Wallace, Amy H. MD; Havrilesky, Laura J. MD, MHSc; Valea, Fidel A. MD; Barnett, Jason C. MD; Berchuck, Andrew MD; Myers, Evan R. MD, MPH
The U.S. population is both increasing and aging. The Bureau of the Census states that “the nation will be … much older by midcentury.”1 The population aged 65 years and older is projected to more than double to 88.5 million by 20501 both because of increases in life expectancy and the aging of the baby boom generation.2 As the number of older people increases, the number of cases of diseases that are more common in the elderly, including gynecologic cancers, is also expected to rise over the next few decades. The number of gynecologic oncologists in practice may also need to increase over this time period to provide optimal care for the growing population.
Comparing estimates of the number of practicing gynecologic oncologists with the expected number of gynecologic cancer cases may provide insight into the number of trainees that will be needed over the coming decades. Using existing data on the U.S. population, gynecologic cancer incidence, and Society of Gynecologic Oncologists membership, we sought to 1) estimate the number of cases of gynecologic cancers in the United States over the next 40 years; 2) estimate the number of gynecologic oncologists who will be in practice over the same period; 3) estimate the current ratio of gynecologic cancer cases per practicing gynecologic oncologist; and 4) estimate this ratio over the next 40 years to help inform decisions about future training needs.
MATERIALS AND METHODS
The projected population of women in the United States from 2009 through 2050 was obtained using the Middle Series of National Population Projections (released 2008 based on Census 2000) from the U.S. Census Bureau.3 The equation was used to project the annual number of new cancer cases for the five major gynecologic cancer types (ovarian, uterine corpus, cervical, vulvar, and vaginal) from 2010 to 2050.
Here, “Population” is equal to the number of women in a specific age group, “AG,” from U.S. Census data.3 “Crude Rate” is the incidence of each gynecologic cancer per age group from the Surveillance, Epidemiology and End Results (SEER) database,4 and “i” represents the age group of the population. The age groups used were 0, 1–4, 5–9, 10–14, 15–19, 20–24, 25–29, 30–34, 35–39, 40–44, 45–49, 50–54, 55–59, 60–64, 65–69, 70–74, 75–79, 80–84, and 85 years and older.
Equation (Uncited)Image Tools
To generate estimates of cervical cancer cases, we also took into account the effects of the human papillomavirus (HPV) vaccine. We estimated that 30% of preadolescents would receive the HPV vaccine in our base case scenario. Currently, the vaccine is 95% effective against 70% of cancers.5 Assuming lifetime duration of vaccine efficacy, we therefore estimated a 20% overall reduction in cervical cancer cases (0.95*0.70*0.30)=0.20.
The equation also was used to project the annual number of cancer deaths using SEER mortality data. For mortality calculations, “Crude Rate” represents the mortality rate for each gynecologic cancer per age group from the SEER database4 and “i” represents the age group of the population. Identical age groups were used.
Other than vaccination against oncogenic HPV, we assumed no changes in other factors that might affect gynecologic cancer incidence such as rates of hysterectomy and oophorectomy for benign disease, oral contraceptive use, or obesity.
The number of practicing gynecologic oncologists was projected from 2009 through 2050. We used the 2009 Society of Gynecologic Oncologists membership of 873 full (completed fellowship training and board-certified) and candidate (completed fellowship training, eligible for board certification) members as baseline data based on information supplied by the Society of Gynecologic Oncologists from membership surveys. We then distributed the 873 members into 5-year age groups using age distribution data as presented in the 2005 Society of Gynecologic Oncologists Practice Survey.6 We used information from the American Board of Obstetrics and Gynecology gynecologic oncology fellowship and oral examination results concerning the number and sex distribution of graduating fellows. Taking into account the current percentage of women passing gynecologic oncology oral board examinations (41% in 2009, Stephen Rubin, MD, personal communication) and the higher percentage of female graduates from obstetrics and gynecology residencies (78% in 2008), we assumed for the base case that, on average, 50% of graduating fellows would be female, and this assumption was varied for sensitivity analysis (American Board of Obstetrics and Gynecology. Gynecologic Oncology fellowship match results and Board certification data. Data compiled and supplied by SGO front office and by personal communication with Stephen Rubin, MD).7 We assumed the following age distribution for graduating fellows based on 2005 Society of Gynecologic Oncologists Practice Survey data regarding time in practice and authors' estimates of the average ages at which residents complete residency and fellowship training: 12% age 33 years, 16% age 34 years, 33% age 35 years, and 40% age 36 years. Using the known age distribution of current Society of Gynecologic Oncologists full and candidate members in 2005 as well as the known number and estimated age distribution of current graduating fellows, a model was created using Excel that allowed members of each age cohort to progress to the next highest age group at 1-year intervals. Each year, we assumed that the same number of graduating fellows with a similar age and sex distribution would be added to the practicing physician pool. At yearly intervals, we applied sex- and age-specific mortality rates from the most recently available U.S. life tables to estimate the percentage of members that would survive to the next age group.8 After retirement, physicians were no longer part of the practicing physician pool. We modeled different retirement ages for men and women based on information from the 2005 Society of Gynecologic Oncologists Practice Survey as follows. At the time of the survey, responding male gynecologic oncologists had been in practice an average of 18 years and planned to practice on average an additional 14 years (32 years total practice time); female gynecologic oncologists had been in practice an average of 11 years and planned to practice on average an additional 16 years (27 years total practice time).6 We therefore assumed a 5-year difference in total planned time in practice between sexes. With regard to the choice of retirement age for the model, we added the average expected time in practice (32 years for men, 27 years for women) to the average expected age of fellowship completion (33 years), which resulted in an average retirement age of 65 years for men and 60 years for women. We then made the assumption that all MDs will not retire at exactly the same age, and this resulted in the following distribution of retirement ages: 5% of female members retire by age 55 years, 50% retire by age 60 years, and 100% retire by age 65 years. For male members, 50% retire by age 65 years and 100% retire by age 70 years.
We performed a sensitivity analysis to account for uncertainty in the projected U.S. population over the next 40 years. For estimation of the “low end” scenario number of cancer cases, we multiplied Lower Series National Population Projections by the lower 95% confidence interval (CI) of Crude Rate (Equation). For estimation of the “high end” scenario, we multiplied the Upper Series National Population Projections by the upper 95% CI of Crude Rate. This exercise resulted in the creation of “error bars” to surround estimates of cancer cases and the ratio of cancer cases per gynecologic oncologist.
We next examined a scenario in which the annual number of fellowship graduates increases by 25% from the current average of 43 to 54. We chose the trend of an increase in fellowship positions based on an upward trend in the number of fellowships as well as the number of positions offered per fellowship program over the past 10 years. We also varied the sex distribution of fellowship graduates from 50% female to extremes of 35% and 75% female.
We next addressed the potential effect of obesity on the number of cases of endometrial cancer. Using the projected prevalence of obesity through 2050 and the relative risk of endometrial cancer associated with that risk factor, we revised our projections of the incidence of endometrial cancer9,10 and used this incidence to revise estimates of endometrial cancer cases.
Finally, we varied estimates of HPV vaccine coverage for sensitivity analysis to determine the estimates number of cervical cancer cases if vaccination coverage were increased to 100%.
The number of cases of each major gynecologic cancer will increase over the next 40 years (Fig. 1). Annual projected totals for 2010 and 2050 are: ovarian cancer, 26,870 and 44,130; uterine cancer, 48,339 and 77,980; cervical cancer, 14,020 and 20,208; vulvar cancer, 4,044 and 7,642; and vaginal cancer, 1,384 and 2,524. The relative proportions of each cancer type will not change substantially. Uterine cancer accounts for slightly over 50% of cases with ovarian cancers comprising another 30%. The total number of cervical cancer cases decreases by 2050 when accounting for HPV vaccine coverage. Assuming 30% HPV vaccine coverage, the number of new cervical cancer cases per year decreases from 153 to 70 per year.
The projected number of gynecologic oncologists in practice will rise from 873 in 2009 to 1,133 in 2050 (Fig. 2). The number of gynecologic oncologists appears to plateau in the last decade. This is because we assume that the age distribution for fellows entering the pool does not change. After the current fellows reach retirement in 2040, the population is in a steady state with the number of fellows entering the practicing physician cohort equal to the number of gynecologic oncologists retiring.
At constant fellowship training rates, the annual number of new gynecologic cancer cases per practicing gynecologic oncologist will rise from 112 in 2010 to 133 in 2050, representing a 19% increase (Fig. 3).
Assuming no HPV vaccine, the projected total number of cervical cancer cases increases at a constant rate of 153 cases per year (Fig. 4). If 100% HPV vaccine coverage is assumed, the number of cervical cancer cases decreases yearly by over 100 cases (Table 1), although these changes are not seen until at least 2025 when the preadolescents who received the vaccine reach the age at which the incidence of cervical cancer begins to increase. The total number of cancer cases per gynecologic oncologist increases by 14% to 128 by 2050 if we assume 100% HPV vaccine coverage compared with a ratio of 133 (19% increase) under base case assumptions.
The error bars in Figure 3 represent the uncertainty in the U.S. population as projected by the Census Bureau over the next 40 years as well as uncertainty regarding the incidence of gynecologic cancers. Using the lower end estimate of gynecologic cancer cases, the ratio of cancers per practicing gynecologic oncologist will remain relatively stable at 107 by 2050, whereas use of the high end estimate of cancer cases results in an increase in this ratio by 43% to 160.
For sensitivity analysis, if the annual number of fellowship graduates increases by 25%, the projected ratio of new cancer cases per gynecologic oncologist decreases by 5% to 106 (Fig. 5). This estimate is lower than the lowest estimate (represented by error bars) of the projected ratio at current fellowship training rates. If we reduce the sex distribution of fellows entering practice from 50% (base case estimate) to 35% female, the projected number of practicing gynecologic oncologists rises (Fig. 2) and the ratio of new cancer cases per gynecologic oncologist increases by 13% to 127 by 2050. If the sex of fellows entering practice increases to 75% female, the projected number of practicing gynecologic oncologists drops (Fig. 2) and the ratio of new cancer cases per gynecologic oncologist will increase by 28% to 143 by 2050.
Assuming an increased incidence of endometrial cancer based on a higher future prevalence of obesity, the projected total number of uterine cancer cases increases by almost 25% compared with our base case assumptions (Fig. 6). This translates into a 33% increase in cancer cases per gynecologic oncologist in 2050.
As the U.S. population continues to grow and age, cancer cases will inevitably increase as well. Bodies that regulate and oversee the training of gynecologic cancer specialists should be aware of the demand for their services. Our model suggests that, at constant fellowship training and obesity rates, the annual number of new cancer cases per practicing gynecologic oncologist will increase by almost 20% over the next 40 years. Case load projections are even higher for the following scenarios: 1) continued increase in the prevalence of obesity; and 2) lower than 30% levels of HPV vaccine coverage. These estimates are bounded by wide CIs, mostly owing to uncertainty in the growth of the U.S. population. However, even at the lower population estimates, the ratio of cancer cases per gynecologic oncologist is likely to increase slightly.
Attention to the number of gynecologic oncologists who are being trained is important when considering both the adequacy of experience trainees receive and the ability of women with cancer to access practitioners of this subspecialty. Access to care is critical; in one study, approximately one-third of women with ovarian cancer nationally were treated by a gynecologic oncologist.11 Besides access to care, the number of surgical cases performed by a physician may also play a vital role in the quality of treatment a woman receives. Two recent studies have proposed a link between cancer case volume and outcomes. Regarding uterine cancer, one study defined the performance of at least 100 uterine cancer surgeries over 12 years as high volume and associated higher volume with lower short-term mortality, decreased risk of in-hospital death, and improved long-term survival.12 For ovarian cancer, a higher annual surgeon case volume of 10 or more cases has been associated with optimal tumor cytoreduction, decreased risk of in-hospital death, and improved survival outcomes.13 Thus, attention to the number of cancer cases available to trained gynecologic oncologists is critical to the quality of care women receive.
In considering the appropriateness of increasing the number of fellowship training positions, it is also important to foresee possible changes in referral practices by the general gynecology community. In an era of increasing specialization, precancerous lesions, low-grade endometrial cancers, and complicated benign lesions may be increasingly managed by gynecologic oncologists.14 Changes in referral practice may serve to further increase the demand for the services of gynecologic oncologists and should be considered in addition to the number of women diagnosed with a gynecologic malignancy. If it is assumed that an increasing proportion of future complicated benign gynecologic cases will be referred to gynecologic oncologists, the case load for this subspecialty will likely increase further than our study indicates. Nonetheless, because cancer serves as the basis for the gynecologic oncology subspecialty, cancer cases may serve as an indicator of anticipated trends in the demand for gynecologic oncologists.
One area of uncertainty in our estimation of projected numbers of gynecologic oncologists in practice is sex distribution. Because the average planned time in practice was 5 years less for a female than for a male gynecologic oncologist at the time of the 2005 Society of Gynecologic Oncologists Practice Survey, increasing the percentage of practicing women in the model reduced the pool of practicing physicians over time, thus increasing the future ratio of cancer cases per gynecologic oncologist. In 2009, 41% of those who passed the gynecologic oncology oral examination were female. Given the increasing percentage of female residents in obstetrics and gynecology (78% in 2008), we assumed for the base case that on average 50% of fellowship graduates will be female in the future.7 For sensitivity analysis, we examined a scenario in which as many 75% of future fellowship graduates are female, and this resulted in a 29% increase in cancer cases per gynecologic oncologist. This estimate still may be conservative because it does not account for other potential sex-related practice effects such as maternity leave on the female gynecologic oncologist workforce.
We acknowledge that these estimates are subject to numerous other uncertainties. The Census projections themselves have wide CIs, particularly in later years. The Society of Gynecologic Oncologists 2005 Practice Survey, which we used for our base case assumptions on sex and age distribution, was limited by a low response rate of 41%.6 Likewise, differences in practice patterns by geographic location are not addressed by our study. For example, in the 2005 Society of Gynecologic Oncologists Practice Survey, gynecologic oncologists in the western region of the United States plan to remain in practice for an average of 13 years longer as compared with 18 years for gynecologic oncologists in the southern region and 17 years for gynecologic oncologists in the northeast.
Finally, we did not account for all factors that may influence cancer incidence and mortality over the coming 40 years. For example, our modeling of the effect of HPV vaccines on invasive cervical cancer incidence and mortality was simplistic; we did not attempt to comprehensively model varying vaccine coverage, potential herd immunity, potential changes in screening behaviors, or other factors that may well affect the effect of the vaccine on cancer.15–18 However, because cervical cancer is a relatively small proportion of the case volume for gynecologic oncologists, rapid increases in HPV vaccine coverage are unlikely to result in substantial decreases in overall case volume over the next 20–40 years. Likewise, we did not include the effects of future technologic advances that may lead to earlier diagnosis or prevention of cancers, changes in rates of tobacco abuse, trends in healthcare coverage that may affect compliance with cancer screening, changes in child-bearing patterns, prevalence of oral contraceptive use, or rates of procedures that reduce the incidence of gynecologic cancers such as hysterectomy and oophorectomy. Although all of these factors have a potential to affect the projected number of gynecologic cancers, their prediction is beyond the scope of this study.
The ideal number of gynecologic oncologists in practice is unknown. However, under some scenarios incorporating population growth and the demographics of gynecologic oncologists, changes in the number of trainees may be needed to strike a reasonable balance between timely access to care and the minimum volume of cases necessary to ensure good patient outcomes. Continued research into the effects of all of the factors affecting this balance is needed to rationally plan for our future patients' needs.
2. Heron M, Hoyert DL, Murphy SL, Xu J, Kochanek KD, Tejada-Vera B. Deaths: final data for 2006. Natl Vital Stat Rep 2009;57:1–134.
3. U.S. Census Bureau. Projected population by single year of age, sex, race, and Hispanic origin for the United States: July 1, 2000 to July 1, 2050. Washington, DC: US Census Bureau; 2008.
4. Surveillance, Epidemiology and End Results (SEER) Program. SEER*Stat Database, November 2007 Sub (1973–2005), National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, Released April 2008, based on the November 2007 submission. Available at: www.seer.cancer.gov
. Retrieved September 15, 2009.
5. Clifford GM, Smith JS, Plummer M, Muñoz N, Franceschi S. Human papillomavirus types in invasive cervical cancer worldwide: a meta-analysis. Br J Cancer 2003;88:63–73.
6. Society of Gynecologic Oncologists 2005 Practice Survey. Prepared by Smith Bucklin Corporation–Market Research & Statistic Group. Chicago (IL): Society of Gynecologic Oncologists; 2006.
7. Leadley J. Women in U.S. academic medicine: statistics and benchmarking report, 2008–2009. Washington, DC: American Association of Medical Colleges; 2009.
9. Wang Y, Beydoun MA, Liang L, Caballero B, Kumanyika SK. Will all Americans become overweight or obese? Estimating the progression and cost of the US obesity epidemic. Obesity 2008;16:2323–30.
10. Bjørge T, Engeland A, Tretli S, Weiderpass E. Body size in relation to cancer of the uterine corpus in 1 million Norwegian women. Int J Cancer 2007;120:378–83.
11. Carney ME, Lancaster JM, Ford C, Tsodikov A, Wiggins CL. A population-based study of patterns of care for ovarian cancer: who is seen by a gynecologic oncologist and who is not? Gynecol Oncol 2002;84:36–42.
12. Díaz-Montes TP, Zahurak ML, Giuntoli RL 2nd, Gardner GJ, Bristow RE. Uterine cancer in Maryland: Impact of surgeon case volume and other prognostic factors on short-term mortality. Gynecol Oncol 2006;103:1043–7.
13. Bristow RE, Zahurak ML, Diaz-Montes TP, Giuntoli RL, Armstrong DK. Impact of surgeon and hospital ovarian cancer surgical case volume on in-hospital mortality and related short-term outcomes. Gynecol Oncol 2009;115:334–8.
14. Im SS, Gordon AN, Buttin BM, Leath CA 3rd, Gostout BS, Shah C, et al. Validation of referral guidelines for women with pelvic masses. Obstet Gynecol 2005;105:35–41.
15. Bauch CT, Li M, Chapman G, Galvani AP. Adherence to cervical screening in the era of human papillomavirus vaccination: how low is too low? Lancet Infect Dis 2010;10:133–7.
16. Regan DG, Philp DJ, Hocking JS, Law MG. Modelling the population-level impact of vaccination on the transmission of human papillomavirus type 16 in Australia. Sex Health 2007;4:147–63.
17. Kulasingam SL, Myers ER. Potential health and economic impact of adding a human papillomavirus vaccine to screening programs. JAMA 2003;290:781–9.
18. Kulasingam SL, Pagliusi S, Myers E. Potential effects of decreased cervical cancer screening participation after HPV vaccination: an example from the U.S. Vaccine 2007;25:8110–3.