Population-Based Patterns and Predictors of Prostate-Specific Antigen Screening Among Older Men in the United States

Abstract

Purpose

For patients who elect to have prostate cancer screening, the optimal time to discontinue screening is unknown. Our objective was to describe rates and predictors of prostate-specific antigen (PSA) screening among older men in the United States.

Methods

Data were extracted from the population-based 2000 and 2005 National Health Interview Survey (NHIS). PSA screening was defined as a PSA test as part of a routine exam within the past year. Demographic, socioeconomic, and functional characteristics were collected, and a validated 5-year estimated life expectancy was calculated. Age-specific rates of PSA screening were determined, and sampling weight-adjusted multivariate regressions were fitted to determine predictors of screening among men age 70 years or older.

Results

The PSA screening rate was 24.0% in men age 50 to 54 years, and it increased steadily with age until a peak of 45.5% among age 70 to 74 years. Screening rates then gradually declined by age, and 24.6% of men age 85 years or older reported being screened. Among men age 70 years or older, screening rates varied by estimated 5-year life expectancy: rates were 47.3% in men with high life expectancies (≤ 15% probability of 5-year mortality), 39.2% in men with intermediate life expectancies (16% to 48% probability), and 30.7% in men with low life expectancies (> 48% probability; P < .001). In multivariate analysis, estimated life expectancy and age remained independently associated with PSA screening (P < .001 for each).

Conclusion

Rates of PSA screening in the United States are associated with age and estimated life expectancy, but excessive PSA screening in elderly men with limited life expectancies remains a significant problem. The merits and limitations of PSA should be discussed with all patients considering prostate cancer screening.

INTRODUCTION

The landscape of prostate cancer diagnosis and treatment, particularly in the United States, has been dramatically altered by the identification¹ and widespread use of prostate-specific antigen (PSA) as a screening tool for prostate cancer.² Over the last 20 years, the proportion of men presenting with metastases has decreased 75%, and mortality rates have declined more than 30%.³ Concomitant with these encouraging trends is the estimated 1.3 million men diagnosed with prostate cancer since the introduction of PSA who do not reap a clinical benefit from treatment, because they would not have otherwise known of their cancer during their natural life span.⁴

For these reasons, intense controversy exists regarding the appropriate role of PSA-based prostate cancer screening. Recently reported randomized trials have not resolved, but rather have fueled, the ongoing debate.^5,6 The Prostate, Lung Cancer, and Ovarian Cancer Screening Trial randomly assigned 76,000 men to annual PSA screening or usual care and did not observe a difference in cancer-specific mortality at 10 years, although a significant proportion of men in the control arm underwent PSA screening (approximately 50%).⁵ The European Randomized Study of Screening for Prostate Cancer evaluated 162,000 men and showed a 20% reduction in cancer-specific mortality for screened men.⁶ However, the screening benefit was restricted to men age 55 to 69 years.

For advocates of screening, the recommended age at which to obtain the initial PSA, along with suggested frequency of screening, widely vary. The American Urological Association and American Cancer Society (ACS) published recommendations that are based on multidisciplinary expert opinion and that emphasize the importance of discussing and offering screening to men with at least 10 years of estimated life expectancy.^7,8 For men who elect to undergo screening, a baseline PSA level between age 40 to 50 years is recommended, depending on individual risk factors.^7,8

When to discontinue annual screening presents a particular challenge. The maximum ages included in the randomized trials were 69 years⁵ and 74 years,⁶ thereby limiting the amount of available data to guide evidence-based decisions for older men. Data from a longitudinal cohort study suggests that men age 75 years or older with a PSA less than 3.0 ng/mL have a negligible risk of prostate cancer–specific death, thereby suggesting screening can safely be excluded in these men.⁹ On the basis of expert opinion recommendation from the American Urological Association, PSA screening in men with less than a 10-year life expectancy, either because of age or comorbidity, is discouraged.⁷ Similarly, the ACS recommends that asymptomatic men who have at least a 10-year life expectancy have an opportunity to make an informed decision.⁸ In distinct contrast, the United States Preventive Services Task Force recently suggested an age-based cutoff to stop routine screening at age 75 years.¹⁰

To additionally inform this important and common clinical situation, our primary objective was to estimate the proportion of contemporary civilian US men age 70 years and older who underwent PSA-based prostate cancer screening. Our secondary objective was to estimate predictors of screening and determine whether estimated life expectancy was associated with screening patterns.

METHODS

Data Source and Participants

Data were extracted from the National Health Interview Survey (NHIS), a public access database containing responses from the noninstitutionalized civilian population of the United States. The NHIS is a cross-sectional household survey administered through a series of face-to-face interviews conducted by employees of the US Census Bureau. The cancer screening survey is administered every 5 years as a supplement to the standard NHIS, and response data from 2000 and 2005 were extracted. Because interviewees are randomly selected from the population, the chance of an individual being included in both survey years (2000 and 2005) is exceedingly unlikely.

Men age 70 years and older were included and separated into 5-year groups (70 to 74 years, 75 to 79 years, 80 to 84 years, and ≥ 85 years of age). Of the 3,305 men age 70 years and older participating in the NHIS survey (adult cancer module), 340 had a history of prostate cancer, 288 had missing values on status of PSA test use (eg, refused to answer, answer was not ascertained, did not know), and 54 had missing values on date of most recent PSA test. After excluding these participants, 2,623 men were included in the analysis, representing approximately 8.1 million men age 70 years and older in the United States in 2000 and 8.0 million men of that age in the United States in 2005. By using the same criteria, men age 40 to 69 years old (n = 11,984) were also included in the analyses as a comparative control.

Data Measurement

In both the 2000 and 2005 surveys, male respondents older than age 40 years who had received a PSA test were asked the main reason for their most recent PSA test. Response categories included part of a routine exam, because of a problem, or some other reason. In the 2000 survey, men could also answer follow-up test for an earlier exam or family history. Our primary outcome, PSA screening test, was defined as receipt of a PSA test as part of a routine exam within the year before the survey.

Our primary factor of interest was life expectancy. We measured 5-year life expectancy by using a validated mortality index based on NHIS data that included age, sex, smoking status, body mass index (BMI), comorbidities, recent hospitalizations, self-perceived health, and functional status measures. The index categorized 5-year mortality risk as high (≥ 48% risk of mortality), intermediate (16% to 47%), or low (≤ 15%).¹¹

We also prespecified and evaluated other factors thought to be associated with PSA screening among older men. These factors were race, ethnicity, place of birth, educational attainment, marital status, health insurance (eg, Medicare, private insurance), visits with a generalist or specialist within the previous year, a family history of any cancer, a family history of prostate cancer, utilization of colorectal cancer screening, vitamin use, alcohol use, and activity level.

Statistical Analyses

After accounting for sampling weight and study design (strata and primary sampling units^11a), PSA screening rate was estimated according to age group, survey year, category of life expectancy, and other potential predictors. We also estimated the total number of men receiving a PSA test according to age group and category of life expectancy after taking account of sampling weight. Logistic regression for survey data was used to compare PSA screening rate by survey year. Because there was no substantial difference in the pattern of screening rate by age group between the two surveys, we pooled data from year 2000 and 2005 surveys to increase efficiency. Sensitivity analysis was done for each survey year, and the results were similar. Logistic regressions were also used to examine predictors of PSA testing among men age 70 years or older. Univariate analyses were done first, followed by age-adjusted analyses, as age was the strongest predictor for PSA screening. All variables found to be significant in age-adjusted analyses were included in a multivariate analysis. Because age is a component of the estimated life expectancy, we adjusted for age to interpret the association between life expectancy and PSA screening on the basis of other component parts of life expectancy. Multicollinearity can cause predictors to compete and make the selection of important variables arbitrary in a stepwise selection procedure. Therefore, we checked and found no important collinearity among variables that were significant in age-adjusted analysis. Then, we used backward stepwise logistic regression procedure to find independent predictors. To understand the overall influence of life expectancy as well as the influences of individual components of life expectancy, we present two final models: one model included estimated life expectancy, and the other model included component parts of the life expectancy calculation. Stata statistical software version 11.0 (Stata, College Station, TX) was used to complete all the analyses, and a P value less than .05 was considered statistically significant.

RESULTS

PSA-Based Prostate Cancer Screening Rates

The overall PSA screening rate in men age 40 years and older was 23.7% in 2000 and was 26.0% in 2005 in the United States. However, it varied dramatically by age (P < .001; Fig 1). The PSA screening rate was lowest in the 40- to 44-year-old age group, and the rate increased with age until a plateau occurred at ages 65 to 79 years (44.6% in ages 65 to 69 years, 45.5% in ages 70 to 74 years, and 44.2% in ages 75 to 79 years). Then, the rate decreased to 24.6% in the ≥ 85-year-old age group. The pattern of screening rate by age group was similar between 2000 and 2005 (P = .32), so we combined data from 2000 and 2005 to increase precision of estimates in the subsequent analyses. An estimated 1.5 million men age 80 years or older reported a screening PSA test within the prior year (Table 1).

Fig 1.

Estimated prevalence of prostate-specific antigen (PSA) screening (with 95% CIs) within the past year by year and age, National Health Interview Survey 2000 and 2005.

Table 1.

Estimated Use of PSA Test for Routine Exam in the Past Year by Age and Estimated Life Expectancy, National Health Interview Survey 2000 and 2005

Characteristic	High Life Expectancy*	Intermediate Life Expectancy†	Low Life Expectancy2021	Total
Age range, years
70-74
Estimated % tested	47.1	44.8	22.3	45.5
95% CI	43.1 to 51.1	38.5 to 51.2	7.0 to 37.6	42.2 to 48.7
Estimated No. tested	2,101,000	881,000	59,000	3,041,000
75-79
Estimated % tested	47.3	41.1	45.4	44.2
95% CI	41.1 to 53.5	35.6 to 46.5	34.2 to 56.6	40.4 to 48.0
Estimated No. tested	953,000	916,000	283,000	2,151,000
80-84
Estimated % tested	49.2	35.2	32.1	35.7
95% CI	33.3 to 65.0	29.7 to 40.6	23.1 to 41.1	31.3 to 40.1
Estimated No. tested	137,000	700,000	241,000	1,078,000
≥ 85
Estimated % tested		28.3	21.8	24.6
95% CI	NA	19.9 to 36.8	15.1 to 28.5	19.3 to 30.0
Estimated No. tested		195,000	194,000	388,000
Total % tested	47.3	39.2	30.7
95% CI	44.0 to 50.6	35.9 to 42.4	25.8 to 35.6
Estimated No. of men tested in United States	3,190,000	2,691,000	777,000

Abbreviations: PSA, prostate-specific antigen; NA, not applicable.

^*Defined as ≤ 15% probability of 5-year mortality.

^†Defined as 16% to 47% probability of 5-year mortality.

^‡Defined as ≥ 48% probability of 5-year mortality.

Among men age 70 years and older, screening rates were associated with estimated 5-year life expectancy: 47.3% of men with high life expectancies (≤ 15% probability of 5-year mortality), 39.2% with intermediate life expectancies (16% to 47%), and 30.7% with low life expectancies (≥ 48%; P < .001; Table 1). The estimated total number of men with ≥ 48% probability of 5-year mortality who received PSA testing was 777,000. No material differences in screening rates were observed between men with low and intermediate life expectancies in each age group of 75 to 79 years, 80 to 84 years, or ≥ 85 years.

After analysis was adjusted for age, estimated life expectancy remained associated with PSA screening, although the strength of association was attenuated. This was expected, because age is a component of predicting life expectancy (P = .034; Table 2). Higher rates of PSA screening in men age 70 years or older were also associated with the following: younger age, higher estimated life expectancy, non-Hispanic white ethnicity, place of birth, socioeconomic status, frequency of physician visits, insurance status, lifestyle choices, and higher functional status measures (Table 2).

Table 2.

Age-Adjusted Analysis of Predictors for PSA-Based Prostate Cancer Screening Among Men Age 70 Years and Older, National Health Interview Survey 2000 and 2005

Characteristic	No. of Participants Interviewed	% of Participants Tested*	Age-Adjusted Odds Ratio	95% CI	P
Age, years
70-74	1,070	45.5	1	Ref	< .001
75-79	756	44.2	0.95	0.78 to 1.16
80-84	507	35.7	0.67	0.53 to 0.84
≥ 85	290	24.6	0.39	0.29 to 0.54
Life expectancy
High	1,063	47.3	1	Ref	.034
Intermediate	1,123	39.2	0.81	0.65 to 1.01
Low	437	30.7	0.66	0.48 to 0.91
Ethnicity
Non-Hispanic white	2,070	43.4	1	Ref	< .001
Non-Hispanic black	237	25.4	0.43	0.30 to 0.61
Hispanic	237	30.0	0.54	0.38 to 0.77
Asian	55	34.5	0.62	0.30 to 1.27
Other	22	46.5	1.12	0.45 to 2.78
Born outside the United States
No	2,343	42.3	1.00	Ref	.004
Yes	279	31.3	0.62	0.45 to 0.85
Education
Less than high school	848	29.3	1	Ref	< .001
High school graduate	748	42.1	1.66	1.31 to 2.11
Some college	447	47.4	2.08	1.59 to 2.72
College graduate	547	51.3	2.39	1.84 to 3.10
Marital status
Single	1,020	30.3	1	Ref	< .001
Married/living with partner	1,599	44.8	1.72	1.43 to 2.08
Family history of cancer
No	1,190	36.0	1	Ref	< .001
Other cancer	1,140	45.3	1.48	1.22 to 1.78
Prostate cance	144	53.3	2.01	1.36 to 2.96
Seen GP
No GP in past year	450	20.0	1	Ref	< .001
GP in past year	2,169	45.3	3.42	2.56 to 4.56
Seen specialist
No specialist in past year	1,417	34.9	1	Ref	< .001
Specialist in past year	1,203	48.0	1.69	1.40 to 2.04
Medicare coverage
No coverage	145	29.7	1	Ref	.02
Coverage	2,471	41.9	1.79	1.15 to 2.79
Private insurance coverage
No coverage	1,034	34.1	1	Ref	< .001
Coverage	1,580	45.2	1.58	1.32 to 1.90
Colorectal cancer screening
Never	1,219	29.1	1	Ref	< .001
Ever	935	47.0	2.12	1.72 to 2.62
Screened in past year	406	63.0	4.12	3.19 to 5.32
Moderate activity participation
None	1,472	35.1	1	Ref	< .001
Participated	1,100	48.7	1.67	1.39 to 2.02
Alcohol drinking
Lifetime abstinence	527	29.2	1	Ref	< .001
Former use	791	39.0	1.53	1.16 to 2.00
Current use	1,276	47.4	2.11	1.65 to 2.70
Vitamin use
No vitamins/minerals	1,178	32.4	1	Ref	< .001
Takes vitamins/minerals	1,427	47.9	1.87	1.56 to 2.24
BMI, kg/m2
< 25	1,031	37.8	1	Ref	.27
25-29.9	1,145	43.5	1.18	0.96 to 1.44
≥ 30	408	44.4	1.14	0.87 to 1.48
Smoking
Never smoked	913	45.5	1	Ref	.007
Former smoker	1,451	50.9	1.17	0.96 to 1.41
Current smoker	253	38.8	0.71	0.49 to 1.02
COPD
No COPD	2,344	41.3	1	Ref	.79
COPD diagnosed	274	40.6	0.96	0.72 to 1.28
Diabetes
No diabetes	2,131	39.8	1	Ref	.039
Diabetes diagnosed	434	47.5	1.29	1.01 to 1.65
Cancer history
No cancer	2,366	41.3	1	Ref	.74
Previous cancer diagnosis	255	40.5	0.95	0.71 to 1.27
Hospitalizations in past year
None	2,098	41.2	1	Ref	.42
One overnight	370	43.3	1.10	0.87 to 1.40
Multiple	152	36.0	0.83	0.57 to 1.22
Self-perceived health
Excellent/very good	1,005	47.1	1	Ref	< .001
Good	919	39.2	0.74	0.61 to 0.91
Fair/poor	692	35.5	0.63	0.50 to 0.80
Daily activities
Does not need help	2,362	42.8	1	Ref	< .001
Needs help	259	23.9	0.48	0.34 to 0.67
Difficulty walking several blocks
No difficulty walking	1,600	44.6	1	Ref	.003
Difficulty walking	1,019	35.8	0.75	0.62 to 0.91

Abbreviations: PSA, prostate-specific antigen; Ref, reference value; GP, general practitioner; BMI, body mass index; COPD, chronic obstructive pulmonary disease.

^*Population estimates were based on National Health Interview Survey sampling design.

In multivariate analysis that adjusted for several predictors of PSA screening, estimated life expectancy maintained its significant association (Table 3; model 1). However, only two factors in the calculation of life expectancy were associated with lower PSA screening rates: older age and help in daily activities (Table 3; model 2). We also found several other independent predictors for higher PSA screening rates: higher education, married or living with a partner, family history of cancer (especially prostate cancer), seeing a general practitioner in prior year, Medicare coverage, recent colorectal cancer screening, participation in moderate physical activity, current or former alcohol use, and vitamin use. As a sensitivity analysis, we fit models by using data from 2000 and 2005 separately, and the findings were consistent (data not shown).

Table 3.

Multivariate Analyses of Predictors for PSA-Based Prostate Cancer Screening Among Men Age 70 Years and Older, National Health Interview Survey 2000 and 2005

Characteristic	Model 1			Model 2
	OR	95% CI	P	OR	95% CI	P
Age, years
70-74				1	Ref	< .001
75-79				0.95	0.75 to 1.20
80-84				0.80	0.60 to 1.05
≥ 85				0.49	0.34 to 0.70
Need help in daily activities
Yes v no				0.56	0.37 to 0.84	.005
Life expectancy
High	1.00	Ref	< .001
Intermediate	0.76	0.61 to 0.94
Low	0.53	0.39 to 0.74
Education
Less than high school	1.00	Ref	.003	1.00	Ref	.003
High school graduate	1.51	1.16 to 1.96		1.46	1.12 to 1.89
Some college	1.65	1.22 to 2.24		1.65	1.22 to 2.22
College graduate	1.56	1.15 to 2.10		1.59	1.18 to 2.14
Marital status
Single	1.00	Ref	.001	1.00	Ref	.005
Married/living with partner	1.44	1.16 to 1.77		1.36	1.10 to 1.67
Family history of cancer
No	1.00	Ref	.027	1.00	Ref	.049
Other cancer	1.27	1.03 to 1.57		1.25	1.01 to 1.55
Prostate cancer	1.58	1.01 to 2.47		1.50	0.96 to 2.34
Seen GP last year
Yes v no	2.46	1.76 to 3.46	< .001	2.45	1.75 to 3.44	< .001
Medicare coverage
Yes v no	1.67	1.02 to 2.75	.043	1.68	1.02 to 2.78	.041
Colorectal cancer screening
Never	1.00	Ref	< .001	1.00	Ref	< .001
Ever	1.53	1.22 to 1.93		1.51	1.20 to 1.90
Past year	2.95	2.20 to 3.96		2.87	2.13 to 3.86
Moderate activity participation
Yes v no	1.27	1.02 to 1.58	.033	1.31	1.05 to 1.63	.018
Alcohol drinking
Lifetime abstinence	1.00	Ref	.016	1.00	Ref	.022
Former use	1.38	1.03 to 1.85		1.33	0.99 to 1.78
Current use	1.50	1.14 to 1.99		1.49	1.12 to 1.97
Vitamin use
Yes v no	1.39	1.13 to 1.71	.002	1.41	1.15 to 1.74	.001

NOTE. Nonsignificant results are not shown. Model 1 includes estimated life expectancy and other significant predictors; model 2 includes component parts of estimated life expectancy calculation and other significant predictors.

Abbreviations: PSA, prostate-specific antigen; OR, odds ratio; Ref, reference value; GP, general practitioner.

DISCUSSION

Estimated life expectancy and absolute age have both been proposed as criteria for considering prostate cancer screening, although virtually all consensus guidelines favor an estimated life expectancy of at least 10 years.^7,12–14 By using a contemporary, population-based, US survey, our study showed that PSA-based prostate cancer screening patterns are discordant with recommended guidelines and that PSA screening is utilized at an unnecessarily high rate with modest consideration of competing medical risks and estimated longevity. Our findings suggest that approximately three quarters of a million men with an estimated life expectancy of approximately 5 years received PSA screening in the year prior. These men are unlikely to reap a meaningful benefit.

There are multiple consequences of inordinate PSA screening. By ordering a PSA test, the physician has presumably chosen to recommend a prostate biopsy if the result is considered abnormally elevated. A prostate biopsy is an invasive procedure that can be associated with pain, stress while awaiting the results,¹⁵ and a 4% risk of serious infection or bleeding requiring hospital readmission.¹⁶ Second, of men who undergo biopsy, a significant minority (approximately 25% to 40%) will be diagnosed with prostate cancer, which leads to anxiety,¹⁷ decision about a management strategy, and exposure to potential treatment-related morbidity.¹⁸ Last, the PSA test, clinic visits, staging, treatment, and post-treatment follow-up cumulatively add individual and societal health care expenses.¹⁹ When considering whether to obtain a PSA, a patient should be informed of these potential risks, intended benefits, and associated implications.

Age was a strong predictor of PSA screening but not in the expected pattern, as men who were age 70 to 79 years underwent PSA screening at nearly twice the rate of men age 50 to 54 years. A recent randomized trial demonstrated a modest overall cancer-specific survival advantage for screened patients, a finding limited to men younger than age 70 years.⁶ On the basis of our observations and of those in the randomized screening trial, the men most likely to benefit from PSA screening are paradoxically being screened at markedly lower rates than men highly unlikely to benefit.

Accurately evaluating life expectancy is critical when considering management options in elderly patients considering prostate cancer screening. Age and comorbidities, as quantified by Charlson index scores or similar indices, are the strongest predictors of non–cancer-specific mortality,²⁰ and, intuitively, patients with a larger number or severity of comorbidities are less likely to benefit from screening. Modestly accurate tools are available to predict life expectancy in patients with localized prostate cancer (concordance index, 0.70 to 0.84) but are clinically underutilized.^21–25 Physicians are notoriously poor at estimating life expectancy,²⁶ which may explain the elevated screening rates of elderly men who are unlikely to survive another 5 years. Patients and physicians may also be uncomfortable discussing estimated life expectancy, although life expectancy is important when considering treatment decisions and clinical outcomes.²⁷ This reluctance may contribute to a tendency to excessively screen elderly patients who have limited life expectancies.

Physician confidence in the benefits of cancer screening may contribute to overutilization of PSA-based screening, as the majority of male physicians older than age 50 years self-reported having had a PSA test for the purposes of screening.²⁸ Elevated screening rates of elderly men with limited life expectancies may also reflect defensive medicine in response to previous litigation stemming from the decision not to offer prostate cancer screening.^29–31

The discordance between PSA screening recommendations and practice patterns has also been identified within the Veterans Affairs (VA) medical system. In 2003, 56% of men ≥ 70 years old received a PSA test, and 36% of men ≥ 85 years old received a PSA test.³² Similar to our findings, marital status, socioeconomic status, and ethnicity were associated with screening rates. Although overall health status in the VA study, as measured by the Charlson score, was independently associated with screening, it did not markedly alter screening rates within each age category.

The Behavioral Risk Factor Surveillance System, a population-based telephone survey in the United States sampling 30,000 men older than 50 years in 2001, found that 57% reported a PSA test within the past year.³³ Approximately two thirds (69%) of men age 70 to 79 years and greater than half (56%) of men age 80 years or older reported a PSA test within the past year. Similar to our findings, increasing age was directly associated with rates of PSA screening. Our estimates from 2005 are substantially lower compared with those of the Behavioral Risk Factor Surveillance System survey in 2001.³³ These differences may stem from an encouraging modification of practice patterns or may reflect different methodologies. However, because we did not observe any temporal changes in screening patterns between 2000 and 2005 within the NHIS, we strongly favor different methodologies as the likely explanation.

The major limitation of our study is the NHIS database is de-identified and not linked to claims or laboratory data. Therefore, verification of the self-reported responses is not possible. For the following reasons, we do not suspect formal verification would markedly alter our findings. First, self-reporting inaccuracies for PSA screening within the NHIS are predominantly under reports.³⁴ Second, in self-reported data from three health maintenance organizations, there were equal amounts of PSA over reporting and under reporting compared with medical record extraction, leading to the overall self-report percentage being within 2% of the percentage identified from medical record extraction.³⁵ Therefore, despite not having verification of PSA testing in our study, it is likely our data are good estimates. Another limitation to be considered is the mortality index utilized¹¹ was constructed and validated on NHIS data from 1997 to 2000 and has not been validated on 2005 data or in a clinical setting.

In conclusion, age and 5-year estimated life expectancy were strong predictors of PSA screening but are not considered in a significant number of men. Excessive screening for prostate cancer in elderly men who have limited life expectancies in the United States results in unnecessary anxiety, diagnoses, overtreatment, treatment-related morbidity, and health care expenditures without meaningful clinical benefit. The merits and limitations of PSA should be discussed with patients considering prostate cancer screening, particularly in older men and in those with short estimated life expectancies.

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Although all authors completed the disclosure declaration, the following author(s) indicated a financial or other interest that is relevant to the subject matter under consideration in this article. Certain relationships marked with a “U” are those for which no compensation was received; those relationships marked with a “C” were compensated. For a detailed description of the disclosure categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.

Employment or Leadership Position: None Consultant or Advisory Role: None Stock Ownership: None Honoraria: Scott E. Eggener, Centocor Ortho Biotech, Astra Zeneca Research Funding: Scott E. Eggener, Visualase, Inc Expert Testimony: None Other Remuneration: None

AUTHOR CONTRIBUTIONS

Conception and design: Dezheng Huo, Scott E. Eggener

Financial support: Scott E. Eggener

Administrative support: Scott E. Eggener

Collection and assembly of data: Michael W. Drazer, Dezheng Huo, Aria Razmaria

Data analysis and interpretation: Michael W. Drazer, Dezheng Huo, Mara A. Schonberg, Aria Razmaria, Scott E. Eggener

Manuscript writing: all authors

Final approval of manuscript: all authors