Professional and economic factors affecting access to mammography: A crisis today, or tomorrow?†
Results from a national survey
The Society of Breast Imaging staff and representatives participated in the design and conduct of the current study; in the collection, analysis, and interpretation of the data; and in review of the article.
Objective data and anecdotal reports have suggested that access to mammography may be declining because of facility closures and difficulty in recruiting and retaining radiologists and radiologic technologists. To gain insight into the practice patterns, use of emerging technologies, and concerns of breast imagers in current practice, the Society of Breast Imaging (SBI) conducted a national survey of breast imaging practices in the U.S.
Between October 2003 and April 2004, the SBI conducted a survey of the SBI membership database, and received completed surveys from 575 breast imaging practices in the U.S. Responses to the survey regarding practice characteristics, the utilization of standard and emerging technologies, staffing, malpractice, finance, and morale were analyzed.
Job vacancies for radiologists who read mammograms were reported in 163 practices (29%), 59 of which (10%) had 2 or more openings. A higher proportion of practices with job openings had long appointment waiting times for asymptomatic women when compared with fully staffed practices. Unfilled fellowship positions also were common, with 41 of 65 practices that offer fellowships reporting 47 openings. Among 554 responding practices, 55% reported that someone in their practice was sued because of a mammography related case within the past 5 years, and 50% of practices reported that the threat of lawsuits made radiologist staffing “moderately” or “a lot” more difficult. Of 521 responding practices, 35% reported financial losses in 2002. One in 5 respondents reported that they would prefer to spend less time in mammography, and fewer than 1 in 3 would recommend a breast imaging fellowship to a relative or friend. Emerging technologies, such as breast magnetic resonance imaging and screening ultrasound, currently are being performed in many practices.
The survey results provide support for anecdotal reports that breast imaging practices face significant challenges and stresses, including shortages of key personnel, a lack of trainees, malpractice concerns, financial constraints, increased workload due to emerging technologies, low appeal of breast imaging as a career specialty, and the steady rise in the population of women of screening age. Cancer 2005. © 2005 American Cancer Society.
In recent decades, there has been significant progress in the contribution of mammography to the control of breast carcinoma.1–3 This progress, along with improvements in therapy and increased awareness, has led to annual reductions in breast carcinoma mortality since 1990.3 Nevertheless, despite evidence of the importance of early detection in reducing the mortality from breast carcinoma, anecdotal reports indicate that a woman's access to mammography may be threatened by professional and economic disincentives that appear to have grown over time.4, 5
In response to these reports, both the U.S. Government Accountability Office (GAO) and the Food and Drug Administration (FDA) conducted assessments of the availability of mammography services.6, 7 The FDA report, which was based on data obtained between 1997–2000, concluded that a declining trend in the number of mammography facilities did not appear to affect access because the number of women reporting recent mammograms had increased.6 The GAO report, based on data obtained to March 2002, was more thorough in its assessment of capacity, evaluating trends in the number of mammography facilities, mammography units, and personnel.7 Although the GAO report concluded that trends in these areas were reasons to be concerned for the future of mammography, the current access problems appeared to be a local rather than systemic phenomenon.
Although the FDA and GAO reports addressed the immediate question of whether there was a current crisis in access to mammography, many questions remain unanswered, such as what factors explain facility closings and the declining numbers of personnel, and whether data from some geographic areas are predictive of future problems that may become more widespread.
Recent data indicate that attracting new radiologists to the field is a growing challenge. A survey of radiology residents conducted in 2000 indicated low interest in specializing in breast imaging, as well as low interest in devoting significant professional time to reading mammograms.8 Residents cited a number of disincentives, including high malpractice exposure, the repetitive nature of screening mammogram interpretation, the challenging nature of the examination, low pay, and other factors. At the current time, the majority of breast imaging fellowship positions in the National Matching Residency Program remain unfilled.6, 9
These same reasons may explain reports that hiring breast imagers has become increasingly difficult. According to a recent survey of 45 facilities in 3 geographic locations (Denver, New Hampshire, and the Pacific Northwest), 44% of practices reported shortages of qualified breast imagers.10 In another study, Saketkhoo et al. reported a fourfold increase in the number of help-wanted advertisements for mammography over the past decade.11 During the period between 1999–2002, the authors identified 2068 breast imaging job advertisements compared with only 496 for the period between 1991–1994. Although these job vacancies might be interpreted as simply symptomatic of a broader pattern of shortages of diagnostic radiologists in the nation, recent data indicate that these shortages have eased for other subspecialty areas of radiology.12 In the study by Saketkhoo et al., the field of mammography had the highest growth rate in the number of help-wanted advertisements within the past decade and the second highest number of help-wanted advertisements among nine radiology subspecialties.11
Residents' concerns regarding malpractice also may be well founded. Although most physicians have underlying concerns about medicolegal liability,13, 14 failure to diagnosis breast carcinoma is currently the leading cause of medical malpractice lawsuits, and the second most expensive condition in terms of total indemnity dollars, with greater than 41% of all breast carcinoma-related claims resulting in an indemnity payment.10
Although to our knowledge few financial analyses of mammography services have been published to date, Enzmann et al. assessed the financial status at seven university-based mammography services through an analysis of revenues, direct and indirect costs, and volume of patients and services and observed that all seven institutions incurred losses on the professional component due to “negative contribution margin” (i.e., revenues did not exceed direct and indirect costs).15 The report by Destouet and Sherman of a large efficient private practice with an annual volume of 115,000 mammograms also reported that costs exceeded revenues.16 Other reports also have indicated that large-volume stand-alone breast imaging centers are facing difficulties.17
At the current time, there is a combination of data, both objective and anecdotal, that portends a worrisome trend for the future of breast imaging. Furthermore, at this time, the incidence of breast carcinoma is increasing, as is the size of the U.S. population of women who should undergo regular screening for breast carcinoma.18 These trends underscore the importance of understanding with measurable confidence the degree to which the capacity to deliver mammography will be adequate to meet the growing need, or may be undermined by one or several factors. To gain greater insight into the challenges and concerns of breast imagers, and to establish objective baseline data for the assessment of trends in access to screening, the Society of Breast Imaging (SBI) conducted a national survey of breast imaging practices in the U.S.
MATERIALS AND METHODS
Between October 2003 and April 2004, the SBI conducted a survey of breast imaging practices in the U.S., using the SBI membership database. There are 1684 general members in the SBI. Retired individuals, nonpracticing physicians, foreign members, and nonphysicians were excluded from the database, resulting in 1572 practicing breast imaging radiologists in the study sample. The survey was distributed via E-mail. Respondents could submit their responses online, by fax, or by postal mail. If an individual from a practice did not respond after three E-mail requests, one final request was sent via postal mail. Those individuals who did not supply an E-mail address received the survey by postal mail. If these radiologists did not respond, the survey was resent for a total of three attempts via postal mail.
We requested one survey per breast imaging practice. Each respondent was asked to provide the names of other breast imagers in their practice, which enabled us to track duplicate surveys (which were excluded from the analysis). In addition, business addresses, phone numbers, and practice names were cross-referenced to identify duplicate surveys. If more than one survey from the same practice was received, the respondents were contacted by phone to determine which survey was the most representative of their practice on objective measures such as those described in Tables 1 and 2.
Table 1. Characteristics of Breast Imaging Practices
| ||South Atlantic||94||16.3|
| ||South Central||78||13.6|
|Practice type||Private practice||464||80.8|
|Annual mammography volume||<5000||86||15.4|
Table 2. Breast Imaging Practices and Patient Population Characteristics
The survey included 57 items, which covered general practice characteristics and trends, workload, staff shortages, appointment waiting times, financial status, malpractice concerns, and morale. Geographic regions were based on the classification used by the U.S. Census Bureau. Data for the questionnaire responses were either nominal or ordinal. Frequency and relative frequency distributions were calculated for the responses. Statistical tests were used to determine whether associations existed between selected responses. If the responses were nominal, chi-square tests were used. If one of the responses was ordinal, Kruskal–Wallis tests for singly ordered rows × columns (R × C) tables were used. If both responses were ordinal, Jonckheere–Terpstra tests for doubly ordered R × C tables were used. If 20% of the cells had expected counts of less than 5, exact P values were calculated by means of data permutation. If the calculation took longer than the StatXact software's default limit of 30 seconds (StatXact®; Cytel Software Corporation, Cambridge, MA), Monte Carlo sampling (10,000 samples) was used to estimate the exact P values. For this study, the α value was set at 0.05. A Bonferroni corrected α value to adjust for the familywise error rate of performing the 15 statistical tests of this study results in an α of 0.003. Unadjusted P values are reported in this article. Statistical analyses were performed using JMP Statistical Software (Release 220.127.116.11; SAS Institute, Inc., Cary, NC) and StatXact Statistical Software for Exact Nonparametric Inference (Version 6; Cytel Software Corporation).
Ethics approval was obtained from the Washington University Medical School Human Studies Committee.
We received surveys from 575 practices, representing 1006 radiologists or 64% of 1572 actively practicing breast imagers in the SBI. This return rate allowed us to report responses with a 99% level of confidence, with a confidence interval of ± 3 percentage points. The characteristics of these practices are summarized in Table 1, and approximate estimates of the characteristics of the population served by the facilities according to insurance status and payment mechanisms are shown in Table 2. One hundred percent of the respondents reported that their practice accepts Medicare patients, and approximately 90% of practices provide some care for low-income women who are receiving Medicaid (Table 2). Compared with private practices, a significantly higher proportion of academic practices care for at least some (> 10%) Medicaid patients (45% vs. 23%; P < 0.0001).
Standard services (such as screening mammography, diagnostic mammography, and diagnostic ultrasound) are performed by nearly all practices. Approximately 19% of respondents reported that their patients wait longer than 1 month to get a screening mammogram appointment, with 8% reporting waiting times of longer than 2 months. Greater than half of the practices with wait times of longer than 2 months are located in the Northeast. Only 6% of practices reported that symptomatic patients wait 2 or more weeks to get an appointment for diagnostic imaging; 9 of these 34 practices reported waiting times of longer than 1 month. Of these nine practices, seven were located on the East coast (Table 3).
Table 3. Waiting Times for Screening and Diagnostic Mammography
|Screening mammography||≤24 h||59||10.6|
| ||2–6 days||164||29.4|
| ||1–2 weeks||100||19.4|
| ||2–4 weeks||118||21.2|
| ||1–2 mo||63||11.3|
| ||>2 mo||45||8.1|
|Diagnostic mammography||≤24 h||209||37.3|
| ||2–6 days||266||47.5|
| ||1–2 weeks||51||9.1|
| ||2–4 weeks||25||4.5|
| ||>1 mo||9||1.6|
According to our survey, 71% of practices reported an increase in volume within the past 2 years. The top 3 factors cited were increased demand (96%), increased space/equipment (47%), and increased technologist staffing (36%). Over the past 2 years, only 7% of practices reported a decrease in volume. Approximately 3% of respondents reported that their practices “definitely will” or “probably will” discontinue mammography services within the next 24 months. Approximately 4% of the respondents were not sure whether their practice would continue offering mammography services in the next 24 months.
Twenty-nine percent of practices reported unfilled positions for radiologists who read mammograms. Among practices with vacancies for breast imagers, 38% had 2 or more openings and 25% reported having open positions for breast imagers for longer than 2 years. A higher proportion of academic institutions reported having unfilled breast imaging radiologist positions compared with private facilities (48% vs. 25%) (P < 0.0001). In addition, a higher percentage of urban practices (35%) reported having unfilled positions compared with suburban (25%) or rural practices (11%) (P < 0.0005). There was no significant variation noted in job vacancies based on geographic region. More practices with unfilled radiologist positions had long waiting times for appointments for symptomatic women (P < 0.0001) compared with fully staffed practices (Table 4).
Table 4. Correlation between Unfilled Radiologist Positions and Appointment Waiting Time for Symptomatic Patients
|P < 0.0001|
Approximately 30% of practices reported unfilled mammography technologist positions. Of these practices, 45% reported 2 or more unfilled positions. There was no significant difference noted among urban, suburban, and rural practices. As was observed among practices with unfilled radiology positions, practices with unfilled technologist positions also had longer waiting times for appointments for symptomatic women (P < 0.0029) when compared with practices with no openings (Table 5).
Table 5. Correlation between Unfilled Technologist Positions and Appointment Waiting Time for Symptomatic Patients
|P < 0.0029|| || |
Only 12% of practices surveyed offer breast imaging fellowship positions. Although less than 20% of the study sample was academic, these facilities account for 77% of the fellowship positions and are more likely to offer fellowship training compared with private practices (50% vs. 3%). Of the 65 practices in the current survey that offered fellowship training, 41 practices reported 47 unfilled fellowship positions during the 2003–2004 academic year.
Emerging technologies will likely increase workload. Although screening breast ultrasound remains under investigation,14 35% of respondents offer this service. The role of magnetic resonance imaging (MRI) in breast imaging continues to evolve. Currently, 12% of practices in the study survey offer screening breast MRI. Diagnostic breast MRI is reported to be available in 51% of practices.
Approximately 55% of practices reported that someone in their practice had been sued in a breast imaging-related case within the past 5 years. The Northeast had the highest proportion of practices with lawsuits (71%) (Table 6). In the study survey, 27% of respondents indicated that the threat of lawsuits decreased their willingness to perform breast imaging “moderately” or “a lot.” In addition, approximately 50% of respondents reported that the threat of lawsuits made radiologist staffing “moderately” or “a lot” more difficult.
Table 6. Percentage of Practices with Lawsuit(s) within the Past 5 Years
| ||South Atlantic||51||56.7|
| ||South Central||34||43.6|
|P < 0.00001|
|P < 0.00001|
Approximately 35% of practices reported financial losses for breast imaging in 2002, whereas 37% reported being profitable and 29% reported breaking even. There was no correlation noted between financial status and factors such as geographic region, practice location (urban vs. rural vs. suburban), hospital versus outpatient imaging, or practice payment method (global vs. professional fees only). However, procedure volume was associated with financial status. Practices with lower overall mammography volumes (diagnostic and screening) were more likely to report financial losses in 2002 (P < 0.0001) (Table 7).
Table 7. Correlation between Mammography Volume and Financial Status of Practice
|P < 0.00001|
Financial status was found to be associated with practice type and patient mix. A higher proportion of academic practices reported financial losses for breast imaging compared with private hospital and nonhospital-based practices (P < 0.0002). In 2002, approximately 50% of academic practices reported losses, compared with 36% of private nonhospital practices and 28% of private hospital practices. Practices that provided care for higher percentages of Medicaid patients also were more likely to report losses in 2002 compared with practices that cared for fewer Medicaid patients (P < 0.0115). For example, 27% of practices with no Medicaid patients reported losses compared with 63% of practices with a population of 30% or more Medicaid patients.
Respondent Perceptions of Breast Imaging as a Career
Respondents were asked a series of questions regarding their own work-related quality of life, and whether specializing in breast imaging was a viable career choice. Approximately 21% of respondents reported that they would prefer to personally devote less time to breast imaging in the future. Those individuals who currently spend the least amount of time (< 25%) or the majority of their time (> 75%) in breast imaging were more likely to want to spend less time in breast imaging.
Only 31% of respondents stated that they would recommend a breast imaging fellowship to relatives or close friends. Conversely, 21% definitely would not recommend fellowship training to relatives or close friends, 27% most likely would not recommend a breast imaging fellowship, and 21% reported that they were unsure. The most commonly cited factors that led to reservations about recommending breast imaging as a specialty were: 1) malpractice (93%), 2) too much stress (78%), 3) too many regulations (73%), and 4) low salary (61%).
The findings of the survey provide support for anecdotal reports that breast imaging practices face many challenges. Shortages of key personnel are related to a lack of new entrants to the field, malpractice concerns, and financial strain. Increases in demand are due to increased volume and possibly the integration of emerging technologies, which add to the workload. It is reasonable to conclude that the future capacity to provide screening and diagnostic mammography services to women in the U.S. will depend on our ability to overcome these challenges, which are unlikely to be met through the invisible hand of market forces.
The results of the survey confirm previous reports and anecdotes that the recruitment of breast imagers at various levels is becoming increasingly difficult. Nearly one in three practices reported at least one job opening for a radiologist to read mammograms. Although 1 opening may or may not significantly impact clinical services, it is troubling that 10% of all of the practices surveyed reported having 2 or more openings and 7% reported unfilled openings for 2 years or longer. The higher burden of job vacancies in academic institutions, including unfilled fellowship positions, is particularly worrisome for breast imaging because it suggests that the pipeline of teachers is diminishing. Because the majority of breast imaging fellowship and residency training programs are based at academic institutions, there is reason to be concerned about the declining capacity to adequately train future breast imagers if these shortages are not resolved. Furthermore, academic facilities serve a significant proportion of Medicaid patients and therefore the impact of workforce shortages on care for the underserved is an additional reason to be concerned about current trends.
Interest in breast imaging as a subspecialty career choice also has declined. In the study survey, the majority of practices (41 of 65 practices) with training positions reported having unfilled breast imaging fellowship positions. A national radiology fellowship match program was initiated in 2003 for fellowships beginning in July 2004. Although overall, 53% of offered positions were filled in the match, only 25% of the breast/women's imaging positions were filled, ranking 8th of 9 subspecialties, exceeding only pediatric radiology.9 The low interest of radiology residents in fellowship training in breast imaging is consistent with the attitudes expressed by residents in the study by Bassett et al.,8 in which 64% of the residents reported that they had no interest in a breast imaging fellowship. Although a third of the residents indicated that they would consider a breast imaging fellowship, few ultimately do, based on the current number of breast imaging fellowship vacancies. For whatever reason, the majority of individuals open to the possibility of a fellowship do not pursue this field as a subspecialty.
The study survey also indicates that breast imagers are pessimistic about their specialty. Nearly half of the respondents stated that they definitely or most likely would not encourage a relative or close friend to obtain fellowship training in breast imaging, and another 20% of the respondents were unsure. Because all our respondents are SBI members, it is particularly discouraging that many of the radiologists who are breast imaging specialists and/or have a special interest in the field would not encourage others to pursue their own subspecialty. According to the results of the survey, the primary factors leading to negative responses were malpractice, stress, and too many regulations, reasons that are similar to disincentives cited by current residents.8
The survey results add to the existing evidence that there are shortages of other key mammography personnel. In the study survey, nearly 30% of practices reported having at least 1 job opening and 13% reported having 2 or more vacancies for qualified mammography technologists. According to D'Orsi et al., 20% of the breast imaging facilities in their survey reported a shortage of mammography technologists.10 The Association of Registered Radiologic Technologists (ARRT) reported a significant decline annually from 1996–2000 in the number of first-time examinees for a mammography certificate.7
There was no evidence in the current survey that longer waiting times for screening or diagnostic imaging are common, a finding that is consistent with other recent investigations.6, 7 However, although to our knowledge little is known regarding the impact of staffing on the quality of the care provided in breast imaging, the results of the study survey demonstrate that long wait times for screening and diagnostic appointments were more likely to be associated with radiologist and technologist vacancies. Therefore, there is reason to be concerned that a worsening trend is possible, influenced by declines in professional and physical capacity, growth in the size of the population of women age older than 40 years, and continued growth in the demand for services reported in this survey. For example, as the spectrum of services and technologies provided by breast imaging facilities expands, specifically ultrasound and MRI, the workload required to meet these demands will increase. Two recent studies have shown that MRI outperforms mammography for the detection of breast carcinoma in younger, high-risk women.19, 20 However, MRI does not supplant the need for complementary mammography. Based on growing evidence supporting the use of MRI in breast imaging in specific circumstances, it is likely that high-risk women and their health care providers will begin to request these examinations in greater numbers. Current workforce trends provide little basis for confidence that increased demand will be followed by increases in personnel.
In the study survey, greater than 25% of respondents believed that the threat of lawsuits decreased their willingness to perform breast imaging “moderately” or “a lot.” In addition, half of the respondents believed that malpractice concerns made staffing more difficult. These concerns about malpractice are not unfounded, as documented by the 2002 Physician Insurers Association of America report.21 The high percentage (> 50%) of respondents reporting a breast imaging-related lawsuit in their practice occurring within the past 5 years provides support for the prevalent perception among residents and practicing radiologists that the risk of malpractice associated with breast imaging is high.
Breast imaging traditionally has been viewed as a source of financial loss in radiology practices. Approximately one-third of facilities reported that the cost of their breast imaging practices exceeded the income earned in 2002. Nearly half of the academic practices in the study survey reported financial losses in 2002. Although higher volume practices were more likely to report that they were profitable in 2002 (suggesting some benefit from economies of scale), even among very large practices nearly 25% reported losses in 2002, indicating that economies of scale alone are insufficient to insure profitability. We did not provide survey respondents with an accounting definition of “earn money,” “break even,” or “lost money,” and therefore their responses likely reflect a range of qualitative and quantitative assessments. However, these data are valuable because they reflect the perceived and perhaps comparative financial value of breast imaging in the overall practice of radiology.
Some services associated with breast imaging are costly, but currently are not reimbursed by payers. For example, compliance with the regulations set forth by the Mammography Quality Standards Act (MQSA) of 1992 is an unreimbursed financial obligation that exceeds the general quality assurance costs that can be assumed to be common across all medical specialties.16 For some costs, high volume only increases losses.15 Other nonrecoverable costs include comparison of current films with prior films, film retrieval from another facility,22 and immediate or “on-line” reporting of screening mammography results to the patient, which is preferred by women but also is considerably more costly than off-line reading.23, 24 Approximately 62% of the facilities surveyed provided this service for some screening patients, and 14% provided immediate results for greater than half of their screening patients.
The data reported herein provide sufficient evidence for concern that a number of interrelated and interdependent factors may be slowly eroding the capacity to provide high-quality mammography to a growing population of women in the U.S. The current survey data provide additional support for earlier findings that this problem is more evident in some geographic regions compared with others. However, the findings also suggest that no region is immune to the potential for reduced access due to the interplay between professional and financial disincentives, and growing demand due to technologic and population pressures. Another worrisome possibility is that these disincentives will have a deleterious effect on quality by reducing the incentive to maintain or improve the interpretive skills necessary to achieve a high level of individual performance in reading mammograms.
Reduced access can take several forms, each with different consequences. In rural areas, the average distance to a facility that can provide mammography may be increasing, inasmuch as the GAO report noted that 121 counties have experienced a decrease of 25% or more in the number of mammography units within the past 3 years.7 In all facilities, personnel shortages may reduce the number of days and hours in which appointments for mammography are available. Women who must cancel an appointment may find that the next available appointment is many months in the future, and likewise, women who call to schedule an appointment for screening also may be frustrated to learn that the next available appointment is many months away. One of the consequences of diminished access may be lower adherence to recommended screening intervals, with the potential for a later stage at diagnosis for women who have had too much time elapse since the last normal screening mammogram.25
To our knowledge to date there have been several assessments of capacity as measured by the ratio of mammography units to population in the context of an estimated average daily volume of examinations.7, 26 Although this type of analysis can provide important insights regarding supply and access, units and population are not normally distributed across the U.S. and, as we have seen, clinical staff to operate equipment and interpret mammograms are equally important inputs to capacity estimates. Therefore, the capacity to deliver screening and diagnostic services needs to be assessed regularly in terms of facilities, units, personnel, and population density so that early signs that may foretell declines in access can stimulate attempts to understand contributing factors and formulate corrective action. What should be of great concern is that no agency is charged with an ongoing, formal assessment of the national and local capacity to provide the screening test that is critical to the control of breast carcinoma.
The last two Institute of Medicine reports on mammography have responded to reports of declining access to mammography and concluded that access issues warrant greater attention, both in terms of assessment and long-range planning, to insure that adequate capacity for breast imaging exists.27, 28 The results of the current survey provide further evidence that current and future estimated capacities must be promptly assessed so that appropriate contingency planning can be made an immediate national priority.
The authors thank Michele Wittling and Carrie Chavis for their assistance in distributing the surveys, Julie Imo for her assistance in tracking duplicate surveys, and Anna Langenberg for her help in the preparation of the article.