behavioral associations between prostate and colon cancer screening

8
Behavioral Associations Between Prostate and Colon Cancer Screening Ruth C Carlos, MD, MS, Willie Underwood III, MD, MS, A Mark Fendrick, MD, Steven J Bernstein, MD, MPH BACKGROUND: Despite the controversy over the utility of routine prostate-specific antigen (PSA) screening in reducing prostate cancer–specific mortality, it has gained widespread use throughout the United States. Although colorectal cancer (CRC) screening reduces CRC mortality and appears to be cost effective, CRC screening adherence is suboptimal.To better understand screening behav- iors among men, the Behavioral Risk Factors Surveillance Survey was used to identify potential relationships that would allow interventions to enhance CRC screening. STUDY DESIGN: We included in our analysis 22,304 men, 50 years and older, who participated in the 2002 Behavioral Risk Factors Surveillance Survey. Chi-square and multivariate analyses were per- formed to determine the independent correlates of adherence to CRC screening. Independent variables evaluated were age, race, educational level, employment status, income, health insur- ance, the presence of a personal physician, self-reported general health, current smoking status, and receiving a PSA test. RESULTS: Men were more adherent with PSA screening than CRC screening (50.4% versus 47.6%; p 0.002). In multivariate analysis, adherence to PSA screening (adjusted odds ratio [OR] 3.24, p 0.001) exerted the largest independent effect on CRC screening adherence. Other positive correlates of adherence to CRC screening were having health insurance (adjusted OR 1.39, p 0.01) and a personal physician (adjusted OR 2.01, p 0.01). Age predicted CRC screening with an inverse-U correlation. Failure to adhere to CRC screening was associated with self-reported good health (adjusted OR 0.87, p 0.01) and being a current smoker (adjusted OR 0.65, p 0.01). Even in men who were compliant with PSA testing, CRC screening remained suboptimal (65%). CONCLUSIONS: More men received PSA testing than CRC screening. Men who received PSA testing were more likely to adhere to CRC screening. Taken together, PSA testing may represent a “teachable moment” for a behavior-related intervention aimed at reducing the risk of colon cancer. Tar- geting men who already accept one form of cancer screening can potentially increase CRC screening adherence. (J Am Coll Surg 2005;200:216–223. © 2005 by the American College of Surgeons) In 2004, it is estimated that there were approximately 29,900 deaths from prostate cancer. 1 There is a great deal of controversy regarding the utility of routine prostate- specific antigen (PSA) screening in reducing prostate cancer–specific mortality. 2-9 Despite this controversy, PSA screening has gained widespread use in the United States. 7-9 In contrast, there is ample evidence demon- strating the effectiveness of colorectal cancer (CRC) screening in reducing colon cancer–specific mortal- ity. 10-16 Although the American Cancer Society predicts that CRC will cause approximately 57,000 deaths in 2004, 17 and despite evidence supporting the cost effec- tiveness of CRC screening, 18-19 CRC screening in men continues to trail that of PSA screening. 9 Heightened public awareness of the risk of prostate cancer contributes to the increased rate of prostate can- No competing interests declared. Supported in part by a National Institutes of Health/National Center for Research Resources grant K12 RR017607-01 (Carlos). Received September 2, 2004; Accepted October 23, 2004. From the Departments of Radiology (Carlos), Urology (Underwood), Inter- nal Medicine (Fendrick, Bernstein), and Health Management and Policy (Fendrick, Bernstein), University of Michigan, and the VA Center for Practice Management and Outcomes Research (Carlos, Underwood, Bernstein), VA Ann Arbor Health Care System, Ann Arbor, MI. Correspondence address: Ruth C Carlos, MD, MS, Department of Radiol- ogy, University of Michigan, 1500 E Medical Center Dr, Ann Arbor, MI 48109-0030. 216 © 2005 by the American College of Surgeons ISSN 1072-7515/05/$30.00 Published by Elsevier Inc. doi:10.1016/j.jamcollsurg.2004.10.015

Upload: ruth-c-carlos

Post on 05-Sep-2016

220 views

Category:

Documents


0 download

TRANSCRIPT

BP

RS

I2o

N

SR

RFn(MACo4

©P

ehavioral Associations Betweenrostate and Colon Cancer Screening

uth C Carlos, MD, MS, Willie Underwood III, MD, MS, A Mark Fendrick, MD,teven J Bernstein, MD, MPH

BACKGROUND: Despite the controversy over the utility of routine prostate-specific antigen (PSA) screening inreducing prostate cancer–specific mortality, it has gained widespread use throughout the UnitedStates. Although colorectal cancer (CRC) screening reduces CRC mortality and appears to becost effective, CRC screening adherence is suboptimal. To better understand screening behav-iors among men, the Behavioral Risk Factors Surveillance Survey was used to identify potentialrelationships that would allow interventions to enhance CRC screening.

STUDY DESIGN: We included in our analysis 22,304 men, 50 years and older, who participated in the 2002Behavioral Risk Factors Surveillance Survey. Chi-square and multivariate analyses were per-formed to determine the independent correlates of adherence to CRC screening. Independentvariables evaluated were age, race, educational level, employment status, income, health insur-ance, the presence of a personal physician, self-reported general health, current smoking status,and receiving a PSA test.

RESULTS: Men were more adherent with PSA screening than CRC screening (50.4% versus 47.6%;p � 0.002). In multivariate analysis, adherence to PSA screening (adjusted odds ratio [OR]3.24, p � 0.001) exerted the largest independent effect on CRC screening adherence. Otherpositive correlates of adherence to CRC screening were having health insurance (adjusted OR1.39, p � 0.01) and a personal physician (adjusted OR 2.01, p � 0.01). Age predicted CRCscreening with an inverse-U correlation. Failure to adhere to CRC screening was associated withself-reported good health (adjusted OR 0.87, p � 0.01) and being a current smoker (adjustedOR 0.65, p � 0.01). Even in men who were compliant with PSA testing, CRC screeningremained suboptimal (65%).

CONCLUSIONS: More men received PSA testing than CRC screening. Men who received PSA testing were morelikely to adhere to CRC screening. Taken together, PSA testing may represent a “teachablemoment” for a behavior-related intervention aimed at reducing the risk of colon cancer. Tar-geting men who already accept one form of cancer screening can potentially increase CRCscreening adherence. (J Am Coll Surg 2005;200:216–223. © 2005 by the American College of

Surgeons)

scPSssit2tc

n 2004, it is estimated that there were approximately9,900 deaths from prostate cancer.1 There is a great dealf controversy regarding the utility of routine prostate-

o competing interests declared.

upported in part by a National Institutes of Health/National Center foresearch Resources grant K12 RR017607-01 (Carlos).

eceived September 2, 2004; Accepted October 23, 2004.rom the Departments of Radiology (Carlos), Urology (Underwood), Inter-al Medicine (Fendrick, Bernstein), and Health Management and PolicyFendrick, Bernstein), University of Michigan, and the VA Center for Practice

anagement and Outcomes Research (Carlos, Underwood, Bernstein), VAnn Arbor Health Care System, Ann Arbor, MI.orrespondence address: Ruth C Carlos, MD, MS, Department of Radiol-gy, University of Michigan, 1500 E Medical Center Dr, Ann Arbor, MI

c8109-0030.

2162005 by the American College of Surgeons

ublished by Elsevier Inc.

pecific antigen (PSA) screening in reducing prostateancer–specific mortality.2-9 Despite this controversy,SA screening has gained widespread use in the Unitedtates.7-9 In contrast, there is ample evidence demon-trating the effectiveness of colorectal cancer (CRC)creening in reducing colon cancer–specific mortal-ty.10-16 Although the American Cancer Society predictshat CRC will cause approximately 57,000 deaths in004,17 and despite evidence supporting the cost effec-iveness of CRC screening,18-19 CRC screening in menontinues to trail that of PSA screening.9

Heightened public awareness of the risk of prostate

ancer contributes to the increased rate of prostate can-

ISSN 1072-7515/05/$30.00doi:10.1016/j.jamcollsurg.2004.10.015

cbst“cimSss

MST(rsbtaPSfnp

ACPbtsfwtutp

rmyebcldo

PPrsDowP

SPbTt

iw5cHHthnuul$

cirvfr(e

217Vol. 200, No. 2, February 2005 Carlos et al Prostate Versus Colon Cancer Screening

er screening despite the continued controversy over theenefits of and appropriate regimen for prostate cancercreening. Given its wide acceptance by men, the pros-ate cancer screening encounter represents a potentialteachable moment” (ie, unique opportunity)20 for edu-ating patients about the risk of colon cancer. Accord-ngly, to better understand screening behaviors among

en, data from the Behavioral Risk Factors Surveillanceurvey were analyzed to identify potential relation-hips that would allow interventions to enhance CRCcreening.

ETHODSurvey design and study populationhe Behavioral Risk Factors Surveillance Survey

BRFSS) is an annual telephone survey of a nationallyepresentative sample of noninstitutionalized adults de-igned to measure preventive health practices and riskehaviors in adults.21 Men 50 years and older who par-icipated in the BRFSS 2002 survey were included in thenalysis. This age group was chosen because both the USreventive Services Task Force and the American Cancerociety recommend that CRC screening begin at age 50or asymptomatic individuals of average risk who haveeither a family history of CRC nor other risk factorsredisposing to CRC.22,23

scertainment of cancer screeningolorectal cancer screeningarticipants were asked if they had ever had a fecal occultlood test, sigmoidoscopy, or colonoscopy. No ques-ions were asked about the use of barium enema for CRCcreening. Sigmoidoscopy and colonoscopy can be usedor both screening and diagnosis; the survey questionsere not asked in a manner that allowed the respondents

o identify the indication for the test. Those who hadndergone one of these tests were asked when they hadheir last test (possible responses ranged from within theast year to more than 5 years ago).

Abbreviations and Acronyms

BRFSS � Behavioral Risk Factors Surveillance SurveyCRC � colorectal cancerDRE � digital rectal examinationOR � odds ratioPSA � prostate-specific antigen

For colorectal cancer, the American Cancer Society h

ecommends fecal occult blood testing annually, sig-oidoscopy every 5 years, or colonoscopy every 10

ears in individuals 50 years and older.23 We consid-red the patient compliant if he had a fecal occultlood test within the last year or sigmoidoscopy orolonoscopy within the last 5 years. We grouped theatter two tests together because the survey did notifferentiate between them, and we used the shorterf the two screening periods.

rostate cancer screeningarticipants were asked if they had ever had a digitalectal exam (DRE) or a PSA test. Those who re-ponded yes were then asked when they had their lastRE or PSA test. After excluding men with a history

f prostate cancer, we considered men up-to-dateith prostate cancer screening if they had a DRE orSA test within the past 12 months.9

ociodemographic correlates of CRC screeningotential sociodemographic correlates of CRC screeningehavior were determined a priori and are shown inable 1. These were ascertained through telephone in-erview as part of the fixed component of the BRFSS.

Age, race, educational level, employment status, andncome were treated as categoric variables. Respondentsere divided into five 10-year age groups ranging from0 to 59 years to 90 years or older. Race categories in-luded: Caucasian non-Hispanic, black non-Hispanic,ispanic, multiracial non-Hispanic, and other non-ispanic. Educational level categories included: at-

ended elementary school or less, attended at least someigh school, and attended at least some college or tech-ical school. Employment status categories included:nemployed, currently employed, student, retired, andnable to work. Income was divided into four categories:

ess than $25,000, $25,000 to $49,999, $50,000 to74,999, and $75,000 or greater.Health insurance, the presence of a personal physi-

ian, self-reported general health, and current smok-ng status were treated as dichotomous variables. Self-eported general health was as an ordered categoricariable in the survey with excellent, very good, good,air, and poor categories. For this analysis, self-eported health was dichotomized into good healthcontaining excellent, very good, and good health cat-gories) and poor health (containing fair and poor

ealth categories).

STCssW

tctle

T

nA

R

I

E

E

HHS

CP

C gen.

218 Carlos et al Prostate Versus Colon Cancer Screening J Am Coll Surg

tatistical analysishe primary outcome was adherence to the Americanancer Society recommendations for colorectal cancer

creening. Prostate cancer screening adherence was as-essed as an independent correlate of CRC adherence.

able 1. Sociodemographic Characteristics of 22,304 Men

Sociodemographic characteristic

Ad

n

10,6ge, yMean (range) 64 (5050 to 59 3,860 to 69 3,570 to 79 2,580 to 89 6�90

aceCaucasian, non-Hispanic 9,2Black, non-Hispanic 5Hispanic 3Multiracial, non-Hispanic 1Other, non-Hispanic 3

ncome, $�25,000 2,825,000 to 49,999 3,650,000 to 74,999 1,7�75,000 2,4

ducational level achievedAttended elementary school or less 5Attended at least some high school 3,8Attended at least some college or technical school 6,2

mployment statusUnemployed 2Employed or self-employed 4,3StudentRetired 5,4Unable to work 6ave health insurance 10,1ave a personal doctor 9,9

elf-reported general healthPoor health 2,7Good health 7,9

urrently a smoker 2,2rostate cancer screening adherencePSA test alone 7,2DRE alone 7,3PSA test � DRE 5,9

RC, colorectal cancer; DRE, digital rectal exam; PSA, prostate-specific anti

e performed a multivariate analysis to assess each po- s

ential variable for its independent relationship withancer screening adherence. We used the variance infla-ion factor to determine if there was significant col-inearity between each of the variables, with particularmphasis on correlation between presence of health in-

Responded to Colon Cancer Screening Questionsnt to CRCeening

Nonadherent to CRCscreening

p Value% n %

11,615

61 (50–99)36.4 6,115 52.6 �0.00133.5 3,205 27.6 �0.00123.5 1,701 14.6 �0.0016.3 546 4.7 �0.0010.3 48 0.4 0.14

86.7 9,544 82.2 �0.0015.0 604 5.2 0.613.7 724 6.2 �0.0011.6 223 1.9 0.063.0 520 4.5 �0.001

26.7 3,981 34.3 �0.00134.1 3,940 33.9 0.7316.5 1,859 16.0 0.3122.7 1,835 15.8 �0.001

5.1 899 7.7 �0.00136.2 4,915 42.3 �0.00158.8 5,801 49.9 �0.001

2.2 460 4.0 �0.00140.4 6,216 53.5 �0.0010.2 28 0.2 0.99

50.8 4,082 35.1 �0.0016.4 829 7.1 0.02

95.4 10,123 87.2 �0.00193.3 9,316 80.2* �0.001

26.0 2,938 25.3 0.2674.0 8,677 74.7 0.2620.8 4,237 36.5* �0.001

68.1 3,986 34.3* �0.00170.0 3,739 32.2 �0.00155.7 2,626 22.6 �0.001

Whoherescr

89

–99)8886147031

6839966917

49496526

416682

36182629809476

751423

837250

urance, presence of a personal physician, and PSA test-

iscs7

RApiCaH$cpouCravab3

(n5mparaaa(iPscgpwmww(

MsAwe

T

A

R

I

E

E

HHS

CA

D*

219Vol. 200, No. 2, February 2005 Carlos et al Prostate Versus Colon Cancer Screening

ng. No significant collinearity was demonstrated, con-istent with independence of each of the variables. Weonsidered a p value of less than 0.05 as statisticallyignificant. All statistical analyses were done using Stata.0 (Stata Corp).

ESULTStotal of 22,617 men aged 50 and older who partici-

ated in the 2002 BRFSS had complete data and werencluded for analysis; 22,304 (98.6%) responded toRC screening items. The respondents, whose average

ge was 63 years, were predominantly white, non-ispanic with a median annual income of $25,000 to

49,999. The majority of men attended at least someollege or technical school. At least 90% of all partici-ants had health insurance. When using PSA test aloner DRE alone as a means of screening, more men werep-to-date with prostate cancer screening than withRC screening (61.8% versus 47.6%, p � 0.001). The

ate of prostate cancer screening adherence with PSA testlone was similar to adherence with DRE alone (50.4%ersus 49.4%, p � 0.05). If we considered men compli-nt with prostate cancer screening if they underwentoth DRE and PSA testing, then compliance dropped to8.7%.Men who were adherent to CRC screening were older

median age range 60 to 69 years) than men who wereot adherent to CRC screening (median age range 50 to9 years, p � 0.001) (Table 1). Although Caucasianen constituted the vast majority in both groups, His-

anics were proportionally less represented in the CRCdherent group (p � 0.001). More CRC-adherent meneported a minimum income of $75,000 (p � 0.001)nd achieved a higher educational level (having attendedt least some college); in comparison, more non-CRC-dherent men reported a maximum income of $25,000p � 0.001) and achieved a lower educational level (hav-ng attended at least some high school, p � 0.001).roportionally more men who were current with CRCcreening were retired compared with men who were noturrent (p � 0.001). More than 80% of men in bothroups reported having health insurance and a personalhysician. Men who were current with CRC screeningere less likely be smokers (p � 0.001). Two-thirds ofen who were current with CRC screening were currentith PSA testing; only one-third of men not currentith CRC screening reported adherence to PSA testing

p � 0.001). i

ultivariate correlates of colorectal cancercreening adherencege significantly predicted CRC screening adherenceith an inverse-U relationship (Table 2). That is, adher-

nce to CRC screening increased as the decade of age

able 2. Correlates of Colon Cancer Screening Adherence

Demographic VariablesAdjusted odds ratio

(95% confidence interval)

ge, y50 to 59 —60 to 69 1.43 (1.32–1.54)*70 to 79 1.60 (1.44–1.76)*80 to 89 1.48 (1.28–1.71)*�90 1.02 (0.63–0.65)*

aceCaucasian, non-Hispanic —Black, non-Hispanic 1.16 (1.01–1.32)*Hispanic 0.68 (0.59–0.78)*Multiracial, non-Hispanic 0.92 (0.74–1.15)Other, non-Hispanic 0.80 (0.68–0.93)*

ncome, $�25,000 —25,000 to 49,999 1.21 (1.12–1.31)*50,000 to 74,999 1.33 (1.20–1.48)*�75,000 1.87 (1.69–2.08)*

ducational level achievedAttended elementary school

or less —Attended at least some high

school 1.25 (1.10–1.42)*Attended at least some college

or technical school 1.53 (1.34–1.74)*mployment statusUnemployed —Employed or self-employed 0.81 (0.68–0.97)*Student 1.30 (0.70–2.41)Retired 1.23 (1.02–1.48)*Unable to work 1.34 (1.08–1.64)*ave health insurance 1.39 (1.24–1.57)*ave a personal doctor 2.01 (1.82–2.21)*

elf-reported general healthPoor health —Good health 0.87 (0.81–0.94)*

urrently a smoker 0.65 (0.61–0.70)*dherence to prostate screeningPSA test alone 3.24 (3.06–3.44)*DRE alone 3.82 (3.60–4.06)*PSA test � DRE 3.51 (3.30–3.73)*

RE, digital rectal exam; PSA, prostate-specific antigen.p � 0.05.

ncreased, peaking at ages 70 to 79 (adjusted OR 1.60,

pclcsoCwe�0sgsrssp

loOpjscsc

DItgasvsabwpDd

cAir

(OiihphiUfwgsssts

ceCicv

idtGaqmieof

dochtCtsab

220 Carlos et al Prostate Versus Colon Cancer Screening J Am Coll Surg

� 0.01); CRC screening adherence subsequently de-reased. As income and education level increased, theikelihood of CRC screening adherence increased. Afri-an Americans as a group were more likely than Cauca-ians to undergo CRC screening after controlling for allther variables; Hispanics were less likely to undergoRC screening. Men who were retired or unable to workere more likely to have undergone CRC screening. As

xpected, having health insurance (adjusted OR 1.39, p0.01) or a personal physician (adjusted OR 2.01, p �

.01) independently increased the likelihood of CRCcreening adherence. Interestingly, men in self-reportedood general health were less likely to adhere to CRCcreening guidelines (adjusted OR 0.87, p � 0.01). Cur-ent smokers were also less likely to adhere to CRCcreening guidelines compared with men who nevermoked or formerly smoked (adjusted OR 0.65,� 0.01).Adherence to prostate cancer screening exerted the

argest independent effect on CRC adherence regardlessf the method used for screening. PSA testing (adjustedR 3.24, p � 0.001), DRE (adjusted OR 3.82,� 0.001), and combined DRE and PSA testing (ad-

usted OR 3.51, p � 0.001) similarly affected CRCcreening. Two-thirds of men up-to-date with prostateancer screening recommendations adhered to CRCcreening guidelines, regardless of method of prostateancer screening.

ISCUSSIONn 2002, fewer male respondents to the BRFSS adheredo CRC screening guidelines than to PSA screeninguidelines when considering PSA testing or DRE alones screening tests, confirming the discrepancy in rates ofcreening for CRC compared with prostate cancer pre-iously identified.9 It is unlikely that the relative inva-iveness of DRE compared with PSA testing influenceddherence because the rates of prostate cancer screeningy either exam are remarkably similar (49.4% comparedith 50.4%). To identify the most patients who mayotentially have prostate cancer, one could perform bothRE and PSA testing but this would lead to an apparent

rop in adherence of 11%.We demonstrated that, in addition to demographic

haracteristics (age, income, and education; beingfrican American; and being retired or unable to work)

ncreasing the likelihood of CRC screening, variables

elated to access to care such as having health insurance f

adjusted OR 1.39) and a personal physician (adjustedR 2.01) also improved CRC screening adherence. Be-

ng Hispanic decreased the probability of CRC screen-ng adherence. Smoking correlated with decreased ad-erence to CRC screening in our study, consistent withrevious analyses.24,25 A general perception of goodealth also correlated with decreased adherence, mirror-

ng the experience of others, including a study in thenited Kingdom where nonadherence to completion of

ecal occult blood testing was associated with feelingell.26 Compliance with prostate cancer screening, re-ardless of the method used, strongly predicted CRCcreening compliance. Our results in a nationally repre-entative sample verify the results of a smaller statewideurvey conducted in Massachusetts, which also reportedhat PSA testing was the single greatest correlate of CRCcreening adherence.25

Despite the association of prostate cancer screeningompliance and CRC screening adherence, the differ-nce in overall screening rates for PSA testing versusRC testing does not reflect the evidence that strongly

ndicates the effectiveness of screening in reducing colonancer mortality27-30 compared with continued contro-ersy over the benefits of prostate cancer screening.9

Moul and colleagues31 and Linton and Hamdy32 havendependently highlighted the lack of convincing evi-ence that mass prostate cancer screening programs leado prolonged survival and improved quality of life.iven the substantial period of time in which most cur-

ble disease can be detected, Carter33 suggests less fre-uent prostate cancer screening, at 5-year intervals foren with a baseline PSA of �1 ng/mL and at 2-year

ntervals for a baseline PSA of 1–2 ng/mL. Despite thisvidence, 68% of men in a community setting and 55%f male practitioners believed that men should be testedor prostate cancer at least every 2 years.34

Prostate cancer screening by PSA testing has beenriven largely by patient demand.35,36 This may be sec-ndary to widespread concern, particularly among Afri-an Americans, that prostate cancer is a major publicealth issue.37 But with regard to cancer screening andhe reduction of cancer-specific mortality, althoughRC screening has a proved benefit in reducing colorec-

al cancer mortality, the benefits of prostate cancercreening are controversial.2-9 Taken together, not onlyre interventions to increase CRC screening warranted,ut a greater understanding of how society can benefit

rom the demand for prostate cancer screening is also

ntwbm

iiatvtdpnvtpiehwl“mticm

tthiurtrptlm

dsdeh

oecineaus

Ctdemtfqomtdtacttre

twwaoeweispt

LTpc

221Vol. 200, No. 2, February 2005 Carlos et al Prostate Versus Colon Cancer Screening

eeded. Given that one-third of men who undergo PSAesting are not adherent to CRC screening criteria, cane make use of the public’s demand for PSA testingy using the PSA testing encounter as a “teachableoment”?Many investigators advocate helping men make fully

nformed decisions regarding prostate cancer screen-ng,34 particularly in relationship to CRC screening.9 In

recently conducted metaanalysis reviewing interven-ions directed at improving colon cancer screening ser-ices, Stone and associates38 identified that organiza-ional change, such as establishment of a separate clinicevoted to screening and prevention activities, use of alanned care visit for prevention, and designation ofursing or clerical staff to identify patients needing pre-ention services or to arrange a physician visit, consis-ently increased use of cancer screening services, com-ared with physician-directed and patient-directednterventions (eg, financial incentives, reminders, andducation). These studies suggest that health systemsave substantial opportunities to enhance adherenceith these potentially life-saving tests. Stone and col-

eagues39 demonstrated that coupling education about aless effective” intervention while a patient underwent aore accepted one produced positive effects. So accep-

ance of a less desired, yet proved, intervention may bemproved if coupled with education around a better ac-epted one, an event often referred to as a “teachableoment.”38

Given that more men routinely undergo PSA testinghan they do CRC screening, we propose that the PSAesting be used as a “teachable moment” for deliveringealth interventions aimed at improving CRC screen-

ng.20 The teachable moment has been extensively eval-ated in smoking cessation interventions. A systematiceview by McBride and coworkers39 on smoking cessa-ion proposed office visits, notification of abnormal testesults, hospitalization, and disease diagnosis asotential teachable moments. They demonstrated rela-ively high cessation rates after teachable moments out-ined previously when compared with an untimed for-

al intervention.There are several possible reasons why an individual

oes not complete cancer screening. Inadequate under-tanding of the health benefits screening provides, nooctor recommendation, poor insurance coverage, andven a psychological dread of the specific procedures

ave been identified as potential factors that lead to sub- F

ptimal adherence with cancer screening. Educationalfforts alone, such as informational brochures, phonealls, or videos in a noncancer setting, only modestlymprove cancer screening in general. Given successes inoncancer clinical settings, we believe that providingducation during a screening “teachable moment” suchs PSA testing will likely increase the acceptance rates ofnderused screening tests, such as colon cancercreening.

Recognition of PSA testing as a venue for improvingRC screening adherence increases the value of PSA

esting itself. In general, diagnostic testing may haveirect and indirect benefits.40 Direct benefits typicallyncompass the clinical value, including contribution toortality reduction, and the economic value, including

he cost effectiveness of a diagnostic test. Indirect bene-its include the contribution of testing to health-relateduality of life, including reassurance, value as a predictorf future health behavior, and value as a “teachable mo-ent” for improving health behavior. In the case of PSA

esting, current evidence does not support significantirect clinical benefit from prostate cancer screening inhe general population when accounting solely forverted deaths from prostate cancer. But, if PSA testingan be used as a venue to improve CRC screening, thenhe value of PSA testing increases when its contributiono averted deaths from colon cancer are considered. Thiseduction in CRC mortality may even increase the costffectiveness of PSA testing programs.

We have additionally identified key differences be-ween men who adhered to CRC screening and menho did not. On average, men who were not compliantith CRC screening were younger, Hispanic, had

chieved a lower educational level or lower income level,r were current smokers. These differences suggest thatducational initiatives may be targeted toward patientsho possess these sociodemographic characteristics. For

xample, CRC screening educational outreach programsn predominantly Hispanic neighborhoods or CRCcreening programs coupled with smoking cessationrograms may further enhance the benefit of the educa-ional intervention.

imitationshere are several limitations to this analysis. First, tem-oral associations between each of the screening testsould not be determined, as the data are cross-sectional.

or example, we could not determine if PSA testing

ptcme

utuwuiC

IMdthlC“agsetcttmo

ASAA

D

CSS

R

1

1

1

1

1

1

1

1

1

1

2

222 Carlos et al Prostate Versus Colon Cancer Screening J Am Coll Surg

receded CRC screening. The direction of the associa-ion between PSA testing and CRC screening adherenceould not be determined. That is, PSA testing adherenceay have effectively increased CRC screening adher-

nce, or the opposite effect might have occurred.Second, the survey did not distinguish between the

se of sigmoidoscopy and colonoscopy for diagnosis andheir use for screening, but if some of these tests weresed for diagnosis, the actual rates of CRC screeningould be even lower than reported here. Finally, becausese of barium enema for CRC screening was not exam-

ned, we might have potentially underestimated overallRC screening adherence.

mplicationsen who were up-to-date with PSA testing recommen-

ations were more likely to get CRC screening thanhose who were not. For the one-third of men who ad-ere to PSA testing guidelines but not to CRC guide-

ines, PSA testing can be an opportunity for improvingRC screening adherence. PSA testing may represent a

teachable moment” for a behavior-related interventionimed at reducing risk of colon cancer. Targeting thisroup of men who already accept one form of cancercreening can potentially increase CRC screening adher-nce to levels approximating those of PSA testing. Givenhat gender-related barriers to CRC screening exist, spe-ifically delivering CRC screening educational informa-ion at a PSA testing visit may yield increased adherenceo CRC screening, particularly in the vast majority ofen who receive prostate cancer screening testing as part

f their routine care.

uthor Contributionstudy conception and design: Carlos, Fendrick, Bernsteincquisition of data: Carlosnalysis and interpretation of data: Carlos, Fendrick,Bernsteinrafting of manuscript: Carlos, Underwood, Fendrick,Bernstein

ritical revision: Carlos, Underwood, Fendrick, Bernsteintatistical expertise: Carlos, Bernsteinupervision: Carlos, Bernstein

EFERENCES

1. Jemal A, Tiwari RC, Murray T, et al. Cancer statistics, 2004. CA

Cancer J Clin 2004:54:8–29.

2. Coley CM, Barry MJ, Fleming C, et al. Early detection of pros-tate cancer. Part II: Estimating the risks, benefits, and costs.American College of Physicians. Ann Intern Med 1997:126:468–479.

3. Smith RA, Cokkinides V, Eyre HJ, American Cancer Society.American Cancer Society guidelines for the early detection ofcancer, 2004. CA Cancer J Clin 2004:54:41–52.

4. Carroll P, Coley C, McLeod D, et al. Prostate-specific antigenbest practice policy—part I: early detection and diagnosis ofprostate cancer. Urology 2001:57:217–224.

5. Hahn DL, Roberts R. PSA screening. J Fam Pract 1994:1:12.6. Labrie F. Screening and hormonal therapy of localized prostate

cancer shows major benefits on survival. Cancer J 2000; [Suppl2]:S182–187.

7. Fowler FJ Jr, Bin L, Collins MM, et al. Prostate cancer screeningand beliefs about treatment efficacy: a national survey of pri-mary care physicians and urologists. Am J Med 1998:104:526–532.

8. Ransohoff DF, McNaughton Collins M, Fowler FJ. Why is pros-tate cancer screening so common when the evidence is so uncer-tain? A system without negative feedback. Am J Med 2002:113:663–667.

9. Sirovich BE, Schwartz LM, Woloshin S. Screening men for pros-tate and colorectal cancer in the United States: does practicereflect the evidence? JAMA 2003:289:1414–1420.

0. Pignone M, Rich M, Teutsch SM, et al. Screening for colorectalcancer in adults at average risk: a summary of the evidence forthe U.S. Preventive Services Task Force. Ann Intern Med 2002:137:132–141.

1. Mandel JS, Bond JH, Church TR, et al. Reducing mortalityfrom colorectal cancer by screening for fecal occult blood. Min-nesota Colon Cancer Control Study. N Engl J Med 1993:328:1365–1371.

2. Kronborg O, Fenger C, Olsen J, et al. Randomised study ofscreening for colorectal cancer with faecal-occult-blood test.Lancet 1996:348:1467–1471.

3. Hardcastle JD, Chamberlain JO, Robinson MH, et al. Ran-domised controlled trial of faecal-occult-blood screening forcolorectal cancer. Lancet 1996:348:1472–1477.

4. Selby JV, Friedman GD, Quesenberry CP Jr, Weiss NS. A case-control study of screening sigmoidoscopy and mortality fromcolorectal cancer. N Engl J Med 1992:326:653–657.

5. Newcomb PA, Norfleet RG, Storer BE, et al. Screening sigmoid-oscopy and colorectal cancer mortality. J Natl Cancer Inst 1992:84:1572–1575.

6. Muller AD, Sonnenberg A. Protection by endoscopy againstdeath from colorectal cancer. A case-control study among veter-ans. Arch Intern Med 1995:155:1741–1748.

7. American Cancer Society. Cancer deaths to decline in 2004.January 2004. Available at: http://www.cancer.org/docroot/NWS/content/NWS 1 1x Cancer Deaths to Decline in 2004.asp. Accessed January 2004.

8. Pignone M, Saha S, Hoerger T, Mandelblatt J. Cost-effectiveness analyses of colorectal cancer screening: a systematicreview for the U.S. Preventive Services Task Force. Ann InternMed 2002:137:96–104.

9. Vijan S, Hwang EW, Hofer TP, Hayward RA. Which coloncancer screening test? A comparison of costs, effectiveness, andcompliance. Am J Med 2001:111:593–601.

0. Carlos RC, Fendrick AM. Improving cancer screening adher-

ence: using the “teachable moment” as a delivery setting for

2

2

2

2

2

2

2

2

2

3

3

3

3

3

3

3

3

3

3

4

223Vol. 200, No. 2, February 2005 Carlos et al Prostate Versus Colon Cancer Screening

educational interventions. Am J Managed Care 2004:10:247–248.

1. Centers for Disease Control and Prevention. Behavioral RiskFactors Surveillance System Web site. Available at: http://www.cdc.gov/brfss/. Accessed April 2003.

2. US Preventive Services Task Force. Screening for colorectalcancer. Recommendations and rationale. July 2002. Agencyfor Healthcare Research and Quality, Rockville, MD. Avail-able at: http://www.ahrq.gov/clinic/3rduspstf/colorectal/colorr.htm. Accessed December 2003.

3. Byers T, Levin B, Rothenberger D, et al. American Cancer So-ciety guidelines for screening and surveillance for early detectionof colorectal polyps and cancer: update 1997. American CancerSociety Detection and Treatment Advisory Group on ColorectalCancer. CA Cancer J Clin 1997:47:154–160.

4. Shapiro JA. Colorectal cancer-screening test and associatedhealth behaviors. Am J Prev Med 2001:21:132–137.

5. Lemon S, Zapka J, Puleo E, et al. Colorectal cancer screeningparticipation: comparisons with mammography and prostate-specific antigen screening. Am J Public Health 2001:91:1264–1272.

6. Hynam KA, Hart AR, Gay SP, et al. Screening for colorectalcancer: reasons for refusal of faecal occult blood testing in ageneral practice in England. J Epidemiol Community Health1995:49:84–86.

7. Trends in screening for colorectal cancer—United States, 1997and 1999. MMWR 2001:50:162–166.

8. Walsh JM, Posner SF, Perez-Stable EJ. Colon cancer screening inthe ambulatory setting. Prev Med 2002:35:209–228.

9. Perez-Stable EJ, Sabogal F, Otero-Sabogal R. Use of cancer-screening tests in the San Francisco Bay area: comparison of Latinos

and Anglos. J Natl Cancer Inst Monogr 1995;18:147–113.

0. Hiatt RA, Perez-Stable EJ, Quesenberry C Jr, et al. Agreementbetween self-reported early cancer detection practices and med-ical audits among Hispanic and non-Hispanic white health planmembers in northern California. Prev Med 1995:24:278–285.

1. Moul JW, Anderson J, Penson DF, et al. Early prostate cancer:prevention, treatment modalities, and quality of life issues. EurUrol 2003:44:283–293.

2. Linton KD, Hamdy FC. Early diagnosis and surgical manage-ment of prostate cancer. Cancer Treat Rev 2003:29:151–160.

3. Carter HB. Rationale for earlier and less frequent prostate can-cer screening. Urology 2001:58:639–641.

4. Livingston P, Cohen P, Frydenberg M, et al. Knowledge, atti-tudes and experience associated with testing for prostate cancer:a comparison between male doctors and men in the community.Intern Med J 2002:32:215–223.

5. Stuart ME, Handley MA, Thompson RS, et al. Clinical practiceand new technology: prostate-specific antigen (PSA). HMOPract 1992:6:5–11.

6. Vogt TM, Holis JF, Lichtenstein E, et al. The medical care sys-tem and previous: the need for a new paradigm. HMO Pract1998:12:5–13.

7. Wingo P, Tong T, Bolden S. Cancer statistics. CA Cancer J Clin1995:45:8–30.

8. Stone EG, Morton SC, Hulscher ME, et al. Intervents thatincrease use of adult immunization and cancer screening ser-vices: a meta-analysis. Ann Intern Med 2002:136:641–651.

9. McBride CM, Emmons KM, Lipkus IM. Understanding thepotential of teachable moments: the case of smoking cessation.Health Educ Res 2003;18:156–170.

0. Carlos RC. The added value of screening mammography in

improved screening for other cancers. JACR 2004:1:597–600.