Prostate Cancer Incidence and PSA Testing Patterns in Relation to USPSTF Screening Recommendations

Importance  Prostate cancer incidence in men 75 years and older substantially decreased following the 2008 US Preventive Services Task Force (USPSTF) recommendation against prostate-specific antigen (PSA)–based screening for this age group. It is unknown whether incidence has changed since the USPSTF recommendation against screening for all men in May 2012.

Objective  To examine recent changes in stage-specific prostate cancer incidence and PSA screening rates following the 2008 and 2012 USPSTF recommendations.

Design and Settings  Ecologic study of age-standardized prostate cancer incidence (newly diagnosed cases/100 000 men aged ≥50 years) by stage from 2005 through 2012 using data from 18 population-based Surveillance, Epidemiology, and End Results (SEER) registries and PSA screening rate in the past year among men 50 years and older without a history of prostate cancer who responded to the 2005 (n = 4580), 2008 (n = 3476), 2010 (n = 4157), and 2013 (n = 6172) National Health Interview Survey (NHIS).

Exposures  The USPSTF recommendations to omit PSA-based screening for average-risk men.

Main Outcomes and Measures  Prostate cancer incidence and incidence ratios (IRs) comparing consecutive years from 2005 through 2012 by age (≥50, 50-74, and ≥75 years) and SEER summary stage categorized as local/regional or distant and PSA screening rate and rate ratios (SRRs) comparing successive survey years by age.

Results  Prostate cancer incidence per 100 000 in men 50 years and older (N = 446 009 in SEER areas) was 534.9 in 2005, 540.8 in 2008, 505.0 in 2010, and 416.2 in 2012; rates began decreasing in 2008 and the largest decrease occurred between 2011 and 2012, from 498.3 (99% CI, 492.8-503.9) to 416.2 (99% CI, 411.2-421.2). The number of men 50 years and older diagnosed with prostate cancer nationwide declined by 33 519, from 213 562 men in 2011 to 180 043 men in 2012. Declines in incidence since 2008 were confined to local/regional-stage disease and were similar across age and race/ethnicity groups. The percentage of men 50 years and older reporting PSA screening in the past 12 months was 36.9% in 2005, 40.6% in 2008, 37.8% in 2010, and 30.8% in 2013. In relative terms, screening rates increased by 10% (SRR, 1.10; 99% CI, 1.01-1.21) between 2005 and 2008 and then decreased by 18% (SRR, 0.82; 99% CI, 0.75-0.89) between 2010 and 2013. Similar screening patterns were found in age subgroups 50 to 74 years and 75 years and older.

Conclusions and Relevance  Both the incidence of early-stage prostate cancer and rates of PSA screening have declined and coincide with 2012 USPSTF recommendation to omit PSA screening from routine primary care for men. Longer follow-up is needed to see whether these decreases are associated with trends in mortality.

Quiz Ref IDRecommendations for prostate-specific antigen (PSA)–based prostate cancer screening in the United States have changed substantially over the years (eTable 1 in the Supplement). In 1992, the American Urological Association (AUA) and the American Cancer Society (ACS) recommended annual screening for men 50 years and older. During this time, PSA screening was widely adopted and was associated with increases in prostate cancer incidence.¹ The US Preventive Services Task Force (USPSTF) recommended against PSA-based screening for men 75 years and older in 2008,² and for all men in a publicly announced draft form in October 2011 and final form in May 2012, concluding that the benefits of PSA-based screening for prostate cancer do not outweigh the harms.³ This decision was widely covered by the media and debated in the scientific literature.^4-7 In contrast to USPSTF, the AUA and ACS currently recommend that PSA testing be offered to asymptomatic men aged 55 to 69 years (AUA) or men older than 50 years with a minimum 10-year life expectancy (ACS) after patients receive information about the harms and benefits associated with screening.^8,9

Previous studies documented decreases in early-stage prostate cancer incidence¹⁰ and in health insurance claims–based PSA testing following the 2008 USPSTF recommendation against screening for men 75 years and older.^11,12 A recent study based on the 2005, 2010, and 2013 National Health Interview Surveys (NHIS) reported decreases in self-reported PSA testing following the 2012 USPSTF recommendation against screening for all men, although it was limited to men who saw a physician in the past year.¹³ In this study, we concurrently examined trends in stage-specific prostate cancer incidence and PSA-based screening for men 50 years and older subsequent to the 2008 and 2012 USPSTF recommendations using the most recent population-based incidence and nationally representative screening data.

Quiz Ref IDInvasive prostate cancer incidence data from 2005 through 2012 were obtained from 18 population-based cancer registries participating in the Surveillance, Epidemiology, and End Results (SEER) program of the National Cancer Institute, which includes about 28% of the US population.¹⁴ The most recent year of SEER data is 2012. Incident cases were categorized into local/regional- or distant-stage disease according to the SEER summary stage, which has been used consistently for cases diagnosed since 2001.¹⁵ We categorized and examined incidence by stage because changes in routine screening practice are more likely to first affect detection of early-stage malignancy.¹⁶ The study was based on deidentified publicly available data, which is considered nonhuman participants research under the US Department of Health and Human Services’ Office for Human Research Protection and does not require institutional review board review or informed consent.¹⁷

Incidence was adjusted for delays in reporting as described by Clegg et al,¹⁸ and 99% CIs were calculated by race/ethnicity (all races, non-Hispanic white, and non-Hispanic black), age (≥50, 50-74, ≥75 years), stage (all stages, local/regional, and distant), and SEER registry grouping (all 18 registries, 9 original registries, 1992 expansion [4 registries], 2000 expansion [5 registries]) using SEER*Stat 8.2.1.¹⁴ Adjusting for delay in reporting corrects current counts for future addition and deletion based on historical reporting data from SEER registries.¹⁸ Incidence is most affected by delay in reporting during the most recent 1 to 2 calendar years and for cancers that are diagnosed in nonhospital settings, such as prostate cancer and leukemia.¹⁸ We generated incidence for groups of registries and for non-Hispanic white and non-Hispanic black individuals because delay-adjusted rates are not available for individual registries and for races/ethnicities other than non-Hispanic white and non-Hispanic black. Incidence in each age category was expressed per 100 000 men and age-standardized to the 2000 US standard population. Incidence ratios (IRs), which compare relative changes in incidence between each consecutive year (eg, 2006 vs 2005), and their 99% CIs were calculated using the Tiwari method, an efficient interval estimate for age-standardized rates.¹⁹

In supplementary analyses, we similarly examined changes in incidence by age for the United States based on SEER and National Program of Cancer Registries combined data from the Centers for Disease Control and Prevention (CDC), covering 99.2% of the US population.²⁰ We adjusted the incidence for delay in reporting using the SEER delay-reporting correction factors²¹ and the counts for delay in reporting and nationwide geographic coverage by first multiplying the observed counts by delay-reporting correction factors and then by a geographic coverage correction factor of 1.008. We also examined changes in incidence from 2005 to 2013 by age in Georgia, one of the 18 registries participating in the SEER program, because of the availability of the 2013 data through a special request. Similarly, incidence rates were adjusted for delay in reporting using the SEER delay-reporting correction factors.

We used data from 2005, 2008, 2010, and 2013 NHIS, a nationally representative household cross-sectional sample survey of noninstitutionalized adults, to estimate the rate of self-reported PSA testing in the past year for screening and for any reason by age (≥50, 50-74, and ≥75 years).²² These years were selected because they are the only years in which men were asked about PSA testing since 2005. We began our analyses in 2005 because the 2005 NHIS survey was the latest NHIS survey with PSA testing information before 2008, when the USPSTF recommended against PSA-based screening in men 75 years and older. For each of the years with PSA information, respondents who reported PSA testing were asked to indicate if the test was part of a routine examination, due to a problem, or for other reasons. For our main analyses, we calculated screening rates for men receiving PSA testing in the past year as part of a routine examination, hereafter referred to as PSA screening. In supplementary analyses, we calculated rates for receipt of PSA testing in the past year for any reason, including unknown reasons. After exclusion criteria (eFigure in the Supplement, eg, men without a history of prostate cancer for primary analyses), the primary analysis of PSA testing for screening was based on 18 385 men 50 years and older and for any reason based on 19 014 men 50 years and older in the United States.²² Screening rate ratios (SRRs), which compare the relative change in PSA screening rates between survey years (eg, 2005 vs 2003), and 99% CIs adjusting for race and age were estimated using logistic regression models with predicted marginal probabilities in SAS-callable SUDAAN version 9.0.3.²³

Figure 1 shows annual prostate cancer incidence per 100 000 men 50 years and older and IRs between each consecutive year from 2005-2012 for all races combined by age in SEER areas (N = 446 009). Incidence in men 50 years and older increased from 534.9 (99% CI, 528.6-541.1) in 2005 to 592.5 (99% CI, 586.2-598.9) in 2007 and decreased afterward (eTable 2 in the Supplement). However, the largest year-over-year decline in incidence per 100 000 occurred between 2011 and 2012, from 498.3 (99% CI, 492.8-503.9) to 416.2 (99% CI, 411.2-421.2), an absolute decline of 82.1 cases per 100 000 men and a relative decline of 16% (IR, 0.84; 99% CI, 0.82-0.85). The second largest decline in incidence occurred between 2007 and 2008, from 592.5 (99% CI, 586.2-598.9) to 540.8 (99% CI, 534.9-546.8), an absolute decrease of 51.7 cases per 100 000 and relative decrease of 9% (IR, 0.91; 99% CI, 0.90-0.93). By 2010, the incidence rate had further declined to 505.0 (99% CI, 499.4-510.6). We found generally similar patterns in the age groups 50 to 74 years and 75 years and older and in non-Hispanic white and non-Hispanic black individuals (eTable 2 in the Supplement). The incidence patterns in the United States based on combined SEER and National Program of Cancer Registries data were similar to those of SEER areas, although rates were slightly lower (eTable 3 in the Supplement). Between 2011 and 2012 in US men 50 years and older, incidence per 100 000 men decreased from 468.7 (99% CI, 466.2-471.1) to 381.2 (99% CI, 379.0-383.3), and the number of men diagnosed with prostate cancer decreased from 213 562 to 180 043; 33 519 fewer men received a diagnosis of prostate cancer in 2012 compared with 2011.

Figure 2 presents prostate cancer IRs between consecutive years for all races combined by age and stage from 2005 through 2012 in the SEER registries. In all age groups, the declines in incidence that began in 2008 were confined to local/regional-stage disease, with the largest decrease occurring between 2011 and 2012. For example, among men 50 years and older, incidence per 100 000 significantly decreased from 447.2 (99% CI, 442.0-452.4) in 2011 to 367.3 (99% CI, 362.7-372.0) in 2012, an absolute decrease of 79.9 cases per 100 000 and a relative decrease of 18% (IR, 0.82; 99% CI 0.81-0.84) (eTable 4 in the Supplement). Incidence for distant-stage disease generally remained unchanged over the study period in all age groups, except in those 75 years and older, in whom rates per 100 000 significantly increased from 57.7 (99% CI, 53.2-62.3) in 2011 to 65.0 (99% CI, 60.3-69.9) in 2012. Similar stage-specific incidence patterns were observed in non-Hispanic white and non-Hispanic black individuals (eTable 4 in the Supplement) and in all 3 cancer registry groups (eTable 5 in the Supplement), with the exception of the increase from 2011 to 2012 in men 75 years and older for distant-stage disease, which was confined to non-Hispanic white men.

Prostate cancer incidence patterns in men 50 years and older in Georgia, with the additional data year (2013), were generally similar to those of all SEER registries (eTable 6 in the Supplement). In all age groups, incidence began decreasing in 2008, with the largest decrease occurring between 2011 and 2012. For example, for men 50 years and older, incidence per 100 000 decreased from 519.0 in 2011 to 436.5 in 2012, a relative decline of 16% (IR, 0.84; 99% CI, 0.80-0.88). Incidence in 2013 further decreased to 408.3, a relative decrease of 6% (IR, 0.94; 99% CI, 0.89-0.98) from the 2012 level.

The Table shows adjusted self-reported PSA screening rate in the past year and SRR between each successive survey year by age (≥50, 50-74, ≥75 years). PSA screening rate in men 50 years and older significantly increased from 36.9% (99% CI, 34.5%-39.1%) in 2005 to 40.6% (99% CI, 37.9%-43.3%) in 2008. Screening then showed a nonsignificant decrease to 37.8% (99% CI, 35.3%-40.2%) in 2010 and a significant decrease to 30.8% (99% CI, 29.0%-32.7%) in 2013, an absolute decrease of 7.0% and a relative decrease of 18% (SRR, 0.82; 99% CI, 0.75-0.89). Similar patterns were found in the age groups 50 to 74 years and 75 years and older, although the change between 2010 and 2013 was not statistically significant among those 75 years and older (Table). The crude screening rates were similar to adjusted screening rates (eTable 7 in the Supplement); changes in PSA testing patterns for any reason were similar except the rates were slightly higher (eTable 8 in the Supplement).

Quiz Ref IDUsing the most recent population-based incidence and nationally representative self-reported PSA screening data, we report reductions in early-stage prostate cancer incidence and PSA-based screening rates in men 50 years and older, coinciding with the 2012 USPSTF recommendation against PSA-based screening. The decreases in incidence were evident in both non-Hispanic white and non-Hispanic black individuals and across regions. We also found declines in incidence and PSA testing preceding the 2012 USPSTF recommendation, although the decreases in PSA screening were not significant.

The decrease in incidence for early-stage prostate cancer occurred within a year after the release of the USPSTF draft recommendation against screening in October 2011,²⁴ which was widely debated in the media and scientific literature.^4-7 Similar changes in occurrence of cancer following new screening recommendations, changes in reimbursement policies, or newsworthy publications have been previously noted for other cancers.^25,26 For example, Medicare’s expanded coverage of screening colonoscopy for all individuals in July 2001 was associated with increased use of colonoscopy and early-stage diagnosis of colorectal cancer within 6 months after the expansion of the coverage.²⁵ Declines in prostate cancer detection rates following the 2012 USPSTF recommending against screening were reported in Toronto, Canada, based on institutional health network data.²⁷

Alternative explanations for the substantial decrease in incidence between 2011 and 2012 are unlikely. There are no known strong risk factors for prostate cancer, for which temporal changes are likely to affect incidence, particularly for early-stage disease. In large clinical trials, 5α-reductase inhibitors have been shown to reduce the risk of prostate cancer,^28,29 but they are not approved for prevention of prostate cancer and there is no evidence for increased prescriptions in the general population.³⁰ Change in access to care is unlikely to be responsible because the percentage of people with health insurance was similar between 2011 and 2012,³¹ PSA-based screening rates significantly declined in men who saw a physician following the 2012 USPSTF recommendation,¹³ and there have been no recent changes in coverage of PSA testing by Medicare.³² Any additional delay over the historical delay of 3.6% of total prostate cancer cases for the first reporting year is unlikely to be large enough to explain the observed relative decline of 16%. The possibility that the decrease may be due to random fluctuation in the underlying trend is also unlikely because the decrease occurred across all age, race, and SEER registry groups.

The potential emerging trends in prostate cancer incidence and PSA-based screening following the 2012 USPSTF recommendation against screening could have significant public health implications. About 33 519 fewer men nationwide received a diagnosis of prostate cancer in 2012 than in 2011. On one hand, overdiagnosis and overtreatment may be reduced in view of the substantial proportion of prostate cancer cases detected through PSA testing that would not cause harm if left undetected.³³ On the other hand, less screening or discontinuing screening may lead to missed opportunities for detecting biologically important lesions at an early stage and preventing deaths from prostate cancer, the ultimate goal of screening. Using a simulation model, Gulati and colleagues³⁴ predicted that compared with continuing screening, completely discontinuing screening in the United States would increase prostate cancer mortality by 13% to 20%. Quiz Ref IDHowever, given the slow progression and long natural history of the disease, any increase in mortality rates may not be seen for several years after the discontinuation of screening. Future studies should examine temporal trends in advanced-stage diagnoses and prostate cancer mortality in order to assess the long-term effects of changes in PSA-based screening practices at the population level.

With respect to the trends in prostate cancer incidence and PSA screening rates preceding the 2012 USPSTF recommendation, both early-stage prostate cancer incidence and screening rates in men 75 years and older declined (nonsignificant for PSA screening) coinciding with the 2008 USPSTF recommendation against screening for this age group, consistent with previous reports.^10-12 However, the incidence and screening patterns appeared to be discordant because of difference in onset of the declines in incidence (2008) and screening (2010). This discrepancy may result from differences in the years being compared, every consecutive year for incidence and every 2 or 3 years for screening. In particular, we could not determine whether PSA testing rate was higher in 2007 than in 2008 because NHIS did not include PSA testing questions in the 2007 survey.

Similar to men 75 years and older, incidence and screening rates decreased among men aged 50 to 74 years between 2008 and 2010, although the magnitude of the decreases was smaller. Contributing factors for these patterns are unknown but may include changes in perception of the importance of PSA screening among providers and the general public following the 2008 USPSTF recommendation against screening (spillover effect). The simultaneous publications of 2 large randomized trials with conflicting results regarding the benefit of PSA testing in 2009^35,36 also was associated with a decline in screening.³⁷ Increased emphasis on informed decision making for screening during the most recent time periods by other public health and professional organizations is also likely to have contributed to these trends.^8,9 Patients who received information about PSA screening have reported less interest in testing.^38,39

Strengths of our study include the use of the most recent population-based cancer registry data and national survey data to report emerging trends in prostate cancer incidence and PSA testing following the 2012 USPSTF recommendation against screening. In addition to the ecological nature of the study, our analysis was limited by a number of other factors. First, we used different populations for the incidence data (SEER, covering 28% of the US population) and PSA data (a nationwide sample). However, prostate cancer incidence patterns between SEER and the United States (SEER and National Program of Cancer Registries data combined) were similar. Second, information on PSA screening was based on self-reported data and subject to recall bias.^40,41 Third, our analyses were based on short follow-up following the 2012 USPSTF recommendation against screening, single-year observation data for both incidence (2012) and screening (2013). However, based on incidence data from Georgia, a member of the SEER cancer registry program, the decrease in incidence appeared to be sustained through 2013. Fourth, we examined changes in prostate cancer incidence but not mortality rates, which require a longer monitoring period after discontinuation of screening because of the long natural history of the disease.

Both the incidence of early-stage prostate cancer and rates of PSA screening have declined and coincide with 2012 USPSTF recommendation to omit PSA screening from routine primary care for men. Quiz Ref IDLonger follow-up is needed to see whether these decreases are associated with trends in mortality.

Corresponding Author: Ahmedin Jemal, DVM, PhD, Surveillance and Health Services Research, American Cancer Society, 250 Williams St NW, Atlanta, GA 30303 (ahmedin.jemal@cancer.org).

Author Contributions: Ms Fedewa and Dr Ma had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Jemal, Fedewa, Brawley.

Acquisition, analysis, or interpretation of data: Fedewa, Ma, Siegel, Lin, Ward.

Drafting of the manuscript: Jemal, Fedewa, Brawley.

Critical revision of the manuscript for important intellectual content: Fedewa, Ma, Siegel, Lin, Ward.

Statistical analysis: Jemal, Fedewa, Ma.

Administrative, technical, or material support: Lin, Brawley.

Study supervision: Brawley, Ward.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.

Funding/Support: This project was supported by the Intramural Research Department of the American Cancer Society.

Role of the Funder/Sponsor: The American Cancer Society had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Disclaimer: The opinions expressed are solely the responsibility of the authors and do not necessarily reflect the official views of the American Cancer Society.

Additional Contributions: We thank Kevin Ward, PhD, MPH, of the Georgia Center for Cancer Statistics, Emory University, for providing the incidence data in Georgia for the period 2005-2013. He did not receive compensation for his contribution.