any other type of facility. A much greater proportion (91%) of Florida physicians who had ownership interests were in specialities that are likely to refer their patients for surgery, diagnostic testing, and other services or equipment than were in specialities (9%) that provide services on a consultation basis and that were classified as nonreferring (psychiatry, radiology, pathology, anesthesiology) [63]. Physicians were also more likely to pursue joint ventures with the types of entities to which their specialty normally referred patients. The prevalence of outside physician ownership in freestanding MR imaging facilities was high in other states as well. More than 85% of freestanding MR imaging facilities in California and 75% of those in New Jersey had ownership arrangements involving referring physicians [64].
Utilization and Costs
A 1989 report by the Office of the Inspector General, United States
Department of Health and Human Services, based on two surveys of health care
providers in eight states, found that Medicare patients of outside physicians
with financial interests in independent diagnostic imaging centers received
13% more imaging tests than other patients
[5]. The General Accounting
Office reported on utilization and costs in relation to outside physician
ownership in Florida [6].
Analyses adjusted for physician specialty indicated that outside physicians
with financial interests in freestanding centers ordered 54% more MR images,
37% more nuclear medicine procedures, 27% more CT scans, 27% more
echocardiograms, 22% more sonograms, and 22% more complex radiographs than
nonowners. Referral rates for simple radiographs were similar for owners and
nonowners. The higher utilization was estimated to correspond to approximately
$10 million in costs annually in Florida. Restricting these types of
self-referrals nationally was projected to save Medicare $350 million and
Medicaid $37 million over 4 years.
The total MR imaging utilization rate per Medicare beneficiary in Florida was 45% greater than the national average, and charges for MR imaging per Medicare beneficiary were 55-65% higher [65]. Most of the difference in utilization was attributable to MR imaging in nonhospital settings, where volumes and total charges were 118% and 92% higher, respectively, than the national average. Higher MR utilization could not be explained by fewer CT scans, because CT utilization was also greater than the national average.
Quality
Quality-related issues emphasized in research on joint ventures include
appropriateness of referral decisions, quality control, and access to
services. Utilization review of workers' compensation claims in California in
1990 found 38% inappropriate MR imaging requests by self-referring physicians
compared with 28% by others, with costs of self-referring physicians' scans
estimated to be 31% higher
[66].
None of the more than 500 joint-venture facilities of all types identifiable in Florida was located outside metropolitan statistical areas [64]. The proportions of Medicaid and self-paying patients served in Florida joint-venture physician practices were smaller, 8.4% and 15.7%, respectively, compared with 14.1% and 21.6%, respectively, among non—joint-venture physician practices [63]. A multivariate analysis sustained these findings and also showed positive associations with joint-venture status of fewer Medicaid patients (odds ratio = 0.98, p < 0.01) and location of the practices in medium (odds ratio = 4.14, p < 0.01) and most (odds ratio = 6.52, p < 0.01) densely populated (urban) areas [63]. Comparable findings were noted for Virginia [67].
Effects of Financial Incentives and Payment Structures on Physician
Behavior
|
TopReviewIntroductionNonradiologist Referral to Own...Nonradiologist Referral to an...Effects of Financial Incentives...Summary of FindingsResponses to Self-ReferralConclusionsReferences |
|---|
The large northeastern United States HMO that established guidelines specifying who is permitted to perform imaging in ambulatory care settings (private offices, imaging centers, multispecialty clinics, and hospital outpatient departments) as a direct means to influence utilization and reduce costs, restricted each specialty's self-performed diagnostic imaging in these settings to its area of expertise. This initiative succeeded in decreasing the volume of radiography, the only modality it has thus far reported, by 20-25% from the previous trend; and the proportion performed by self-referrers decreased from 39% to 15% [2].
|
TopReviewIntroductionNonradiologist Referral to Own...Nonradiologist Referral to an...Effects of Financial Incentives...Summary of FindingsResponses to Self-ReferralConclusionsReferences |
|---|
Principal findings regarding utilization and costs are that utilization of a broad range of modalities in either elderly or younger populations, adjusted for case mix and physician specialty, was at least 1.7-7.7 times as high for self-referral on-site as for non—self-referrers. Costs in the younger population varied from three to 17 times as high, depending on the specialty and clinical presentation. Depending on modality, diagnostic imaging procedures were ordered 22-54% more often by physicians with ownership interests in outside facilities than by others.
Other areas of the literature concerning nonradiologists' on-site imaging services are less extensive, but the existing empiric evidence is nevertheless informative.
Principal findings regarding quality and access are that radiologists' interpretations of radiographs are generally more accurate than those of emergency physicians and most other imagers. However, some studies have found self-referring specialists, such as orthopedic surgeons, performing better at times than radiologists in imaging related to their field. Information on appropriateness is limited but indicates decreased appropriateness with self-referral.
Problems in equipment maintenance and calibration, technical quality of images, radiation shielding, documentation, and records management have been found to be common in self-referral on-site facilities. The location of joint ventures and the insurance mix of their patients show they serve underserved populations less frequently than other facilities do. Joint ventures may, in addition, reduce access to care for these populations by "cream skimming" away well-paying patients, thereby reducing the financial ability of other facilities to serve the underserved [69].
Strikingly little literature is available concerning what happens when financial incentives change. Principal findings are that financial incentives do influence physician behavior, with decreased financial incentives generally reducing the imaging performed or ordered by self-referrers, and vice versa.
Relatively direct evidence as to whether self-referrers' higher utilization constitutes overutilization is also limited. Such evidence consists of the finding of a higher rate of inappropriate imaging by self-referrers (but only one study has measured appropriateness) and the findings on the effects of financial incentives. In addition, there is the less direct evidence of higher utilization when physicians refer to an outside facility in which they have a financial interest, a situation in which higher utilization cannot be explained by the purported convenience and efficiency of having imaging on-site; and radically higher utilization rates are sometimes found in on-site self-referral. Overall, the evidence is suggestive of overutilization.
In short, self-referral constitutes most of nonhospital radiography and sonography and results in utilization and costs usually two or more times as high as in its absence, with overutilization likely; and self-referral generally involves poorer quality.
|
TopReviewIntroductionNonradiologist Referral to Own...Nonradiologist Referral to an...Effects of Financial Incentives...Summary of FindingsResponses to Self-ReferralConclusionsReferences |
|---|
The first type of response is found in federal and state legislation. The 1972 Medicare/Medicaid Anti-Kickback Statute prohibited the paying or receiving of any form of remuneration deliberately intended to induce the referral of Medicare and Medicaid business [70]. Federal physician self-referral laws were enacted in two stages, first in 1989 and then in 1992, in response to perceived abuses of self-referral, including those shown by the Florida studies. These so-called "Stark laws" (informally known by the name of their chief congressional sponsor) took effect in 1990 and 1995, respectively, and prohibit submission of Medicare claims and Medicaid reimbursement to states for certain health services, including diagnostic radiology, when the service provider has a financial relationship with the referring physician. Most states have enacted similar laws [71].
The main private approaches to self-referral are the second type. These approaches consist of insurers requiring providers to be certified in order to ensure the quality of their services, and insurers limiting reimbursement for particular services to certain types of physicians, both to assure quality and to limit on-site self-referral. Examples include a Blue Cross Blue Shield of Massachusetts privileging program [1] and the imaging guidelines of a large north-eastern HMO [2].
Blue Cross Blue Shield of Massachusetts introduced its professional and technical privileging for nonhospital diagnostic imaging providers to ensure that such providers have the appropriate background and training to interpret images and to maintain appropriate standards of care. For the professional component, the organization developed specialty-specific lists of imaging codes that it will reimburse, allowing nonradiologists to perform only imaging closely related to their field of practice. Criteria relevant for inclusion of a code were its recognition by the local specialty society as being integral to its practice, formal supervised training and testing, and credentialing at local facilities accredited by the Joint Commission on Accreditation of Healthcare Organizations. Specific certification or accreditation was required for sonography and nuclear medicine. Technical privileging required facilities to pass site inspection in six areas of quality for the relevant diagnostic imaging modality: staff training and qualifications, equipment specifications and performance, quality control policies and procedures, records management and storage, facility safety, and image audits. Many of those who performed procedures on a small scale did not bother applying for the program, which suggests that the mere existence of a program might help screen out providers whose quality is poor because of their low volume of imaging. Blue Cross Blue Shield of Massachusetts reports the program has led to important quality improvements and has identified the need for increased monitoring, education, and systematic assessment of imaging facilities. The program also resulted in a decline of approximately 2% in overall imaging expenditures during the first year, for an estimated return of more than 10:1 on the cost of its implementation.
Blue Cross Blue Shield of Massachusetts' claims reimbursement policy differed from that of the United Mine Workers of America health and retirement funds program, which had denied payment to nonradiologists for the professional component of all imaging services. By allowing the billing of approved specialty-specific codes, Blue Cross Blue Shield of Massachusetts' policy may have averted the increases in the volume of technical claims seen in the United Mine Workers of America health and retirement fund program.
The large northeastern HMO's radiology utilization program also comprised two major components, professional and facility. This insurer established a multispecialty health advisory committee that formulated professional guidelines regarding who could perform which imaging studies in their office practice, and the HMO evaluated the technical quality of the imaging performed using a questionnaire followed by inspections of some facilities. Professional guidelines were based on each physician's specialty and demonstrated expertise. Technical requirements pertained to a physicist's evaluation of the radiology equipment, output and dose to a patient, and condition of the equipment and to the availability of quality assurance programs, licensed technologists, and patient safety programs. The program identified significant deficiencies of equipment and maintenance and of reporting of examinations by nonradiologists. Evaluators found the program resulted in a significant decrease in the proportion of imaging performed by self-referrers and in the overall number of radiologic examinations. The HMO achieved its primary goals of reducing inappropriate utilization and costs and improving quality.
Previously, quality assurance programs had mainly been those of the Joint Commission on Accreditation of Healthcare Organization, which applied to hospital settings, leaving a deficiency in quality control of imaging in nonhospital settings [1]. That is a core reason the cited examples focus on nonhospital care.
Professional medical associations, including the American Medical Association, the American College of Physicians, and the American College of Radiology, have taken positions against self-referral. However, the American Medical Association's early explicit disapproval of joint-venture self-referral was followed by a more tempered stance [72]. The American Medical Association now recommends and emphasizes public disclosure of physicians' relevant financial interests as the primary precautionary measure.
|
TopReviewIntroductionNonradiologist Referral to Own...Nonradiologist Referral to an...Effects of Financial Incentives...Summary of FindingsResponses to Self-ReferralConclusionsReferences |
|---|
Nonetheless, with self-referral involving outside financial interests having received major legislative attention, most remaining self-referral presumably involves physicians' own on-site services, and a similar level of attention toward curbing the negative consequences seems warranted. Voluntary approaches in the private sector may be the most effective because public policy remedies require intrusion into the actual operations of a physician's practice, in a way largely alien to current public policy, through widespread requirements for credentialing and specialty-specific limitations on scope of practice. Effective strategies could follow the leads of the cited insurers' initiatives and, like them, be tied to reimbursement.
Examples from private insurers show that they can take a number of effective actions. They can limit physicians who are not imaging specialists—that is, nonradiologists—to the kinds of imaging closely related to their specialty. They can demand credentialing or do some of their own. They can also provide quality assurance guidelines and link their quality assurance policies to reimbursement.
An advantage of voluntary credentialing programs over those whose content is legislatively mandated is that flexibility can be built in, and the program can be used as an educational and process improvement tool rather than as a punitive measure. A limitation of voluntary credentialing as a solution to quality issues is that parties who have an interest in less rigid quality assurance requirements may develop credentialing programs that lack serious effect. Also, credentialing and accreditation are more suited to ensuring quality than to addressing utilization and cost issues.
New public policy actions are not on the horizon, so the onus rests on the private sector to be attentive to self-referral and to forestall its negative consequences using appropriate measures.
|
TopReviewIntroductionNonradiologist Referral to Own...Nonradiologist Referral to an...Effects of Financial Incentives...Summary of FindingsResponses to Self-ReferralConclusionsReferences |
|---|
1320a-1327b(b) (1972)
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  A. D. Kaye Radiologist versus Same-Specialty Referral: A Subset of the Issue Radiology, May 1, 2008; 247(2): 594 - 594. [Full Text] [PDF]
|
|
|
|
 |
|
 M. R. Tonelli Conflict of Interest in Clinical Practice Chest, August 1, 2007; 132(2): 664 - 670. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. K. Iglehart The new era of medical imaging--progress and pitfalls. N. Engl. J. Med., June 29, 2006; 354(26): 2822 - 2828. [Full Text] [PDF]
|
|
|
|
<
