A population-based cross-sectional study of colorectal cancer screening practices of first-degree relatives of colorectal cancer patients.

0.41 MB·English

by  Courtney, Ryan J

Courtneyetal.BMCCancer2013,13:13
http://www.biomedcentral.com/1471-2407/13/13
RESEARCH ARTICLE Open Access
A population-based cross-sectional study of
colorectal cancer screening practices of first-
degree relatives of colorectal cancer patients
Ryan J Courtney1,4,7*, Christine L Paul1,4, Mariko L Carey1,4, Robert W Sanson-Fisher1,4, Finlay A Macrae2,
Catherine D’Este3, David Hill5, Daniel Barker3 and Jody Simmons6
Abstract
Background: The aim of this study was to determine the proportions and predictors of first-degree relatives (FDRs)
ofcolorectalcancer (CRC) patients (i) ever receiving any CRC testing and (ii)receivingCRC screening in accordance
withCRC screeningguidelines.
Methods: Colorectal cancer patients and their FDRs were recruited through the population-based Victorian Cancer
Registry, Victoria, Australia. Seven hundred and seven FDRs completedtelephone interviews. Of these, 405FDRs
were deemed asymptomatic and eligible for analysis.
Results: Sixty-nine percent of FDRs had ever received any CRC testing. First-degree relatives ofolder age, those
withprivatehealth insurance, siblings and FDRs who had ever been asked about family history of CRCby a doctor
were significantly more likely than theircounterparts to have ever received CRCtesting. Twenty-five percentof
FDRs “ator slightly above average risk”were adherent to CRC screening guidelines. For this group, adherenceto
guideline-recommended screening was significantly more likely to occur for male FDRs and thosewith a higher
levelof education. For persons at“moderately increased risk” and “potentially high risk”, 47%and 49% respectively
adhered to CRC screening guidelines. For this group, guideline-recommended screeningwas significantly more
likely to occurfor FDRs who were living in metropolitan areas, siblings, those married or partnered and thoseever
asked about family history of CRC.
Conclusions: A significant levelof non-compliancewith screening guidelineswas evident among FDRs. Improved
CRCscreening inaccordancewith guidelinesand effective systematic interventions to increase screening rates
among population groups experiencing inequality are needed.
Trial Registration: Australian and New Zealand Clinical Trial Registry: ACTRN12609000628246
Keywords: Colorectal cancer, Screening, Prevention, Early detection, Family history
Background randomised controlled trials have demonstrated that
Worldwide, colorectal cancer (CRC) isdiagnosed in over CRC mortality can be reduced by 15% to 33% through
one million persons annually and is the fourth leading Faecal Occult Blood Test (FOBT) screening, [3-6] with
cause of cancer death [1]. Staging of disease at diagnosis fewer advanced CRCs detected, compared with patients
is a critical factor affecting survival. When discovered presenting with symptoms, in population-based screen-
early, CRC is highly treatable, with a relative five-year ing [7]. Although the use of colonoscopy to detect
survival rate of 90% for localised CRC [2]. Several right-sided CRCs is under debate, [8,9] case control and
cohort studies of colonoscopy screening suggest a CRC
*Correspondence:[email protected] mortality reduction ranging from 60% to 76% [10] and
1PriorityResearchCentreforHealthBehaviour,SchoolofMedicineand
incidencereductionof76%to90%[11].
PublicHealth,FacultyofHealth,UniversityofNewcastle,Callaghan,Australia
4HunterMedicalResearchInstituteUniversityofNewcastle,Callaghan, Approximately 15% to 25% of persons who develop
Australia CRC will have a first-degree relative (FDR), i.e. a parent,
Fulllistofauthorinformationisavailableattheendofthearticle
©2013Courtneyetal.;licenseeBioMedCentralLtd.ThisisanOpenAccessarticledistributedunderthetermsoftheCreative
CommonsAttributionLicense(http://creativecommons.org/licenses/by/2.0),whichpermitsunrestricteduse,distribution,and
reproductioninanymedium,providedtheoriginalworkisproperlycited.
Courtneyetal.BMCCancer2013,13:13 Page2of11
http://www.biomedcentral.com/1471-2407/13/13
sibling or child, also affected by the disease [12,13]. Per- screening participation in terms of published screening
sons with one FDR diagnosed under the age of 55 years guidelines and across level of risk [28-31]. Further, rela-
or with two FDRs diagnosed at any age have a three- to tively little is known about the factors associated with
six-fold increased risk of developing CRC [14]. The rela- FDRs of CRC patients’ guideline-recommended screen-
tive risk of developing CRC is further increased where a ing compliance [21]. The aim of this study was to exam-
known genetic mutation has been identified [15]. For ine among FDRs of persons diagnosed with CRC and at
persons where a known genetic mutation has been iden- each level of risk (“at or slightly above average risk”,
tified, both earlier onset of CRC and much higher risk “moderately increased risk” and “potentially high risk”),
are apparent. Given the increased risk imposed on FDRs the proportions (i) ever receiving any CRC testing in
of CRC patients, screening for CRC assumes major im- their lifetime and (ii) screened in accordance with
portance. Screening strategies targeting FDRs of affected AustralianCRCscreening guidelines.
cases could contribute to the prevention or early detec- The individual- and provider-level factors associated
tion of 15% to 20% of CRCs [16,17]. Healthcare author- with FDRs ever receiving CRC testing and guideline-
ities and professional societies have published guidelines recommendedscreeningwerealso evaluated.
for the appropriate screening of FDRs of persons
affected with CRC [18-20]. International approaches to Methods
risk classification vary slightly, but all follow the same Settinganddesign
pattern, with risk level determined by the number and Index cases (i.e. persons diagnosed with CRC) and their
type of relatives diagnosed (i.e. first- or second-degree), FDRs were recruited through the population-based
the age at diagnosis and the presence of other high-risk Victorian Cancer Registry (VCR), Victoria, Australia be-
features, i.e. mutation status for cancer-predisposing tween 2009 and 2011. Human research ethics approval
genes if present in the family [18-20]. Screening guide- was obtained from The University of Newcastle and The
lines for persons at higher risk generally recommend CancerCouncilVictoria.
additional types of testing (e.g. colonoscopy rather than,
or in addition to, FOBT), more frequent testing and Procedure
commencement of testing at an earlier age, compared Indexcases(ICs)aged18yearsorolder,within9months
with their average risk counterparts [18-20]. Australian of CRC diagnosis, registered with the VCR and English-
National Health and Medical Research Council speaking were eligible to participate in this study. The
(NHMRC) guidelines recommend that asymptomatic VCR checkedtheFamilialAdenomatous Polyposis (FAP)
persons “at or slightly above average risk” commence Registerto excludepersons with FAP.TheVCR wrote to
screening at the age of 50 years and receive FOBT the clinicians of eligible patients to advise them of the
screening every two years or consider sigmoidoscopy study and request consent to approach the index cases.
(preferably flexible) every five years [18]. In Australia, Clinicians were asked to notify theVCR if there was any
[18] contrary to other international guidelines, [19,20] reason why a person should not be invited to participate
colonoscopy screening is endorsed only for asymptom- in the study. Patients for whom the treating clinicians
atic persons at higher levels of risk (i.e. “moderately did now allow consent to patient approach from the
increased risk” or “potentially high risk”). For persons at VCR were not contacted. The VCR wrote to the
“moderately increased risk”, colonoscopy is endorsed remaining patients seeking permission to release their
every five years starting at age 50 years or at ten years contact details to the research team. Index cases who
earlier than first diagnosis in the family, whichever agreed to provision of their contact details to the
comes first [18]. Endoscopy screening for persons at researchers were contacted by the research team via
“potentially high risk” is recommended at least on afive- mail and asked to participate in the study. To accommo-
yearly basis in the Australian guidelines. However, age at date index case preferences about the mode of approach
screening commencement, test type and repeat testing to their relative, consenting index cases provided the
intervalaredependentonthetypeoffamily-specificmu- details of FDRs aged 18 years or older for the purposes
tation identified[18]. of contacting them via either (i) the research team (with
Despite the elevated risk associated with having a fam- their permission) who sent a study invitation by post on
ily history of CRC, the little available evidence suggests their behalf or (ii) a study invitation mailed to the pa-
that adherence to recommended screening for FDRs of tient who passed this invitation on to the relative(s).
CRC patients is low [18]. While FDRs of CRC patients Eligibility criteria for FDRs’ participation were (1) Eng-
are more likely to be screened, compared with those lish-speaking,aged18years orolder,and(2)noprevious
without a family history of CRC, [21-23] screening com- history of advanced adenoma, CRC, ulcerative colitis,
pliance for this group is sub-optimal, at between 21% Crohn’s disease, inflammatory bowel disease or FAP.
and 78% [16,21,24-27]. Scant literature exists related to First-degree relatives meeting these criteria were eligible
Courtneyetal.BMCCancer2013,13:13 Page3of11
http://www.biomedcentral.com/1471-2407/13/13
to complete the baseline telephone interviews. Assess- risk” or “potentially high risk” (colonoscopy every five
ment of first-degree relatives CRC screening behaviour years). Over-screening was not assessed in this study as
occurred at approximately nine to twelve months post information was only obtained on the most recent test-
index cases diagnosis, however, this could have fluctu- ing for each test type. CRC screening undertaken
ated dependent on the time taken to recruit the index before index case CRC diagnoses was included in
case or their first-degree relatives into the study. They the analysis. Associations between ever-tested and
were classified as asymptomatic and eligible for CRC guideline-recommended screening were explored for
screening if they had not undertaken FOBT, sigmoidos- the following items: socio-demographic characteristics
copy or colonoscopy due to a symptom episode in the (i.e. age, gender, education, marital status, Australian
previous five years. A diagrammatic representation of born, employment situation, private health insurance);
therecruitment protocolispresented inFigure1. geographical location (Accessibility/Remoteness Index
of Australia); relationship to index case (parent, child,
Quantifyingriskbasedonfamilyhistoryofcolorectal sibling); quality of life (Euro-Qol EQ-5D, VAS score);
cancer [32] worry about bowel cancer (Worried/Not wor-
Index cases were asked about family history of CRC, in- ried); and ever asked about family history of bowel
cluding all first- and second-degree relatives and their cancer by doctor/health professional (Yes/No). Logis-
ages at diagnosis, if relevant. Index cases’ages at diagno- tic regression modelling in a generalised estimation
sis were obtained from VCR data. The FDRs of index equation framework was used to adjust for multiple
cases were allocated to a level of risk in accordance with FDRs within families. Simple associations were exam-
screeningguidelines(SeeTable 1). ined first before all covariates were added to a mul-
tiple logistic regression model. Variables with p-value
Colorectalcancerscreeninghistory <.10 were retained in the final model. Parents of
FDRs were asked separately whether they had ever index cases were excluded from multiple regression
undertaken any of the following CRC tests: FOBT; sig- models in all analyses due to the small number of
moidoscopy; or colonoscopy. Respondents indicating parents recruited in the sample.
“Yes” to any of these tests were asked to specify how
long ago their most recent test was undertaken and the Results
reason for the test, to establish whether the respondent Of the 748 index cases interviewed, 98% had living
wasasymptomatic atthe timeoftesting. FDRs. Of the index cases with living FDR(s), 88% gave
consent for the research team to send a study invitation
Eligibilityforscreening to at least one FDR. A total of 2376 study invitations
Asymptomatic FDR respondents “at or slightly above were sent to FDRs, with 707 (30%) FDRs participating in
averagerisk”wereeligibleforscreeningiftheywereaged the study. A total of 405 FDRs were deemed asymptom-
50 years or older. For respondents at “moderately atic and eligible for analysis. The proportions of asymp-
increased risk”, in accordance with guidelines, eligibility tomatic FDRs recruited per family were as follows: 56%
forCRCscreeningwasdeterminedonthebasisof“start- (107) of families had one FDR recruited; 23% (43) had
ing at age 50 years or at an age 10 years younger than two FDRs recruited; 15% (29) had three FDRs recruited;
the age of first diagnosis of bowel cancer in the family, and 6.3% (12) had more than three FDRs recruited.
whichever comes first”. Asymptomatic respondents at Table2describes the characteristics oftheasymptomatic
“potentially high risk” were eligible for screening if they study population.
were aged18yearsorolder.
Everreceivedcolorectalcancertesting
Statisticalanalysis Overall, 69% (278/405) of FDRs had ever received any
The proportions of FDRs ever receiving CRC testing (i.e. CRC testing (i.e. FOBT, sigmoidoscopy or colonoscopy)
FOBT, sigmoidoscopy or colonoscopy) overall and by in their lifetime. The proportions of FDRs ever under-
levelofrisk(i.e.“atorslightlyaboveaveragerisk”,“mod- taking CRC testing within each risk category wereas fol-
erately increased risk” and “potentially high risk”) were lows: “at or slightly above average risk” (70%, 116/166);
calculated by the number of FDRs reporting receiving “moderately increased risk” (70%, 30/43); and “poten-
any CRC test, divided by the total number of FDRs. The tially high risk” (67%, 131/195). No significant differ-
proportion of FDRs screened in accordance with Austra- ences in CRC testing across risk category were identified
lian screening guidelines [18] was assessed according to (χ2 = 0.3, df = 2, p = .86). Multiple logistic regression
level of risk as follows: “at or slightly above average risk” analyses (see Table 3) revealed that older FDRs, those
(FOBT every two years, consider sigmoidoscopy, prefer- with private health insurance, siblings of the index case
ably flexible, every five years); and “moderately increased and FDRs who had ever discussed family history of CRC
Courtneyetal.BMCCancer2013,13:13 Page4of11
http://www.biomedcentral.com/1471-2407/13/13
Figure1Flowchartrepresentingselectionandrecruitmentofasymptomaticfirst-degreerelativesofcolorectalcancerpatients.
Courtneyetal.BMCCancer2013,13:13 Page5of11
http://www.biomedcentral.com/1471-2407/13/13
Table1DescriptionofriskcategoriesandtheirrespectivescreeningrecommendationsinaccordancewithNational
HealthandMedicalResearchCouncilcolorectalcancerscreeningguidelines
Riskcategory Riskfeatures Screeningrecommendation
Atorslightlyaboveaveragerisk (cid:129)Nopersonalhistoryofbowelcancer FOBTeverysecondyearfromtheageof
50years.
(cid:129)Eithernocloserelativeswithbowelcanceroronefirst-
degreeorsecond-degreerelativewithbowelcancer Considersigmoidoscopy(preferablyflexible)
diagnosedatage55yearsorolder. everyfiveyears.
Moderatelyincreasedrisk (cid:129)Onefirst-degreerelativediagnosedbeforetheageof55 Colonoscopyeveryfiveyearsstartingatage50,
years(withoutpotentiallyhigh-riskfeatureslistedbelow),or oratanage10yearsyoungerthantheageof
(cid:129)Twofirst-degreerelativesoronefirst-andonesecond- firstdiagnosisofCRCinthefamily,whichever
comesfirst.
degreerelative(s)onthesamesideofthefamily
(withoutpotentiallyhigh-riskfeatureslistedbelow).
Potentiallyhighrisk (cid:129)Threeormorefirst-degreeoracombinationoffirst-degree Dependentonpresenceandtypeoffamilial
andsecond-degreerelativesonthesamesideofthefamily cancer.
diagnosedwithbowelcancer(suspectedHNPCC*),or
Atleastcolonoscopyevery5years.
(cid:129)Twoormorefirst-degreeorsecond-degreerelativesonthe Ageofscreeningcommencementdependent
samesideofthefamilydiagnosedwithbowelcancer, onfamilialcolorectalcancersyndrome
includinganyofthefollowinghigh-riskfeatures: identified***
-bowelcancerbeforetheageof50years
-multiplebowelcancersintheoneperson
-atleastonerelativewithcanceroftheendometrium,ovary,
stomach,smallbowel,renalpelvis,ureter,biliarytractorbrain
-atleastonefirst-degreerelativewithalargenumberof
adenomasthroughoutthelargebowel(suspectedFAP)**
-somebodyinthefamilyinwhomthepresenceofahigh-risk
mutationintheadenomatouspolyposiscoli(APC)geneor
oneofthemismatchrepair(MMR)geneshasbeenidentified.
*HNPCC:Hereditarynon-polyposiscolorectalcancerorLynch’sSyndrome.**FAP:FamilialAdenomatousPolyposis.***SeeguidelinesforsyndromespecificCRC
screeningrecommendation.
with a doctor were at significantly greater odds of ever First-degreerelativesat“moderatelyincreasedrisk”
receivingCRCtesting. Of the 43 respondents at “moderately increased risk”,
47% (20/43) were screened in accordance with NHMRC
Colorectalcancerscreeninginaccordancewithguideline screening recommendation (i.e. colonoscopy screening
recommendations every five years). Thirty percent (13/43) of respondents
First-degreerelatives“atorslightlyaboveaveragerisk” had never undertaken any CRC testing (see Figure 2). Of
Of the 166 FDRs “at or slightly above average risk”, 25% the remaining 23% (10/43), half (5/10) had undertaken
(42/166) were screened in accordance with NHMRC FOBTscreeningintheprevious twoyears.
screening guideline recommendation (See Figure 2). All
respondents screened in accordance with guidelines had First-degreerelativesat“potentiallyhighrisk”
undertaken FOBT screening. Thirty percent (50/166) of Of the 195 respondents at “potentially high risk”, 49%
persons “at or slightly above average risk” had never (95/195) were screened in accordance with NHMRC
undertaken any CRC testing in their lifetime. The screening recommendation (i.e. colonoscopy screening
remaining persons “at or slightly above average risk” every five years). Thirty-three percent (64/195) of
(45%, 74/166) had either undertaken FOBT screening respondents had never undertaken any CRC testing. The
outsidethe recommendedguideline timeframe (i.e.every remaining 18% (36/195) of respondents had undertaken
two years) or had undertaken colonoscopy screening, a CRC screening not in accordance with guideline recom-
test type not endorsed in screening guidelines for per- mendation (see Figure 2). Of these, 61% (22/36) had
sons “at or slightly above average risk”. For persons “at undertakenFOBTscreeningintheprevious twoyears.
or slightly above average risk”, the number of colonos-
copies resulting from positive FOBT was not obtainable, Factorsassociatedwithscreeninginaccordancewith
although the number of such cases is likely to be small, guidelines
as 59% (37/63) of respondents “at or slightly above aver- Multiple logistic regression models for factors associated
agerisk”whohadundertakencolonoscopyscreeninginthe with screening guidelines across level of risk are pre-
previous five years had previously never undertaken sented in Table 4. FDRs “at or slightly above average
guideline-recommendedFOBTorsigmoidoscopytesting. risk” with higher levels of education were significantly
Courtneyetal.BMCCancer2013,13:13 Page6of11
http://www.biomedcentral.com/1471-2407/13/13
Table2Characteristicsofasymptomaticfirst-degree Table3Multiplelogisticregressionanalysisoffactors
relativesofcolorectalcancerpatients(n=405) associatedwitheverreceivingcolorectalcancertesting
Characteristic n %* N(%)* OR(95%CI) pvalue
Riskcategory Privatehealthinsurance
Atorslightlyaboveaveragerisk 166 41 Yes 201(71) 2.05(1.26,3.33) 0.0039
Moderatelyincreasedrisk 43 11 No 77(63) 1
Potentiallyhighrisk 195 48 Relationshiptoindexcase
Sex Sibling 166(78) 2.19(1.32,3.63)
Male 165 41 Child 92(54) 1 0.0024
Female 240 59 Everaskedaboutfamily
historyofbowelcancerby
Highestlevelofeducation
doctor/healthprofessional
Universitydegree 164 41
Yes 125(85) 4.78(2.80,8.18) <.0001
TradeorTAFECertificate/Diploma 72 18
No 153(59) 1
Secondaryschoolingcompleted 44 11
Mean(SD) OR(95%CI) pvalue
Secondaryschoolingnotcompleted 123 31
Age
Maritalstatus
Yes 58.8(11.01) 1.04(1.02,1.07) 0.0002
Married/partnered 315 78
No 51.3(15.64) 1
Notpartnered 90 22
*Percentageofresponses(excludinganymissingvalues).
BorninAustralia
No 33 8
more likely to be screened in accordance with guideline
Yes 372 92 recommendations, compared with FDRs with lower
Employmentsituation levels of education. Further, male FDRs “at or slightly
Full-time 172 42 above average risk” were significantly more likely to be
Part-time 87 21 screened in accordance with guidelines, compared with
Notworking 114 28 female FDRs. Due to the small number of respondents
in the “moderately increased risk” group (n = 43), both
Other 32 8
“moderately increased risk” and “potentially high risk”
Privatehealthinsurance
groups were combined in the analysis of screening inac-
No 122 30
cordance with screening guidelines. Persons at “moder-
Yes 283 70 ately increased risk” and “potentially high risk” who
Geographicallocation(ARIA)
were married or partnered, living in a major city or
Majorcities(urban) 211 54 urban area, sibling of the index case and ever asked
Regional/remote 182 46 about family history of CRC by a doctor were signifi-
Relationshiptoindexcase cantly more likely to be screened in accordance with
Parent** 22 5 guidelinerecommendations.
Sibling 212 52
Discussion
Child 171 42
Rates of ever receiving CRC testing for FDRs of CRC
Worryaboutbowelcancer
patients were relatively high in the current study, with
Notworried 175 43
approximately 70%of FDRs across each level ofrisk ever
Worried 230 57
receiving CRC testing. The rate of FDRs ever receiving
Everaskedaboutfamilyhistoryofbowel
CRC testing was not found to be significantly higher
cancerbydoctor/healthprofessional
among population groups at higher relative risk of CRC
No 258 64
(i.e. persons at “moderately increased risk” and “poten-
Yes 147 36 tially high risk”, compared with persons “at or slightly
Mean SD above averagerisk”).InanAustraliancontext,ratesofpar-
Age(years) 56.5 13 ticipation in CRC testing or testing undertaken in
Qualityoflife(EQ-5DVASscore) 80.8 13 adherence to guidelines among FDRs of CRC have previ-
*Percentageofresponses(excludinganymissingvalues).**Parentsremoved ouslyonlybeenevaluatedintwootherstudies[28,31].
fromfurtherregressionanalysesduetosmallcellsize(n=22).
The current study highlighted low levels of CRC
screening in accordance with guideline recommenda-
tions across varying levels of risk: “at or slightly above
Courtneyetal.BMCCancer2013,13:13 Page7of11
http://www.biomedcentral.com/1471-2407/13/13
Figure2First-degreerelatives’colorectalcancerscreeningstatusinaccordancewithguidelinesbylevelofrisk.
averagerisk”(25%);“moderatelyincreasedrisk”(47%);and and timing of repeat testing vary across countries
“potentiallyhighrisk” (49%).Internationalcomparisonsof [18,33,34]. FewstudieshaveassessedCRCscreeninginac-
risk-appropriate screening in accordance with guideline cordancewithguidelinerecommendationsforpersonswith
recommendations are difficult to ascertain, given that an affected FDR with CRC [28-30]. This study, to our
healthcareauthorities’ endorsement of screeningmodality knowledge,isthefirstAustralianpopulation-basedexamin-
ation of CRC screening participation among FDRs of CRC
Table4Multiplelogisticregressionmodeloffactors
patients
associatedwithfirst-degreerelatives’screeningin
accordancewithguidelines
Screeningoffirst-degreerelatives“atorslightlyabove
“Atorslightlyaboveaveragerisk” N(%)* OR(95%CI) pvalue averagerisk”inaccordancewithguidelines
Gender
The low rate of screening in accordance with guideline
Male 27(35) 2.74(1.32,5.68) .0068 recommendation for FDRs “at or slightly above average
Female 15(17) 1 risk” identified in this study is comparable to that of the
Highestlevelofeducation
general population in Australia [35,36]. Two separate
Secondaryschoolingcompleted 3(19) 0.34(0.09,1.20) 0.0941
population-based evaluations among persons over
Secondaryschoolingnotcompleted 10(18) 0.41(0.18,0.91) 0.0288
50 years of age have indicated that 33% of respondents
TradeorTAFECertificate/Diploma 4(14) 0.21(0.07,0.65) 0.0070 at “average risk” had undertaken FOBTscreening in the
Universitydegree 25(38) 1
previous five years [36] and 18.4% in the previous two
“Moderatelyincreasedrisk”and N(%) OR(95%CI) pvalue
years [35]. Themost recent investigation of CRC screen-
“Potentiallyhighrisk”
ing participation among at-risk persons (i.e. those aged
Maritalstatus
over 55 years) since the National Bowel Cancer Screen-
Married/partnered 97(55) 3.68(1.72,7.88) 0.0008
ing Program’s introduction in 2006 indicated that
Notpartnered 18(30) 1
screening in accordance with guideline recommendation
Geographicallocation(ARIA)
for persons “at or slightly above average risk” was 20 per
Majorcities(urban) 68(56) 2.26(1.27,4.03) 0.0056
cent [24]. Although community-based studies in the
Regional/remote 41(39) 1
United States (US) have generally established rates of
Relationshiptoindexcase
FOBT screening in accordance with guideline recom-
Sibling 74(62) 5.15(2.28,11.67) 0.0001
mendations among FDRs of CRC patients of between
Child 29(30) 1
9% and 42%, [37-40] the studies with the highest screen-
Everaskedaboutfamilyhistoryofbowel
cancerbydoctor/healthprofessional ing rates had recruited participants through advertise-
Yes 65(71) 5.08(2.55,10.11) .0001 ments [38,39]. Such samples are unlikely to be
No 50(34) 1 representative of FDRs of CRC patients in general, and
*Percentageofresponses(excludinganymissingvalues). are likely to be biased as they include participants more
Courtneyetal.BMCCancer2013,13:13 Page8of11
http://www.biomedcentral.com/1471-2407/13/13
likely to engage in screening, [23] thus reducing the CRC screening among FDRs of CRC patients aged over
findings’ relevance for indicating population-based 40 years selected through the Alberta Cancer Registry
estimates. indicated that 60% were appropriately screened for CRC
screening (i.e. FOBT within one year, barium enema or
Screeningoffirst-degreerelativesatelevatedrisk sigmoidoscopy within five years, or colonoscopy within
(“moderatelyincreasedrisk”and“potentiallyhighrisk”) 10 years) [41]. In summary, the available evidence sug-
inaccordancewithguidelines gests that, worldwide, the screening rates of FDRs of
Current study findings indicated that 47% of FDRs of CRC patients in accordance with guideline recommen-
CRC patients at “moderately increased risk” and 49% of dations rarely exceed 50 percent, well short of rates
persons at “potentially high risk” were screened in ac- likely to be necessary for reducing CRC incidence and
cordance with the Australian guideline recommendation mortality onapopulation basis.
(i.e. colonoscopy screening every five years). This is
much higher than other Australian data on risk- Factorsassociatedwithcolorectalcancertestingand
appropriate CRC testing among FDRs, which had identi- adherencetocolorectalcancerscreeningguidelines
fied three FDRs among 225 were screened in accordance The present study identified a number of socio-
with Australian guidelines [28]. The most recent study demographic and provider-level factors impacting upon
in Victoria among first and second-degree relatives CRC testing and screening in accordance with guideline
selected from case and population control probands in recommendations. It is well-established that older age is
the Melbourne metropolitan area identified that 6% (70/ a consistent predictor of CRC screening [23,42,43]. The
1236) of persons at “moderately increased risk” were current study identified that the odds of receiving CRC
screened in accordance with guideline-recommendation testing increased with increasing age of FDRs. This find-
[31]. For relatives of probands at “potentially high risk” ing is consistent with previous literature among the FDR
1% (3/389) were adherent to guideline- recommended population [25,41,44]. For persons “at or slightly above
screening[31]. average risk”, adherence to screening guidelines was sig-
For the most part, the rate of screening in accordance nificantly more likely to occur for male compared with
with guideline recommendations for FDRs of CRC female FDRs. This is contrary to other studies of FDRs
patients at elevated levels of risk (i.e. “moderately that have, on the whole, largely indicated no association
increased risk” and “potentially high risk”) in the current betweengenderandscreening behaviour[23].
study is similar to rates identified in the general popula- Previous literature has indicated that having medical
tion [24,35]. The most recent Australian community- insurance is significantly correlated with CRC screening
based study of CRC screening participation found that adherence [36,43,45]. Consistent with this literature, the
among persons over 55 years of age, 45% of those at current study identified that the likelihood of ever re-
“moderately increased risk” and “potentially high risk” ceiving CRC testing was at significantly higher odds for
were screened in accordance with guideline CRC screen- FDRs with private health insurance. This suggests that
ing recommendation (i.e. colonoscopy screening every the costs of medical consultation or screening itself rep-
five years) [24]. Another population-based investigation resent significant barriers to CRC screening rates among
among persons aged over 50 years at “above average persons without private health insurance.
risk” indicated that 30% had undertaken endoscopy (col- The current study also found that siblings compared
onoscopy or sigmoidoscopy) in the previous five years with children of the index case were at significantly
[35]. Data available from a US population-based increased likelihood of ever receiving CRC testing and
National Health Interview Survey sample of persons 41- receiving CRC screening in accordance with screening
75 years of age with a FDR diagnosed with CRC indi- guideline recommendations for persons at “moderately
cated that 28% of FDRs had undertaken colonoscopy increased risk” and “potentially high risk”. Previous lit-
screening in the previous ten years [21]. For this study, eraturehaslargelyanalysedeitherCRCscreeningamong
direct compliance with guideline screening recommen- siblings only [29,46] or FDRs combined and not sepa-
dationwasnotascertained.Anotherstudyrecruitingsib- rated into type of FDRs (i.e. parent, child or sibling)
lings of CRC patients diagnosed younger than 56 years [37,41,47]. A previous Australian investigation found
of age through four cancer centres in the US indicated that being a sibling of a CRC patient, consistent with
that 57% ofsiblings older than 34 years were screened in current study findings, was significantly associated with
accordance with guidelines [29]. Study findings from an- increased likelihood of previous participation in CRC
other investigation of FDRs of CRC patients participat- testing [28]. Further, a multi-centre nation-wide study in
ing in a free FOBT screening program indicated that Spain identified a higher rate of adherence to colonos-
22% of respondents were screened in accordance with copy screening for siblings and children, compared with
guidelines [30]. In Canada, the most recent evaluation of parents, when offered screening by a gastroenterologist
Courtneyetal.BMCCancer2013,13:13 Page9of11
http://www.biomedcentral.com/1471-2407/13/13
[26]. Current data suggest that screening compliance history information, automation of familial risk stratifi-
may be lower among CRC patients’ children who are cation and risk-appropriate screening advice, [56,57]
deemed eligible for screening in accordance with screen- should be considered in the primary and secondary
ing guidelines compared to their siblings. There is a healthcaresystems.
pressing need to ensure equality in CRC screening up- Despite the evidence that physician recommendation to
take across at-risk FDRs e.g. children, siblings and screen has a powerful motivating effect on CRC screening
parents. uptake, [58] FDRs are most often not informed by physi-
Thisstudyalsofoundthatpersons“atorslightlyabove ciansabouttheneedforCRCscreening[59].Asignificantly
averagerisk”withahigherlevelofeducationwereat sig- higher rate of colonoscopy screening can result among
nificantly increased odds of receiving screening in ac- those siblings where the index cases were aware of their
cordance with guideline recommendations. Previous FDRs’increasedrisk[27].ImprovementsinCRCscreening
findings related to CRC testing and educational attain- among the higher-risk populations (i.e. persons at “moder-
ment for persons with a family history of CRC are atelyincreasedrisk”and“potentiallyhighrisk”)relyonphy-
mixed, with some studies indicating a positive trend sicians’ active involvement in discussions with index cases
[29,30,48] and others indicating no association between and their FDRs surrounding their families’ increased risk
CRC testing and level of education [28,38,47]. For per- andneedforCRCscreening.
sons at “moderately increased risk” and “potentially high
risk”, screening according to guideline recommendations Strengthsandlimitationsofthestudy
was significantly more likely to occur for FDRs who Ininterpreting studyfindings,severallimitationsshouldbe
were married or partnered, compared with those not considered. The response rates among index cases and
partnered. For the most part, marital status has been FDRswerelow.Giventhatpeopleinterestedintheirhealth
identified as a significant factor influencing screening are more likely to participate in health research, it is likely
participation in the general population [43,49-51] and that this has resulted in an over-estimate rather than an
among FDRs of CRC patients, [21,47] with increased under-estimate of the true screening rates. It should be
screening compliance commonly found for married and noted that the index case response rate achieved in the
partneredpersons. current study is comparable to the only other Australian-
The role of geographical barriers to CRC screening based investigation of CRC screening among FDRs
was also evident in the current study, as screening in ac- that had adopted a Cancer Registry based recruitment
cordancewithguidelinerecommendations for personsat method[28].
“moderately increased risk” and “potentially high risk” Personal history of adenomatous polyps was not investi-
was significantly more likely to occur for persons resid- gatedinthisstudy,makingitdifficulttocalculateprecisely
ing in metropolitan areas, compared with regional or re- theproportionofpersons“atorslightlyaboveaveragerisk”
mote areas. This finding is plausible, given the high undertakingcolonoscopyscreening(i.e.inthepreviousfive
concentration of colonoscopy services in major hospitals years) not in accordance with guideline recommendation.
in large cities in Australia [52]. Future attention to Nonetheless, the available data are suggestive of significant
equality in access to CRC screening services for those at over-use of colonoscopy, given that a large proportion of
increased levels of risk (i.e. “moderately increased risk” persons “at or slightly above average risk” had received
and“potentiallyhighrisk”)isclearly required. recent colonoscopy without any other CRC screening
The FDRs of CRC patients in the current study who (i.e.FOBTorsigmoidoscopy)beforehand.
had ever been asked about their family history of CRC Family history and CRC screening behaviour were self-
by a doctor were also at significantly higher odds of ever reported, rather than objectively assessed. Nonetheless,
undertaking CRC testing and undertaking testing in ac- studies have indicated that self-reported family history of
cordancewithguidelinerecommendations for personsat CRC is moderately accurate [60,63]. The level of risk
“moderately increased risk” and “potentially high risk”). obtainedfor FDRswasallocated onthebasisofindexcase
This finding highlights the need for assessment and on- self-reported information about family history, thus allow-
going monitoring of family history of CRC within both ing only a potential estimate. It was not practical to obtain
the primary and specialist healthcare settings. Although a full family history from each FDR of CRC patients for
physicians have been found to follow appropriate guide- thisstudy.
line recommendations for CRC screening once increased
risk has been identified, the family history information Conclusions
gathered is often insufficient for risk stratification pur- In summary, the current study identified a significant
poses[53-55].Thewiderincorporationofrecentlydevel- level of under-screening among a high-risk population
oped cancer risk assessment tools, showing both and a substantial level of colonoscopy screening not in
feasibility and effectiveness in the collection of family accordance with screening guidelines for persons “at or
Courtneyetal.BMCCancer2013,13:13 Page10of11
http://www.biomedcentral.com/1471-2407/13/13
slightly above average risk”. There is an urgent need for 2. AlketruseSF,KosaryCL,KrapchoM,NeymanN,AminouR,WaldronW,Ruhl
enhanced physician and patient education about risk- J,HowladerN,TatalovichZ,ChoH,MariottoA,EisnerMP,LewieDR,Cronin
K,ChenHS,FeuerEJ,StinchcombDG,EdwardsBK:Nationalcancerinstitute.
appropriate screening for FDRs of CRC patients, and for SurveillanceepidemiologyandEndresults.Cancerstatisticsreview.1975–2007.
further descriptive research to identify the barriers to http://seer.cancer.gov/csr/1975_2007/.
CRC screening among this population group. Effective 3. MandelJS,BondJH,ChurchTR,SnoverDC,BradleyGM,SchumanLM,
EdererF:Reducingmortalityfromcolorectalcancerbyscreeningfor
systematic interventions on a population basis are
fecaloccultblood.Minnesotacoloncancercontrolstudy.NEnglJMed
required to improve CRC screening participation of 1993,328:1365–1371.
FDRsofCRCpatients. 4. KronborgO,FengerC,OlsenJ,JorgensenOD,SondergaardO:Randomised
studyofscreeningforcolorectalcancerwithfaecaloccultbloodtest.
Lancet1996,348:1467–1471.
Ethicalapproval 5. HardcastleJD,ChamberlainJO,RobinsonMH,MossSM,AmarSS,Balfour
This study was approved by the University of Newcastle TW,JamesPD,ManghamCM:Randomisedcontrolledtrialoffecaloccult
bloodscreeningforcolorectalcancer.Lancet1996,348:1472–1477.
(No 2008-0047) and Cancer Council Victoria (No 0810)
6. MandelJS,ChurchTR,BondJH,EdererF,GeisserMS,MonginSJ,SnoverDC,
ethics committee, and all participants provided written SchumanLM:Theeffectoffecaloccult-bloodscreeningontheincidence
consent. ofcolorectalcancer.NEnglJMed2000,343:1603–1607.
7. AnandaSS,McLauglinSJ,ChenF,HayesIP,HunterAA,SkinnerIJ,SteelMC,
JonesIT,HastieIA,RiegerNA,SheddaS,CompstonDJ,GibbsP:Initial
Competinginterests
impactofAustralia'snationalbowelcancerscreeningprogram.MedJ
AllauthorshavecompletedtheUnifiedCompetingInterestformatwww.
Aust2009,191:378–381.
icmje.org/coi_disclosure.pdf(availableonrequestfromRJC)anddeclarethat;
8. BrennerH,Chang-ClaudeJ,SeilerCM,RickertA,HoffmeisterM:Protectionfrom
(2)[RJC,CLP,RWSF,MLC,FAM,CD,DH,DBJS]havenorelationshipswith
colorectalcanceraftercolonoscopy:apopulationbasedcasecontrolstudy.
anycompaniesthatmighthaveaninterestinthesubmittedworkinthe
AnnIntenMed2011,154:22–30.
previous3years;(3)theirspouses,partners,orchildrenhavenofinancial
9. RansohoffDF:Howmuchdoescolonoscopyreducecoloncancer
relationshipsthatmayberelevanttothesubmittedwork;and(4)[RJC,CLP,
mortality?AnnInternMed2009,150:50–52.
RWSF,MLC,FAM,CD,DH,DBJS]havenonon-financialintereststhatmaybe
10. CrespiM,StiglianoV,AssisiD:Currenttrendsandsecondaryprevention
relevanttothesubmittedwork.
ofcolorectalcancer.Hepatogastroenterology2001,48:1635–1640.
ConflictofInterestDisclosure:Nopotentialconflictsofinterest.
11. WinawerSJ,ZauberAG,HoMN,O'BrienMJ,GottliebLS,SternbergSS,Waye
JD,SchapiroM,BondJH,PanishJF:Preventionofcolorectalcancerby
Authors’contribution
colonoscopicpolypectomy.Thenationalpolypstudyworkgroup.NEngl
Allauthorsparticipatedintheacquisitionandanalysisofdataandcritical JMed1993,329:1977–1981.
revisionofthemanuscript,haveseenandapprovedthefinalversion,had
12. WinawerSJ,FletcherRH,MillerL,GodleeF,StolarMH,MulrowCD,Woolf
fullaccesstostudydata,andwerejointlyresponsibleforthedecisionto
SH,GlickSN,GaniatsTG,BondJH,RosenL,ZapkaJG,OlsenSJ,Giardiello
submitforpublication.RJCisguarantorforthepaper.
FM,SiskJE,VanAntwerpR,Borwn-DavisC,MarciniakDA,MayerRJ:
Grantsupport:ThisresearchwasfundedbyagrantfromtheNational
Colorectalcancerscreening:clinicalguidelinesandrationale.
HealthandMedicalResearchCouncil(ID:510776)andreceivedinfrastructure Gastroenterology1997,112:594–642.
supportfromtheHunterMedicalResearchInstitute.TheFundershadnorole
13. Australiancancernetworkandcommonwealthdepartmentofhealthand
inthedesign,implementation,analysisorpreparationofthefinal
ageing.Familialaspectsofbowelcancer:aguideforhealthprofessionals.
manuscript.Allresearcherswereindependentfromthefundingsource.
http://www.cancerscreening.gov.au/internet/screening/publishing.nsf/
content/1F35A75DC194E59CCA2574EB007F7532/$File/familial-guide.pdf.
Authors’information 14. StJohnDJB,McDermottFT,HopperJL,DebneyEA,JohnsonWR,Hughes
Datasharing:Allauthorshadfullaccesstodataandtakefullresponsibility ES:Cancerriskinrelativesofpatientswithcommoncolorectalcancer.
fortheintegrityandaccuracyofanalysis.ThedatasetisavailablefromRyan AnnInternMed1993,118:785–790.
[email protected]. 15. LindorNM,PetersenGM,HadleyDW,KinneyAY,MiesfeldtS,LuKH,Lynch
P,BurkeW,PressN:Recommendationsforthecareofindividualswith
Acknowledgements inheritedpredispositiontoLynchSyndrome.JAMA2006,
Wethankallcolorectalcancerpatientsandtheirfirst-degreerelativesfor 296:1507–1517.
participatinginthisresearch. 16. ParienteA,MilanC,LafonJ,FaivreJ:Colonoscopicscreeninginfirst-degree
relativesofpatientswith'sporadic'colorectalcancer:acase-controlstudy.
Authordetails Gastroenterology1998,115:7–12.
1PriorityResearchCentreforHealthBehaviour,SchoolofMedicineand 17. BoultronMC,FaivreJ,QuipourtV,SenesseP,MichaelsC:Familyhistoryof
PublicHealth,FacultyofHealth,UniversityofNewcastle,Callaghan,Australia. colorectaltumoursandimplicationsfortheadenoma-carcinomasequence:a
2DepartmentofColorectalMedicineandGenetics,TheRoyalMelbourne case-controlstudy.Gut1995,37:830–834.
Hospital,Melbourne,Australia.3TheCentreforClinicalEpidemiologyand 18. AustralianCancerNetworkColorectalCancerGuidelinesRevision
Biostatistics,FacultyofHealth,UniversityofNewcastle,Callaghan,Australia. Committee:Clinicalpracticeguidelinesfortheprevention,earlydetectionand
4HunterMedicalResearchInstituteUniversityofNewcastle,Callaghan, managementofcolorectalcancer.http://www.nhmrc.gov.au/guidelines/
Australia.5ProfessorialFellow,UniversityofMelbourne,Melbourne,Australia. publications/cp106.
6CentreforBehaviouralResearchinCancer,CancerCouncilVictoria, 19. AmericanCancerSociety:Americancancersocietyrecommendationsforcolorectal
Melbourne,Australia.7PriorityResearchCentreforHealthBehaviour, cancerearlydetection.http://www.cancer.org/Cancer/ColonandRectumCancer/
UniversityofNewcastle,Room269a,Level2,DavidMaddisonBuilding, MoreInformation/ColonandRectumCancerEarlyDetection/colorectal-cancer-early-
Callaghan2308,Australia. detection-acs-recommendations.
20. UnitedStatesPreventiveServicesTaskForce:Screeningforcolorectal
Received:15August2012Accepted:27December2012 cancer:U.S.Preventiveservicestaskforcerecommendationstatement.
Published:10January2013 AnnIntenMed2008,149:627–637.
21. LongacreAV,CramerLD,GrossCP:Screeningcolonoscopyuseamong
References individualsathighercolorectalcancerrisk.JClinGastroenterol2006,40:490–496.
1. BoyleP,LevinB:Internationalagencyforresearchoncancer.Worldhealth 22. RamjiF,CotterchioM,MannoM,RabeneckL,GallingerS:Associationbetween
organisation.Worldcancerreport.2008.http://www.iarc.fr/en/publications/ subjectsfactorsandcolorectalcancerscreeningparticipationinOntario,
pdfs-online/wcr/. Canada.CancerDetectPrev2005,29:221–226.