2007 Survey of Radiologists: Practice Characteristics, Ownership, and Affiliation With Imaging Centers
2008; Elsevier BV; Volume: 5; Issue: 9 Linguagem: Inglês
10.1016/j.jacr.2008.03.011
ISSN1558-349X
Autores Tópico(s)Ultrasound in Clinical Applications
ResumoThe ACR conducts periodic surveys to assess the status of radiologists and their practices. The most recent survey was in 2007. One purpose of the survey was to obtain up-to-date data on top-line parameters on radiologists and their practices. That is the focus of this article, the first of two JACR articles highlighting the survey's findings. The 2007 survey was a stratified random-sample telephone quota survey of ACR radiologist members and practice leaders conducted in May and June 2007. In total, 601 responses were received from currently practicing radiologists. The response rate was 20%, and the margin of error was ±4.3% at the 95% confidence level. Survey results presented in this article fall into 3 areas, which include characteristics of radiologists and their practices, ownership of imaging equipment, and the affiliation of radiologists with nonhospital imaging centers. Some highlights are as follows: interventional radiology and neuroradiology were the most frequently reported subspecialties; the mean practice size was 20 full-time radiologists; radiologists worked an average of 50 hours per week, took 8 weeks of vacation per year, and performed 14,000 imaging studies annually; slightly more than half of radiologists owned imaging equipment, with magnetic resonance and computed tomographic scanners being the most likely equipment to be owned; nearly one-half said that they had equity interest in centers, 36% provided services to centers under contract, fewer than 5% were employees of centers, and the remaining nearly one-third reported no affiliation of any kind. The ACR conducts periodic surveys to assess the status of radiologists and their practices. The most recent survey was in 2007. One purpose of the survey was to obtain up-to-date data on top-line parameters on radiologists and their practices. That is the focus of this article, the first of two JACR articles highlighting the survey's findings. The 2007 survey was a stratified random-sample telephone quota survey of ACR radiologist members and practice leaders conducted in May and June 2007. In total, 601 responses were received from currently practicing radiologists. The response rate was 20%, and the margin of error was ±4.3% at the 95% confidence level. Survey results presented in this article fall into 3 areas, which include characteristics of radiologists and their practices, ownership of imaging equipment, and the affiliation of radiologists with nonhospital imaging centers. Some highlights are as follows: interventional radiology and neuroradiology were the most frequently reported subspecialties; the mean practice size was 20 full-time radiologists; radiologists worked an average of 50 hours per week, took 8 weeks of vacation per year, and performed 14,000 imaging studies annually; slightly more than half of radiologists owned imaging equipment, with magnetic resonance and computed tomographic scanners being the most likely equipment to be owned; nearly one-half said that they had equity interest in centers, 36% provided services to centers under contract, fewer than 5% were employees of centers, and the remaining nearly one-third reported no affiliation of any kind. IntroductionThe ACR conducts periodic surveys to assess the status of radiologists and their practices. The timing of the latest survey in 2007 was propitious, because it occurred as radiology practices were adapting to the latest round of Medicare payment reductions. The Deficit Reduction Act of 2005 became law early in 2006. One of the act's provisions caps the technical component of Medicare payment for physician office imaging to the lesser of the Hospital Outpatient Prospective Payment System or Medicare fee schedule payment, effective January 1, 2007. This provision, combined with other cuts, reduces payment for numerous, necessary imaging studies.Purpose of the SurveyOne purpose of the ACR survey was to obtain up-to-date data on top-line parameters on radiologists and their practices. That is the focus of this first of two JACR articles highlighting survey findings. Another purpose of the study was to assist the ACR in assessing the impact of the Deficit Reduction Act on radiologists by collecting information on the extent to which radiologists and their patients were affected by the implementation of this act. That will be the principal subject of the second article.Methodology and SampleThe 2007 survey was a stratified random-sample telephone quota survey of ACR radiologist members and practice leaders conducted in May and June 2007. Radiation oncologists were excluded, as were trainees and retirees. The strata were the states of Arkansas, California, Georgia, Iowa, Louisiana, Michigan, Montana, New Jersey, New York, North Dakota, Texas, West Virginia, and all other states grouped by the 4 census regions (Northeast, Midwest, South, and West). For the selected states, the sample included minimums of 30 responses for New York, 42 responses for California, and 20 responses for each of the other states on the list. The remaining states were treated as 4 groups by census region, and the minimum sample size required was proportionate to the population of these states in the census region. For example, the states in the South not listed above had 66% of the population of the South, and the minimum sample size collectively required from these states was 66.A pretest was conducted to assess the flow and interpretation of the questions. An e-mail message from the ACR board chair with a link to the Web survey was sent to radiologists with e-mail addresses. The remaining radiologists received a letter and practice manager worksheet by regular mail. An e-mail reminder was sent to e-mail sample nonrespondents. Follow-up phone calls were made to verify and correct information.Respondents were promised confidentiality; to further ensure confidentiality, the survey was conducted by an outside contractor, dmrkynetec (St Louis, Missouri), and the data set delivered to the ACR was stripped of all individual identifiers.In total, 601 responses were received from currently practicing radiologists, 488 from the ACR member list and 113 from the practice presidents list. By survey mode, the Web survey contributed 457, and another 144 were obtained via computer-assisted telephone interview. The response rate was 20% (601/[3,596 contacted − 512 disconnected phone numbers or wrong numbers − 95 not qualified]). The margin of error was ±4.3% at the 95% confidence level.Survey sample results were weighted to yield statistics representative of the entire population of active, posttraining radiologists in the United States. The weighting matched survey respondents to population estimates on group size, practice type (academic, private, etc), location (large metropolitan or small metropolitan area, main city or suburb, etc), and state. Using American Medical Association data for all radiologists by state and ACR data on ACR members by state, all weights were rescaled so that sum of weights added up to the population numbers of radiologists and members in 2007.ResultsTable 1 summarizes practice characteristics of radiologists in 2007. About one-third resided in the South, with the remainder being equally distributed throughout the other regions. One-half practiced in large metropolitan areas and another one-third in the main cities of small metropolitan areas. Fewer than 10% were in rural areas. About one-half of radiologists were in private, radiology-only practices, and 20% were in private, multispecialty practices. Slightly fewer than one-fifth were in academic practices. Five percent were in solo practices.Table 1Radiology practice characteristicsCharacteristic% DistributionCensus region Northeast22.6 Midwest21.3 South34.2 West21.9Practice location Main city of a large metropolitan area30.2 Suburb of a large metropolitan area20.9 Main city of a small metropolitan area33.4 Suburb of a small metropolitan area5.4 Nonmetropolitan location7.7 Varied locations0.2 Don't know/no answer2.2Practice type Private radiology, nuclear medicine, interventional radiology52.2 Private, multispecialty20.0 Primarily academic18.5 Solo practice4.9 Government practice1.0 Hospital0.9 Other0.3 Don't know/no answer2.2Subspecialty Interventional19.8 Neuroradiology16.4 Body imaging11.1 Breast imaging10.9 Musculoskeletal8.3 Abdominal imaging8.1 MRI7.9 Nuclear medicine6.0 PET/CT3.0 Ultrasound2.8 Pediatric radiology1.2 Other1.9 Don't know/no answer2.5Note: Percentages may not add to 100 because of rounding. MRI = magnetic resonance imaging; PET/CT = positron emission tomography/computed tomography. Statistics are weighted to represent the entire population of active, posttraining radiologists in the United States. Open table in a new tab Radiologists were distributed among numerous subspecialties, as shown in Table 1. Interventional radiology and neuroradiology topped the list at 20% and 16%, respectively. Percentages for the other subspecialties declined according to no apparent pattern.The number of radiologists in a practice and measures of physician work effort are shown in Table 2. In the survey, the number of full-time radiologists in a practice ranged up to 120; the mean was 20.1, and the median was 13. There was a wide distribution: one-fourth were in practices of 6 or fewer radiologists, and another one-fourth were in practices with more than 27 radiologists. As expected, practices typically had far fewer part-time radiologists, the average being 3.4.Table 2Number of radiologists in a practice and work effortVariableMean25th PercentileMedian75th PercentileFull-time radiologists per practice20.161327Part-time radiologists per practice3.4024Hours worked per week50.1455055Weeks of vacation per year7.86810Annual studies per full-time radiologist⁎In thousands.13.97.514.218.3Note: Statistics are weighted to represent the entire population of active, postresident radiologists in the United States. In thousands. Open table in a new tab The number of hours worked per week by radiologists averaged around 50. The interquartile range (ie, the difference between the 25th and 75th percentiles) of 10 was reasonably tight. Radiologists took an average of 8 weeks of vacation a year, with half getting between 6 and 10 weeks.An average of about 14,000 studies per radiologist were performed in 2007. There was considerable variation around the average, however. One-quarter did no more than 7,500, whereas another one-quarter did at least 18,000.The mean annual number of studies performed in a practice in the past 12 months was 208,000. Figure 1 depicts a wide and skewed distribution. One-half of practices did fewer than 150,000. Ten percent did 500,000 or more. The number of studies can be expressed on a per full-time radiologist basis (Table 2). On average, radiologists performed about 14,000 studies annually. There was, however, considerable variation around the average, with more than a two-fold difference between the 25th and 75th percentiles.Survey participants were asked, Do you yourself, or you as part of your group, have an ownership interest or equity partnership in any radiology office or imaging center facility or other entity that owns diagnostic imaging equipment used to perform tests outside the hospital setting? This could be ownership through your practice or through a separate partnership, joint venture, corporation, etc.Fifty-six percent of respondents answered “yes,” and 44% said “no.” Those who were owners by virtue of responding in the affirmative were asked which imaging modalities they owned, with multiple mentions being allowed. Figure 2 graphically displays the percentage of radiologists owning each modality type. Computed tomography and magnetic resonance imaging were the most frequently owned modalities, at about 80% for both. Positron emission tomographic scanners were the least likely to be owned.Fig 2Modality ownership. Multiple mentions were allowed. CT = computed tomography; MRI = magnetic resonance imaging; PET = positron emission tomography.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Table 3 presents detailed responses by modality and location. Nationally, the average modality ownership rate was 52%, which means that radiologists who were equipment owners were about as likely to own a given modality as not. For most locations, the modality ownership rate departed from the national average by only a few percentage points. Main cities tended to have higher rates than suburbs, but the differences were small. In the main cities of large metropolitan areas, ownership rates for computed tomography and magnetic resonance approached 90% and were significantly greater than the national average; on the other hand, mammography ownership was substantially below average in large cities. Magnetic resonance ownership was more likely to be found in main cities than in suburbs of those cities. Radiologists in smaller cities had above-average ownership of ultrasound, mammography, and interventional radiology modalities.Table 3Type of imaging equipment owned by radiologist equipment owners (percentage who own)ModalityAll LocationsMain City of Large Metro AreaSuburb of Large Metro AreaMain City of Smaller Metro AreaSuburb of Smaller Metro AreaNonmetro or RuralCT80.789.579.376.267.491.1MRI80.388.869.081.567.3100.0PET26.035.832.614.940.636.3PET/CT31.841.633.627.737.226.3Other nuclear med33.926.128.041.337.579.7Ultrasound73.567.564.980.560.991.2Mammography56.832.556.774.656.581.0Interventional29.327.620.939.99.660.4All modalities51.551.248.154.647.170.8Note: Statistics are weighted to represent the entire population of active, posttraining radiologists in the United States. Sample was restricted to respondents who owned at least one type of imaging modality. Multiple mentions were allowed. CT = computed tomography; MRI = magnetic resonance imaging; PET = positron emission tomography. Open table in a new tab The nonmetropolitan and rural locations stand out from the others. The overall ownership rate was 71%, 20 percentage points higher than the average for all locations. Furthermore, ownership rates for nearly all individual modalities were substantially above their national averages.Significant regional differences in ownership existed for some modalities, as shown in Figure 3. Radiology practices in the West, for example, owned significantly more magnetic resonance imaging scanners and ultrasound machines than did their counterparts in other regions. Those in the Northeast owned more positron emission tomographic and other nuclear medicine equipment than average, whereas radiologists in the Midwest tended to own fewer positron emission tomographic and positron emission tomographic/computed tomographic scanners than their counterparts in other regions.Fig 3Regional differences in modality ownership. Multiple mentions were allowed. CT = computed tomography; MRI = magnetic resonance imaging; PET = positron emission tomography. *Different from the national average at the .01 level. †Different from the national average at the .05 level. ‡Different from the national average at the .10 level.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Survey participants were asked if they were affiliated in any of several ways with diagnostic imaging centers that perform imaging studies outside the hospital setting, with multiple responses being allowed. Figure 4 summarizes their responses. Nearly one-half said that they had equity interest in centers. Thirty-six percent provided services to centers under contract. Fewer than 5% were employees of centers. The remaining nearly one-third of respondents reported no affiliation of any kind. As shown in Figure 5, one-half of respondents had one type of affiliation. Another 19% had 2 types, a high majority of which were “equity partnership” and “contractor.” All 3 types were reported by only a fraction of respondents.Fig 4Affiliation with a diagnostic imaging center that performs imaging studies outside the hospital setting. Multiple mentions were allowed.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Fig 5Number of affiliations with imaging centers.View Large Image Figure ViewerDownload Hi-res image Download (PPT)The interface between imaging center affiliation and radiology practice ownership status illuminates some interesting relationships. Table 4 presents the percentage of respondents who fell into a series of five 2 × 2 grids, with rows representing affiliation or nonaffiliation and columns representing ownership or nonownership. The percentages in the 4 cells sum to 100%. For example, radiologists with some type of affiliation were nearly 3 times more likely to be practice owners than nonowners (52% vs 18%). This was especially true in the case of equity partnerships, for which the odds of someone with an equity interest in a center being a practice owner instead of a nonowner were about 46 to 3. An imaging center contractor was about twice as likely to own a practice as not own a practice. In contrast, relatively few radiologists were salaried employees of imaging centers, and there was no practical difference in the likelihood of practice ownership compared with nonownership among these imaging center employees.Table 4Imaging center affiliation and practice ownershipAffiliation With Imaging CenterRadiology Office or Imaging Center OwnershipOwnerNot OwnerSome affiliation with imaging center52.2%17.7%No affiliation with imaging center3.526.6Equity partner in imaging center45.82.6Not an equity partner in imaging center9.941.7Contractor to imaging center23.612.8Not a contractor to imaging center32.131.5Salaried employee of imaging center1.81.9Not a salaried employee of imaging center42.553.8Other affiliation with imaging center0.41.0Not other affiliation with imaging center55.343.3Note: Statistics are weighted to represent the entire population of active, postresident radiologists in the United States. Multiple affiliation mentions were allowed. Open table in a new tab DiscussionThe most recent ACR survey of radiologists provides a snapshot overview of radiologists and their practices in 2007. A 2003 survey of radiologists contained some questions similar to those in the 2007 survey, which permits a 4-year trend comparison [1Sunshine J.H. Lewis R.S. Bhargavan M. A portrait of interventional radiologists in the United States.AJR Am J Roentgenol. 2005; 185: 1103-1112Crossref PubMed Scopus (52) Google Scholar].Geographically, in 2007, radiologists were more highly concentrated in the Northeast in comparison with the US population (23% vs 18%, respectively); the proportions of radiologists in the other regions were all 1 to 2 percentage points lower than those for the US population [2US Census BureauTable 1: annual estimates of the population for the United States, regions, states, and Puerto Rico: April 1, 2000 to July 1, 2007.http://www.census.gov/popest/states/tables/NST-EST2007-01.xlsGoogle Scholar].The regional distribution of radiologists in 2007 did not shift in significant ways from the situation that existed in 2003. The percentage in the West increased about the same as the percentage in the Midwest fell. That does not imply that a direct migration occurred from Midwest to West. More likely, there were back-and-forth flows between all regions. It so happened that the net outflow from the Midwest roughly matched the net inflow to the West, while there were no net inflows or outflows for the Northeast and South. Did radiologists follow their relocating patients? Apparently not in a systematic way. The US population shifted out of the Northeast and Midwest into the South and West. The Midwest's share of the US population dropped only 0.5 percentage points between 2003 and 2007, whereas the share of radiologists fell by 3%. Furthermore, the share of radiologists in the West rose by 2 percentage points more than the share of the US population in that region did over that period.American radiologists were more urbanized than the rest of the citizenry. Ninety percent of radiologists practiced in urban areas in 2007, whereas the percentage of the US population residing in urban areas was closer to 80% [3US Census BureauThe U.S. statistical abstract: Table 29: urban and rural population by state: 1990 and 2000.http://www.census.gov/prod/2007pubs/08abstract/pop.pdfGoogle Scholar]. Moreover, urbanization of radiologists increased from 2003 to 2007. The fraction of radiologists in large metropolitan areas and the main cities of small metropolitan areas increased by 8 percentage points between 2003 and 2007, with associated decreases in small metropolitan suburbs and metropolitan areas.The types of practices in which radiologists work have changed significantly in recent years. Whereas 40% worked in private, radiology-only practices in 2003, slightly more than half did so in 2007. This shift was accompanied mainly by a reduction of comparable magnitude in the percentage working in multispecialty practices. The other practice types that experienced share reductions over the 4-year period were sole practitioners and “others” (locum tenens, government, and hospital). The share in academic practices rose slightly. The trend toward larger practice sizes could reflect increasing pressure on radiology practices to become more efficient by scaling up and to gain more market power with payers and hospitals in an era of ever tightening public and private payer reimbursements. It is also likely due in part to the proliferation and diffusion of teleradiology and the ability to transmit digital images to a centralized location. Rural sites now need only one on-site radiologist for contrast procedures, and the remainder of images can be sent to a metropolitan center or other location for interpretation.There was a large variation in the number of imaging studies performed in a practice in 2007. Not surprisingly, the number of radiologists in a practice accounts for much of the variation. The correlation between the number of studies and number of full-time radiologists in a practice was 0.56 (P < .001). Not only did the number of radiologists in a practice affect the total number of studies, but the number of studies done by each radiologist played a role as well. As previously discussed, the number of studies per full-time radiologist varied considerably in 2007, with a more than two-fold difference between the 25th and 75th percentiles.The most frequently radiologist-owned imaging modalities were computed tomography and magnetic resonance. One reason is that these modalities tend to be among the most frequently performed in the nonhospital setting [4Cohen D. Radiology 2005: state of the industry.Imaging Econ. 2005; 18: 24-34Google Scholar]. Modality ownership and modality utilization are not the same thing but are probably highly correlated. Another reason could be that ownership of these expensive, complex scanning machines can yield an attractive financial rate of return compared with other types of investments [5Armstrong D. MRI and CT centers offer doctors way to profit on scans.The Wall Street Journal. 2005; (May 2)Google Scholar]. Radiologists' ownership of mammographic scanners in large cities is well below that in other locations. Mammography reimbursement is relatively low, compared with other modalities. Digital startup costs are high, and malpractice suits represent an above-average risk. Therefore, it is somewhat distressing that ownership of mammographic equipment in large cites is significantly below ownership rates in other locations, if low utilization correlates with low ownership. The implications are that some women in large cities may not be getting mammograms at recommended intervals because of long waits for appointments. This could be especially true for lower income women in urban cores who lack health insurance or knowledge of their mammography needs.It was noted that equipment ownership rates in nonmetropolitan and rural areas were much higher than in urban locations. This may signal that imaging utilization is at least adequate. However, given the long distances that patients may have to travel in rural America, high radiologist ownership does not necessarily imply adequate access in those areas. This is particularly true for positron emission tomography and positron emission tomography/computed tomography, both of which had lower radiologist ownership rates relative to the national average and were much lower than rates for other imaging modalities in nonmetropolitan and rural areas. Low radiologist ownership in large metropolitan areas does not necessarily imply poor imaging utilization in those locations. Ownership in metropolitan areas could be lower in part because hospitals tend to own imaging equipment, in which case utilization is probably adequate despite radiologist ownership being low.Differences by region in the types of modalities owned were found, for example, more magnetic resonance imaging and ultrasound ownership in the West. Such differences undoubtedly had something to do with variations in practice styles, among other factors.It is not surprising that among respondents who were owners of radiology offices or imaging centers, a large majority (82%) had equity interest in imaging centers, because of the nearly tautological definitions. What is perhaps more surprising is that a substantial portion of practice owners (42%) also were contractors of imaging centers. These could be contracts with centers other than ones owned by such individuals, or contracts with imaging centers by radiologists who own the practices in which they work but do not own imaging centers. The main finding is that the interface between practice ownership and affiliation with imaging centers is not a straightforward one.Study Strengths and LimitationsThe ACR's 2007 survey of radiologists yielded important information on several topics of interest to the profession that either had not been studied before or had not been updated in several years. Strengths include the fact that the data are from a moderately sized, carefully conducted survey that achieved a good response rate. Weighting adjusted for nonresponse bias (ie, differences between respondents and nonrespondents) in the characteristics used in the weighting.Nonetheless, the survey had some limitations. Survey timing and resource constraints limited the ability to provide meaningful statistics for detailed breakouts. As with almost any survey, statistics drawn from the 2007 survey of radiologists may have inaccuracies from several sources: sampling variability, nonresponse bias, and incorrect or illogical responses. IntroductionThe ACR conducts periodic surveys to assess the status of radiologists and their practices. The timing of the latest survey in 2007 was propitious, because it occurred as radiology practices were adapting to the latest round of Medicare payment reductions. The Deficit Reduction Act of 2005 became law early in 2006. One of the act's provisions caps the technical component of Medicare payment for physician office imaging to the lesser of the Hospital Outpatient Prospective Payment System or Medicare fee schedule payment, effective January 1, 2007. This provision, combined with other cuts, reduces payment for numerous, necessary imaging studies.
Referência(s)