|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Originally published as JCO Early Release 10.1200/JCO.2006.07.2462 on November 6 2006 © 2006 American Society of Clinical Oncology. Validation of a Self-Administered, Computerized Tool for Collecting and Displaying the Family History of Cancer
From the Department of Family Medicine, Comprehensive Cancer Center, Department of Reproductive Biology, Department of Epidemiology and Biostatistics, Department of Sociology, Department of Genetics, and Center for Human Genetics, Case Western Reserve University; and University Hospitals of Cleveland, Cleveland, OH Address reprint requests to Louise S. Acheson, MD, MS, Department of Family Medicine, Case Western Reserve University and University Hospitals of Cleveland LC 5036, 11100 Euclid Ave, Cleveland, OH 44106-5036; e-mail: louise.acheson{at}case.edu
PURPOSE: A detailed family history is important for cancer risk assessment, but obtaining it is time consuming and infrequently accomplished in practice. The Genetic Risk Easy Assessment Tool (GREAT) conducts a computer-administered family history interview and immediately generates a pedigree diagram in digital form. The purpose of this study was to validate family cancer histories produced by patients using the computer tool in comparison with pedigrees made by genetic counselors. METHODS: Patients scheduled for genetics consultation recorded their family histories using the GREAT, separate from their genetic counseling session. The presence of each relative; presence, type, and age at diagnosis of cancers; and cancer geneticist's risk assessment were compared for 120 pairs of pedigrees produced by counselors versus computer tool.
RESULTS: The automated telephone interview took a mean of 33.5 minutes and was highly acceptable to respondents. Ninety-four percent of first-degree relatives, 67% of second-degree relatives, and 38% of third-degree relatives were identical on paired pedigrees; computer-generated pedigrees included additional relatives. Sixty-three percent of all cancers were identified by both family histories, with 90% agreement on the type of cancer. There was very good agreement ( CONCLUSION: The GREAT computer-administered questionnaire provides an acceptable, reliable, and valid way of collecting an unverified but extensive family history of cancer and displaying it as a pedigree, in an entirely automated process.
As knowledge of genetic susceptibility to cancers increases, family history can be used to recognize families with hereditary cancer susceptibility, to tailor cancer screening and prevention, and to screen for eligibility to participate in cancer genetic research.1-5 A family history of cancer is among the genetic issues most commonly encountered in primary health care of adults.6-11 The US Preventive Services Task Force recommends referring women for genetic counseling if their family history is consistent with increased risk of hereditary breast or ovarian cancer.3 The American Cancer Society and the American Society of Clinical Oncology recommend recording a three-generation family tree as part of cancer risk assessment for everyone.12,13 However, obtaining a detailed family history of cancer is time consuming and labor intensive, which has made it impractical for the majority of physicians. Collecting the information to draw a pedigree takes 15 to 30 minutes or more,14,15 longer than the average face-to-face time for the doctor and primary care patient.16-18 In clinical practice, family history documentation is low; pedigrees, which allow rapid recognition of family cancer syndromes, are uncommon. In direct observation of 4,454 patients seen by family physicians, half of new patient visits and 22% of visits by established patients involved gathering family history information.19,20 However, although 37% of the medical records contained some notation of positive or negative family history of breast and colon cancers, pedigree diagrams were present in the medical records of only 11% of patients.19 Through systematic family history taking, Frezzo et al11 found that 9% of patients were at increased familial risk for cancers; for most, the risk had not been recognized previously by their internists. Others have found that even for patients with cancer, only 36% of medical records contained accurate family history information21 and only 14% noted any familial cancer risk assessment.22 Lacking systematic ascertainment, cancer risk counseling and management based on family history remain underused. Digital communication promises efficient means for patient-initiated family history collection or assessment.23-30 The Genetic Risk Easy Assessment Tool (GREAT) uses informatics technology to enable patients to record their detailed family history of cancer using a computer-administered questionnaire, quickly producing a pedigree in digital form. However, the validity and reliability of such systems have not been widely reported. Therefore, this study was conducted to validate the family cancer histories and pedigrees produced by the GREAT, by comparing them with the information recorded by genetic counselors in face-to-face interviews during genetics consultations.
Development of the Computer-Administered Questionnaire The GREAT computer-administered family history questionnaire prompted users to enumerate first-degree relatives (FDRs) and second-degree relatives (2DRs) and first cousins. For any family member with cancer, it recorded the primary site(s) (24 types of cancer) and age(s) at diagnosis. The automated interview adapted to the user's responses. The digital pedigree unambiguously denoted twins and triplets, half-siblings, aunts, uncles, and first cousins (Fig 1). More distant relatives were shown only if they had cancer. In addition to the family history, the questionnaire collected information about personal risk factors for cancer, such as smoking, ethnicity, colon polyps, reproductive history, and breast biopsy results, needed for application of risk prediction models (for example, CancerGene, Version 4.0; University of Texas Southwestern Medical Center, Dallas, TX).31,32 Confidentiality was maintained by assigning each respondent an identification code that was the only personal identifier collected.
The family history interview was pilot tested by being read over the telephone to several patients before their visit to the Cancer Genetics Clinic at the Center for Human Genetics. Then the script was revised, programmed, and recorded to produce an automated telephone interview. The computerized system was tested using more than 50 large pedigrees entered by the researchers, and then with 74 students and lay people who recorded their own family cancer histories. Participants in this study called a toll-free telephone number, heard a recorded voice, and answered questions by pressing buttons on the telephone keypad.
Sample At the beginning of this study, standards for recording the family history of cancer during all genetics consultations (whatever their purpose) were agreed on by the cancer geneticist (G.L.W.) and all genetic counselors who obtained the pedigrees. Genetic counselors drew pedigrees by hand during face-to-face consultations. This protocol was approved by the University Hospitals of Cleveland Institutional Review Board for Protection of Human Subjects.
Data Collection
Data Analyses Planned subgroup analyses—comparing outcomes by reason for consultation (cancer genetic risk assessment v prenatal and pediatric) and order in which the family histories were recorded—were performed using unpaired t tests. We hypothesized that agreement between the computer and the counselors would be higher for the cancer consultations. We hypothesized that the first interview would prepare respondents to find out more information about their family history, resulting in a tendency for the second interview to provide more complete data and fewer unknown responses.
The genetic counselors' hand-drawn pedigrees were transcribed by trained research assistants into a computer format similar to the GREAT pedigrees (Progeny Software, South Bend, IL; www.Progeny2000.com). Batches of digital pedigrees in random order were reviewed by a cancer geneticist (G.L.W.) for a clinical interpretation. The geneticist attempted on the basis of the pedigree to stratify the respondent's familial risk of breast, colorectal, and other cancers into categories of average, moderately increased, or possible high-risk family with a hereditary cancer pattern. The geneticist's interpretation was compared for the counselors' and the computer-generated pedigrees, using the
Test-retest reliability of the GREAT questionnaire was measured by comparing the presence and cancer diagnoses of FDRs and 2DRs in the initial pedigrees versus repeat pedigrees recorded by the same patient. The Statistical analyses were performed using SPSS versions 10 to 13 (Statistical Package for the Social Sciences; SPSS, Chicago, IL). We calculated that a sample of 130 pedigrees would provide 90% power to detect a 10% difference between the GREAT questionnaire and the counselors' interviews in mean numbers of relatives or cancers per family, with 95% confidence (NQuerry Advisor, Version 4.0; Statistical Solutions Ltd, Los Angeles, CA).
Study Participants Twenty percent of eligible Center for Human Genetics outpatients (151 of 755) used the GREAT system. Only two respondents did not complete the interview. Participants and nonparticipants did not differ by age or sex. Patients consulting for cancer risk assessment were more likely to participate in the study than patients with other indications for genetics consultation (26% v 18% participated). Table 1 shows the characteristics of participants and their evaluations of the computer questionnaire. For most participants, the GREAT was acceptable and easy to understand. The time and duration of the telephone call were known for 77 participants. Eighty-one percent of calls to complete the computer questionnaire occurred after hours (ie, not between 8 AM and 5 PM on Monday to Friday). The mean duration was 33.6 minutes (range, 8 to 55 minutes). Pedigrees ranged in size from six to 108 members, with a mean of two parents, four grandparents, 2.0 children, 2.9 siblings, 6.7 aunts and uncles, and 14.4 first cousins.
Comparisons of Family Histories Recorded by Patients Using the Computer Tool With Those Obtained by Genetic Counselors Of the 149 participants who used the GREAT, a matched pedigree from a genetic counselor was not available for 29 participants because 19 patients did not keep their appointments at the Genetics Center and 10 consultation records were not located. For 120 participants, pedigrees from the GREAT were compared with pedigrees from the genetic counselor. Thirty-nine percent (n = 47) completed the GREAT questionnaire before and 32% (n = 39) after their genetic counseling session. The order could not be determined with certainty for 34 participants because of incomplete data on time of the GREAT questionnaire. The proportion of data in agreement did not differ depending on the order in which counselors' or computer-generated pedigrees were obtained. Pedigrees produced second did not identify significantly more family members (t = 0.72; P = .48) or relatives with cancer (t = 0.43; P = .67) than those from the first interview. Therefore, GREAT pedigrees obtained before and after genetic counseling were combined for additional analyses.
Family Composition
Cancers in the Family Tables 4 and 5 show the number of cancers (excluding nonmelanoma skin cancers) shown on GREAT pedigrees compared with the genetic counselors' pedigrees. Overall, 63% of 707 cancers were shown on both pedigrees (Table 4); 21% were recorded only using the computer tool and 16% were recorded only by the genetic counselors. For 445 cancers reported by both, the primary site agreed for 97% of FDRs, 86% of 2DRs, and 64% of third-degree relatives (90% overall; data not shown). When there was disagreement, most commonly the GREAT identified the cancer as primary bone, brain, or liver cancer, whereas the counselor showed it as metastatic from some other primary site. When age at diagnosis was recorded on both pedigrees, it agreed within 5 years for 83% of cancers (data not shown). The age at diagnosis was recorded substantially more often on computer pedigrees (74% of cancers) than by the genetic counselor (59%), except for cancer genetic consultations, for which 84% of cancers recorded by counselors showed the age at diagnosis.
In the subgroup of participants consulting for cancer genetics (Table 5), computer and genetic counselors identified similar proportions of the cancers in FDRs (85% v 91%, P = .31), but genetic counselors recorded a significantly higher proportion of the cancers in 2DRs (P = .04), third-degree relatives (P = .03), and distant relatives (P = .04).
Geneticist's Pedigree Interpretation
Test-Retest Reliability Sixty-one participants completed the GREAT questionnaire twice, 24 hours to 2 weeks later. Table 7 shows that test-retest reliability was high, with 97% of FDRs, 93% of 2DRs, and 98% of cancers identical on initial and repeat testing. The type of cancer disagreed in only one case (0.3%), whereas for 37 cancers (10%) the primary site was unknown in one iteration of the GREAT questionnaire, but was specified in the other.
The GREAT provided an acceptable, valid, and reliable means for collecting a detailed family history of cancer and displaying it as a pedigree, in an entirely automated process. This self-administered tool enables an initial family history to be recorded outside the time of patients' visits with clinicians, providing access to more extensive family history than would otherwise be feasible in many settings. This method can potentially increase the efficiency of genetic cancer risk assessment. Our data suggest that the computer-administered GREAT questionnaire and a genetic counselor's interview have complementary strengths. The computer tool collects the family cancer history with systematic thoroughness, whereas an expert interviewer can focus selectively on aspects of the history, and verify potentially significant data. For example, the computer tool recorded many more unaffected cousins than did counselors, albeit with little effect on the familial cancer risk assessment; conversely, genetic counselors recognized metastatic cancer in several cases when patients using the GREAT reported bone, brain, or liver cancer. Limited data are available on the reliability and validity of a family history of cancer. The population-based Connecticut Family Health Study33 found that respondents reported not knowing whether a relative had cancer for 1% of FDRs and 6% of 2DRs, especially on the father's side. The primary site of cancer was unknown for approximately 20% of affected relatives, especially older, more distant relatives. Validation against medical records and relatives' histories showed that family history of cancer collected by personal interviews is highly sensitive for FDRs (eg, 95% for breast, 83% for ovarian, 90% for colorectal, and 79% for prostate cancer) but less so for 2DRs (eg, 82% for breast, 44% for ovarian, 58% for colorectal, and 67% for prostate cancer).34 In both clinical and research settings, it is essential that the computer-generated family history be verified. The GREAT questionnaire allowed for updating whenever new information was obtained. This study was conducted with a sample of volunteers consulting a geneticist; these participants are likely to be more motivated and knowledgeable than average about their family histories. Almost all of the adult consultands (and 93% of our respondents) were women. The sample contained limited variation in age, ethnicity, and educational level. Caution should be used in generalizing this validation study to different populations and to different modes of self-administering the computerized questionnaire (eg, Internet v telephone). Given that 36% of a population-based sample35 had no FDR or 2DR with cancer, a preliminary question about relatives with cancer is desirable to select people who may benefit from recording a detailed family cancer history. The GREAT has a unique combination of features compared with other digital family history tools being developed. 23,36-44 As far as we are aware, it is the only one designed to be self-administered by a layperson, to obtain a precise, three- to four-generation family history of many types of cancer, to automatically display a pedigree, and to allow family history updates. Additional automated processes based on the family history can include risk assessment and decision support, as pioneered by several groups.24,28,29,38,39,44,45 Several digital family history tools collect information about a variety of diseases in addition to cancers, but record less extensive family information to assess hereditary cancer risk.37-40 Ideally, once recorded, family history should be portable, private, yet accessible in various electronic medical record systems.46 Secure Internet technology can facilitate this goal, especially once there is widespread adoption of standard data formats for digitally communicating family history and cancer risk information.30 In response to our study participants' preference, we have recently programmed a computer questionnaire using the GREAT script in a Web-based format. The Web-based family history questionnaire can interface with cancer risk calculations, databases, and tailored prevention messages to automate fully the initial process of familial cancer risk assessment. In summary, more evidence is needed about the effects of using informatics to make familial cancer risk assessment more available. The GREAT may be efficient in primary care to focus families on risk-appropriate cancer screening and primary prevention efforts. In oncology practice it may prompt cancer prevention recommendations for family members. It could help to identify high-risk families for genetic cancer susceptibility counseling and testing, and may streamline genetics consultations by collecting family history beforehand. Researchers might also use this tool to find families eligible for genetic epidemiology and gene-finding research. We plan clinical studies to measure the effects of the automated family history questionnaire for patients, families, clinicians, and researchers.
The author or immediate family members indicated a financial interest. No conflict exists for drugs or devices used in a study if they are not being evaluated as part of the investigation. For a detailed description of the disclosure categories, or for more information about ASCO’s conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.
Katherine Lynch, MS, Deborah Fruit, and Elizabeth Ruff contributed to data collection and data entry. Willa Eisele helped with manuscript preparation. Mike Brammer at Progeny Software LLC, and the members of DFR Group Inc, programmed, tested, and operated the computerized telephone system for this research.
published online ahead of print at www.jco.org on November 6, 2006. Supported by the American Cancer Society, Cuyahoga Branch, Pilot Research (L.S.A. and G.L.W.); the Behavioral Measurement Core of the Case Comprehensive Cancer Center (S.J.Z.); the National Cancer Institute: K-12 postdoctoral fellowship (L.S.A.); K07 career development award CA86958 (L.S.A.); K24 mid-career development award (K.C.S.); an HRSA Development of Departments of Family Medicine training grant; and the Center for Human Genetics at University Hospitals of Cleveland. Presented in part at the American Society of Human Genetics Annual Meeting, October 17, 2002, Baltimore, MD; and American College of Medical Genetics Annual Meeting, March 13-17, 2002, New Orleans, LA. Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
1. Kutner SE: Breast cancer genetics and managed care. Cancer 86:2570-2574, 1999 (suppl)[CrossRef][Medline] 2. Burke W: Taking family history seriously. Ann Intern Med 143:388-389, 2005 3. US Preventive Services Task Force: Genetic risk assessment and BRCA mutation testing for breast and ovarian cancer susceptibility: Recommendation and statement. Ann Intern Med 143:355-361, 2005 4. Whittemore AS: Genetically tailored preventive strategies: An effective plan for the twenty-first century? Cancer Epidemiol Biomarkers Prev 8:649-658, 1999 5. Rubinstein W, O'Neill S, Peters JA, et al: Mathematical modeling for breast cancer risk assessment: State of the art and role in medicine. Oncology (Williston Park) 16:1082-1094, 2002[Medline] 6. Acheson L, Stange K, Zyzanski S: Clinical genetics issues encountered by family physicians. Genet Med 7:501-508, 2005[Medline] 7. Carroll J, Brown J, Blaine S, et al: Genetic susceptibility to cancer: Family physicians' experience. Can Fam Physician 49:45-52, 2003 8. Emery J, Watson E, Rose P, et al: A systematic review of the literature exploring the role of primary care in genetic services. Fam Pract 16:426-445, 1999 9. House W, Sharp D, Sheridan E: Identifying and screening patients at high risk of colorectal cancer in general practice. J Med Screen 6:205-208, 1999 10. Leggatt V, MacKay J, Yates J: Evaluation of questionnaire on cancer family history in identifying patients at increased risk in general practice. BMJ 319:757-758, 1999 11. Frezzo TM, Rubenstein WS, Dunham D, et al: The genetic family history as a risk assessment tool in internal medicine. Genet Med 2003:584-91, 2003 12. Gould RL, Lynch HT, Smith RA, et al: Cancer and Genetics: Answering Your Patient's Questions: A Manual for clinicians and their patients. Huntington, NY, American Cancer Society, 1997 13. American Society of Clinical Oncology. Cancer Genetics and Cancer Predisposition Testing. Alexandria, VA, American Society of Clinical Oncology; 1998 14. Rogers J, Durkin M: The semi-structured genogram interview: I. Protocol, II—Evaluation. Family Systems Medicine 2:176-188, 1984 15. Rogers JC, Rohrbaugh M: The SAGE-PAGE trial: Do family genograms make a difference? J Am Board Fam Pract 4:319-326, 1991[Medline] 16. Stange KC, Zyzanski SJ, Jaen CR, et al: Illuminating the ‘Black Box’: A description of 4454 patient visits to 138 family physicians. J Fam Pract 46:377-389, 1998[Medline] 17. Mechanic D, McAlpine D, Rosenthal M: Are patients' office visits with physicians getting shorter? N Engl J Med 344:198-204, 2001 18. Yarnall K, Pollack K, Ostbye T, et al: Primary care: Is there enough time for prevention? Am J Public Health 93:635-641, 2003 19. Medalie JH, Zyzanski SJ, Langa D, et al: The family in family practice: Is it a reality? J Fam Pract 46:390-396, 1998[Medline] 20. Acheson LS, Wiesner GL, Zyzanski SJ, et al: Family history-taking in community family practice: Implications for genetic screening. Genet Med 2:180-185, 2000[Medline] 21. Church J, McGannon E: Family history of colorectal cancer: How often and how accurately is it recorded? Dis Colon Rectum 43:1540-1544, 2000[CrossRef][Medline] 22. Sweet K, Bradley T, Westman J: Identification and referral of families at high risk for cancer susceptibility. J Clin Oncol 20:528-537, 2002 23. Emery J, Walton R, Coulson A, et al: Computer support for recording and interpreting family histories of breast and ovarian cancer in primary care (RAGS): Qualitative evaluation with simulated patients. BMJ 319:32-36, 1999 24. Emery J, Walton R, Murphy M, et al: Computer support for interpreting family histories of breast and ovarian cancer in primary care: Comparative study with simulated cases. BMJ 321:28-32, 2000 25. Westman J, Hampel H, Bradley T: Efficacy of a touchscreen computer based family cancer history questionnaire and subsequent cancer risk assessment. J Med Genet 37:354-360, 2000 26. Acheson L: Fostering applications of genetics in primary care: What will it take? Genet Med 5:63-65, 2003[Medline] 27. Hampel H, Sweet K, Westman J, et al: Referral for cancer genetics consultation: A review and compilation of risk assessment criteria. J Med Genet 41:81-91, 2004 28. Emmons K, Wong M, Puleo E, et al: Tailored computer-based cancer risk communication: Correcting colorectal cancer risk perception. J Health Commun 9:127-141, 2004[Medline] 29. Emery J: The GRAIDS Trial: The development and evaluation of computer decision support for cancer genetic risk assessment in primary care. Ann Hum Biol 32:218-227, 2005[CrossRef][Medline] 30. Shabo A, Hughes K: Family history information exchange services using HL7 clinical genomics standard specifications. Int J Semantic Web Information Syst 2005:142-65, 2005 31. Gail MH, Brinton LA, Byar DP, et al: Projecting individualized probabilities of developing breast cancer for white females who are being examined annually. J Natl Cancer Inst 81:1879-1886, 1989 32. Euhus D: Understanding mathematical models for breast cancer risk assessment and counseling. Breast J 7:224-232, 2001[CrossRef][Medline] 33. Wideroff L: The Connecticut family history of cancer validation study: Methodologic and analytic issues. Presented at Family History Tools to Improve the Public's Health, Atlanta, GA, November 17, 2005 34. Ziogas A, Anton-Culver H: Validation of family history data in cancer family registries. Am J Prev Med 24:190-198, 2003[CrossRef][Medline] 35. Grant D: Developing a family history module—California Health Interview Survey. Presented at Family History Tools to Improve the Public's Health, Atlanta, GA, November 17, 2005 36. Sweet K: Cancer risk assessment and pattern recognition—JamesLink. Presented at Family History Tools to Improve the Public's Health, Atlanta, GA, November 17, 2005 37. Hampton T: Family health history. JAMA 294:3075, 2005 38. Stein C: Evaluating prevention messages—Your Disease risk. Presented at Family History Tools to Improve the Public's Health, Atlanta, GA, November 17, 2005 39. Jorgensen C: Developing and testing health messages—Family Healthware. Presented at Family History Tools to Improve the Public's Health, Atlanta, GA, November 17, 2005 40. Cohn W: Health Heritage (HH) a Web-based family health history tool. Presented at Family History Tools to Improve the Public's Health, Atlanta, GA, November 17, 2005 41. Braithwaite D, Sutton S, Mackay J, et al: Development of a risk assessment tool for women with a family history of breast cancer. Cancer Detect Prev 29:433-439, 2005[CrossRef][Medline] 42. O'Neill S: Communicating familial risk—MyGenerations. Presented at Family History Tools to Improve the Public's Health, Atlanta, GA, November 17, 2005 43. Hughes K: Breast/ovarian cancer risk assessment tool (AVON). Presented at Family History Tools to Improve the Public's Health, Atlanta, GA, November 17, 2005 44. Euhus D, Smith K, Robinson L, et al: Pretest prediction of BRCA1 or BRCA2 mutation by risk counselors and the computer model BRCAPRO. J Natl Cancer Inst 94:844-851, 2002 45. Scheuner MT, Wang SJ, Raffel LJ, et al: Family history: A comprehensive genetic risk assessment method for the chronic conditions of adulthood. Am J Med Genet 71:315-324, 1997[CrossRef][Medline] 46. Rich E, Burke W, Heaton C, et al: Reconsidering the family history in primary care. J Gen Intern Med 19:273-280, 2004[CrossRef][Medline] Submitted May 1, 2006; accepted September 25, 2006.
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2006 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
= 0.70; correlation = 0.77) between the geneticist's breast cancer risk assessments based on computer versus counselors' pedigrees. In a subsample of 61 users, test-retest reliability for the computer-administered questionnaire was high (
= 0.94 for cancers in first-degree and 
for correlation of ordinal data. 
