| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Diagnostic Radiology, Yale University School of Medicine, 333
Cedar St., TE-2, New Haven, CT 06510.
2 Research Department, American College of Radiology, 1891 Preston White Dr.,
Reston, VA 20191.
3 Department of Diagnostic Radiology, Memorial Sloan-Kettering Cancer Center,
1275 York Ave., New York, NY 10021.
4 Department of Economics, Yale College, Yale University, New Haven, CT
06510.
5 School of Management, Yale University, New Haven, CT 06510.
Received February 14, 2003;
accepted after revision February 25, 2003.
Address correspondence to H. P. Forman
(howard.forman{at}yale.edu).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. All diagnostic radiology positions advertised in the American Journal of Roentgenology and Radiology from January through December 2002 were coded by practice type, geographic location, and subspecialty. Data were compared with the previously published results from 1991 through 2001.
RESULTS. From January through December 2002, 5117 positions were advertised for diagnostic radiologists, representing a 10.4% decrease from 2001. The 12-month rolling average of job advertisements peaked at 488 in February 2002 and dropped to 432 by December 2002, a level not seen since August 2000. In 2002, 42.7% of positions advertised were academic, representing a steady increase from 34.0% in 1999. When comparing the 4-year periods 1999–2002 and 1995–1998, a statistically significant proportional increase was found for positions advertised in the Midwest. Statistically significant relative increases were also observed for the subspecialties of abdominal radiology, mammography, neuroradiology, pediatric radiology, chest radiology, and nuclear medicine. Statistically significant relative decreases were seen in general and vascular and interventional positions.
CONCLUSION. The absolute demand for diagnostic radiologists nationwide appears to have stabilized during 2002, albeit at a level much higher than in previous years, and may be decreasing. A split seems to have appeared in the market between academia and private practice. Current policy should be directed toward increasing the supply of diagnostic radiologists, especially academicians.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Multiple studies have tried to describe and understand this shortage of radiologists. In 1998, the American College of Radiology (ACR) conducted a survey of hiring that reported a nationwide undersupply of 600 radiologists [2], and demand projections performed by Bhargavan et al. [3] in 2001 estimated the shortage at 5% of approximately 26,000 practicing clinicians. Also, since 1996, surveys of residency program directors have consistently shown decreasing levels of unemployment among graduates, greater ease in finding jobs, and increasing employment in jobs that match graduate training and goals [4–6]. Surveys of these newly trained radiologists, meanwhile, have indicated increases in the number of job offers and improved initial job satisfaction [6, 7]. Finally, the ACR's Professional Bureau, the largest placement service in diagnostic radiology, reported an increase in the ratio of job listings per job seeker from 1.3 in 1998 to 3.8 in 2000 [8]. Conflicting opinions now exist, though, as to whether the employment market has begun to reverse itself over the past 2 years or if, alternatively, the radiologist shortage continues unabated [3, 9].
This study is a follow-up of our previous job market article (reporting data through 2001), which suggested that the nationwide shortage of diagnostic radiologists might have been stabilizing [9]. The goal of this paper is to present the updated help wanted index data for 2002 and to compare it with our results from 1991–2001 [7, 9, 10].
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
The collection, presentation, and statistical analysis of all the data were also performed in a manner similar to our previous three studies [7, 9, 10] and are described in detail in our original article [10]. For some data analysis, we pooled the new data, those from January–December 2002, with the data from 1999–2001. Pooling of old and new data was performed merely for statistical comparison, allowing three 48-month sample periods to be compared. When comparing geographic location and subspecialty data across different time periods, we performed statistical analysis using a two-tailed comparison of proportions. Values for p of less than 0.05 were considered statistically significant.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The monthly peak of 599 advertisements occurred in December 2001, whereas the monthly nadir of 37 advertisements was noted in July 1995 (Fig. 2). Although the absolute number of advertisements peaked during December 2001, the 12-month rolling average, which was calculated to minimize seasonal variation, peaked 2 months later (February 2002) at 488 advertisements. From February through December 2002, the rolling average declined with increasing momentum, to 432 advertisements by the end of the year, a level not seen since August 2000 (Fig. 2). In addition, the total of 5117 advertisements for 2002 represented a 10.4% decrease from the total for 2001.
|
When the percentage of change in the rolling average data from month to month is calculated, (equivalent to the first derivative of the rolling average plot seen in Fig. 2), three clear inflection points are identified (Fig. 3). In June 1992, the percentage of change switches from positive to negative (indicating a market top); in December 1995, the percentage of change switches back from negative to positive (indicating a market bottom); and recently, around November 2001, the percentage of change switches again from positive to negative (indicating another market top). This November 2001 date correlates well with the absolute peak in the rolling average noted during February 2002 (Fig. 2). The three inflection points are also marked in Figure 2 for comparison.
|
Type of Practice
Thirty-eight percent of jobs advertised from January 1999 through December
2002 were for academic positions. This result represents a statistically
significant increase (p < 0.01) from 36.1% during the previous
48-month period from January 1995 through December 1998, but it is
significantly down from 46.6% during the period of January 1991 through
December 1994. The widening of the gap between private practice and academic
advertisements (absolute numbers) during the past several years, which was
most pronounced during 2001 (Fig.
4), reversed itself during 2002. This phenomenon is best seen in
Figure 5. During the course of
2002, the 12-month rolling average for private positions decreased from 301 to
247 advertisements, whereas the rolling average for academic positions
remained steady at approximately 184 advertisements
(Fig. 5). The proportional (not
absolute) gap between academic and private position has, in fact, been closing
for the past 4 years as academic positions have increased steadily from 34.0%
in 1999 to 36.8% in 2000, 37.2% in 2001, and 42.7% in 2002.
|
|
Geographic Location
Several geographic trends are noted when the data are compared with the
results from the previous help wanted index studies
[7,
9,
10]
(Table 1 and
Fig. 6). A statistically
significant increase is seen in the relative number of jobs advertised in the
Midwest, which accounted for 29.7% of total advertisements from January 1999
to December 2002 (up from 22.4% and 15.9% in the middle and earliest 48-month
periods, respectively). The Midwest now accounts for the greatest percentage
of positions advertised, having surpassed the Southeast (25.9%) and Northeast
(23.0%) during the most recent 4-year period. Although the absolute numbers of
advertisements for all six geographic regions increased substantially in
1999–2002 compared with the previous 4-year periods, the percentages of
advertisements in the Northeast (at 23.0%, compared with 25.8% and 27.4%) and
Southwest (at 7.8%, compared with 9.8% and 12.4%) both exhibited statistically
significant decreases. The percentage of advertisements also decreased in the
Southeast and California, but these changes were not statistically
significant. Finally, the Northwest experienced an increase in its percentage
of advertisements during the most recent 4-year period versus the previous one
(6.7% vs 6.5%), but again this change is not statistically significant and
still represents a decrease from the initial 4-year period (7.5%, in
1991–1995). When comparing the 1-year data for 2002 with those for 2001,
the absolute number of positions decreased in all geographic regions except
California and the Northwest, where the numbers rose from 270 to 346 and from
229 to 471, respectively (Fig.
7).
|
|
|
Subspecialty
Several subspecialties, including abdominal imaging, mammography,
neuroradiology, pediatric radiology, nuclear medicine, and chest radiology,
all displayed statistically significant relative increases in demand for the
period 1999–2002, compared with the previous two 4-year periods
(Table 2 and
Fig. 8). Abdominal imaging
accounted for 15.4% of advertisements (up from 10.2% and 13.7%); mammography,
10.1% (up from 8.3% and 5.9%); neuroradiology, 9.7% (up from 8.5% and 8.2%);
pediatric radiology, 4.6% (up from 2.7% and 4.4%); nuclear medicine, 3.5% (up
from 2.9% and 1.6%); and finally, chest radiology, 3.1% (up from 2.1% during
the previous 48-month period). On the other hand, statistically significant
percentage decreases in demand were observed in general radiology (down to
32.5%, from 39.7% and 43.4%) emergency radiology (down to 0.7%, from 1.0% and
1.1%), and vascular and interventional radiology (down to 13.4%, from 16.4%
during the previous 4-year period). No statistically significant changes were
seen in the percentages for musculoskeletal radiology, but just as with all
the other subspecialties, the absolute number of advertisements increased
substantially for this subspecialty from the previous 4-year period to the
current one (Table 2). Finally,
advertisements for MR imaging and sonography, which were included in the
abdominal imaging category, accounted for 3.6% and 0.6% of total positions,
respectively, listed from January 2000 through December 2002.
|
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Second, there appears to be a split in the market between academia and private practice. For the first time in several years, the absolute numbers of private and academic advertisements did not move in the same direction; rather, the number of private position advertisements decreased and the number of academic position advertisements increased during 2002. The growing shortage in academic radiology is no surprise. Reportedly, academic departments have experienced heavy turnover during the past few years. Our data suggest that the shortage of academic radiologists is indeed severe, because academic advertisements composed 42.7% of the total coded in 2002, even though only 20% of all radiologists work in academic settings [11]. Along the same lines, a recent study of academic radiology chairpersons reported that nearly 600 radiology department chair, directorship, and section leader positions are vacant in the United States, placing the number of overall faculty openings at an average of five for each institution nationwide (Maynard CD, presented at the Radiological Society of North America meeting, December 2002).
Third, a worsening shortage is occurring in the Midwest relative to the rest of the country. For the most recent 4-year period (1999–2002), the absolute number of advertisements in the Midwest more than quadrupled and the relative proportion almost doubled (from 15.9% to 29.7%) as compared with the original period, 1991–1994. At the same time, the relative proportions in all other areas in the country have declined, although some declines were not statistically significant (Table 1). As we mentioned in our previous article, it is not uncommon for many private and academic hospitals in the Midwest to be advertising to restaff almost entire departments [9].
Finally, a shift has occurred in positions advertised from general radiology to the radiology subspecialties. The proportion of general advertisements has declined steadily over the past three 4-year periods (from 43.4% to 39.7% to 32.5%, Table 2) and decreased to a 12-year low of 27.2% in 2002. This phenomenon has continued to occur despite the apparent difficulties many radiology groups are finding in obtaining radiologists of any type. The decline may reflect the demands of technologic advancements and specialization. It may also reflect the apparent decline in the number of radiology graduates who are performing fellowships, which would make subspecialists rarer and, in turn, lead to more advertising for them. According to the help wanted index, the subspecialists in greatest demand are abdominal imagers, mammographers, pediatric radiologists, and nuclear medicine physicians.
Help Wanted Index Characteristics and Limitations
We believe that our analysis of help wanted index data usefully portrays
the current radiologist employment situation. A help wanted index has been
used for the general economy for decades; its development is based on the
original work of Malcolm S. Cohen and the Nobel Prize–winning economist
Robert Solow
[12–15].
Furthermore, the value of the help wanted index and its validity for the
radiology job market in particular have been outlined in our previous articles
[7,
10]. However, substantial
caution should be taken when using help wanted index data to analyze the
employment market and predict long-term trends. Several characteristics of the
help wanted index must be weighed carefully.
First, a help wanted index is not an exact, but instead an indirect, measure of demand pressure in labor markets. Such an index does not quantify the actual number of vacant positions or unemployed job seekers, but rather tracks the number of job advertisements. These advertisements are sometimes repeated and sometimes outdated, and whether to advertise is often left to the discretion of the employer. Thus, indirect indicators, like the help wanted index, have no natural scaling. Absolute numbers mean little; only general trends can be described [12].
Second, the number of advertisements is related to the size of a given field as well as to the demand for employees. As a field grows, the help wanted index should rise, independent of labor demand, with the number of job seekers and number of establishments advertising [12]. This phenomenon should not be an issue for the radiology market as a whole, because the number of entrants has been relatively fixed by the number of annual residency positions (900–1000) for the past several years. This effect may have an impact, though, on the subspecialty data, because the proportional increases in advertisements noted for fields such as abdominal imaging, mammography, and nuclear medicine may be due in part to the overall growth of these fields in addition to shortages of these subspecialty-trained radiologists.
Third, the help wanted index tends to magnify market trends. As unemployment decreases, the number of advertisements required to fill a single vacancy should increase at an accelerating pace. This curvilinear relationship between the help wanted index and unemployment is especially pronounced at very low levels of unemployment (such as that seen in the radiology job market today) and has been documented by Burch and Fabricant [13] through historical review of the manufacturing sector of the economy. As a result, trends noted in this paper probably exaggerate the true details of the market. Large proportional increases in certain geographic areas (the Midwest) and specific subspecialties (pediatrics, mammography) may be partially explained by this effect. On the other hand, this sensitivity allows our help wanted index study to highlight subtle trends that may not otherwise be identified by alternative forms of analysis (surveys and modeling).
Fourth, the help wanted index tends to lag the true job market by several months. Cohen and Solow [12] documented that more help wanted advertising occurs late in an upswing (rather than early) and less advertising late in a downswing. This lag is mainly because formal advertising is a last-resort method of recruitment. During upswings, such advertising costs money and hurts prestige and so is usually avoided until absolutely necessary. During downswings, there is often belated recognition of labor market conditions, and advertisements tend to remain in print for extended periods before being pulled [12]. Taking this phenomenon into account, the labor market for diagnostic radiologists probably peaked sometime before the November 2001 date that was reported in our study (Fig. 2). Nevertheless, help wanted indexes provide a more current picture of the labor market than do surveys, which sometimes take months to collect and analyze.
Finally, our help wanted index may not be completely comprehensive. Electronic job advertising techniques have grown in recent years throughout the economy. For radiologists, as for others, electronic advertisements may be partly replacing print advertising. Second, anecdotally we hear that radiology practices are making increasing use of headhunters, given the extreme difficulty of successful recruiting in the current period of shortage. Both of these phenomena would tend to produce a decline in our index even if the shortage were not easing.
An overestimation of the decline in the private job market by our index may well be happening. Because of concerns about equal employment opportunity laws, all academic positions tend to be advertised in print, making the help wanted index for academic jobs more reliable than that for private jobs. A decline in private advertising when academic position advertising is growing has not been seen before, and it is hard to understand how the academic shortage would be worsening if the much larger private job market is easing. Thus, we acknowledge the possibility that the decline in private job advertisements may represent the phenomena described in the previous paragraph rather than a decrease in the actual number of private job positions. A survey being carried out by the ACR during 2003 should cast further light on this issue.
Conclusion
The current shortage of diagnostic radiologists poses a significant threat
to the viability of our specialty. Without adequate numbers of properly
trained radiologists, we stand to yield turf to competing specialties,
sacrifice the quality of our academic training programs, lose radiologists
available to perform research, and, most important, compromise our ability to
provide optimal patient care. Although this study indicates that the overall
radiologist shortage may be easing, it does not provide a firm conclusion; and
even a mild easing does not mean that the crisis has passed. The academic
shortage is at its most severe point ever
(Fig. 4), and the recent
private market stabilization, if real, may be temporary and derived from
one-time productivity increases (as a result of PACS [picture archiving and
communication systems], teleradiology, and computer-aided diagnosis) or
increased individual radiologist work-hours
[16]. Enhancing the
productivity argument is a recent report in the Radiology Business
Management Association Bulletin that showed that academic radiologist
productivity—as measured in relative value units for each full-time
equivalent radiologist—grew by 11.5% in 2001 as compared with 2000 and
by 34.6% since 1997 [17].
Meanwhile, the demand for image interpretation continues to increase rapidly and to outpace the supply of radiologists available to perform the work [3]. Many forces remain to ensure that this situation continues: the growth and aging of the population has increased the demand for imaging (and medical care in general); the ebbing of managed care has shifted power from primary care physicians to specialists; the increased pressure for more indepth 24-hr radiology coverage has reduced radiologist availability; and the effects of cutbacks in the number of residency positions during the mid 1990s are just now beginning to be felt as fewer trainees graduate. Specifically, during the past decade the number of procedures has grown by 4.5% annually and the complexity of procedures by 1.75% annually. This growth has caused relative value units to increase at a yearly rate of 6.0%, four times greater than the rate of radiologist growth (1.5% each year) [3, 18]. At these current rates of growth, Bhargavan [3] et al. project a deficit of 10,000–15,000 radiologists by 2015 (almost half the number of radiologists nationwide today).
Clearly, changes need to be instituted to alleviate this situation. An all-radiology leadership task force convened by the ACR in 2001 studied possible changes and made recommendations. The ACR has put many of these changes into effect. We hope that this article has presented a useful overview of the current employment market conditions and provided sufficient details to allow continued evolution of an effective course of action. The preservation of our specialty depends on such policy.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  P. M. Colletti The Diagnostic Radiology Exam of the Future: The Law of Unintended Consequences Meets the Law of Supply and Demand Am. J. Roentgenol., May 1, 2008; 190(5): 1147 - 1148. [Full Text] [PDF]
|
|
|
|
|
|
A. Licurse, D. D. Saket, J. H. Sunshine, C. D. Maynard, and H. P. Forman Update on the diagnostic radiology employment market: findings through 2005. Am. J. Roentgenol., September 1, 2006; 187(3): W249 - W254. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Merewitz and J. H. Sunshine A Portrait of Pediatric Radiologists in the United States Am. J. Roentgenol., January 1, 2006; 186(1): 12 - 22. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Muchantef and H. P. Forman Cost Accounting in Radiology: New Directions and Importance for Policy Am. J. Roentgenol., December 1, 2005; 185(6): 1404 - 1407. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. D. Saket, C. C. Nwanze, C. D. Maynard, J. H. Sunshine, and H. P. Forman Update on the Diagnostic Radiologist Employment Market: Findings Through 2004 Am. J. Roentgenol., December 1, 2005; 185(6): 1408 - 1415. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. H. Sunshine, C. D. Maynard, J. Paros, and H. P. Forman Update on the Diagnostic Radiologist Shortage Am. J. Roentgenol., February 1, 2004; 182(2): 301 - 305. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
