With the turn of the new millennium, many experts, professional pundits, and scholars reviewed the events of the past century, and attempted to offer predictions of what may happen for the next 10 decades. Here at Decisions in Axis Imaging News, we have made a contribution of our own by conducting a census of sorts. Our scope is a bit more narrow, but the spirit is the same-where is radiology now? Where might the field be going? What can we draw from indicators of the recent past? What trends might we see?

In an effort to provide a pool of telling data, Decisions turned to the American College of Radiology (ACR), Reston, Va; the American Hospital Association (AHA), Chicago; as well as a coterie of radiology management and consulting companies. These included HealthHelp, Houston; National Imaging Associates (NIA), Hackensack, NJ; and IMV Medical Information Division (formerly Technology Marketing Group), Des Plaines, Ill. In publishing this data, it is not our intention that it be used for benchmarking purposes. The data and information that these organizations willingly provided allow a spot check on the ebb and flow of the radiology profession. But even with this limited view, several trends of interest emerged, which the readership may also find of interest and perhaps even utility.

Radiologists and Subspecialties

Figure 1. Distribution of radiologist subspecialization based on 1995 survey. Bar shows distribution of general diagnostic radiologists vs. subspecialists. Reprinted with permission: AJR 174:1203-1209; 2000 Sunshine J, Crewson P. Diagnostic radiologists’ subspecialization and fields of practice.

Probably the best place to begin a census is with the basic building blocks, the radiologists. According to ACR estimates, there are about 25,000 working radiologists in the United States (excluding residents, fellows, and retirees). The number does not include nuclear medicine physicians, however. According to Jonathan Sunshine, PhD, director of research at ACR, “Most nuclear medicine specialists did not train in radiology, are not ABR [American Board of Radiology] certified, and do not have links to radiology such as being ACR members.” Despite that, Sunshine says, “My very rough guess is that about one-third of nuclear medicine practitioners have any of these ties to radiology.” An additional fact is that the number of radiologists is growing. Five years ago, the total estimated number of radiologists was 23,000, and ACR data indicate that the ranks are swelling per annum with about 500 new radiologists. Sunshine projects that there will be approximately 27,500 in 5 years, and about 30,000 in the profession in a decade’s time. These general estimates of the overall profession show a paced growth.

The profession’s growth notwithstanding, the number of radiologists is not adequate to accommodate current demand,1 as previously reported in this journal.

Equally as interesting are numbers about differentiation in diagnostic radiology. In a recent paper coauthored by Crewson and Sunshine,2 the researchers found that based on a 1995 survey of more than 1,200 practicing diagnostic radiologists, more than a quarter (28%) of those surveyed were specialists (figure 1, page 22). The majority of these physicians who subspecialized were in interventional or vascular radiology (28%), followed by neuroradiology (20%), MRI (12%), mammography (9%), CT (9%), sonography, pediatric radiology (both 7%), and nuclear medicine (6%).

Furthermore, Crewson and Sunshine noted that subspecialists were, on average, younger than generalists (45 vs 49 years of age), and had fewer years of experience (11 vs 15 years). The researchers also found significant variations in practice and environmental statistics: subspecialists were as a group more likely than generalists to work full time (95% vs 88%), practice in an academic institution (43% vs 9%), and practice in a larger group (an average of 19 vs 10 radiologists per group). Additionally, subspecialists as a group were more likely than generalists to work in a center city (44% vs 17%), and less likely to work in a nonmetropolitan or rural area (4% vs 21%). Although 40% of the generalists had fellowship training, such training was much more common for subspecialists (68%). Facilities where the radiologists were employed also provided interesting data: although most generalists (58%) practiced in one or two locations, those specialists in MRI (51%) and neuroradiology (52%) practiced in three or more locations.

The paper also noted that while women radiologists accounted for only 11% of generalists, they accounted for more than 30% of those in subspecialties-pediatric radiology (38%), mammography (36%), CT (34%), and sonography (32%).

The authors report that the trends in the data that they had looked at between 1990 and 1995 indicate that “subspecialization is likely to be greater today than was reported from the 1995 study.”

Departments, Centers, groups

Figure 2. Cost comparison of 1998 and 1999 imaging for more than 3 million non-Medicare individuals. Utilization increases, percentage of utilization, and percentage of spending also noted. Source: NIA

According to data from the AHA, there are 6,247 hospitals in the United States, with 4,956 of those termed community hospitals (non-Federal, nonpsychiatric, and short-term stays). AHA statistics also show that of those hospitals that reported, 3,683 provide ultrasound, 1,497 single-photon-emission computed tomography (SPECT), 224 positron-emission tomography (PET), 2,174 MRI, and 3,619 CT scanner services.

But Sunshine points out that the actual number of radiology groups in this country is much smaller than the number of hospitals, for many groups staff multiple hospitals. In 1995, the ACR estimated that there were about 2,600 physician groups in the United States with multiple diagnostic radiologists. “This includes all types of groups-multi-specialty, academic and government (Veterans Affairs, Department of Defense), as well as private radiology. The number of groups is probably somewhat less now, as there’s been some consolidation and merger activity, but not a great deal,” comments Sunshine.

As for solo-practice radiologists, ACR data indicated in 1995 that there were about 1,700 such physicians. Although Sunshine says this quantification probably exhibits a downward trend in the intervening 5 years, he guesses that the numbers have not declined in any great manner. Additionally, Sunshine points out that “some of these solos provide the radiology services at multiple small rural hospitals.”

With freestanding imaging centers, the ACR in its data gathering uses the term “nonhospital offices” to include both imaging centers and these offices. The most recent statistical data come from 1992, when about 3,200 were counted, including radiation oncology. Sunshine thinks as a very rough estimate that there were 2,000 diagnostic radiology nonhospital practice sites in that year. Although no studies have taken place in the interim, Sunshine says it is possible that these numbers may have changed-possibly a slight increase.

utilization leaps

NIA’s database contains diagnostic radiology utilization and cost data on a time-series (1996-2000) basis for more than 70 million individuals across the United States. Analysis of this information reveals the major cost and utilization trends currently driving diagnostic radiology cost spending higher.

According to NIA’s data, one of the fastest growing markets for diagnostic radiology utilization is the Northeast. A review of data from 1999 for more than 3 million non-Medicare individuals across New York, New Jersey, Connecticut, Pennsylvania, and Massachusetts shows that the total cost for diagnostic imaging services was up more than 10% in the year (figure 2, page 25).

A deeper look shows that this growth in spending is being fueled by utilization increases in principally four imaging modalities: bone densitometry, MRI and MR angiography (MRA), CT, and nuclear cardiology. While these modalities typically account for less than 16% of the total volume of diagnostic imaging services in NIA’s universe of data, they represent more than 50% of total spending (figure 3).

“It is the disproportionate increase of these more expensive imaging modalities that are placing upward pressure on total costs,” says Robert LaGalia, senior vice president for strategy and development, NIA. For instance, data from one Northeast plan shows total costs rose 13.5% in 1999 vs costs in 1998. Breaking out MRI/MRA, CT, and nuclear cardiology as a group shows that these costs were up more than 20.5% vs only 8.3% for all other imaging modalities combined.

Examining spending and utilization by modality more closely from the nine plans provided by NIA, the year-on-year per member per month (PMPM) change-in-cost increases focus most strongly on bone densitometry (52%), nuclear cardiology (26.1%), MRI/MRA (19.9%), and CT (17.6%). Collectively for 1998, CT, MRI, and MRA accounted for 11.4% of the volume of utilization in that year, yet these three modalities were responsible for between 60% and 75% of the cost increases seen in NIA databases in 1999. Additionally, it should be noted that while radiography was utilized more than any modality, the cost increase in NIA’s universe was only 4.3%, and female-oriented modalities (such as obstetric ultrasound and mammography) together accounted for increases over the previous year of 11.6% and 4.1%, respectively (figure 4, page 26).

Figure 3. Distribution of utilization by modality illustrates that 15 percent of procedures drive 75 percent of cost increase. Source: NIA

Based on data from this past year, the company’s outlook is that costs will continue to accelerate even beyond the pace set in 1999. “We are seeing total cost inflation for diagnostic radiology services coming in at double-digit rates for 2000 over 1999, again fueled by the increased usage of MRI, CT, and nuclear cardiology procedures,” declares LaGalia. “Alarmingly, the pace is beginning to pick up, with MRI spending in some key Northeast states up more than 40% in the year.” The Northeast is not the only area of the country put under pressure by rising costs. According to LaGalia, areas of the Midwest have reported similar cost increases with one organization noting MRI cost inflation of more than 50%.

In dollar terms, the effects of the disproportionate increase in the use of MRI/MRA, CT, bone mineral densitometry, and nuclear cardiology procedures are more stunning. LaGalia points out that the data show these modalities drive between 60% and 75% of the total incremental dollars spent for diagnostic radiology services. LaGalia also believes that although the data is site-specific to the NIA universe, it is also representative of the cost increases in radiology as a whole. According to NIA’s special report for Decisions in Axis Imaging News, “Overall, per member per month (PMPM) costs were up $0.86 PMPM, or 10.3% in 1999 compared to 1998. Of this increase, a group of imaging procedures-bone mineral densitometry, CT, MRI, MRA, diagnostic nuclear medicine, and nuclear cardiology–representing less than 16% of total health plan utilization drove 75% or $0.64 of the cost increase.” Interestingly, while there have been utilization increases across the board from 1998 to 1999 in the data provided by NIA, the costs of other modality imaging in aggregate were up only 4.7% in the same time frame.

Looking at the utilization of modality imaging per 1,000 patients (both inpatient as well as outpatient) across all nine of the plans in the commercial sector, radiography usage garnered the highest rates (averaging 46.7% of total volume by our calculation), followed by general ultrasound at 12.8%, mammography at 12.3%, and CT, with an average of 7.7% of total usage rates (although in some plans, CT usage significantly exceeded mammography). In the Medicare sector, similar usage patterns can be found (with radiography usage averaging about 53.3% of the total utilization rates, yet at 6.8% for ultrasound), but with CT usage (12.16%) exceeding mammography (8.3% of image utilization).

figure 4. An analysis of 1999 cost increase by imaging modality. Source: NIA

In viewing the NIA data that break down utilization rates via gender, trends emerge that echo and amplify the overall tendencies. For example, as the population becomes elderly, the utilization of radiography increases, where patients 65 years of age and older require the most coverage. Also, obstetric ultrasound utilization spikes for females in the 20 to 44 age range and then declines to single digits once again, while general ultrasound utilization also spikes, but has a more gradual trailing off in numbers. Additionally, image utilization rates for mammography are highest among women in the 45 to 64 age range, although the utilization rate for the same modality for women 65 and over drops only minimally. For men, the highest rate usage outside of plain films is CT, with a tripling of imaging from males in the age range of 45 to 64, to those 65 years of age or older (figure 5, page 26).

Modalities, Reading Rates, More

Figure 5. Imaging utilization rates per 1000 patients by modality organized by age groups. Source: NIA

HealthHelp’s contribution to this article includes an analysis of data from eight plans (six had a Medicare line of business) from different areas of the country, including: Florida, rural states of the Northwest and Southwest, and metropolitan areas in the Northeast and the central United States. The purpose was to project diagnostic imaging utilization patterns of Medicare-aged individuals in the Baby Boom population in an effort to project the volume and dollar increase experienced nationally when this age group reaches age 65. While there was not space in this article to outline the precise mechanics of this statistical analysis, the complete report on the HealthHelp study will appear online when the January/February issue is posted at www.imagingeconomics.com , embedded in the online version of this story.

Figure 6. Utilization based on analysis of Medicare claims data. Source: HealthHelp

In their study, HealthHelp researchers determined that a Medicare-aged individual had approximately 2.5 diagnostic imaging procedures annually (or nearly 2,500 images per 1,000 people) on average (figure 6, page 28). But in more specificity, the breakdown for utilization rates via modality (per annum) for patients 65 and older included: general x-ray (1,300 times per 1,000 people), mammography (280), CT (251), ultrasound (202), nuclear medicine (107), MRI (78), fluoroscopy (71), and bone densitometry (64), the least utilized.

A cost PMPM was calculated using the plan’s allowed amount per procedure (figure 7, page 28). The plan’s allowed amount was also used in determining the average paid per procedure. In an attempt to address fee-for-service variations in each plan’s reimbursement schedule, the 1999 unadjusted Medicare Allowable was applied to produce a per member per month figure (MCA PMPM).

Figure 7. Cost per Medicare member per month was calculated using plans’ per member per month PMPM cost; average age amount paid per procedure PPP; and 1999 unadjusted Medicare Allowable MCA rate. Source: HealthHelp

Via a special statistical manipulation and utilizing age bands from the 1999 US Census, HealthHelp researchers enacted an age progression on the population group of Baby Boomers who are presently in the 45 to 59 age range. This group was aged until the year 2020 when the youngest of the individuals in this group reached age 65 (figure 8).

The results are eye-opening regarding the financial impact of the aging population bulge as it moves through the Medicare system with an increase in spending of more than 140% by the year 2020 that more than doubles the 1999 annual expenditure. In the year 2010, utilization is projected to increase by 37%, and by 2015, it is estimated to jump by 85%. In 2020, the volume of radiology-related procedures will more than double the 1999 volume, with an estimated 140% increase in annual imaging procedures. In 2000, all diagnostic imaging volume was nearly 85.35 million images. When those presently in the 55 to 59 age bracket hit retirement in 2010, the volume of imaging is projected to be 31.81 million procedures annually for them alone; when those currently 50 to 54 turn 65 in 2015, the associated volume of diagnostic imaging is estimated to be 40.64 million images.

Figure 8. Imaging volume projections for aging Baby Boom population groupings broken down by modality during their Medicare years. Source: HealthHelp

When the Baby Boomers hit 65 in 2020, HealthHelp’s researchers project 47.83 million images taken annually. While not adjusted for mortality, the total projected number of images for all persons 65 and older is a staggering 205.63 million images annually.

HealthHelp researchers also attempted to calculate the impact on imaging utilization rates of various disease states in Medicare-aged men and women. Utilization rates for diagnostic imaging performed for the diagnosis and management of selected diseases were calculated first for all Medicare-aged men and Medicare-aged women, sick or well. Those utilization rates were then divided by the utilization rates for all diagnostic imaging on Medicare-aged men and all Medicare-aged women to arrive at a percentage of total imaging contributed by each disease state.

Figure 9. Percentage of all diagnostic imaging within study population performed on Medicare-aged men for the diagnosis and management of five disease states. Source: HH

For instance, 17% of all diagnostic imaging done on Medicare-aged women is for the diagnosis and management of breast disease. And 6% of all diagnostic imaging done on Medicare-aged men is for the diagnosis of heart disease. The image utilization rates per 1,000 individuals show a modestly lower rate among females for heart disease and stroke, but dramatically higher rates for osteoporosis and breast disease. Imaging utilization among females with lung cancer was slightly higher than that for males with lung cancer (figures 9 and 10).

Using volume projections for aging population groupings, all disease states not surprisingly show image utilization growth (figures 11 and 12, page 30). Examples of large increases in utilization include both female heart disease (from the present 1.17 million images per annum to a projected 2.62 million in 2020) and male heart disease (from the present near 916,000 to a projected 2.43 million in 2020). Female osteoporosis imaging levels are projected to climb from 991,000 images presently to a projected 2.2 million images in 2020. In fact, total volume of imaging utilization for all female disease states in the 65 and over age band was higher than that for their male counterparts, possibly due to a higher mortality rate for men.

Figure 10: Percentage of all diagnostic imaging within study population performed on Medicare-aged women for the diagnosis and management of five disease states. Source: HH

HealthHelp also attempted cost projections, which proved to be problematic. According to the HealthHelp report, “…a forecast of specific dollar expenditures was not possible using the data in this study. The $26 per member per month (PMPM) figure generated from the claims data was apparently much higher than the approximate $9.43 PMPM paid by HCFA for imaging in 1999. Projecting the study PMPM would greatly overstate the forecasted expenditure. Using that $9.43 PMPM [benchmark], however, one can show an increase of more than $5 billion over the next 20 years with an annual [cost] of almost $9.5 billion in 2020 [figure 13].”

This exercise offers insight into the financial impact of the aging Baby Boomers as well as the demands that will be placed on the radiology profession in the years to come.

Equipment Age and Planned BUYS

With more than 4,874 hospitals and 4,274 independent sites performing imaging or therapeutic services in its marketing database, IMV’s contribution is national in scope. The information that IMV supplied covers modalities ranging from radiation oncology to ultrasound.

Figure 11. Volume projections for female Baby Boomer imaging utilization by disease state during Medicare years. Source HH

The core therapeutic and diagnostic services in the chart in figure 14, page 34, represent an older technology base in comparison to the newer modalities.

Fluoroscopic-based imaging, nuclear medicine, and radiation oncology represent that core group of older modalities. “X-ray data are not collected by IMV but could generally be included in this core technology and age group,” believes Mitchell Goldburgh, vice president and general manager.

The IMV data show that there is strong interest in the newer nonradiographic-based imaging modalities, according to Goldburgh. This has created a demand for investment in modalities such as MRI and ultrasound. Additionally, CT, although x-ray based, is changing its technology fundamentals with the introduction of multislice capability and the speed of scanning. Additionally, there is also a push toward replacing the R/F installed base as this modality is converted to digital technology. Because of these technology advancements, the projected 3-year purchase plans reflect a healthy ratio of planned purchases to sites responding.

Figure 12. Volume projections for male Baby Boomer imaging utilization by disease state during Medicare years. Source HH

A highlight in the chart is the planned new CT purchases. The IMV universe of sites (locations performing CT procedures) was 5,940 with 6,417 CT scanners identified as installed. Among these sites, the respondents to the IMV survey indicated that 1,946 sites had plans to purchase new CT units in the years 1999-2001. These potential planned purchases are given impetus by the age of the installed CT scanners (in 1999, 53% of these units were reported to be at least 5 years old) and the technological shift with the introduction of multislice capability.

Another interesting facet of the IMV data shows that the picture archiving and communications systems (PACS) (multiple modalities, but sharing the same common network, workstations, and archive) and miniPACS (one modality, but sharing the same archive, network, or workstations) appear in 1999 to still be an emerging market. Respectively, 28% and 13% of the hospitals and imaging sites surveyed planned to purchase this technology. “The decrease in technology costs, and potential operational benefits from this investment, may very well change this emergence dynamic as IMV surveys the market in 2000,” comments Goldburgh.

Figure 13. Imaging volume projections of Medicare-aged Baby Boomers. Source: HH

Even though there are many installations approaching their typical replacement cycle of 5-7 years, Goldburgh cautions that a fraction of this total will be replaced as the total that radiology can spend in each budget cycle is limited. As for costs shown in the chart for each modality, Goldburgh cautions that the prices are generalizations, and that these average numbers are rapidly changing.

Toward an Expanding Presence

Figure 14. Analysis of the installed imaging technology base at the respondents’ sites to produce amount of possible planned purchases and replacement buys. Source: IMV

Probably of all threads running through the data presented here, the ones that are the most pronounced are that the field and the numbers of practitioners of radiology were growing. Total cost of imaging to payors-particularly MR- and CT-related imaging-is increasing substantively. Subspecialties in the field are proliferating. So too is the need for radiologic services, as an aging Baby Boomer population reaches retirement age and beyond.

Related Articles for Download:

Aging Study Summary of Findings
Aging Study Synopsis

Peter Pesavento is associate editor of Decisions in Axis Imaging News.


  1. Pesavento P. Where have all of the radiologists gone? Decisions in Axis Imaging News. 2000;13(5):48-53, 62.
  2. Crewson PE, Sunshine JH. Diagnostic radiologists’ subspecialization and fields of practice. AJR. 2000;174:1203-1209.