USPSTF guidelines may have decreased the number of patients presenting with prostate cancer for radiation oncology care, particularly those with low-risk disease, without a short-term increase in higher-risk disease.
About the lead author:
Kevin Su, BS
Mr. Su is a 4th year medical student currently attending Yale University School of Medicine
Timur Mitin, MD, PhD
Department of Radiation Medicine
Oregon Health & Science University
Prostate cancer represents a spectrum of disease that ranges from nonaggressive, slow-growing disease that may not require treatment to aggressive, fast-growing disease that does. Death rates for prostate cancer in the United States have fallen dramatically, largely due to early detection and improved treatment. Nevertheless, the screening of men for prostate cancer is probably the most hotly debated clinical topic in this country. In 2008, the US Preventive Services Task Force (USPSTF) recommended against prostate-specific antigen-based screening program for men over age 75, and in 2012, for all men. Other expert physician organizations, including the American Urological Association (AUA) and the National Comprehensive Cancer Network (NCCN), support prostate cancer screening with balanced and proper patient education preceding the PSA draw.
Significant flaws of the US randomized screening study, which largely formed the basis for the USPSTF recommendations, have been addressed by many leaders in the field of prostate cancer. The effect of this recommendation on the well-being of US men will take more years to become apparent; however its effect on the healthcare system is already obvious, with fewer patients being diagnosed with prostate cancer, leading to fewer referrals to urologists, medical, and radiation oncologists.
This manuscript documents a significant decrease in the volume of patients seen across all radiation oncology and urology clinics in Lee County in southwest Florida. The analysis does not prove that this trend is directly related to the USPSTF recommendations. Nevertheless, these findings are important for healthcare budgeting, physician training, and workforce allocation. It will be interesting to compare this trend to those seen at academic institutions across the country. With a longer follow-up, the authors are encouraged to re-evaluate the percentage of patients presenting with locally advanced and metastatic disease. Modeling studies will then be necessary to determine if these clinical outcomes are related to the USPSTF recommendations. One day, we will know whether the experts on the USPSTF panel got it right or wrong. Time will tell.
Background: In 2008, the US Preventive Services Task Force (USPSTF) recommended against prostate-specific antigen-based screening for prostate cancer in men age ≥75, and in 2012, recommended against screening in all men. It is unknown how these guidelines have impacted patient presentation to radiation oncology care.
We conducted a retrospective study of patients presenting with localized prostate cancer to 21st Century Oncology radiation oncology clinics in southwest Florida, which supply nearly all radiation oncology services for the region. We assessed trends in patient volume and the risk distribution of a random sample of patients (using National Comprehensive Cancer Network risk criteria) to estimate the total patient volume within each risk group.
Results: From 2007 to 2013, the number of patients presenting with prostate cancer decreased from 1,860 to 1,209 (p <.001). The estimated number of low-risk patients decreased from 858 (779-939) to 354 (307-405), as did the number of intermediate-risk patients, from 657 (583-736) to 506 (454-560). There was no significant change in the number of patients with high-risk (323 [267-389] to 312 [267-361]) or metastatic disease (22 [10-48] to 37 [23-61]). Similar trends held in all age groups, except intermediate-risk disease decreased significantly for men ≥75 years, but not younger men.
USPSTF guidelines may have decreased the number of patients presenting with prostate cancer for radiation oncology care, particularly those with low-risk disease, without a short-term increase in higher-risk disease. Future research should assess the longer-term impact of these guidelines nationally.
Prostate cancer is the most common cancer among males in the United States, with an estimated 233,000 new cases in 2014.1 Because a large proportion of men with prostate cancer do not experience symptoms during their lifetime, the significant increase in the incidence of prostate cancer since the introduction of prostatespecific antigen (PSA)-based screening has raised concerns of overdiagnosis and overtreatment.2 In 2008, the US Preventive Services Task Force (USPSTF) issued a recommendation against PSA-based screening for prostate cancer in men age ≥ 75,3 which was expanded in 2012 into a recommendation against screening in all men.4 These guidelines have generated substantial controversy, with some professional societies and patient advocacy groups arguing that screening can reduce mortality.5-8
Several studies have evaluated the impact of the 2008 guidelines, and most have found only a small decrease in the use of screening PSA.9-15 Decreased screening presumably decreases the detection rate of prostate cancer and could result in some men progressing to higher-risk disease before they present with symptoms. However, only one study has evaluated whether the guidelines have changed the number or risk profile of patients presenting with prostate cancer. A study of Surveillance, Epidemiology, and End Results (SEER) found that the number of patients presenting with prostate cancer declined from 2005 through 2009, with a more pronounced reduction in early than late-stage disease, but did not have sufficient follow-up time after the 2008 guideline change to fully assess its clinical impact.16 As the time it takes for asymptomatic prostate cancer to progress to clinically detectable disease can vary,17 it is important to understand the longer-term impact of the guidelines on patient presentation. Furthermore, no studies have assessed the impact of the guidelines on specialty care.
Patient presentation for radiation oncology care is of particular interest for several reasons. First, changes in patient volume and risk profile at radiation oncology clinics can be considered markers for changes in the detection of prostate cancer because radiation therapy is one of the most common curative treatments for prostate cancer, especially for the elderly.18 Second, changes in patient volume may affect the projected demand for radiation oncology services since the treatment of prostate cancer accounts for nearly 20% of radiation therapy administered.19 Third, the volume of patients presenting with prostate cancer may affect rates of both treatment-related benefits and harms. Treatment with radiation therapy may decrease mortality for some diagnosed patients,20 but could also increase overall healthcare costs and lead to unnecessary adverse events in patients who would not otherwise have had progressive disease.21 Trends in patient presentation at radiation oncology clinics highlight the downstream effects of reduced screening and will inform future debates regarding the consequences of the guidelines on actual clinical practice.
a P-value is for chi-squared test of difference in the distribution of cancer characteristics across years.
b denotes a significant change (p <.05) in the proportion of patients in a risk group, comparing 2007 versus 2013 using two sample test of proportions
To address these gaps in knowledge, we studied the numbers and risk profiles of patients presenting with prostate cancer to radiation oncology practices from 2007 to 2013. Because more than 5 years have passed since the 2008 guidelines, one might expect a change in patient presentation for men age ≥75, for whom the guidelines apply.
Although the 2012 guidelines apply to men of all ages, it is unclear whether enough time has passed for changes to manifest. We focused our study on a single county in southwest Florida, where every radiation oncology practice is affiliated with 21st Century Oncology, a specific radiation oncology provider. Thus, a review of patients presenting to 21st Century Oncology clinics encompassed the majority of radiation oncology referrals within this region. Although our study focused on radiology oncology, we also assessed patient volume at urology clinics in the region to determine whether changes in radiation oncology patient volume were due to substitution by urologic care. We conducted a retrospective study of patients presenting to 21st Century Oncology radiation oncology clinics in Lee County, Florida. The largest radiation oncology organization is 21st Century Oncology; it is also one of the largest groups of urologists in the United States.22 All radiation oncology practices in southwest Florida and all urology practices within Lee County are affiliated with 21st Century Oncology. We restricted our analysis to 12 radiation oncology clinics that were part of the 21st Century Oncology system from at least 2005-2013; this ensured that changes in the numbers of patients presenting were not due to the acquisition or loss of centers during the study period. In addition, we included 4 21st Century Oncology urology clinics in the region as a secondary sensitivity analysis to assess whether the USPSTF guidelines affected the use of competing treatment modalities—for instance, whether decreased patient volume at radiation oncology clinics corresponded with increased patient volume at urology clinics. However, cancer characteristic data were accessible only at radiation oncology clinics for our analysis of patient risk profiles.We identified all patients age ≥18 who presented with newly diagnosed prostate cancer between 2007 and 2013 at one of 12 radiation oncology practices or 4 urology practices (Table). Most patients with newly diagnosed prostate cancer, including those ultimately managed with watchful waiting or active surveillance, are evaluated at least once by a radiation oncologist or urologist. Hence, we included all patients presenting to specialty care, regardless of subsequent treatment or lack thereof, to fully capture trends in prostate cancer detection. For each calendar year, we excluded any patients who had been seen in the practice in the prior year. We also randomly sampled a subgroup of patients presenting to radiation oncology clinics in 2007 (before USPSTF recommendations) and 2013 (after USPSTF recommendations); we conducted a chart review on this subgroup to determine cancer characteristics at presentation.
Patient characteristics include age at diagnosis (≥ 75 years or <75 years) and year of diagnosis. We noted the specific 21st Century Oncology practice and type of practice (radiation oncology or urology) at which the patient first presented; if a patient presented to multiple practices, we only counted the first.
For a random subgroup of patients in 2007 and 2013, risk at presentation was categorized by T-stage, PSA, Gleason score (by biopsy), or National Comprehensive Cancer Network (NCCN) risk strata.23 T-stage was categorized into low (T1-T2A), intermediate (T2B-T2C), or high (T3-T4). PSA was categorized into low (<10 ng/ml), intermediate (10-20 ng/ml), or high (>20 ng/ml). Gleason score was categorized into low (<6), intermediate (7), or high (>8). NCCN risk was categorized into low (low-risk PSA and Gleason and T-stage), intermediate (intermediate risk PSA or Gleason or T-stage), high (high-risk PSA or Gleason or T-stage), or metastatic (N1 or M1). We conducted a chart review for patients classified as metastatic to assess whether the use of advanced imaging contributed to the detection of metastatic disease.To assess trends in patient volume, we calculated the number of patients age ≥75 years and <75 years presenting with prostate cancer to radiation oncology or urology practices in 2007 through 2013, and analyzed yearly trends using Poisson regression. We tested for an interaction between age group and year to assess whether changes in patient volume differed by age group. For the random subgroup of patients in 2007 and 2013 we categorized risk at presentation as outlined above, and calculated 95% confidence intervals (CIs) for the proportion of patients within each risk category, as well as P-values using 2 sample proportions. We compared differences in proportions across risk categories using chi-squared tests for all patients and for patients ≥75 years and <75 years, separately. We multiplied the proportion of patients within each risk category by the total patient volume in 2007 and 2013 to estimate the number of patients presenting within each risk category.
Similarly, we transformed the 95% CIs of the proportions by multiplying them by the total patient volume. We performed statistical analyses using SAS version 9.4 (SAS Institute Inc, Cary, North Carolina), Stata 13 (College Station, Texas) and Microsoft Excel. The Yale Human Investigations Committee and the Western Institutional Review Board deemed this study exempt from review.We identified 9334 patients who presented to radiation oncology clinics with prostate cancer from 2007 to 2013. Over our study period, the annual number of patients decreased by 35%, from 1860 to 1209 (Figure 1; P <.001). From 2007 to 2013, the number of patients age ≥ 75 years decreased 41% (from 689 to 408), while the number of patients age <75 years decreased 32% (from 1171 to 801). For both age groups, increasing year was associated with decreasing patient volume (P <.001). The magnitude of the decrease in patient volume was greater among patients age ≥ 75 than for patients <75 years (P-value for age group-year interaction <.001).
We observed similar trends for the 3818 patients presenting to urology clinics in the region (Figure 2). The annual number of patients decreased 45% over our study period, from 778 to 428 (P <.001). The number of patients age ≥75 years decreased 49% from 376 to 191, while the number of patients age <75 years decreased 41%, from 402 to 237. For both age groups, increasing year was associated with decreasing patient volume (P <.001). There was no significant difference in the magnitude of decrease in patient volume between age groups (P value for age groupyear interaction = .13).
The distribution of cancer characteristics at radiation oncology clinics changed over time (Table 1). The proportion of patients presenting with NCCN low-risk disease decreased significantly from 46.1% to 29.3% (P <.01). There was a trend toward increased proportions of intermediate-risk (35.3% to 41.9%), high-risk (17.4% to 25.8%), and metastatic (1.2% to 3.1%) disease, although none of these reached significance (P = .08, .06, and .39, respectively). We found a significant difference in the overall NCCN risk distribution across years (P <.001). We also found a significant reduction in low-risk disease when we categorized risk by T-stage (P <.01), PSA (P =.04), or Gleason score (P <.01).
The estimated annual number of patients presenting within each NCCN risk category also changed over time (Figure 3a). The annual number of low-risk patients decreased significantly, from 858 (95% CI, 779-939) to 354 (307-405) patients per year. The number of intermediate- risk patients decreased to a lesser though still significant degree, from 657 (95% CI, 583-736) to 506 (454-560). There was no significant change in the number of high-risk or metastatic patients.
We estimated a decreasing number of lower-risk patients and a stable number of higher-risk patients for both men ≥ 75 years and <75 years (Figures 3b and 3c). For men ≥75 years, the annual number of low-risk patients decreased significantly from 275 (234-317) to 84 (64-108), while the number of intermediate-risk patients decreased significantly from 250 (210-292) to 165 (139-192). We found no significant changes in the number of high-risk and metastatic patients. For men <75 years, the number of low-risk patients also decreased significantly from 610 (539-679) to 293 (248- 342). However, we found no significant changes in the number of intermediate-risk, high-risk, or metastatic patients.We found a significant reduction in the number of patients presenting with prostate cancer to radiation oncology practices. Over our study period, we observed a significant decrease in the number of men with low-risk disease, a modest decrease in intermediate-risk disease, and no significant changes in high-risk or metastatic disease. Our findings are consistent with our hypothesis that the 2008 USPSTF guidelines decreased the detection rate of prostate cancer, primarily through the reduction of low-risk disease, and that this has translated into reduced patient volume at radiation oncology practices.
Prior studies have assessed the impact of the 2008 guidelines, with most finding only marginal decreases in PSA screening.9-15 However, a study of national tumor registry data found that the number of patients age ≥75 years presenting with early-stage prostate cancer decreased 25.4% from 2007 to 2009.16 Our findings build upon this prior work by highlighting the impact of the guidelines on specialty care up through 2013. Because prostate cancer is currently the second most common indication for radiation therapy, the reduction in patient volume observed in our study suggests that future demand for radiation oncology services may be less than previously estimated.19 We considered whether practice trends other than decreased disease detection could have accounted for this reduction in patient volume. Specifically, we did not find evidence of increasing use of competing treatment modalities (eg, use of robotic surgery), as we also observed a parallel decrease in patient volume at urology clinics. Although the use of radiotherapy for prostate cancer may be decreasing in favor of watchful waiting or active surveillance,24 this would be unlikely to affect patient volume, as patients managed with these strategies would still receive an initial referral to specialty care.
Interestingly, we found a decrease in patient volume for men age <75 years, suggesting that our findings may reflect a secular decrease in the detection of prostate cancer independent of adherence to the 2008 guidelines.
For instance, it is possible that our study may have captured growing skepticism regarding the use of PSA screening that predates the 2008 guidelines; indeed, national tumor registry data suggest that the incidence of prostate cancer has decreased from 2000-2010, although this remains somewhat uncertain as incidence has also varied widely year to year.1 However, it is also possible that we are observing an indirect effect of the guidelines, with primary care physicians becoming increasingly skeptical of PSA screening for men of all ages. Lastly, factors outside the scope of our study, such as changes in benefit design, could also potentially shift practice patterns. Although major health plans currently cover radiotherapy for prostate cancer, it remains possible that patients are deferring active treatment as high deductible plans become more common. A decline in 21st Century Oncology market-share could contribute to the observed decrease in patient volume; however, we selected a geographic region in which all radiation oncology care was affiliated with 21st Century Oncology to minimize the effects of changing market share.
Although a reduction in PSA screening may prevent overdiagnosis and overtreatment, it could potentially increase prostate cancer-specific mortality by decreasing the detection and treatment of early-stage disease before it progresses.21 Two large randomized controlled trials evaluating screening PSA have reported conflicting findings regarding the effects of screening on risk at presentation and mortality.25,26 The US Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial reported no mortality benefit from screening and found a similar number of stage III/IV diagnoses in the screening and control arms at 10 years, though more patients in the control arm presented with high-grade disease.25 However, the European Randomized Study of Screening for Prostate Cancer found that screening reduced mortality by 20% and identified a reduction of stage III/IV and high-grade diagnoses in the screening arm.26 Furthermore, the impact of the USPSTF guidelines on risk at presentation also remains unclear, although one recent abstract found an increasing proportion of highrisk disease shortly after 2011.27 In contrast, 5 years after the 2008 guidelines, we found that the detection rate of prostate cancer has decreased without an increase in the incidence of higher-risk disease. While our study describes preliminary trends in prostate cancer presentation following the USPSTF guidelines, an increase in the prevalence of high-risk disease may take a longer period of time to manifest. Future studies should assess the long-term impact of the guidelines on prostate cancer epidemiology and mortality.
Our study has several limitations. First, our study is geographically limited by design and may not be generalizable to national trends or to academic centers. Although our sample is highly comprehensive for radiation oncology practice in the region, we cannot account for patients who left the region for treatment. Second, we did not collect cancer characteristics data at urology clinics; we only assessed patient risk at radiation oncology clinics. Third, the small number of metastatic patients in our sample limits our ability to make definitive conclusions regarding the impact of the guidelines on the incidence of metastatic disease. Although we note a nonsignificant increase in the number of men presenting with metastatic disease in our study, this does not appear to be a direct effect of the 2008 guidelines, as the increase was restricted to men age <75 years. We considered whether increased use of advanced imaging could have increased detection of metastatic disease by 2013; however, upon chart review, only 1 out of 15 patients presenting with metastatic disease in 2013 had 18F-Fluoride PET imaging at the time of diagnosis.
Our findings suggest that the 2008 USPSTF guidelines may have decreased the number of patients presenting with prostate cancer to radiation oncology care, particularly those with low-risk disease, without a short-term increase in the incidence of higher-risk disease. The decrease in patient volume at treatment centers may reflect a reduction in overtreatment and its associated costs, while the stable volume of high-risk disease suggests that expenditures related to treating advanced disease have yet to increase. The long-term impact of the guidelines on specialty care nationally, and on patient outcomes, are important areas for future investigation.
About the Authors
Cancer Outcomes, Public Policy, and Effectiveness Research (COPPER) Center, Yale Comprehensive Cancer Center and Yale University School of Medicine (KWS, JH, PRS, JBY, SBE, CPG). Section of General Internal Medicine, Department of Internal Medicine, Yale University School of Medicine, New Haven, CT (JH, PRS, CPG). 21st Century Oncology (RR, DED), Fort Myers, FL. 21st Century Oncology, (TDS) Providence, RI. Department of Therapeutic Radiology, Yale University School of Medicine (JBY, APD, SBE, RHD), New Haven, CT. 21st Century Oncology Translational Research Consortium (SEF), Scottsdale, AZ.
Corresponding Author: Cary P. Gross, 333 Cedar Street, PO Box 208025, New Haven CT 06520-8025, Tel: 203-688-8588, Fax: 203-688-4092, Email: firstname.lastname@example.org
Acknowledgements: We would like to thank 21st Century Oncology.
Conflicts of interest disclosures: Dr. Hall receives research funding from 21st Century Oncology. Mr. Ross is employed by and receives research funding from 21st Century Oncology. Ms. Soulos receives research funding from 21st Century Oncology. Dr. Shafman has is employed by and receives ownership interests in 21st Century Oncology. Dr. Yu has received research funding from 21st Century Oncology. Dr. Arie Dosoretz receives research funding from 21st Century Oncology, and reports that an immediate family member is employed by 21st Century Oncology in a leadership role and has ownership interests in 21st Century Oncology. Dr. Daniel Dosoretz is employed by and holds a leadership role in 21st Century Oncology. Dr. Finkelstein is employed by 21st Century Oncology, and reports a consulting or advisory role with Bayer, Dendreon, Medivation, Spectrum Pharmaceuticals, and Blue Earth; he participates in the speaker’s bureau for Bayer, Dendreon, and Medivation; he also receives research funding from Dendreon and Bayer, and holds patents pending for novel radiation technology. Ms. Salenius is employed by 21st Century Oncology. Dr. Decker has ownership interests in Bristol-Myer Squibb, honoraria with AMAG Pharmaceuticals and Leidos Biomedical Research, and research funding from 21st Century Oncology and Merck. Dr. Gross receives research funding from Medtronic, Johnson and Johnson, and 21st Century Oncology, has received honoraria from the American Society for Radiation Oncology, and serves on the Scientific Advisory Board for Fair Health Inc. No other conflicts of interest were disclosed.