GENETIC TESTING Volume 6, Number 4, 2002 © Mary Ann Liebert, Inc.

Patient vs. Physician as the Target of Educational Outreach about Screening for an Inherited Susceptibility

to Colorectal Cancer

STARLENE LOADER,L2 CLEVELAND SHIELDS, 3.4 JEFFREY C. LEVENKRON,4 RICHARD FISHEL,' and PETER T. ROWLEy L2

ABSTRACT

Are patients identified from a cancer registry better educated directly or via their physician about screening for an inherited susceptibility for colorectal cancer? Of 974 patients diagnosed with colorectal cancer at ::560 years from 1987 to 1999 in a five-county area including Rochester, the physicians of 651 patients (67%) for­warded a cancer family history survey to their patient; 459 (71 %) completed the survey. Of these 459, 167 (36%) reported having at least one first- or second-degree relative with colon cancer and were sent a set of questionnaires and a more detailed family cancer history form. Of the 167, a total of 101 (60%) continued to qualify by returning the questionnaires. These 101 qualifying patients were randomized to either the patient­education or physician-education group. Of the 101, a total of 47 (47%) came for a free genetic evaluation. Individuals were more likely to accept evaluation if they were parents (p = 0.001), had more cancers of all kinds in their families (p = 0.02), and had a larger social network (p = 0.04). Of the 47 counseled, 36 (77%) chose to have DNA testing at no cost. Of these 47, individuals were more likely to choose DNA testing if they had more cancers in the family (p = 0.04) and fewer symptoms of depression (p = 0.05). Of the 36 tested pa­tients, 6 (20%) were found to have mutations. In summary, acceptance of genetic services was related to the magnitude of the threat (more cancers in the family), perceived ability to deal with the threat (perceived good health and a supportive network), and a desire to inform relatives (being a parent). The two approaches to educating patients, viz. direct patient education vs. education via their physician, did not significantly differ in terms of percentages of patients receiving counseling (42 % vs. 51 %, respectively) or the percentage choos­ing DNA testing (32% vs. 37%, respectively).

INTRODUCTION

T RADlTlONALL Y, the same recommendations for cancer screen­ing practices are made for everyone . For colorectal cancer,

for exampl e, screening is recommended to begin at age 50. However, advances in the genet ics of cancer have revealed that , for many types of cancer, certa in segments of the population are genet ica lly predisposed. Detect ion of individuals with an inherited susceptibility to cancer is important for four reasons. First, such individuals are at very high ri sk. Second, these in ­di vidual scharacter istically have cancer diagnosed decades ea r-

li er than the rest of the population; therefore, the traditionally recommended starting age for screening will mi ss many of lhe cancers . Third , special types of sc reening may be indicated that are not required for the population at large. Finally and of spe­c ial importance, detection of such an individual identifies a whole family at ri sk.

Hered itary nonpol ypos is colorectal cancer (HNPCC) is a case in point. In men, the ri sk to age 70 is 82% for de veloping colorectal cancer; in women the ri sk is 54% for colorectal can­cer and 60% for e ndometrial cancer (Aarnio et aI. , 1999). The average age at which colon cancer is diagnosed (40-45 years)

JDepartment s of Medi cine, 3Famil y Medi cine, and 4Psychiatry, and 2Division of Genetics, Universi ty of Rochester School of Medicine, Roch­es ter, NY 14642.

' Kimmel Cancer Institute, Thomas Jefferso n University, Philadelphia, PA 19 107.

281

282

in these individuals is about two decades earli er than for spo­radic colon cancer (Aaltonen, 2000). Individual s whose ri sk is no t identifi ed before cancer appears ma y not be diagnosed at a c ura ble stage (Hakala et aI., 199 1). Because most of the colon cancers ari se on the ri ght s ide, colonoscopy is the sc reening method of choice, not the s igmoidoscopy traditionally used for the general population. Women with HNPCC have a ri sk of en­do metrial cancer even grea ter than that of colon cancer and merit additional surve illance measures .

Identifi cationof indi viduals with HNPCC is criti cal. A study in Finland showed that, for individuals detec ted to have HN PCC, the proportion surviving 15 years was 90% for those hav ing colonoscopy every 3 years, but onl y 74% for those de­clining colonoscopy (Jarvinen et a I., 2000). A dec ision analy­s is indicated that, compared with no inte rvention , mutation de­tec tion and colonoscopic survei lIance begun at age 25 increases li fe expectancy by 13.5 yea rs (Syngal et a I., 1998) .

Unfortunately, colon cancer families are often reluctant to undergo geneti c testing. When 258 individuals without cancer from known or suspec ted HNPCC famili es were offe red DNA testing, onl y 30% accepted (Codori et a I., 1999). When 208 famil y members without cancer from four HNPCC families with known mutations we re offe red testing, only 43% accepted (Lerman et a I., 1999). According to Aktan-Collan et a l. (2000), "The challenging task is to convey the initial offer in such man­ner that individuals of different soc io-economic backgrounds become interested ... Research is needed on the barriers to par­ti c ipation in the test. "

Our major goal was to conduc t a randomized stud y to as­certain whether more patients would seek geneti c services if we , as a tertiary medical cente r (l ) mailed them relevant in ­fo rmation in layman 's te rms or (2) sent the same information in profess ional language to their physicians . We had hypothe­sized that educating the patient directl y would result in a larger proportion of pati ents receiving geneti c se rvices, because the pati ent mi ght fee l a greater responsi bi lity to ale rt relatives than would the phys ician and because the physic ian might be re luc­tant to rai se genetic issues because of unfamiliarity with ad ­vances in cancer geneti cs . Subjec ts qualifying for se rvice we re the re fore randomized between these two methods of outreac h.

MATERIALS AND METHODS

Subject ide ntification

We obtained information on all individuals diagnosed with carc inoma of the colon or rectum at age 60 or less from 1987 through 1999 in a fi ve-county area of New York including Rochester. We asked the physician of record or c urrent phys i­c ian for consent to send the patie nt a brief cancer famil y hi s­tory questionnaire to be returned to us. From the questionnaire, we identifi ed patients who had at least one firs t- or second-de­gree relati ve with colorectal cancer and sent them a set of ques­ti onnaires for completion. These included a more detailed fam ­ily hi story form, the 36-ltem Health Survey 1.0 (SF-36) (Hays et a I., 1993), the Beck Depress ion Inventory (Beck, 1967), the Intrusion Subscale of the Impact of Events Scale (Horowitz et aI. , 1979) adapted to a concern about colon cancer, the Family Emotional Involvement and Criti cism Scale (Shields et a I.,

LOADER ET AL.

1992; Fisce lla et a I., 1997), the Soc ial Support Ques tionnaire (Sarason et a I., 1983), and the Behavioral Style Scale (Miller, 1987). The Behavioral Style Scale is intended to class ify sub­j ects as hi gh monitors (i. e., those who reac t to a threat by seek­ing information) or as low monitors (i. e., those who react to a threat by avoiding information and seeking a source of di s­traction , al so called high blunters) . This request was accompa­nied by the offer of $25 for completion , but not with the offe r of geneti c counse ling or testing, so as not to bias the return in fa vor of persons with an interest in receiving geneti c informa­tion.

Those returning questionnaires and still qualifying by fam ­il y hi story we re then randomized to one of two groups. One group was sent a brochure about inherited susceptibility to col ­orec tal cancer written in lay language and an invitation to come for geneti c counse ling and the offer of DNA testing. For the othe r group, the subject's phys ician was sent information about inherited susceptibil ity to colorectal cancer in profess ional lan­guage and invited to refe r the patient for geneti c counse ling and the offer of DNA testing. For both groups, counseling and DNA testing were offe red without charge.

When the subject came for counse ling, the famil y hi story of cancer was verified, the importance of earl y detection of col ­orec tal cancer for optimal outcome was emphasized, and the surve illance recommended for persons at increased ri sk re­viewe d. All subjec ts coming for counseling were offe red DNA testing. The benefit s, ri sks, and limitations of DNA testing for an inherited susceptibility were explained. The benefit s li sted included recognition of an inc reased ri sk of a second colorec­tal cancer and of cancer in other s ites, the adoption of preven­

ti ve measures, and the opportunity to ale rt relati ves to their ri sk and to refer them for genetic counseling and DNA testing. Ris ks li sted included anxi ety, depress ion, guilt about poss ible trans­mi ss ion to children, stigmatization, and di sc rimination regard­ing insurance and employment. Limitations included the im­perfect sensitivity of the tes t and the poss ibility of obtaining a result of unknown c linical significance . DNA testing was o f­fe red without a recommendation to accept or decline .

Patients were sent home with a copy of the consent form and mailed a summary of the conte nt of the counse ling sess ion. A patient c hoosing DNA testing had to make another appointme nt for compl etion of the informed consent process, have blood drawn , and agree to a third visit to receive the test result. The patient was notifi ed when results we re available . The result and it s s ignifi cance were then explained in person. Those in whom a mutation was found were urged to follow the recommenda­tions of NCI's Cancer Studies Consortium (Burke et aI. , 1997) . These include colonoscopy eve ry 1-3 years beginning at age 20-25, transvaginal ultrasound or endometrial aspiration and poss ibly hysteroscopy annuall y beginning at age 25-35, a low­fat , high-fiber diet, with an adequate intake of vegetables or fruit , the avoidance of known carc inogens such as tobacco, and the options of prophylacti c subtotal col ec tom y, hysterectomy, and oophorectomy.

Mutation detection

Blood from subjects choosing DNA testing was sent initia11y to Genz yme (Framingham, MA) for amplifi cation of all exons and adjacent intronic reg ions of hMLHi and hMSH2 by dena-

SCREENING FOR INHERITED COLON CANCER

turing gradient gel elec trophoresis (DGGE). Any abnormality found was pursued by sequencing. Later samples were sent to Huntington Medical Research Laboratories (Pasadena , CA), to City of Hope (Duarte CA), or to Myriad Genetics (Salt Lake City, UT) for direc t sequencing of all exons and adjacent in ­tronic reg ions of these two genes.

Post-results counseling

When DNA test results were available, patients were noti ­fi ed and all returned to receive them in persoll. Those found to have a mutation were offered DNA testing for re lati ves with­out charge . Recommendations about surveillance were made for those with or without detected mutations.

This study was approved by the Research Subject<; Review Board of the University of Rochester.

RESULTS

The des ign of the study is shown in Figure I.

Physician response

Of 974 colorectal cancer patient s, the phys ician of record or c urrent phys ic ian agreed to the contac ting of 65 1 patients (67%) . In the remaining cases, the physician of record had died , moved without indicat ion of the identit y of the patient 's new physician , regarded the pati ent as unsuitable for participation ,

283

or did not respond to our request, or the patient had di ed or moved without leaving a forwardin g address.

Patient response

Of the 651 patie nt s sent a cancer famil y hi story quest ion­naire, 459 (71 %) returned them. Of the 459, 167 (36%) reported at least one first- or second-degree relative with colorectal can­cer. Of the 167 sent an offer of a genet ic evaluation and asked to compl ete and return a more detailed famil y hi story ques­tionnaire and additional quest ionnaires, 123 (74%) responded. Of these 123, 101 (60% of 167) still qualified by famil y hi s­tory. Their c haracte ris tics are shown in Table I. Of these 101 , 47 (47%) have come for genet ic counse ling.

Factors affecting coming fo r counseling

Coming for counseling was not signifi cantl y re lated to be­ing informed about the poss ibility of having an inherited sus­ceptibility to colorectal cancer by us versus by their own physi­c ian. Factors affecting coming for counseling using uni variate analysis are shown in Tables 2 and 3. Table 2 shows the results of I-tests on continuous va riables . Those who came for coun­seling reported more total cancers in their famil y (p = 0.03) and larger soc ial networks (p = 0.03) . Social network s ize , a c haracte ri stic deri ved from the Social Support Questionnaire , re fers to the pe rceived numbe r of persons available to provide e motional support. Those who came for counselingalso showed a tre nd toward greater pe rce ived hea lth (p = 0.06) and a beljef

974 Patients diagnosed with colorectal cancer at age s 60 from 1987 through 1998 in a

5-county area of NY including Rochester from the cancer reg istry

I 6511974 (67%)

Physicians consented to sending patients a famil y hi story quest ionnaire

I 459/651 (7 1%)

Patients returned completed quest ionnaire

I 167/459 (36%)

Patient s reported at least one 1st or 2nd degree relati ve with colon cancer

I 101 /167 (60%)

Patients continued to qualify by returning a more detail ed famil y hi story and other quest ionnaires

I 471101 (47%)

Pati ents came for geneti c counse ling

I 36/47 (77%)

Patients chose DNA testing

I 6/3 6 (17%)

Patients had a mutation detected

FIG.1. Results of contacting colorectal cancer patients identifi ed from a reg ionill cancer reg istry.

284

TABLE I. CHARACTERISTICS OF Q UALIFYING P ARTICIPANTS

Sex Male 47 Female 54

Race White 94 African-American 6 Asian I

Marital Status Married 78 Widowed 10 Single 7 Separated 2 Divorced 2 Unknown 2

Pare ntaJ Status Hav ing children 8 1 Having no children 20

Number oj cololl cancers

Number oj Percentage oj

I 2 3 4 5 6

2 3 4 5 6 7 8 9 10 II

part icipal1ts participants

Total numbe r of colon cancers in first­and second-degree relatives

9 51 24 II 5 I

Total number of cancers in firs t-and second-degree relatives

6 17 18 12 20 II 6 6 2 3

9.8 50.5 23 .8 10.9 5.0 1.0

5.9 16.8 17.8 11.9 19.8 10.9 5.9 5.9 2.0 3.0

LOADER ET AL.

that surve illance for colon cancer was effect ive (p = 0 .08) . Table 3 shows X2 tests on di chotomous va riables. Being a par­e nt was stron gly re lated to coming for counse ling (Fisher's ex­

act test, p = 0.00 1). There was a tre nd toward ret irees being more like ly to come for counsel ing (p = 0 .07) .

Six months after indi v iduals who qualified were inv ited to

come for counsel ing, we wrote to those who had not come in

to ask their reason for not coming . Five responded by coming in at that time . Others gave various reasons. Five sa id their

needs were being met by the ir own phys ic ian . Smalle r numbers

li ved far from Roc hester, sa id their relati ves were already alerted , or reported prev ious ge netic counse ling, partic ipating in researc h e lsewhere, or being too ill.

Factors affecting coming for counsel ing us ing multi variate

analysis are shown in Table 4 . We pe rformed a logist ic re­gress ion us ing stepwise procedures because we did not ha ve

prior hypotheses indicatin g in what order variables should be e ntered into the model. We e ntered va riables from Tables 2 and

3 that were s ignifi cant or showe d a trend toward s ignifi cance . The final model containe d three variables: being a parent , total

number of famil y cancers, and soc ial network size . Parents were

more than 10 times more likely to come for counsel ing than were non-parents . For every additional cancer in the fami Iy, in ­di viduals were 29% more likel y to come for counse ling . For

eve ry additiona l person in the soc ial ne twork, individuals were

30% more likel y to come for counsel ing .

Factors affecting choosing DNA testing

Reasons patients gave for c hoos ing DNA testin g are li sted

in Table 5 . Of the 36 patients, 16 (44%) gave a conce rn for re l­

atives as the most important reason. Ten (28%) said they chose

testin g because they thought they had an inhe rited susceptibil ­it y . Four (11 %) chose testing to help research. Three (8 .3%)

said a pos iti ve test would assist their medical ca re . One each said they chose testing to sa ti sfy an e motional need and because

T ABLE 2. FACTORS AFFECTING COMING FOR COUNSE LING (11 101 ): U NIVAR IATE A NALYSIS (I-TESTS)

Variable

Age at diagnosis Age when contac ted Years of education Number of colon cancers in famil y Number of all cancers in famil y Colon cance r knowledge Pe rceived seriousness Pe rceived suscepti bility Pe rceived effec ti veness of surveillance Social network s ize (SSQ) Pe rceived health (SF36) Hjgh monitor (MBSS) Intrus ive thoughts (lES) Depress ion (Beck Scale)

Not couflseled mean (SD)

49 .1 (72) 57 .3 (69) 14. 2 (28) 2.5 (09) 5.0 (22)

52 .7 (173) 3.5 (II) 3.1 (09) 4.2 (II )

13.8 (87) 63 .3 (253) 9.6 (39)

14.7 (65) 7.8 (72)

Counseled mean (S D)

51.1 (66) 59 .2 (65) 14.6 (26) 2.7 (1.1 ) 5.9 (22)

53 .6 (1 73) 3.5 (II ) 3.2 (09) 4.5 (08)

19.0 (1 3.4) 71.5 (17 .3)

9.5 (3.4) 13.2 (56) 7.5 (73)

- 1.39 - 1.43 - 0.67 - 0.95 - 2.24 - 0.23

0.04 - 0.29 - 1.74 - 2.25 - 1.93

0.03 1.14 0.22

p

0 .17 0 .1 6 0 .50 0 .34 0 .03 ' 0 .8 1 0 .96 0 .77 0 .08-0 .03 ' 0 .06-0 .97 0 .25 0 .82

SSQ, Social Support Questionnaire; SF36, 36- ltem Health Survey; MBSS, Monitor sca le of Mi ll e r Beha vioral Style Sca le; IES, Intrus ion subscale of Impac t of Events Scale adapted to a conce rn about colon cance r.

Unreferenced items were obtained from a se lf-des igned questionnaire . Perce ived se riousness, perceived susceptibility, and pe rceived e ffect iveness of surve illance all refer to colon cance r; each characteri st ic was based on four to e ight quest ions .

*Identifi es variabl es with p s 0 .05 . - Identifies variables with 0 .09 > P > 0 .05 .

SCREENING FOR INHERITED COLON CANCER 285

T ABLE 3. FACTORS AFFECTING COMING FOR COUNSELING (11 10 1): UNIVARIATE A NA LYS IS (x2 TESTS)

Variable Not couflseled Coullseled X' p

Gender Male Female

Marr ied No Yes

Pare nt No Yes

Employed Not Yes

Re tire d No Yes

Educat ion V ia physician Direct to patient

3Fisher's exact test due to small ce ll s ize . *Identifies variables with p :s 0 .05 . - Identifies variables with 0 .09 > P > 0 .05 .

29 25 II 43 18 2

20 34 44 10 25 29

the ir phys ic ian recommended test ing . F inall y, one said a nega­ti ve test result wou ld be reassuring even though we stressed that a negat ive tes t wou ld be uninformati ve .

Reasons pati ents gave for c hoosing no DNA testing were co ncern about c onfidentiality, d isc rimination regarding insur­ance or e mployment , and anx ie ty about rece iving bad news or about having to transmit it to relat ives .

Fac tors affecting c hoos ing DNA test ing us ing uni variate ana lysis are li sted in Table 6. Subject<; choosing DNA test ing re ported more cancers in the famil y (p = 0.02) and better per­ceived hea lth (p = 0 .04). They showed a trend to have less de­press ion (p = 0 .07) and more colon cancer knowl edge (p =

0 .09). We also examined the same di chotomous variables that we examined for coming for counse ling, but none were s ignif­icant predictors of c hoosing test ing.

Factors affect ing choosing DNA test ing using multi variate ana lysis are shown in Table 7. Agai n we used s tepwise logis­ti c reg ression. Two var iables re mained in the model: total fam ­il y cancers (p = 0.04) and a lower score on the Beck depres­s ion scale (p = 0 .05). For each add itional cancer in the famil y, individuals were 68% more likely to c hoose testing. For each s tandard deviation increase in the Beck depression inventory, individuals were 36% less likely to c hoose tes ting.

Choos ing test ing was not signifi cantl y related to be ing in ­formed about the poss ibili ty of having an inherited susceptibil ­ity to colorectal cancer by us ve rsus by their own phys ician. A]so not signifi cantl y related to e ither coming for counse ling or choosing testing were age at diagnosis, age when contacted , sex, marital status, num ber of years of education , being em-

18 2.40 0 .1 2 29 12 0 .38 0 .54 35 36 0 .00 l3* 45 23 1.45 0 .23 24 3 1 3.1 6 0 .07-16 26 0 .8 1 0 .37 2 1

ployed, c lassificat ion as monitor versus blun ter on the Miller Behavioral Style Scale, in trusive thoughts about colon cancer, perce ived seriousness of colon cancer, or pe rce ived suscept i­bility to colon cancer. The e ffect of having health insurance could not be eva luated because all participmlts had health in­surance.

Mwations de tected

Of the 36 individuals choosing DNA testing, 6 had abnor­mal result s, as shown in Table 8 . All were missense mutations. Two were reported to us as delete rious. One pati ent was a 63-yea r-old man who had had colon cancer diagnosed at age 60. Colon cancer had been diagnosed in hi s father at 72 and prob­ably in a paternal aunt at 66. The proband had a L6 18A muta­tion in hMLH1 . Th is mutation has been reported prev iously (Farr ington et aI. , 1998).

The other patient was a 44-year-old woman who had had colon cancerdiagnosedat 42. Her mother had a mali gnant polyp diagnosed at 58, a maternal au nt had colon cancer at 39, and a maternal grandmother had bladder cancer at 68 . The proband had a A636P mutation in hMSH2. This mutation has also been prev iously reported (Yuan et aI. , 1999).

The abnormali ties in 4 other pe rsons were reported as of un­known clinical signifi cance. One patient was a 52-year-old woman who had had colon cancer at age 46. Colon cancer had been diagnosed in a brother at 53 and in a paternal au nt at 80 and metastatic cancer from an unknown primary s ite was di s­covered in a sis te r at 34 . The proband was found to have an

T AB LE 4 . F ACTORS A FFECTING COMING FOR CO UNSELING: M ULTIVAR IATE A NALYS IS3

Variable

Be ing a parent Number of all cancers in famil y Social network size

aStepwise logistic reg ress ion. *Identifies variabl es with p s 0 .05 .

Odds ratio

10.08 1.29 1.30

95% cOl1fidence limits X' p

2.06-49.43 8.1 3 0 .00 1* 1.04-1. 62 5.1 8 0 .02 * 1.0 1-1.67 4.29 0 .04 *

286 LOADER ET AL.

TABLE 5. R EASONS GI VEN FOR CHOOSING DNA TESTING (11 = 36)

Reasol1

I. Conce rn for relat ives 2 . I think I have an inherit ed susceptibility 3 . To help research 4 . A positi ve test w ill ass ist my medi cal ca re 5 . To sat isfy an e motional need 6 . My phys ician recommends testi ng 7. A nega ti ve test will reassure me

L9 11 R mutation in hMSH2. Four s isters who li ve in Ireland re­

quested testi ng wh ile v is iting the United States; I of the 4 sis­te rs was found to share the proband 's mutation. Subsequentl y, microsate llite instabilit y was found to be high in the proband 's tumor and , because this mutation is in a reg ion c riti cal for in ­te raction with hMSH6 (Guerrette et aI. , 1998), the patient was told that the mutation is probabl y dele terious . All 4 of the pa­

ti e nt 's children, ages 18- 27, also requested testing, and 2 of the 4 share the ir mother 's mutation.

A second patient, a 5 1-year-old woman, had had cance r in the region of the il eocecal valve d iag nosed at age 37 . Colon cancer had also been diagnosed in her fath er at 44 and in a pa­te rnal uncle at 50 and endometrial cance r diagnosed in that un ­cle's daughter at 40. The proband was found to have a G I62R mutation in hMSH2. M icrosatelli te instability was high and im­munohi stochemis try showed a loss of express ion of hMSH2,

but normal express ion of hMLH1. Consequently she was told

that her mutation was like ly to be del eteri ous . She referred her daughter for testing and he r daughter was found to share th is mutation.

The onl y two African-A meri can patients in the stud y who were found to ha ve a mutation had the same mutation, H7 18Y in hMLH1. They are not known to be related . One, a 57-year­

old woma n, had been diagnosed with a cance r in the anal re­g ion diagnosed at age 49 . Colon cance r had been diagnosed in

Number of subjects

16 10 4 3 I I I

%

44 .4 27 .8 11.1 8.3 2.8 2.8 2.8

her fath er at 59, in a paternal uncle at 78, in a mate rnal uncle at 88, and in a maternal firs t cous in once re moved. Uterine can­cer had been diagnosed in her paternal grandmother in her 50s. The other patient, a 6 1-year-old man , had been diagnosed with a ca rc inoma of the ascending colon diagnosed at age 55 . Colon cancer had been d iag nosed in a brother at 35, in a s ister at 60 and in hi s father at 70. Thi s mutation had not been reported at the time, but s ince has been reported (Weber et aI. , 1999) in 3 unrelated African-Americans with earl y-onse t colorecta l can­cer. The product of the mutation has been shown to be de fi ­c ient in fun ction in a yeast assay (Weber et aI. , 200 l ) . There­

fore, these 2 patients were also told that their mutat ion was probabl y deleter ious .

Tested patient s in whom no mutation was detected were re­

minded that an inherited suscepti bility to colon cancer had not been ruled out and the ir surve ill ance program was rev iewed. The long-term impact of findin g or not findin g a mutation will be the subject of a future report.

DISCUSSION

Overview

In patients with colorectal cance r at a younger than average age who responded , I in 3 reporte d colorectal cance r in a close

T ABLE 6. F ACTORS A FFECTING CHOOSING DNA TE STING (n = 47): U NIVARIATE A NALYS IS

Not tested Tested Va riable meal1 (S D) meal1 (S D) p

Age at diagnosis 50.0 (58) 5 14(68) - 0 .6 1 0 .54 Age whe n contacted 59 .4 (62) 59 .2 (67) 0 .07 0 .94 Years of e ducation 13.5 (28) 14.9 (25) - 1.52 0 .14 Number of colon cancers in famil y 2.1 (07) 2.8 (12) - 1.35 0 .1 8 Number of all cancers in famil y 46 (17) 6.3 (22) - 2.33 0 .02 ' Colon cance r knowledge 46 .1 (20.5) 56.1 (1 5.7) - 170 0 .09-Pe rceived seriousness 3.3 (1.1 ) 3.5 ( 1.1 ) - 04 1 0 .68 Pe rceived susceptibility 2.8 (09) 3.3 (08) - 140 0 .1 6 Pe rceived effec ti veness 4 .2 ( 1 0) 4 .6 (07) - 1.35 0 .1 8

of surve ill ance Social network s ize 23 .5 (15.0) 17.5 ( 12.8) 1.3 1 0 .1 9 Pe rceived health (SF36) 62 .5 (206) 74 .2 ( 154) - 2.02 0 .04 ' Hjgh monitor (MBSS) 9.2 (30) 9.6 (35) - 0 .39 0 .69 In trus ive thoughts (JES) 13.9 (79) 13.0(47) 0 .35 0 .73 Depress ion (Beck Scale) 13.1(11.9) 5.8 (41) 2.00 0.07-

*Identifi es variabl es with p s 0 .05 . - Identifies variables with 0 .09 > P > 0 .05 .

SCREENING FOR INHERITED COLON CANCER 287

T ABLE 7. F ACTORS AFFECTING CHOOSING DNA TE STING (11 = 47): M ULTIVAR IATE A NALYS 1S3

Variable Odds ratio

Number of all cancers in famil y 1.675 Depress ion (Beck Scale)b 0 .359

aStepwise logistic reg ress ion. bOdds ratio based on s tandard dev iation units of change . *Identifies variabl es with p s 0 .05 .

re lat ive . When offered a genet ic evaluat ion , nearl y half of those qualifying accepted. Following nond irecti ve education about DNA testing, 3 out of 4 requested testing. Of patients c hoos­in g DNA testing, I in 6 were abnormal. Acceptance of geneti c services was related to ha ving children, the total number of c an­cers in the famil y, soc ial network s ize, pe rce ived good hea lth , and a low depression score . Ne ither coming for counsel ing nor c hoosing DNA testing were related to being contacted by us versus by the ir own phys ic ian .

Most studi es of sc reening for a genet ic susceptibility for col­orec tal cancer have been undertaken to e lucidate it s mutational bas is or to evaluate screening methodology . Most studies of re­ceptivity to sc reening ha ve involved hi gh-ri sk cl inics without we ll -defin ed e ntry criteria. To our knowledge, thi s is the first population-based study of receptiv ity to the offer of genetic se r­vices to colorecta l c ance r patients in the United States.

Design considerations

In the in terest of miss ing as few c ases as poss ible , we chose quite li bera l c riteria for offering genetic se rv ices . Our c rite rion for proband age at diagnosis was 60 or less, whereas the mod­ified Bethesda criteria for HNPCC require an age less than 50. Our criterion for famil y hi story was also li bera l, requiring onl y one first- or second-degree relati ve with colorectal cancer. Nat­urall y, the more li bera l the se lect ion criteria , the lower per­

centage of subj ects will be found to have a mutation. It is note­worthy that the detec ted alterat ions were all missense ami no ac id changes . These alterat ions ma y di splay lower pe netrance, accounting for the smalle r number of affected re latives .

95% confidence limits x' p

1.034- 2.714 4.3933 0 .04* 0 .100-0.8 19 3.9253 0 .05 *

Because of the limi ted size of the population served by our c ancer reg istry , we went back to 1987 to obtain a suitable num­ber of colorecta l c ancer cases for stud y. As expected, a num­ber of patients had di ed , moved away, or c hanged phys icians in the in ter im. These fac tors contr ibuted to the fact that invita ­tions to partic ipate could be sent with phys ic ian approva l to onl y 65 1 of the 974 recorded colorec tal cancer pati ents. Ini ­tia1l y, we sent the invitations to the pati ent once we received the permiss ion of the physic ian . However, when a patient com ­plained that her phys ic ian should have checked with her before permitting us to contact her, our Institutional Review Board (lRB) insisted that the phys ic ian , not we, should contac t the pa­ti ent. Th is require ment decreased patient recruitment furth er.

In obta ining a cancer famil y hi story from patients, we d id not want to bias re turns for an in terest in rece iv ing genet ic ser­vices. Therefore , we did not mention our in tent to offer gene ti c services until after patie nts had provided a famil y hi story. This may have contri buted to the fact that, of the 65 1 pati ents sent famil y hi story inquiries, onl y 459 pati ents returned them. To have the required time (60- 90 min ) at the first face-to-fa ce ses­

sion for confirming the famil y hi story and e duca ting the patie nt about an inherited susceptibility to colorectal c ancer and DNA testing, we asked the patient to c omplete the questionnaires dea ling with demographics, knowle dge, and att itudes in ad ­vance of the vis it. We wanted thi s information from subjecl-;

unbiased for in terest in geneti c serv ices, thus we tr ied to moti ­vate them to complete the questionnaires by offer ing them a cash incent ive ($25) rather tha n by offer ing them free gene ti c services.

TABLE 8 . !!MUll AN D !!MSH2 MUTATIONS FOUND IN SUBJECTS CHOOS ING DNA T ESTING (n = 34)

Age at Clin ical Gene Mutation diagnosis Family history significance

hMLHl K61 8A 60 Colon F 72, 1PA 66 Dele terious

H71 8Y 49 Colon F 59, PU 78, MU 88 Probably M cousin ? age delete rious

Uterus PGM 50's

H7 18Y 55 Colon B 35, S 60, F 70 Probably delete rious

hMSH2 A636P 42 Colon M 58, MA 39 Dele terious Bladder MGM 68

L91 1R 46 Colon B 53, PA 80 Probably Unknown primary S 34 delete rious

G I62R 37 Colon F44, PU 50 Probably Uterus P cousin 40 delete rious

F, Father; M, mother; B, brother; S, sister; U, uncl e; A, au nt ; GM , grandmother; P, paternal ; M, mate rnal.

288

The attrition from an initial pool of 974 colon cancer pati ents to only 101 offered genet iccounsel ingmay rai se concerns about the representati veness of the data obtained . To summarize, the dec rease from 974 tota l patients to 651 contacted resulted from the requirement to go through the physician and the fact that cancer diagnoses as far back as 1987 were included. Regard­in g 459 responses from 65 1 patients contacted, a 7 1 % response rat e is a quite acceptable response rate to a mail ed survey . The dec rease from 459 responders to 167 reporting an affected re l­ative is consistent with other reports on the frequency of a pos­itive fa mily hi story in colon cancer patients (Ivanovich et aI., 1999). The decrease from 167 initial qualifi ers to 10 I final qual ­ifiers reflects, in part, our requiring ex tensive quest ionnaire completion in advance of counse ling to avoid an overlong ini ­ti al counse ling vis it.

To encourageca reful considerationof DNA testing and avoid impulsive decision making, we required the patient to take home information about test ing and the consent form, and , if they c hose DNA test ing, to make another appointment for phle­botomy . Nevertheless, of the 47 who came for genet ic coun­sel ing, 36 (77%) returned to complete the informed consent pro­

cess and to have blood drawn for testing .

Educating the patient directly versus via the physician

A major goal of thi s stud y was to ascertain whether more patients would seek genetic services if we mailed them relevant information in layman's terms or sent the same information in

professional language to the ir physicians. We found no s ignif­icant diffe rence between these two methods in terms of the per­

centage of patients coming for counsel ing or in the percentage of those counse led who chose DNA testing . We had hypothe­sized that educa ting the pati ent directl y mi ght yield highe r ac­ceptance rates for the reasons given in the Introduction. How­ever, it ma y be that the patients rece iv ing a letter from their personal physician about the availability of genet ic counse ling rega rded it as a recommendatiol1 to seek genetic counse ling. In many studies of patients' adherence to health practices, physi­

c ian recommendation has been found to be influential. In fact, Esplen et al. (200l) found that the recommendation of a phys i­c ian or genet ic counselor was ve ry important or somewhat im­portant in 79% of patients' motiva tion to pursue genet ic test­

in g for HNPCC. Thus, perceived physician recommendation may have counterbalanced the benefit we had hypothesize d for educat ing patient s directl y.

Other fac tors affecting patient receptivity

Among factors predicting coming for counsel ing, be ing a pare nt was by far the most influential (p = 0.00 I ). The re­sponsibility to inform relatives of a cancer ri sk was give n as a reason for c hoosing DNA testing more frequentl y than any other reason (Table 5). Many studies of testing for an inherited ri sk for colorectal cancer have found that informing relati ves is a leading motivat ion (Lerman et aI. , 1996; Lynch et aI. , 1999 ; Vernon et aI. , 1999 ; Aktan-Collan et aI. , 2000; Esplen et aI. , 200 1).

A large numbe r of cance rs of all types in the fami ly was also associated with coming for counsel ing (p = 0 .02) (Table 4) .

This is readil y inte rpretable as a measure of the magnitude of

LOADER ET AL.

the perce ived cancer threat. It has not been noted as a dete r­

minant of accepting genet ic se rvices in this context in previous studies.

The larger a pati ent 's social network size , the more like ly was the patient to come for counsel ing (p = 0.04) (Table 4) . This findin g could represent eithe r a greate r perce ived obliga­tion to inform famil y membe rs of ri sk, the availability of oth­e rs for emotional support, or both. In our prev ious stud y of re­

cepti v ity to screening for breast-ovarian cance r genet ic susceptibility, fami Iy closeness was a predictor of choosing ge­netic testing (Loader et aI., 1998) .

Factors predicting choosing DNA testing included numbe r of famil y cancers (p = 0 .02) (Table 6), better perceived heahh (p = 0 .04), and fewer symptoms of depression (p = 0 .05) (Table 7) . The negat ive effec t of depress ion on accepting test­ing has been noted before (Vernon et aI. , 1997 ; Lerman et aI. , t999 ; Esplen et aI., 200 t ), but perceived health has not been previously reported in this connec tion. The relationship between not seeking genet ic se rvices and poorer general health , a higher depress ion score, and a less supporti ve famil y suggests that the process of invest igat ing an inherited cancer suscepti bility is suf­fi cientl y stressful that impaired hea lth or deficient soc ial sup­

port di scourage it.

HNPCC versus fa milial adenomatous polyposis

None of the pati ents in this stud y had a personal or famil y hi story suggesti ve of familial adenomatous polyposis. This fact ma y be explaine d by its greate r rarity and the poss ibility that such a famil y might have been previously diagnosed and coun­seled and hence not come forward to rece ive genet ic services at our invitation.

Screening in a service versus research setting

In a sc reening program in a primary ca re sett ing des igned for serv ice purposes rather than for research, the fami Iy hi story could be obtained at the initial work-up and only one physician

mi ght be involved, s implify ing the process and probably in­c reas ing the percentages of appropriate patient s counseled and tested .

Should all cases of colorectal cancer be eva luated for the poss ibility of an inherited susceptibil ity?This is a matter of fre­quency, cost, and patient recept ivity. Regarding frequency, many studies have been published involving screening for HNPCC. Especially notable are the landmark studies in Finland by Mecklin et al. ( 1995), Jarvinen et al. (1995), Aaltonen et al. (l998) , Loukola et al. (l999), Jarvinen et al. (2000) , Salovaara et al. (2000), and Loukola et al. (200 I ) involv ing screening con­

secuti ve cases of colorecta l carc inoma for a sugges ti ve famil y hi story and microsatelli te instab ility in the tumors . However, these studies provide a limited guide to the feas ibility of such screening in the United States because, in Fin land, founder mu­tations increase the frequency of certain mutations and because a national hea lth scheme obviates patient concerns about di s­

c rimination in health insurance. Among U.S. s tudies, some are based on high-ri sk clinics

(Syngal et aI. , 1999, 2000) or used diffe rent se lection criteria (e .g. , Amsterdam) (Peel et aI. , 2000). Two studies (Cunning­ham et aI., 200 t ; de la Chapelle, 200 1) performed microsa tel -

SCREENING FOR INHERITED COLON CANCER

lite instability testing (MSI) on consecuti ve colorectal cancer without famil y hi story criteria. The large r study (de 1a Chapelle, 2001) found that 3 .3% had germ-line mutations in hMLH1 , hMSH2, or hMSH6. Although 3.3% may seem like a small per­centage for consideration of routine population screening, each patient diagnosed identifi es a whole famil y at ri sk.

Regarding cost, te sting is now expensive because mutation id entificat ion for the index case c urrentl y involves the DNA se­quenc ing of at least two large genes . If a mutation is found in the index patient, relati ves can be offered testing at a much lowe r cost. Once the majority of mutations responsible for HNPCC are identifi ed, DNA microarra y technology may make mutation detect ion more affordable. Lower-cost methodologies for genomic sequencekleletion analysis may al so become avail ­able.

With regard to patient receptivity, it is noteworth y that less than half the patients offered free genetic counse ling ca me in and that only one in three of those in whom a mutation was de­tec ted were responsible for relatives taking advantage of free DNA testing. Admittedly the latte r proportion is based on ve ry small numbers. Codori et at. (l999) found that predictors of c hoosing DNA test ing among first-degree relati ves of colon cancer patients were increased ri sk perception , greater per­ceived confidence in ability to cope with unfa vorable genet ic information , more frequent cancer thoughts, and having had at least one colonoscopy.lt may be an unfortunate feature of test­in g for an inherited susceptibility to cancer us ing a tes t of im­pe rfec t sensitiv ity that a famil y member diagnosed with cancer must be tes ted first because that indi vidual has the least to ga in from testing since their own cancer ri sk has already been shown to be high by their cancer diagnosis. What may be needed is an educa tional method that encourages perceiving such genet ic testing as a long-term benefit in terms of protecting famil y hea lth , rather than as a short-term stress of being the bearer of bad news to re lati ves .

Limitations of swdy

Several limitations of thi s study should be noted. We may have underes timated pati ent recepti vity for several reasons. First, many colorectal cancer patients had only a s ingle affected re lative and therefore may not have been too concerned about the ir ri sk of an inherited susceptibility. Second, patients were contacted only with the pe rmi ss ion of their own phys ician and in some cases the physic ian did not forward our invitation. Fi­nally, some pati ents may have been di scouraged from pro­ceeding by our request to complete a se t of quest ionnaires, notwithstanding the cash incenti ve .

On the other hand , we ma y have overestimated receptivity. First, the partic ipants were re latively well -educa ted, having completed an average of 2 yea rs of college . Some studies of geneti c sc reening have shown that test acceptors are better ed­ucated. Second , we offered genetic counseling and DNA test­in g free of charge to spare subject hassles with insurers over coverage. Third , all participants had health insurance and there­fore were not immediate ly threatened with di scrimination for in surance purposes .

In any case, the hea lth benefits of detec ting famili es with in ­herited susceptibility to colorec tal cancer are substantial. Reap-

289

ing these benefits will require a better understanding of their reluctance to be identified.

ACKNOWLEDGMENTS

We thank Pamela Babcock, Sandra LaBella , Cassandra Sul ­li van , Mary True, Sue Burkholder, and Deb Rose for theirclin­ical and technical ass istance . This s tudy was supported by Cen­ters for Disease Control (CDC) grant UR6/CCU217194 (P.R. ) and National Institutes of Health (NIH) grant CA92027 (R.F. ).

REFERENCES

AALTONEN, L. A. (2000) . Hereditary intestinal cancer. Cancer BioI. 10, 289- 298 .

AALTONEN, L. A., SAALOVAARA, R., KRISTO, P. , CANZIAN, F. , HEMMINKI, A., PELTOMAKI, P. , CHADW ICK, R.B. , PERCE­SEPE, A., KAARIAINEN, H., AHTOLA, H. , ESKELI NEN, M. , HARKONEN, N ., JULKUNEN, R. , KANGAS, E. , OJALA, S., TU­LlKOURA, J., VALKAMO, E., JARVINEN, H., MECKLlN, J. -P. , and DE LA CHAPELLE, A. (1998) . Incidence of hereditary no n­polyposis colo rectal cancer and molecular screening for the disease . N. Engl.J. Med.1338, 148 1- 1487.

AARNlO, M., SANKILA, R. , PUKKALA, E., SALO VAARA, R., AALTONEN, L.A., DE LA CHAPELLE, A., PELTOMA KI, P. , MECKLl N, J.-P. , and JARVINEN, H.J. ( 1999). Cancer ri sk in mu­tation carriers of DNA-mismatch-repair genes . In!. J. Cancer 81, 2 14-2 18.

AKTAN-COLLAN, K. , MECKLlN, J. -P. , JARVINEN, H., NYS­TROM-L AIHTI, M., PELTOMAKI , P. , SODERLlNG, I. , UUTEL A, A., DE LA CHAPELLE, A., and KAARIAINEN, H. (2000) . Pre­dicti ve genetic testin g for heredit ary non -polyposis colorectal can ­cer: Uptake and long- term satisfaction. Int. J. Cancer 89, 44- 50.

BECK, A.T. ( 1967). Depre.uion: Clinical, Experimental and TheoreT­ical AspeCTS . New York: (Harper & Row, New York ).

BURKE, W., PETERSEN, G., LYNCH, P. , BOTKIN, J. , DALY, M. , GARBER, J., KAHN, MJ., MCTIERNAN, A., OFFIT, K. , THOM­SON, E. , and VARRICCHIO, C. ( 1997). Recommendations fo r fo l­low-up care of indi viduals with an inherited predisposition to can ­cer. J. Am. Med. Assn. 277, 9 15-9 19.

CODORI, A.M., PETERSEN, G.M., MIGLIORETTI, D.L., L ARKIN, EX, BUSHEY, M.T., YOUNG, C., BRENSINGER, J.D., JOHN­SON, K. , BACON, J.A. , and BOOKER, S.V. (1999). Altitudes to­ward co lo n cancer gene testin g: fac tors predictin g test uptake. Can­cer Epidemiol. Biomarkers & Prevo 8, 345- 35 1.

CUNNINGHAM, J.M., KIM, C.-Y., CHRISTENSEN, E.R., TESTER, OJ., PARC, Y., BURGART, L.l. , HALLING, K.C., MCDONNELL, SK, SCHAID, OJ., VOCKLEY, CW ., KUBLY, V., NELSON, H., MICHELS, V.V., and THIBODEAU, S.N. (2001 ). The frequency of hereditary defective mismatch repair in a prospective series of un ­selected colorectal carcinomas. Am. J. Hum. Genet. 69, 780- 790.

DE LA CHAPELLE, A. (200 1). Hereditary nonpolyposis co lorectal cancer: natural history, mutation detection, and molecular screen ing. Presented at the American Society of Human Genetics, San Diego, CA

ESPLEN, M.J., MADLENS KY, L. , BUTLER, K., Mc KINNON, W., BAPAT, B. , WONG, J., ARONS ON, M., and GALLI NGER, S. (2001 ). Mo tivati ons and psychosocial impact of genetic testing fo r HNPCC. Am. J. Med. Genet. 103, 9- 15.

FARRINGTON, S., LIN-GOERKE, J., LING, J., WANG, Y., BUR­CZAK, J.D., ROBBINS, D.J. , and DUNLOP, M.G. (1998). System-

290

alic analysis of hM SH2 and hMLH I in yo un g colon cancer patients and controls. Am. J. Hum. Genet. 63,749- 759 .

f[SCELLA, K., FRANKS, P. , and SHIELDS, e.G. ( 1997). Perceived family criticism and primary care utilization psychosocial and bio­medical pathways. Famil y Proc. 36, 25-4 1.

GUERRETTE, S., WILSON, T., GRADlA, S., ""d FISHEL, R. (1998) . Interactions of human hMSH2 wi th hMSH3 and hMS H2 with hMS H6; examination of mu tations found in hereditary nonpolyposis coloreclal cancer. Mol. Cell BioI. 18,6616-6623.

HAKALA,T., MECKLlN, J. -P. , FORSS, M., JARVINEN, H. , and LEHTO VIRTA, P. ( 199 1). Endomelrialcarcinoma in the cancer fam­ily synd rome. Cancer 68,1656- 1659.

HAYS, R.D., SHERBOURNE, CD., ""d MAZEL , R.M. (1993). The Rand 36- ltem Health Survey 1.0. Health Economics 2, 2 17- 227 .

HOROW ITZ, M., WILNER, N., and ALVAREZ, W. ( 1979). Impact of Event Scale: a measure of subjecti ve stress. Psychosomat. Med. 41, 209- 2 18 .

JARVINEN, H.J., AARN IO, M., MUSTONEN, H. , AKTAN­COLLAN, K. , AALTONEN, L. A., PEL TOMAKI , P. , DE LA C HAPELLE, A., and MECKLlN, J. -P. (2000). Controlled 15-year trial on screening fo r colorectal cancer in families with hereditary nonpolyposis colorectal cancer. Gastroenterology 118, 829--834 .

JARVINEN, H. , MECKL IN, J. -P., and SISTONEN, P. ( 1995). Screen­ing reduces colorectal cancer rate in families with heredit ary non­polyposis colorectal cancer. Gas troenterolo gy 108, 1405- 14 1 I.

LERMAN, C., MARS HALL, J., AUDRA IN, J., and GOMEZ­CAMlNERO, A. ( 1996). Genetic testin g fo r colon cancer suscepti­bility : an tic ipated reactions of patients and challenges to providers . Int. J. Cancer 69, 58-6 1.

LERMAN, C., HUGHES, C., TROCK, B.1., MYERS, R.1., MAIN, D., BONNEY, A., ABBASZADEGAN, M.R., HARTY, A.E., FRANK­LIN, B.A., L YNCH, J.F. , and LYNCH, H.T. ( 1999). Genetic testing in families wi th hereditary nonpolyposiscolon cancer. J. Am. Med.

Assn. 281, 16 18- 1622. LOADER, S., LEVENKRON, J.C., ""d ROWLEY, P.T. (1998). Ge­

netic testin g fo r breast-ovarian cancer susceptibility: a reg ional trial. Gene t. Test. 2, 305- 3 13.

LOU KOLA, A., DE LA CHAPELLE, A., and AALTO NEN, L. A. ( 1999). Strategies for screenin g fo r hereditary non-pol yposis co l­orectal cancer. J. Med. Genet. 36, 8 19- 822 .

LOU KOLA, A., EKLIN, K. , L AIHO, P. , SALOVAARA, R., KRISTO, P. , JARVINEN, H., MECKLlN, J.P. , LAUNONEN, V., and AA L­TONEN, L.A. (2001 ). Microsatellite marker analysis in screening fo r hereditary nonpolyposisco lorectal cancer (HNPCC) . Cancer Res. 61, 4545-4549.

LYNCH, H.T., WATSON, P. , SHAW, T.G., LYNCH, J.F., HARTY, A.E., FRANKLIN, B.A., KAPLER, C.R., TINLEY, S.T., Ll U, B., and LER MAN, C. ( 1999) . Clinical impact of molecu lar genetic di­agnosis, genetic co unseling, and management of hereditary cancer. Part 11: Hereditary nonpol yposis co lorectal carcinoma as a model. Cancer Suppl. 86, 2457- 2463 .

MECKLlN, J.P., JARVINEN, H.1., HAKKILUOTO, A., HALLIKAS, H. , HILTUNEN, K.M., HARKONEN, N., KELLOKUMPU, I. , L AITI NEN, S., OVASKA, J., and TULIKOURA, J. ( 1995). Fre­quency of hereditary nonpolyposiscolorectal cancer. A prospecti ve multicenter stud y in Finland. Dis . Colon Rectum 38, 588- 593.

M ILLER, S.M. ( 1987). Monitoring and blunting: validation of a ques­tionnaire to assess styles of information seeking under threat. J. Per­so nality Soc. Psychol. 53, 345- 353 .

PEEL, D.J. , Z IOGAS, A., FOX, E.A., GILDEA, M. , L AHAM, B. , C LEMENTS, E., KOLODNER, R.D., and ANTON-CUL VER, H. (2000). Characte ri zation of hereditary nonpol yposiscolorectal can­cer families from a population-based seri es of cases . J. Natl. Cancer Inst. 92,151 7- 1522.

LOADER ET AL.

SALO VAARA, R., LOUKOLA, A., KRISTO, P. , KAARIAINEN, H., AHTOLA, H., ESKELINEN, M., HARKONEN, N., JULKUNEN, R. , KANGAS, E. , OJALA, S., TULIKOURA, J., VALKAMO, E. , JARVINEN, H. , MECKLlN, J.-P., AALTONEN, L.A., and DE LA CHAPELLE, A. (2000). Popu lation-based molecular de tection of hereditary nonpolyposis co lorectal cancer. J. Clin. Onco l. 18, 2 193- 2200.

SARASON, I.G., LEVINE, H. M., BASHAM, R.B. , and SARASON, B.R. ( 1983). Assessing social support: The Social Support Ques­tionnaire. J. Personalit y Soc . Psychol. 44,1 27- 139.

SHIELDS, C., FRANKS, P., HARP, J., MCDAN IEL, S., and CAMP­BELL, T. ( 1992). Development of the family emotional involvement and criticism scale (FEICS): a self report scale to measure exp ressed emo tion. J. Marital Family Ther. 18, 395-407 .

S YNGAL, S., WEEKS, J .c., SCHRAG, D. , GARBER, J .E., and KUNTZ, K.M. ( 1998). Benefit s of colonoscopic surveillance and prophylac tic colectomy in patients with hereditary colorectal cancer mu tati ons. Ann. Int. Med. 129, 787- 796 .

SYNGAL, S., FOX, E.A., L1 , C., DOVIDIO, M., ENG, C, KOLOD­NER, R.D., and GARBER, J .E. ( 1999). Interpre tati on of genetic test results fo r hereditary nonpolyposiscolorectal cancer: implications fo r clinical predisposition testing. J. Am. Med. Assn. 282, 247- 253 .

SYNGAL, S., FOX, E.A., ENG, c., KOLODNER, R.D., and GAR­BER, J.E. (2000). Sensitivi ty and specific ity of clinical c rite ria fo r hereditary non -polyposis colorectal cancer associated mutation s in MS H2 and M LH I. J. Med. Genet. 37, 64 1-645.

VERNON, SW., GRITZ, E.R., PETERSON, S.K., AMOS, C. I. , PERZ, C .A., BAILE, W.F., and LYNCH, P.M. ( 1997). Co rrelates of psy­chologic di stress in colorectal cancer patient s undergoin g gene tic testing for hereditary co lon cancer. Health Psychol. 16, 73-86.

VERNON, S.W., GRITZ, E.R., PETERSON, S.K., PERZ, CA., MARANI, S., AMOS, C. I., and BAILE, W.F. ( 1999). Intention to learn results of genetic testing fo r hereditary colon cancer. Cancer

Epidemiol. Biomarkers Prev o 8, 353- 360. WEBER, TK, CIflN, W.-M., RODRIGUEZ-BIGAS, M., KEITZ, B.,

GILLIGAN, R., O'MALLEY, L. , URF, E., DlBA, N ., PAZ1K, J., and PETRELLI , N.J. ( 1999). Novel hMLH I and hMS H2 germline mu tati ons in African American s with colo rectal cancer. J. Am. Med. Assn. 281, 23 16- 2320.

WEBER, TK, YUAN, Z., MCDERMOTT, D., PETRELLI, N ., RODRIGUEZ-BIGAS, M. , KEITZ, B., and RAVIKUMAR, T.S. (2001 ). Functional significance of a novel H7 18 Y hMLH I alte rati on associated with increased colorectal cancer ri sk among African­Americans. Am. J. Hum. Genet. 169, 255.

YUAN, A., WONG, N., FOULKES, W.D., ALBERT, L., MANGA­NARO, F., ANDREUTTI-ZAUGG, C., IGGO, R., ANTHONY, K., HSIEH, E., REDSTON, M., PINS KY, L. , TRIFIRO, M. , GORDON, P.H. , and LASKO, D. ( 1999). A missense mutation in both MS H2 and APC in an Ashkenazi Jewish HNPCC kindred : implication fo r clinical screenin g. J. Med. Genet. 36, 790--793.

Address repr int requests to : Peter T. Rowley, M. D.

Divisiol1 oj Genetics, Box 641 University oj Rochester School oj Medicine

Rochester, NY 14642

E-mail ." peter_rowl ey@urmc .lOCheste r. edu

Received for publication March 28, 2002; accepted August 2, 2002 .

This article has been cited by:

1. N . Dekker, R. P. M. G. Hermens, A. R. Mensenkamp, W. A. G. van Zelst-Stams, N . Hoogerbrugge. 2014. Easy-to-use online referral test detects most patients with a high familial risk of colorectal cancer. Colorectal Disease 16:1, 026-034. [CrossRef]

2. N. Dekker, R. P. M. G. Hermens, F. M. Nagengast, W. A. G. van Zelst-Stams, N. Hoogerbrugge. 2013. Familial colorectal cancer risk assessment needs improvement for more effective cancer preve ntion in relatives. Colorectal Disease 15:4, eI75-el85. [CrossRef]

3. Kathy E Watkins, Christine Y Way, Deborah M Gregory, Holly M LeDrew, Valerie C Ludlow, Mary Esplen, Jeffrey J Dowden, Janet E Cox, G William N Fitzgerald, Patrick S Parfrey. 2013. Deve lopment and preliminary testing of the psychosocial adjustment to hereditary diseases scale. BMC Psychology 1:1, 7. [CrossRef]

4. Deborah Cragun, Teri L. Malo, Tuya Pal, David Shibata, Susan T. Vadaparampil. 2012. Colorectal Cancer Survivors' Interest in Genetic Testing for Hereditary Cancer: Implications for V niversal Tumor Screening. Genetic Testing and Molecular Biomarkers 16:6,493-499. [Abstract] [Full Text HTML] [Full Text PDF] [Full Text PDF with Links]

5. Susan T. Vadaparampil, Gwendolyn Quinn, Teri L. Malo, Angelic Alvarez, Cheryl A. Miree, David Shibata. 2010. Knowledge About HereditaryColorectal Cancer Among Colorectal Cancer Survivors. Genetic Testing and Molecular Biomarkers 14:5, 603-609. [Abstract] [Full Text HTML] [Full Text PDF] [Full Text PDF with Links]

6. C. Carlsson, M. N ilbert . 2007. Living with Hereditary Non-polyposis Colorectal Cancer; Experiences from and Impact of Genetic Testing. Journal of Genetic Counseling 16:6, 811-820. [CrossRef]

7. Starlene Loader, Cleveland Shields, Peter T. Rowley. 2005. Impact of Genetic Counseling and DNA Testing on Individuals with Colorectal Cancer with a Positive Family History: A Population-Based Study. Genetic Testing 9:4, 313-319. [Abstract] [Full Text PDF] [Full Text PDF with Links]

8. Kirsty Challen, Hilary J. Harris, Claire Julian-Reynier, Leo P. ten Kate, Vlf Kristoffersson, Irmgard Nippert, J??rg Schmidtke, Caroline Benjamin, Rodney Harris. 2005. Genetic education and nongenetic health professionals: Educational providers and curricula in Europe. Genetics in Medicine 7:5, 302-310. [CrossRef]

9. Peter T. Rowley. 2005. Inherited Susceptibility to Colorectal Cancer. Annual Review of Medicine 56:1, 539-554. [CrossRef]

10. Peter T. Rowley. 2004. Screening for an Inherited Susceptibility to Colorectal Cancer. Genetic Testing 8:4, 421-430. [Abstract] [Full Text PDF] [Full Text PDF with Links]