Clin Genet 2000: 57: 48–55Printed in Ireland. All rights reser6ed

Original Article

Assessment of education and counsellingoffered by a familial colorectal cancer clinic

Collins V, Halliday J, Warren R, Williamson R. Assessment of educa-tion and counselling offered by a familial colorectal cancer clinic.Clin Genet 2000: 57: 48–55. © Munksgaard, 2000

We have evaluated whether or not client expectations, in terms ofeducation and information needs, have been met by a multi-disciplinaryfamilial colorectal cancer clinic. The study used a pre- and post-clinicquestionnaire design and 126 (84 women, 42 men) clients of the clinicparticipated. The most common reason for coming to the clinic is to‘find out whether there is a gene for colorectal cancer in the family’,followed by ‘to reduce risk for bowel cancer’ and ‘concern for chil-dren’s risk’. Clients would have preferred to receive more informationbefore attendance at the clinic to help with preparation. Informationgiven during the clinic increased knowledge of bowel cancer geneticsand had a positive effect on the accuracy of some clients’ perceptionsof their risk of developing cancer. In multivariate analysis, higher riskperceptions, higher education level and greater baseline knowledge pre-dicted post-clinic knowledge of bowel cancer genetics and an increasein knowledge. Client reports of the presence of a genetic counsellor ormedical geneticist in the session also predicted post-clinic knowledgeand an increase in knowledge. Most participants felt they receivedenough information during the clinic session on various aspects of fa-milial colorectal cancer, but the desire for more written information onprevention, including lifestyle actions, was expressed by many.

Veronica Collins,Jane Halliday,Rosemary Warren andRobert WilliamsonThe Murdoch Institute, Melbourne, Australia

Key words: clinic evaluation – familialcolorectal cancer – genetic counselling –information needs

Corresponding author: Veronica Collins,The Murdoch Institute, Royal Children’sHospital, Flemington Rd, Parkville 3052,Melbourne, Australia. Tel: +61 3 83416260; fax: +61 3 9348 1391

Received 8 October 1999, revised and ac-cepted for publication 10 November 1999

Colorectal cancers are amongst the most com-monly occurring cancers in both men and womenin Western countries (1). The hereditary forms ofthese cancers – familial adenomatous polyposis(FAP) and hereditary non-polyposis colorectalcancer (HNPCC) – make up between 2 and 10%of all colorectal cancers, and the HNPCC syn-drome also involves extra-colonic cancers includ-ing the uterus and ovaries.

Over the past 10 years, there has been increasinginterest in the contribution of genetics to the pre-vention and treatment of familial cancers, particu-larly those of the breast and colon (2, 3). Due tothe cloning of the APC gene for FAP and genespredisposing to HNPCC, it is now possible toperform DNA testing for mutations in high-riskfamilies (4, 5). Along with these developments,there has been a great deal of discussion about theneed for comprehensive cancer genetic serviceswith an emphasis on the necessity of providinggenetic counselling, particularly pre- and post-ge-netic testing (6–8).

Hereditary colorectal cancers are particularlysuited to dedicated familial cancer clinics for sev-eral reasons. The relevant genes have a high pene-trance (\90% for FAP and 85–90% lifetime riskof colorectal cancer in HNPCC) and they affectboth men and women. There is also potential forpreventive strategies to be implemented throughscreening and subsequent removal of premalignantpolyps (9, 10) or prophylactic colectomy (11),which can be targeted more appropriately if ge-netic information is available.

Many cancer clinics have now been establishedto serve the needs of individuals and families athigh risk of familial cancers. As genetic testing maynot be possible for individuals in some families(where a mutation is not identifiable), much of thecounselling involves risk assessment, appropriatesurveillance advice, referrals to other agencies andaddressing psychosocial issues. There have onlybeen a few studies reporting evaluations of familialcancer services, particularly the assessment of pa-tient needs and satisfaction with the service (12–

48

Assessment of a familial colorectal cancer clinic

14). Here we report an evaluation of a familialcolorectal cancer clinic focussing on client’s expec-tations before coming to the clinic, changes inknowledge of familial colorectal cancer and percep-tions of personal risk of developing cancer afterattendance at the clinic and satisfaction of theclient’s needs for information.

Subjects and methodsThe clinic

A family cancer clinic dedicated to the needs ofindividuals with colorectal cancer and their familiesis based at a major public hospital in Melbourne(Royal Melbourne Hospital). For the past 2.5 years,this clinic has been a part of a government-fundedcollaborative pilot familial bowel cancer projectbetween the gastroenterology department at thehospital, the state clinical genetics service (VCGS)and the Anti-Cancer Council of Victoria. The clinicservice provides detailed risk assessment, geneticcounselling, DNA testing if appropriate, and liaisonwith other cancer specialists and cancer registriesfor management and/or surveillance of patients andtheir families.

Referrals to the clinic come from GPs, otherfamily members, the VCGS and self-referral. Someclients had previously attended a hospital clinic thatprovided routine bowel cancer screening but noeducation nor counselling services. Most of thoseattending the family bowel cancer clinic are atmoderate to high risk of colorectal cancer, althougha proportion at lower risk come because of anxiety.

The clinic sessions generally last for 1 h and thereis always a gastroenterologist present. In addition,there are often either a medical geneticist or geneticcounsellor (or both) and a nurse or data managerwho has collected the pedigree information. At thetime of making appointments, clients are given aquestionnaire asking for details of their familyhistory and all details are checked for accuracy withcancer registries and hospital records before theclinic appointment. This allows most of the time atthe session to be spent discussing the implications ofthe family history rather than collecting the data atthat time. Those considered likely to be part ofHNPCC families are given information booklets onthis condition. All clients receive letters after theclinic outlining the issues discussed in the session.Based on family history, a risk category is desig-nated for all clinic attendees for the purposes offollow-up and screening recommendations. The cat-egories range from definite HNPCC familieswith the most stringent screening protocol to stan-dard population risk with routine populationscreening.

Study design

The clinic evaluation was done using a pre- andpost-clinic questionnaire design. The Royal Mel-bourne Hospital Research Committee granted ap-proval for the study. Pre-clinic questionnaires weresent out with appointment letters to all prospectiveclinic clients, along with a study information sheetemphasising that the evaluation study was separatefrom the clinical service and participation in thisstudy would not affect clinical care in any way. Ifthe initial questionnaire was not returned, this wasaccepted as a refusal to participate in the study andno follow-up was attempted. Approximately 3weeks after attendance at the clinic, all those whohad returned a pre-clinic questionnaire were sent apost-clinic questionnaire. If the post-clinic ques-tionnaire had not been returned within 4 weeks ofposting, a reminder letter was sent with anothercopy of the questionnaire.

In the pilot phase of the study, participants wereasked to comment on the questions and severaldrafts of the questionnaires were assessed. A finalversion of the questionnaire was agreed upon andused for the main phase of the evaluation, whichwas done between February 1998 and March 1999.

Response

Of 193 baseline questionnaires sent during thisperiod, 157 were completed (81%). A post-clinicquestionnaire was completed by 126 of these par-ticipants (80%) resulting in an overall response rateof 65%. These 126 individuals comprised the studypopulation. Information on age and sex was avail-able for 28 out of the 36 clinic attendees who didnot participate in the study at all. The sex distribu-tion was very similar to the main study group, butthe non-participants were slightly younger than thestudy group (29% of the non-participants were lessthan 35 years old). Only 12% of the non-partici-pants were definitely or likely to be members ofHNPCC families.

The non-responders who answered the pre-clinicbut not the post-clinic questionnaire (n=31)showed similar demographic characteristics to thestudy population, except that they tended to beyounger (40% of the non-responders were under 35years of age compared with 18% of the responders,p=0.03). Baseline perceptions of cancer risk werevery similar between the two groups (63% of non-responders perceived their risk to be higher ormuch higher than the general population com-pared with 61% of responders). However, 46% ofnon-responders to the post-clinic questionnairewere classified by the clinic as definitely or likely to

49

Collins et al.

be part of an HNPCC family (based on Amster-dam criteria) compared with 30% of those in thestudy group (p=0.10).

Questionnaires

Both questionnaires consisted of closed questionsallowing quantitative analysis. These were either‘yes’, ‘no’, ‘don’t know’ or scaled responses, wherethe respondent expresses their opinion by choosinga response on a scale from one extreme to theother, such as from strongly disagree to stronglyagree. Several questions also allowed open re-sponses and longer comments.

The questionnaires covered many areas, includ-ing both information aspects of the clinic service aswell as feelings about the clinic experience andworries and concerns about cancer. As this paperfocuses on the educative and information aspectsof the service, including risk perceptions, only rele-vant sections are outlined here. The baseline (pre-clinic) questionnaire included: demographic details– age, sex, education level (six categories rangingfrom 1) ‘not completed secondary’ to 6) ‘post-grad-uate’), marital status and number of children; pre-vious diagnosis of bowel cancer, other cancers orserious illnesses; questions on how individualsheard about the clinic; reasons for attending; andthe amount of information wanted about differentaspects of cancer genetics, surveillance andmanagement.

Both pre- and post-clinic questionnaires in-cluded questions about perception of risk of devel-oping bowel cancer compared to the generalpopulation for those who did not have a history ofbowel cancer. This involved choosing from thefollowing descriptions: ‘much higher’, ‘higher’,‘slightly higher’, ‘the same’, ‘slightly lower’,‘lower’, ‘much lower’ and ‘don’t know’. To assessknowledge of bowel cancer genetics, nine items(modified from other cancer studies (15)) weredeveloped and included in both questionnaires (seeAppendix for items). Each item required a ‘true’,‘false’ or ‘don’t know’ response. This measure hashigh internal consistency (coefficient a=0.73).

The post-clinic questionnaire asked about theorganisation of the clinic, the amount of informa-tion received, recommendations for follow-up orsurveillance and thoughts about the clinic letterand information booklet.

Statistical analysis

Data were entered into an SPSS for Windows(version 9.0) database and all analyses were doneusing SPSS. Pre- and post-clinic total knowledge

scores were calculated by adding the number ofcorrect responses to the nine knowledge items,giving a score between 0 and 9. Frequency datawere generated to examine demographic character-istics of study participants, reasons for attendingthe clinic and how participants heard about theclinic. To measure perceptions of the adequacy ofthe amount of information received at the clinic,those who indicated before the clinic that theywould like ‘very detailed’ information about vari-ous topics related to bowel cancer were selected,and the frequencies of amount of information re-ceived (‘none’, ‘not enough’, ‘enough’, ‘too much’)were listed for each item.

Paired t-tests were used to assess the significanceof differences in mean knowledge scores from pre-to post-clinic, and a t-test for independent groupswas used to compare mean knowledge scores be-tween men and women. To assess the independenteffect of age (continuous), sex (female/male), edu-cation level (six categories) and perceived risk level(higher-risk categories coded with higher numbers)on knowledge, multiple linear regression was per-formed with knowledge score as the dependentvariable. Separate models were calculated for pre-clinic knowledge score, post-clinic knowledge scoreand change in knowledge (post-clinic–pre-clinicscores). The effects of reading the HNPCC bookletand whether a geneticist or genetic counsellor wasreported to be present at the session were alsoincluded in the post-clinic knowledge and changein knowledge regression models.

Results

The demographic characteristics of the studygroup are shown in Table 1. About twice as manymen as women had a previous diagnosis of bowelcancer. When asked where they first heard aboutthe clinic, 25% of participants had heard fromfamily and 6% from friends, 42% had heard fromdoctors (GP or specialist) and 15% from otherclinics. Of all the possible reasons for coming tothe clinic, the one most frequently chosen as ‘veryimportant’ was ‘to find out if there is a gene forcancer in my family’ (75%) (Table 2). This wasfollowed by ‘to reduce my risk of getting bowelcancer’ (69%) and concern for children’s risk ofcancer (63%). In the post-clinic questionnaire,about half of the sample reported they were eligi-ble for DNA testing (n=62), and of these, 79%were interested in going ahead with it.

Table 3 shows the proportion of participantsindicating that they wanted ‘very detailed’ (ratherthan less detailed) information about various as-pects of bowel cancer, and their assessment of the

50

Assessment of a familial colorectal cancer clinic

Table 1. Characteristics of the study population

Men Women Total

42 (33%) 84 (67%) 126Number (% of total)

Age47.7 46.3 46.7Mean23–78 16–79Range 16–79

Marital status71.4 76.2Married/de facto (%) 74.616.7 7.1 10.3Have been married (%)11.9 16.7Single (%) 15.1

Education39.0 35.7Not completed high school (%) 36.829.3 31.0 30.4High school complete or post-high school (non-university) (%)31.7 33.3University graduate (%) 32.8

26.2 13.8 18.0Previous diagnosis of bowel cancer (%)

amount of information received. Most participantsindicated they wanted ‘very detailed’ informationabout each topic (56–81%) and the majoritythought they received ‘enough’ information.

Perceptions of the risk of developing bowel can-cer (for those without a previous diagnosis) beforeand after the clinic are presented in Fig. 1. Beforethe clinic, 24% of participants thought that theirrisk was ‘much higher’ than the general populationand 38% thought that their risk was ‘higher’. Afterthe clinic, risk perceptions had decreased for morethan a third of participants, reducing to 15 and25% those who thought their risk was ‘muchhigher’ and ‘higher’, respectively. It was not possi-ble to individually validate risk perception withrisk assessed by clinic staff, but the overall propor-tions in the study group could be compared. Forinstance, 14% of participants were definitely partof HNPCC families, which would correspond to‘much higher’ than population risk (see Fig. 1). Inaddition, 21% were assigned general populationrisk and 65% were somewhere between slightlyhigher than population risk but not as high asthose in HNPCC families.

The proportion of individuals correctly answer-ing the nine knowledge items (see Appendix)ranged from 23 (item 5) to 90% (item 8) before theclinic, and from 42 (item 3) to 98% (item 8) afterthe clinic. For each item, there was an increase inthe proportion of participants choosing the correctresponse after attendance at the clinic. Mean totalknowledge score increased significantly after atten-dance at the clinic for both men and women (pB0.001 for each sex), and women had higherknowledge scores than men after the clinic (p=0.016).

To assess the independent effects of variablesfound to be significantly associated with knowl-

edge score in bivariate analyses, multiple linearregression analyses were performed (Table 4). Ageand sex were not independently associated withknowledge score or change in knowledge. Educa-tion level was independently associated with bothpre- and post-clinic knowledge, but was lessstrongly associated with change in knowledge. Af-ter controlling for the other confounders, increas-ing post-clinic risk perception was positivelyassociated with increasing post-clinic knowledgeand with a positive change in knowledge score.Having read a booklet on HNPCC was only sig-nificantly associated with post-clinic knowledge

Table 2. Important reasons for coming to the clinic

Reason for coming to clinic Participants who chose‘very important’ for eachreason

%n

To find out if there is a gene for cancer in 91 74.6my family

To reduce my risk of getting bowel 82 68.9cancer

I am concerned about my children’s risk 62.876of cancer

To find out about screening for my 51.761children

To aid research/science 61 51.3To get the latest information on cancer 57 49.6To support other people with a family 48.358

history of cancerTo help with general planning for the 48 43.6

futureDoctor’s advice 51 43.2I am interested in having a blood test 49 40.5Reassurance because I am worried 33 28.2To clarify things I am confused about 26 22.6

20.2I was told by my family/friends 2312 11.2I have symptoms of bowel cancer

51

Collins et al.

Table 3. Participants’ perceptions of how much information they received at the clinic, for those who indicated on the pre-clinic questionnaire that they wanted‘very detailed’ information

Number wanting ‘veryType of information Information received (%)detailed’ information

Number % None Some but not enough Enough Too much

Personal risk of cancer 87 73 1.1 13.8 85.1 0.082 71Children’s risk of cancer 4.9 13.4 81.7 0.067 56 9.0Seriousness of cancer 14.9 73.1 3.0

Things to prevent cancer 89 75 2.2 14.6 82.0 1.1Gene testing for self 86 71 4.7 16.3 79.1 0.0

72 64 6.9Gene testing for children 13.9 79.2 0.097 81 0.0 8.2 89.7Screening 2.186 76 11.6Early warning signs of cancer 27.9 59.3 1.2

score and change in knowledge when risk percep-tion was not included in the models (data notshown). Those who reported having a geneticcounsellor or medical geneticist in the clinic sessionwere more likely to have a higher post-clinicknowledge score and a greater increase in knowl-edge. A lower knowledge score at baseline pre-dicted a greater increase in knowledge score.

Very few participants (8%) indicated they wouldneed to go back to the clinic for more information,although 61% said they would like more informa-tion about HNPCC. The preferred medium for thisinformation was as written material (81%) in pref-erence to video or audio tape. When asked aboutthe letter sent after the clinic outlining the issuesdiscussed in the session, 17% said it was ‘extremelyuseful’, 46% said it was ‘very useful’ and 34% saidit was ‘quite useful’ in terms of the information itprovided. If new information came to hand, 106participants (86%) said they would like to go backto the clinic to hear about it. A question on thepost-clinic questionnaire asked if it would be help-ful to have an additional session just with a geneticcounsellor (the question referred to a non-medicaltrained genetic counsellor, but may have been in-terpreted by the participants as referring to either agenetic counsellor or a medical geneticist). Thirty-four percent thought either a face-to-face or tele-phone session would be helpful.

Discussion

Evaluation of new health services is necessary toensure that the needs of the clients to which theservice is directed are being met, and that the aimsof the service are appropriate and achievable.Methods for evaluation of genetic counselling ser-vices have been much debated (16). Genetic coun-selling differs from other medical services in that itis concerned with giving information and facilitat-ing decision making with people who are generally

healthy, rather than providing treatment for ill-ness. A review of these issues has concluded thatsimple outcome measures, such as the ability toquote risk figures or reproductive decisions, arenot adequate and we must attempt to measuresatisfaction with the service after the client hascompleted the process (16, 17).

This study has evaluated aspects of a familialbowel cancer clinic concerned with giving and re-ceiving information. Although there are many aimsof a service such as this (6), the ability to makeinformed choices about how to deal with cancerrisk and how to communicate with family mem-bers will depend on the provision of appropriateinformation and the way in which it is understoodby the client. We have looked not only at howknowledge has been changed by attending theclinic, but also at what expectations clients had interms of information needs, and how they per-ceived these needs have been met.

The mean age of those attending the clinic wassimilar to that reported for other cancer clinics (13,14). More women than men attended and this islikely to be related to the general reluctance of men

Fig. 1. Perceived risk of developing bowel cancer, comparedwith the general population, for those with no previous historyof bowel cancer. Proportion (%) of participants choosing eachlevel of risk before and after the clinic. Proportions in riskcategories assigned by the clinic are also included correspond-ing to the most likely category of perceived risk.

52

Assessment of a familial colorectal cancer clinic

Table 4. Multiple linear regression analyses of knowledge score and change in knowledge score

Predictors BetaDependent variable Standard error of beta p-value

Age −0.025 0.0161. Pre-clinic knowledge 0.128Sex 0.236 0.445 0.597Education level 0.319 0.106 0.003Pre-clinic risk perception 0.142 0.157 0.368

Age2. Post-clinic knowledge −0.005 0.012 0.659Sex 0.310 0.360 0.391Education level 0.260 0.082 0.002Booklet read 0.532 0.362 0.145Seen genetic counsellor* 1.041 0.345 0.003Post-clinic risk perception 0.483 0.182 B0.001

Age B0.0013. Change in knowledge 0.012 0.962Sex 0.282 0.340 0.410Education level 0.168 0.082 0.043Booklet read 0.502 0.343 0.147Seen genetic counsellor* 0.875 0.329 0.009Post-clinic risk perception 0.465 0.173 0.009Pre-clinic knowledge −0.764 0.065 B0.001

* Participant reported seeing either a genetic counsellor or a clinical geneticist in the clinic session.

to use health services (18), as men were more likelyto come after symptoms had already occurred.

Reasons for coming to the clinic were similar tothose reported in other studies with respect toreducing risk (12, 19) and concern for children (12,15, 20). Interest in finding out about a gene forbowel cancer is not surprising given the nature ofthe clinic and this interest was borne out by thehigh proportion indicating they would go aheadwith DNA testing. Whether or not this actuallytranspires remains to be seen. Another study ofHNPCC testing suggests that the proportion tak-ing up testing was lower than expected (21).

Helping research or science was an importantmotivation for attendance for over half of theparticipants. Some individual comments suggesteda desire to help others in the same predicament,but there were one or two people who viewed thisas a negative thing as they felt they were helpingthe doctors more than the doctors were helpingthem. An American study also reported this im-pression in a small number of clients (14), and it issomething clinic staff need to be aware of.

Results suggest that many are coming to theclinic with inflated perceptions of their risk ofdeveloping bowel cancer. Before attending theclinic, the main influence on those perceiving them-selves at high risk was family history. It appearsthat some people assume their family history has agreater impact on their risk than it really does andthis impression may have been aided by advicefrom other health practitioners (personal commu-nication with clinic staff). These risk perceptionsare also likely to be related to the participants’

personal experience of cancer in relatives and theiranxiety about developing cancer themselves (22).Other studies have also found that many partici-pants come to clinics believing their risk to behigher than it is (13, 14, 22).

Post-clinic perception of risk was more closelyaligned with the risk categories defined by theclinic. Of particular interest was the fact that 24%perceived a ‘much higher’ risk pre-clinic, but thisreduced to 15%, which corresponds with the pro-portion categorised as ‘definite HNPCC’ families.Participants indicated that post-clinic risk percep-tions were influenced by ‘what the doctors said’ aswell as ‘family history’, and some indicated thatgene tests would clarify risk. It is difficult to deter-mine how the ‘higher’ and ‘slightly higher’ percep-tions of risk correspond to the clinic categories, butthe general reduction in risk perception and ap-proximately correct proportions in these clinic cat-egories suggests that most are not maintaininginflated risk perceptions, as shown in a breastcancer study (22).

The information given at the clinic clearly in-creased participants’ knowledge of bowel cancergenetics. The positive association between educa-tion level and knowledge score was not surprising,but the fact that education level was only weaklyassociated with change in knowledge score sug-gested that increases in knowledge occurred acrossthe spectrum of education. The association be-tween the presence of a genetic counsellor or med-ical geneticist in the session with post-clinicknowledge and increase in knowledge needs to beinterpreted with caution, as this information is not

53

Collins et al.

validated. However, it does support the impor-tance of having a person trained in dealing withgenetic information as part of the clinic team.Moreover, a third of participants were also inter-ested in another session with a genetic counsellor.

The strong association of post-clinic risk percep-tion with post-clinic knowledge and change inknowledge suggests that those for whom theknowledge of bowel cancer is more relevant arebeing given and are retaining this information.However, how this knowledge will manifest in be-haviour remains to be determined. Knowledgealone will not necessarily result in a change inbehaviour, but knowledge is one of the majorfactors necessary for change to occur (23). Thereappear to be no currently available data demon-strating that knowledge mediates lifestyle changes,such as screening behaviour, in high-risk individu-als. However, knowledge of the effectiveness andbenefits of screening has been found to be associ-ated with adherence to colorectal screening in pop-ulation-risk individuals (24). One benefit ofattendance at the clinic is that it allows clients tobe entered into a surveillance program appropriateto their level of risk, at least allowing the opportu-nity for regular surveillance.

Even those wanting ‘very detailed’ informationwere satisfied with the amount of information re-ceived at the clinic. However, many expressed adesire for more written information to have as aresource outside of the clinic setting. The observa-tions that over a quarter of participants indicatedthat they did not get enough information about‘early warning signs’, that many said they wantedto know about screening and prevention, and thatseveral people in open questions said they wouldlike more advice about diet and lifestyle suggestthat many would like to be proactive in taking careof their health.

Several unsolicited comments were made aboutthe usefulness of answering the study question-naires in terms of knowing what to expect whenattending the clinic. These comments indicate thatinformation provided before the clinic session maybe a very useful addition to the service. The needfor information before attendance at a familialcancer clinic has been noted elsewhere (25), and ithas been shown that receiving written informationbefore attendance at a clinic can also help reduceanxiety (26).

The results presented here need to be consideredin the light of the limitations of the study. We haveonly assessed this particular clinic with these clientsand service providers and the results could bedifferent in another situation. We acknowledgethat there are many other aspects of the service

that could be assessed, such as cost effectiveness,as well as long-term outcomes, like effects on mor-bidity and mortality, that will need to be addressedover time.

Ideally, the response rate could be higher butfrom the information available, particularly withrespect to those who responded to the first but notthe second questionnaire, the non-responders weresimilar to the responders. The variable most likelyto influence results is perceived risk of developingbowel cancer and this was very similar between thetwo groups. Although not statistically significant, ahigher proportion of non-responders to the post-clinic questionnaire were classified by the clinic asdefinitely or likely to be part of HNPCC families.Non-response amongst individuals in this high-riskgroup may have biased the results if these peoplewere more likely to be dissatisfied due to the infor-mation they received about their risk.

When using a pre–post design, it is difficult toknow how much of the observed changes can beattributed to the clinic alone. However, it is likelythat the changes in knowledge are due either to thespecific information provided by the clinic, or tofurther reading or discussion clients have had thatwere prompted by attendance at the clinic.

AcknowledgementsThis study was supported by funding from the PublicHealth Division of the Department of Human Services,Victoria, Australia. We thank the staff of the Royal Mel-bourne Hospital Family Bowel Cancer Clinic for theircooperation and assistance. The participation of the clientsof the Family Bowel Cancer Clinic is also greatly appreci-ated. Drs Mac Gardner, Clara Gaff, Mark Rogers andBettina Meiser gave helpful comments on the manuscript.

References

1. Austoker J. Screening for colorectal cancer. Br Med J1994: 309: 382–386.

2. Parry D, Berg K, Mulvihill J, Carter C, Miller R. Strate-gies for controlling cancer through genetics: report of aworkshop. Am J Hum Genet 1987: 41: 63–69.

3. Lynch H, Lynch J. Familial factors and genetic predisposi-tion to cancer: population studies. Cancer Detect Prev1991: 15: 49–57.

4. Groden J, Thliveris A, Samowitz W et al. Identificationand characterisation of the familial adenomatous polyposiscoli gene. Cell 1991: 66: 589–600.

5. Nicolaides N, Papadopoulos N, Liu B et al. Mutations oftwo PMS homologues in hereditary nonpolyposis coloncancer. Nature 1994: 371: 75–80.

6. Peters J, Biesecker B. Genetic counseling and hereditarycancer. Cancer 1997: 80 (Suppl.): 576–586.

7. Menko F, Whijnen J, Vasen H, Oosterwijk M. Geneticcounseling in hereditary nonpolyposis colorectal cancer.Oncology 1996: 10: 71–76.

8. Lynch H, Lynch J. Genetic counselling for hereditarycancer. Oncology 1996: 10: 27–34.

54

Assessment of a familial colorectal cancer clinic

9. Atkin W, Cuzick J, Northover J, Whynes D. Prevention ofcolorectal cancer by once only sigmoidoscopy. New Engl JMed 1994: 341: 736–740.

10. Atkin W, Morson B, Cuzick J. Long term risk of col-orectal cancer after excision of rectosigmoid adenomas.New Engl J Med 1992: 326: 658–662.

11. Jagelman D. Clinical management of familial adenoma-tous polyposis. Cancer Surv 1989: 8: 159–167.

12. Bleiker E, Aaronson N, Menko F et al. Genetic counselingfor hereditary cancer: a pilot study on experiences ofpatients and family members. Patient Educ Couns 1997:32: 107–116.

13. Julian-Reynier C, Eisinger F, Chabal F et al. Cancergenetics clinics: target population and consultee’s expecta-tions. Eur J Cancer 1996: 32A: 398–403.

14. Stadler M, Mulvihill J. Cancer risk assessment and geneticcounseling in an academic medical centre: consultands’satisfaction, knowledge, and behaviour in the first year. JGenet Couns 1998: 7: 279–297.

15. Lerman C, Narod S, Schulman K et al. BRCA1 testing infamilies with hereditary breast-ovarian cancer. JAMA1996: 275: 1885–1892.

16. Clarke A, Parsons E, Williams A. Outcomes and processin genetic counselling. Clin Genet 1996: 50: 462–469.

17. Kessler S. Process issues in genetic counselling. Birth De-fects: Orig Article Ser 1992: 28: 1–10.

18. Verbrugge L. Gender and health: an update on hypo-theses and evidence. J Health Soc Behav 1985: 26: 156–182.

19. Richards M, Hallowell N, Green J, Murton F, Statham H.Counseling families with hereditary breast and ovariancancer: a psychosocial perspective. J Genet Couns 1995: 4:219–233.

20. Lerman C, Marshall J, Audrain J, Gomez-Caminero A.Genetic testing for colon cancer susceptibility: anticipatedreactions of patients and challenges to providers. Int JCancer 1996: 69: 58–61.

21. Lerman C, Hughes C, Trock B et al. Genetic testing infamilies with hereditary nonpolyposis colon cancer. JAMA1999: 281: 1618–1622.

22. Lerman C, Lustbader E, Rimer B et al. Effects of individ-ualised breast cancer risk counselling: a randomized trial.JNCI 1995: 87: 286–292.

23. Becker MH, Maiman L. Sociobehavioral determinants ofcompliance with health and medical care recommenda-tions. Med Care 1975: 13: 10–24.

24. Myers R, Ross E, Jepson C et al. Modeling adherence tocolorectal cancer screening. Prev Med 1994: 23: 142–151.

25. Hallowell N, Murton F, Statham H, Green J, Richards M.Women’s need for information before attending geneticcounselling for familial breast or ovarian cancer: a ques-tionnaire, interview, and observational study. Br Med J1997: 314: 281–283.

26. Austoker J, Ong G. Written information needs of womenwho are recalled for further investigation of breast screen-ing: results of a multicentre study. J Med Screen 1994: 1:238–244.

Appendix: Knowledge items in pre- and post-clinicquestionnaires

Bowel cancer is always inherited (F).1.Everyone who has a gene for bowel cancer will2.develop bowel cancer (F).

3. A gene for bowel cancer can also increase therisk for other cancers (T).A person who does not have a gene for bowel4.cancer can still develop bowel cancer (T).

5. There is more than one gene that can causebowel cancer (T).Faecal occult blood tests (FOBT) will always6.detect bowel cancer (F).If a person looks like or has the personality of7.a relative who has had bowel cancer, they arelikely to have inherited the gene from thatrelative (F).A colonoscopy is very likely to detect bowel8.cancer if it is present (T).

9. In a family where a gene for bowel cancer hasbeen found, those without the gene have thesame risk of getting bowel cancer as the gen-eral population (T).

55