Journal of Genetic Counseling, Vol. 8, No. 5, 1999
Are General Practitioners Willing and Able to Provide Genetic Services for Common Diseases? Marc A. Suchard,1,2 Patricia Yudkin,3 and Janet S. Sinsheimer124
Primary care in the United Kingdom has been advocated as an optimal location for the provision of genetic services for common diseases. Little, however, is known about general practitioners' own views toward this suggestion or the possible demand for such services from patients. To assess general practitioners' attitudes to providing genetic services for common diseases, and to estimate the demand from patients for these services, we used a single-page postal questionnaire survey of all 359 general practitioners registered with the Oxfordshire Health Authority; 339 (94%) responded. These physicians reported that a mean 4.1 patients (95% CI, 3.3—4.9) out of every 1,000 consulting them were concerned about their own risk of a common disease associated with a diagnosis in a family member. Half of the general practitioners (95% CI, 45—56%) counseled about genetics in the last year. A majority of general practitioners accept the need to provide genetic services, but far fewer are competent to do so. Although 60% (95% CI, 55-65%) believed they should be involved with genetic screening for common diseases, only 29% (95% CI, 24-34%) felt sufficiently prepared to take family histories and draw pedigrees, and only 15% (95% CI, 11-19%) felt sufficiently prepared to counsel patients about their genetic test results. Given the necessary training and information, 63% (95% CI, 58-68%) and 64% (95% CI, 59-69%) were willing to provide these services. Even with training and information, not all would
1 Wellcome
Trust Centre for Human Genetics, University of Oxford, Windmill Road, Headington, OX3 7BN, United Kingdom. 2Present affiliation: Department of Biomathematics, UCLA Medical School, Los Angeles, CA. 3Imperial Cancer Research Fund General Practice Research Group, Division of Public Health and Primary Health Care, Institute of Health Sciences, University of Oxford, Oxford, OX3 7LF, United Kingdom. 4Correspondence should be directed to Janet S. Sinsheimer, Department of Biomathematics, UCLA Medical School, Los Angeles, CA 90095-1766; e-mail: [email protected].
301 1059-7700/99/0800-0301516.00/1 © 1999 National Society of Genetic Counselors, Inc.
302
Suchard, Yudkin, and Sinsheimer
be willing to provide these services, and lack of time may be a major deterrent. "Practice-enabling" strategies, such as computerized aids in genetics, may be useful. KEY WORDS: Genetic screening; primary care; common diseases.
INTRODUCTION Advances in human genetics are making possible predictive testing for genetic susceptibility to many common diseases (Genetics Research Advisory Group, 1995), such as Alzheimer's disease, breast cancer, and cardiovascular disease. Leading genetics enthusiasts and several key health policy reports in the United Kingdom have advocated primary care as an ideal location for the provision of these genetic services (Welsh Health Planning Forum, 1995; Harper, 1995; Harris and Harris, 1996; Kinmonth et al., 1998). This recommendation is, partially, prompted by a foreseen shortfall in the number of genetic specialists. Currently, in the United Kingdom, there are only one or two consultant geneticists per million people (Kinmonth et al., 1998). Additionally, Bell (1998) has conjectured that pressure for genetic screening from patients is likely to increase steadily in the coming years, far outpacing the resources of genetic specialists. In the United Kingdom, general practitioners provide the bulk of primary care, acting as gatekeepers to all further specialties. Possible roles for general practitioners, as genetic screening for common diseases moves to general practice, include making initial contact with concerned patients, taking family histories to assess risk, referring patients more frequently for genetic testing and counseling, and counseling patients directly about their genetic risks (Qureshi and Raeburn, 1993; Hoskins et al., 1995). However, there has been little attempt to assess general practitioners' own views toward these postulated changes. Specifically, we do not know how much demand general practitioners perceive for these services from patients, general practitioners' attitudes toward involvement, or their ability to take family histories and give appropriate advice. We conducted a survey to better understand general practitioners' views about providing genetic screening for common diseases like Alzheimer's disease, breast cancer, and cardiovascular disease. SUBJECTS AND METHODS We developed a single-page questionnaire with input from a local general practitioner and a health psychologist. We piloted the questionnaire on 20 randomly selected general practitioners from Berkshire, a neighboring county, and revised the questionnaire in the light of their comments or if answers appeared ambiguous. The final form (Table I) consisted of 14 questions (attitudinal = 5 and multiple choice = 9) and provided space for open-ended comments. The
GPs Providing Genetic Services
303
Table I. General Practitioners' Views on Genetic Screening for Common Diseases Questionnaire
1. Patients sometimes present family history information about noncommunicable, common diseases
2. 3. 4. 5. 6. 7. 8. 9. 10. 11.
12. 13. 14.
which may run in their families, for example, "I saw a TV program about breast cancer and my mother had breast cancer, so what should I do?" In the last month, how many patients have consulted you concerning their health because of a diagnosis made in a family member? (Number) In the last year, have you referred any patients for genetic testing or ordered a test yourself? (Yes/no) In the last year, have you counseled any patients about their genetic test results? (Yes/no) I believe GPs should be involved with genetic screening for common diseases like Alzheimer's disease, breast cancer, and cardiovascular disease. (Strongly agree to strongly disagree) I feel sufficiently prepared to take family histories from my patients and draw pedigrees. (Strongly agree to strongly disagree) I feel sufficiently prepared to counsel my patients about their genetic test results. (Strongly agree to strongly disagree) Given the necessary training and information, I am willing to take family histories from my patients and draw pedigrees. (Strongly agree to strongly disagree) Given the necessary training and information, 1 am willing to counsel my patients about their genetic test results. (Strongly agree to strongly disagree) I would like to learn more about genetic screening for susceptibility to common diseases. (Strongly agree to strongly disagree) In what year did you qualify? (Number) What is your gender? (Female/male) Have you ever take a course in medical genetics (including medical school)? (Yes/no) How many patients do you see in total during an average week? (Number) Are you interested in participating further in our research about the genetics of common disease in general practice? (Yes/no) If not, why not? (Not enough time/not interested in genetics/no incentive offered/other)
attitudinal questions measured agreement with statements regarding the provision of genetic services in general practice, using 5-point ordinal scales ("strongly agree" to "strongly disagree"). The remaining questions asked about perceived patient demand, past referrals and counseling involving genetics, workload, year of qualification, gender, previous coursework in genetics, and further participation in the project. Between November 1996 and January 1997, we mailed the questionnaire to all 359 general practitioners registered with the Oxfordshire Health Authority on 1 November 1996. We sent two follow-up mailings to nonresponders. To insure accuracy, we double-entered the data. Further, we independently classified all open-ended comments into five subject groups (yes/no). Interrater agreement was close in all cases (k > 0.85) (Landis, 1977). We used SPSS 6.1 for the Macintosh computer (SPSS Inc., 1994) for all statistical analyses. We estimated overall perceived patient demand in two ways. First, we calculated the simple arithmetic mean of the self-reported genetic consultation numbers in the last month. Second, we allowed for imprecise recall of the number of these consultations when greater than zero. Under the latter model, we assumed that the actual number of consultations in the last month was difficult to recall; however, information about the complete absence or presence of such consultations was
304
Suchard, Yudkin, and Sinsheimer
accurate. We then modeled the data as a Poisson distribution using the number of zero versus nonzero reported observations. Let Xi equal the number of genetic consultations in the last month for General Practitioner i. Then,
In general, we made comparisons between means of two groups using the t test and between proportions using the x2 test (with continuity correction). We scored answers to attitudinal questions on a scale of 1 to 5 and compared results between two groups using the Mann-Whitney test and between more than two groups using the Kruskall-Wallis test, correcting for ties in both cases. We made no corrections to p values for multiple testing.
RESULTS The vast majority of general practitioners (339/359 [94%]) returned questionnaires, but not all respondents completed every question on the form. Response rates to the attitudinal questions ranged from 328/339 (97%) to 332/339 (98%); the lowest responses (303/339 [89%] and 316/339 [93%]) were to questions about the number of concerned patients seen in the last month and the total number of patients seen in an average week. Characteristics of Responders Table II outlines the background information collected. Overall, 50% (95% confidence interval [CI], 45-56%) of the general practitioners reported counseling patients about their genetic test results in the last year, and 76% (95% CI, 72-81%) had referred patients for genetic testing or ordered a genetic test themselves. Only 38% (95% CI, 33-43%) of the general practitioners had taken a course in medical genetics. The mean length of time since qualifying for a general practitioner was 19.5 (SD 8.2) years. General practitioners who had taken a course in medical genetics qualified more recently than those who had not. Mean( SD) years since qualification were 15.9 (6.3) years and 21.8 (8.4) years, respectively, a difference of 5.9 years (95% CI, 4.1-7.6, P < 0.001). Frequency of Concerned Patients General practitioners reported that a mean of 2.2 (median 1.5, interquartile range 1-2.5) patients had consulted them in the last month about their risk of a
GPs Providing Genetic Services
305
Table II. Background Information from Responding General Practitioners % (No.) of general practitioners (n = 339) Gender Female Male Referred patient to medical genetics unit in last year Yes No Counseled patient about genetics in last year Yes No Ever taken a course in medical genetics Yes No Year of qualification Before 1972 1972-1977 1978-1983 After 1983
35(116) 65(212) 76(251) 24 (78) 50(162) 50(161) 38(124) 62 (203) 24(80) 25(85) 24(81) 26(88)
common disease associated with a diagnosis in a family member. In total, 62/303 (20%) respondents reported no such consultations in the last month, yielding a Poisson mean frequency of 1.6. Respondents who counseled patients about genetic test results in the last year were more likely to perceive a patient demand higher than the median. Of those with counseling experience, 86/147 (59%) perceived a higher demand compared to 62/148 (42%) without experience (p = 0.006). Demand was independent of the general practitioner's genetic coursework (p = 0.504) and gender (p = 0.595). The mean number of patients seen in an "average" week was 134 (median 130, interquartile range 100-150), giving an estimated 580 consultations per calendar month. We estimated the frequency of concerned patients for each general practitioner. The mean frequency was 4.1 (95% confidence interval 3.3-4.9) per 1,000 consultations. Median frequency was 2.5 (interquartile range 1.2-4.7) per 1,000 consultations. Excluding the four cases in which general practitioners reported seeing more than 250 patients in an average week made little difference to these results.
General Practitioner Attitudes Table III shows the levels of agreement general practitioners reported for each of the attitudinal statements. The majority of general practitioners, 60% (95% CI,
306
Suchard, Yudkin, and Sinsheimer Table III. Attitudes of General Practitioners About Genetic Screening for Common Diseases Statement
Believe in involvement for diseases like Alzheimer's disease, breast cancer, and cardiovascular disease Feel sufficiently prepared to take family histories and draw pedigrees Feel sufficiently prepared to counsel about genetic test results Given the necessary training and information, are willing to take family histories and draw pedigrees Given the necessary training and information, are willing to counsel about genetic test results Want to learn more about genetic screening for susceptibility to common diseases
Strongly agree
Agree
Neutral
10 (32) |50 (165)| 24 (80) 4 (14) I 25 (83) 1 (3)
14 (47)
Disagree
Strongly disagree
(32)
6 (19)
10
22 (72) | 40 (132)| 21 (69)
50 (166)
9 (29) 14 (47)
9 (30) |54 (178)| 19 (62)
14
(47)
4 (13)
9 (30) |55 (182)| 17 (55) I
14
(46)
6 (19)
19 (61)
15 (49) 59(195)|
5 (16)
2 (8)
Notes: Values are Percentages (Numbers) of General Practitioners, Bold and normal boxes highlight median and interquartile ranges, respectively.
55-65%), agreed or strongly agreed they should be involved with genetic screening for common diseases, and 78% (95% CI, 73-82%) wanted to learn more about genetic screening. However, only 29% (95% CI, 24-34%) felt sufficiently prepared to take family histories and draw pedigrees, and only 15% (95% CI, 11-19%) felt sufficiently prepared to counsel patients about their genetic test results. General practitioners who had referred or counseled patients were only slightly more prepared. Of those respondents who had referred to a medical genetics unit in the last year, only 79/246 (32%) felt prepared to take family histories, and of those who had counseled about genetics, only 33/160 (21%) felt prepared to do so, indicating that exposure alone does not sufficiently increase preparedness. Given the necessary training and information, 63% (95% CI, 58-68%) and 64% (95% CI, 59-69%) of the whole sample were willing to provide these services.
Willingness to Provide Genetic Services Willingness to take family histories and draw pedigrees and to counsel patients about their genetic test results was greater among those general practitioners who reported having counseled in the last year or who reported higher than average demand from patients concerned about their risk of a common disease because of a diagnosis in a family member (Table IV). Having taken a course in medical genetics increased willingness to counsel but did not significantly increase willingness to take family histories and draw pedigrees. There was no association between willingness and year of qualification or gender.
GPs Providing Genetic Services
307
308
Suchard, Yudkin, and Sinsheimer
Of the responding general practitioners, 163/339 (48%) agreed to participate in further research. Those who agreed to participate were more willing to provide genetic services; for example, 132/162 (81%) agreed or strongly agreed that they were willing to take family histories and draw pedigrees, compared to 76/168 (45%) respondents who did not want to participate. Also, 133/163 (82%) agreed or strongly agreed that they were willing to counsel about genetics, compared to 79/169 (47%) not wanting to participate. Of those not wanting to participate in further research, 158/176 (90%) gave "Not enough time" as a reason. Open-Ended Comments A substantial minority of respondents (20%, 68/339) provided additional comments. From the 68 comments received, 38 (56%) mentioned a need for more time and funding to provide genetic services. Eight (12%) identified a lack of referral/information guidelines and the need for further training for general practitioners in genetics. One general practitioner wrote, "I commonly see women with strong family histories of 'premenopausal' breast cancer and am not sure where to refer them." A substantial group of general practitioners were concerned about the ethical implications of genetic screening. Twenty-one (31%) general practitioners raised ethical issues about genetic screening for common diseases, including increased anxiety among patients, screening for untreatable diseases, confidentiality, insurance discrimination, and disclosing genetic information among family members. A general practitioner commented, "We frequently have to write medical reports on our patients, and an increased risk of a disease may increase premiums or even make them uninsurable/unable to have a mortgage." DISCUSSION The general practitioner respondents to the survey reported that approximately 4 patients per 1,000 inquire about their own risk of a common disease because of a family history. In other words, roughly every 2 weeks, a general practitioner is forced to consider the genetics of common diseases. General practitioners who reported a higher than average demand were more willing than others to take family histories and to counsel their patients about their genetic test results, suggesting that general practitioners will be responsive to an increasing demand. Unlike single-gene disorders like Huntington's disease with a population prevalence of 7.6 per 100,000 in South Wales (Walker et al., 1981), at least one common disease such as Alzheimer's disease, breast cancer, and cardiovascular disease is likely to impact on most patients seen in general practice. The prevalence of first-degree relatives affected with cancer was investigated during a randomized trial of health checks in general practice (Imperial Cancer Research Fund
GPs Providing Genetic Services
309
OXCHECK Study Group, 1994). Johnson et al. (1995) estimate that a typical general practitioner would have 18 women aged 35-64 with a history of breast cancer in a first-degree relative under the age of 70, putting them at a higher risk of the disease. Women aged 20-44 with an affected first-degree relative are approximately twice as likely to develop breast cancer as women without a family history (Claus et al., 1991; Thompson, 1994). Given the prevalence of family histories for common diseases, growing media attention to genetics is likely to increase patient demand for genetic services, forcing general practitioners to take a greater interest in genetic screening and other related services. This is the United Kingdom's first study focusing on general practitioners' attitudes to genetic screening for common diseases. We achieved a remarkably high response (94%). Surveys of general practitioners generally achieve poor responses rates (McAvoy and Kaner, 1996), some as low as 18% (McDonald, 1993). A survey of general practitioners regarding presymptomatic testing for Huntington's disease in Scotland achieved a response of 74% (Mennie et al., 1990), but in the United States, only 62% of family physicians responded to a survey about cystic fibrosis screening (Tambor et al., 1993). The short length of our survey and the inclusion of an "opt-out" question may be responsible for some of the improved response over these surveys; but Lydeard (1996) maintains that the most important factor in determining response is the general practitioners' perception of the value and relevance of the project to their patients. If true, genetic screening for common diseases is an important and topical issue for general practice. The most disturbing results of our survey conscerned general practitioners' self-reported preparedness for genetic counseling. Many general practitioners are compelled to provide genetic counseling but feel they are not well prepared to do so. While 50% of the respondents indicated they counseled patients about genetics in the last year, even in this group of general practitioners only 21% felt sufficiently prepared to counsel. Similar findings have been reported previously. For example, from a study of physicians' use of commercial APC gene test results in the United States, as many as 32% misinterpret the results for their patients (Giardiello et al., 1997). There is clearly a need to improve the skills of general practitioners in this area; however, there are at least two potential barriers. Conventional methods of training are of limited effectiveness (Davis et al., 1995). Moreover, in our study, a substantial group of general practitioners (20%) indicated that they would not be willing to take family histories, or to counsel about genetics even given training and information; their responses indicated that lack of time may be a major deterrent. "Practice-enabling" strategies (Davis et al., 1992), such as computerized aids to help determine familial risk, may improve performance without demanding too much time from general practitioners. The ethical issues raised by genetics cannot be eliminated as a strong deterrent to general practitioners. While our survey did not specifically address ethical issues, respondents expressed many unsolicited comments addressing concerns such as increased anxiety among patients, screening for untreatable diseases,
310
Suchard, Yudkin, and Sinsheimer
insurance discrimination, and disclosing genetic information among family members. These responses are consistent with the attitudes of obstetrician-gynecologists in the United States toward genetic breast cancer screening (Rowley and Loader, 1996). The long-term relationship between the general practitioner and the patient/ family makes primary care a valuable but difficult place to provide genetic services, especially when considering general practitioners' workloads, their lack of preparedness, nondirectiveness in genetic counseling, and the ethical issues raised. All of these concerns must be addressed if genetic services are to be provided in general practice. ACKNOWLEDGMENTS We thank all the general practitioners who kindly replied to our questionnaire. We also thank Drs. Godfrey Fowler, Peter Rose, Paul Salkovskis, Mike Murphy, and Tim Lancaster for their critical comments and support, and Thiru Thangarajah and Hannan Lamlum for technical assistance. MAS is supported by a predoctoral fellowship from the Howard Hughes Medical Institute. JSS is partially supported by NCI grant CA16042 and AIMH grant 28697.
REFERENCES Bell J (1998) The new genetics: The new genetics in clinical practice. BMJ 316:618-620. Claus EB, Risch N, Thompson WD (1991) Genetic analysis of breast cancer in the Cancer and Steroid Hormone Study. Am J Hum Genet 48:232-242. Davis DA, Thomson MA, Oxman AD, Haynes B (1992) Evidence for the effectiveness of CME: A review of 40 randomized controlled trials. JAMA 268:1111-1117. Davis DA, Thomson MA, Oxman AD, Haynes B (1995) Changing physician performance: A systematic review of the effect of continuing medical education strategies. JAMA 274:700-705. Genetics Research Advisory Group (1995) The genetics of common diseases: A second report to the NHS Central Research and Development committee on the new genetics. London: Department of Health. Giardiello MF, Jill D, Brensinger JD, Gloria M, Petersen GM, Luce MC, Hyline LM, et al. (1997) The use and interpretation of commercial APC gene testing for familial adenomatous polyposis. N Eng J Med 336:823-827. Harper PS (1995) Genetic testing, common diseases, and health service provision. Lancet 346: 1645-1646. Harris R, Harris H (1996) Genetics in primary care. J Med Genet 33:346-348. Hoskins KF, Stopfer JE, Calzone KA, Merajver SD, Rebbeck TR, Garber JE, Weber BL (1995) Assessment and counseling for women with a family history of breast cancer: A guide for clinicians. JAMA 273:577-585. Imperial Cancer Research Fund OXCHECK Study Group (1994) Effectiveness of health checks conducted by nurses in primary care: Results of the OXCHECK study after one year. BMJ 308:308-312. Johnson N, Lancaster T, Fuller A, Hodgson SV (1995) The prevalence of a family history of cancer in general practice. Fain Pract 12:287-289. Kinmonth AL, Reinhard J, Bobrow M, Pauker S (1998) The new genetics: Implications for clinical services in Britain and the United States. BMJ 316:767-770.
GPs Providing Genetic Services
311
Landis JR (1977) The measurement of observer agreement for categorical data. Biometrics 33:159-174. Lydeard S (1996) Avoid surveys masquerading as research [commentary]. BMJ 313:733-734. McAvoy BR, Kaner EPS (1996) General practice postal surveys: a questionnaire too far? BMJ 313:732-733. McDonald P (1993) Response rates in general practice studies [letter]. Br J Gen Pract 43:484. Mennie ME, Holloway SM, Brock DJH (1990) Attitudes of general practitioners to presymptomatic testing for Huntington's disease. J Med Genet 27:224-227. Qureshi N, Raeburn JA (1993) Clinical genetics meets primary care: Patients should benefit. BMJ 307:816-817. Rowley PT, Loader S (1996) Attitudes of obstetrician-gynecologists toward DNA testing for a genetic susceptibility to breast cancer. Obstet Gynerol 88:611-615. SPSS Inc. (1994) SPSS 6.1 User's Guide. Chicago: SPSS Inc. Tambor ES, Chase GA, Faden RR, Geller G, Hofman KJ, Holtzman NA (1993) Improving response rates through incentive and follow-up: The effect on a survey of physicians' knowledge of genetics. Am J Public Health 83:1599-1603. Thompson WD (1994) Genetic epidemiology of breast cancer. Cancer 74:279-287. Walker DA, Harper PS, Wells CEC, Tyler A, Davies K, Newcombe RG (1981) Huntington's chorea in South Wales: A genetic and epidemiological study, Clin Genet 19:213-221. Welsh Health Planning Forum (1995) Genomics: The Impact of the New Genetics on the NHS. Cardiff.
Are General Practitioners Willing and Able to Provide Genetic Services for Common Diseases? Marc A. Suchard,1,2 Patricia Yudkin,3 and Janet S. Sinsheimer124
Primary care in the United Kingdom has been advocated as an optimal location for the provision of genetic services for common diseases. Little, however, is known about general practitioners' own views toward this suggestion or the possible demand for such services from patients. To assess general practitioners' attitudes to providing genetic services for common diseases, and to estimate the demand from patients for these services, we used a single-page postal questionnaire survey of all 359 general practitioners registered with the Oxfordshire Health Authority; 339 (94%) responded. These physicians reported that a mean 4.1 patients (95% CI, 3.3—4.9) out of every 1,000 consulting them were concerned about their own risk of a common disease associated with a diagnosis in a family member. Half of the general practitioners (95% CI, 45—56%) counseled about genetics in the last year. A majority of general practitioners accept the need to provide genetic services, but far fewer are competent to do so. Although 60% (95% CI, 55-65%) believed they should be involved with genetic screening for common diseases, only 29% (95% CI, 24-34%) felt sufficiently prepared to take family histories and draw pedigrees, and only 15% (95% CI, 11-19%) felt sufficiently prepared to counsel patients about their genetic test results. Given the necessary training and information, 63% (95% CI, 58-68%) and 64% (95% CI, 59-69%) were willing to provide these services. Even with training and information, not all would
1 Wellcome
Trust Centre for Human Genetics, University of Oxford, Windmill Road, Headington, OX3 7BN, United Kingdom. 2Present affiliation: Department of Biomathematics, UCLA Medical School, Los Angeles, CA. 3Imperial Cancer Research Fund General Practice Research Group, Division of Public Health and Primary Health Care, Institute of Health Sciences, University of Oxford, Oxford, OX3 7LF, United Kingdom. 4Correspondence should be directed to Janet S. Sinsheimer, Department of Biomathematics, UCLA Medical School, Los Angeles, CA 90095-1766; e-mail: [email protected].
301 1059-7700/99/0800-0301516.00/1 © 1999 National Society of Genetic Counselors, Inc.
302
Suchard, Yudkin, and Sinsheimer
be willing to provide these services, and lack of time may be a major deterrent. "Practice-enabling" strategies, such as computerized aids in genetics, may be useful. KEY WORDS: Genetic screening; primary care; common diseases.
INTRODUCTION Advances in human genetics are making possible predictive testing for genetic susceptibility to many common diseases (Genetics Research Advisory Group, 1995), such as Alzheimer's disease, breast cancer, and cardiovascular disease. Leading genetics enthusiasts and several key health policy reports in the United Kingdom have advocated primary care as an ideal location for the provision of these genetic services (Welsh Health Planning Forum, 1995; Harper, 1995; Harris and Harris, 1996; Kinmonth et al., 1998). This recommendation is, partially, prompted by a foreseen shortfall in the number of genetic specialists. Currently, in the United Kingdom, there are only one or two consultant geneticists per million people (Kinmonth et al., 1998). Additionally, Bell (1998) has conjectured that pressure for genetic screening from patients is likely to increase steadily in the coming years, far outpacing the resources of genetic specialists. In the United Kingdom, general practitioners provide the bulk of primary care, acting as gatekeepers to all further specialties. Possible roles for general practitioners, as genetic screening for common diseases moves to general practice, include making initial contact with concerned patients, taking family histories to assess risk, referring patients more frequently for genetic testing and counseling, and counseling patients directly about their genetic risks (Qureshi and Raeburn, 1993; Hoskins et al., 1995). However, there has been little attempt to assess general practitioners' own views toward these postulated changes. Specifically, we do not know how much demand general practitioners perceive for these services from patients, general practitioners' attitudes toward involvement, or their ability to take family histories and give appropriate advice. We conducted a survey to better understand general practitioners' views about providing genetic screening for common diseases like Alzheimer's disease, breast cancer, and cardiovascular disease. SUBJECTS AND METHODS We developed a single-page questionnaire with input from a local general practitioner and a health psychologist. We piloted the questionnaire on 20 randomly selected general practitioners from Berkshire, a neighboring county, and revised the questionnaire in the light of their comments or if answers appeared ambiguous. The final form (Table I) consisted of 14 questions (attitudinal = 5 and multiple choice = 9) and provided space for open-ended comments. The
GPs Providing Genetic Services
303
Table I. General Practitioners' Views on Genetic Screening for Common Diseases Questionnaire
1. Patients sometimes present family history information about noncommunicable, common diseases
2. 3. 4. 5. 6. 7. 8. 9. 10. 11.
12. 13. 14.
which may run in their families, for example, "I saw a TV program about breast cancer and my mother had breast cancer, so what should I do?" In the last month, how many patients have consulted you concerning their health because of a diagnosis made in a family member? (Number) In the last year, have you referred any patients for genetic testing or ordered a test yourself? (Yes/no) In the last year, have you counseled any patients about their genetic test results? (Yes/no) I believe GPs should be involved with genetic screening for common diseases like Alzheimer's disease, breast cancer, and cardiovascular disease. (Strongly agree to strongly disagree) I feel sufficiently prepared to take family histories from my patients and draw pedigrees. (Strongly agree to strongly disagree) I feel sufficiently prepared to counsel my patients about their genetic test results. (Strongly agree to strongly disagree) Given the necessary training and information, I am willing to take family histories from my patients and draw pedigrees. (Strongly agree to strongly disagree) Given the necessary training and information, 1 am willing to counsel my patients about their genetic test results. (Strongly agree to strongly disagree) I would like to learn more about genetic screening for susceptibility to common diseases. (Strongly agree to strongly disagree) In what year did you qualify? (Number) What is your gender? (Female/male) Have you ever take a course in medical genetics (including medical school)? (Yes/no) How many patients do you see in total during an average week? (Number) Are you interested in participating further in our research about the genetics of common disease in general practice? (Yes/no) If not, why not? (Not enough time/not interested in genetics/no incentive offered/other)
attitudinal questions measured agreement with statements regarding the provision of genetic services in general practice, using 5-point ordinal scales ("strongly agree" to "strongly disagree"). The remaining questions asked about perceived patient demand, past referrals and counseling involving genetics, workload, year of qualification, gender, previous coursework in genetics, and further participation in the project. Between November 1996 and January 1997, we mailed the questionnaire to all 359 general practitioners registered with the Oxfordshire Health Authority on 1 November 1996. We sent two follow-up mailings to nonresponders. To insure accuracy, we double-entered the data. Further, we independently classified all open-ended comments into five subject groups (yes/no). Interrater agreement was close in all cases (k > 0.85) (Landis, 1977). We used SPSS 6.1 for the Macintosh computer (SPSS Inc., 1994) for all statistical analyses. We estimated overall perceived patient demand in two ways. First, we calculated the simple arithmetic mean of the self-reported genetic consultation numbers in the last month. Second, we allowed for imprecise recall of the number of these consultations when greater than zero. Under the latter model, we assumed that the actual number of consultations in the last month was difficult to recall; however, information about the complete absence or presence of such consultations was
304
Suchard, Yudkin, and Sinsheimer
accurate. We then modeled the data as a Poisson distribution using the number of zero versus nonzero reported observations. Let Xi equal the number of genetic consultations in the last month for General Practitioner i. Then,
In general, we made comparisons between means of two groups using the t test and between proportions using the x2 test (with continuity correction). We scored answers to attitudinal questions on a scale of 1 to 5 and compared results between two groups using the Mann-Whitney test and between more than two groups using the Kruskall-Wallis test, correcting for ties in both cases. We made no corrections to p values for multiple testing.
RESULTS The vast majority of general practitioners (339/359 [94%]) returned questionnaires, but not all respondents completed every question on the form. Response rates to the attitudinal questions ranged from 328/339 (97%) to 332/339 (98%); the lowest responses (303/339 [89%] and 316/339 [93%]) were to questions about the number of concerned patients seen in the last month and the total number of patients seen in an average week. Characteristics of Responders Table II outlines the background information collected. Overall, 50% (95% confidence interval [CI], 45-56%) of the general practitioners reported counseling patients about their genetic test results in the last year, and 76% (95% CI, 72-81%) had referred patients for genetic testing or ordered a genetic test themselves. Only 38% (95% CI, 33-43%) of the general practitioners had taken a course in medical genetics. The mean length of time since qualifying for a general practitioner was 19.5 (SD 8.2) years. General practitioners who had taken a course in medical genetics qualified more recently than those who had not. Mean( SD) years since qualification were 15.9 (6.3) years and 21.8 (8.4) years, respectively, a difference of 5.9 years (95% CI, 4.1-7.6, P < 0.001). Frequency of Concerned Patients General practitioners reported that a mean of 2.2 (median 1.5, interquartile range 1-2.5) patients had consulted them in the last month about their risk of a
GPs Providing Genetic Services
305
Table II. Background Information from Responding General Practitioners % (No.) of general practitioners (n = 339) Gender Female Male Referred patient to medical genetics unit in last year Yes No Counseled patient about genetics in last year Yes No Ever taken a course in medical genetics Yes No Year of qualification Before 1972 1972-1977 1978-1983 After 1983
35(116) 65(212) 76(251) 24 (78) 50(162) 50(161) 38(124) 62 (203) 24(80) 25(85) 24(81) 26(88)
common disease associated with a diagnosis in a family member. In total, 62/303 (20%) respondents reported no such consultations in the last month, yielding a Poisson mean frequency of 1.6. Respondents who counseled patients about genetic test results in the last year were more likely to perceive a patient demand higher than the median. Of those with counseling experience, 86/147 (59%) perceived a higher demand compared to 62/148 (42%) without experience (p = 0.006). Demand was independent of the general practitioner's genetic coursework (p = 0.504) and gender (p = 0.595). The mean number of patients seen in an "average" week was 134 (median 130, interquartile range 100-150), giving an estimated 580 consultations per calendar month. We estimated the frequency of concerned patients for each general practitioner. The mean frequency was 4.1 (95% confidence interval 3.3-4.9) per 1,000 consultations. Median frequency was 2.5 (interquartile range 1.2-4.7) per 1,000 consultations. Excluding the four cases in which general practitioners reported seeing more than 250 patients in an average week made little difference to these results.
General Practitioner Attitudes Table III shows the levels of agreement general practitioners reported for each of the attitudinal statements. The majority of general practitioners, 60% (95% CI,
306
Suchard, Yudkin, and Sinsheimer Table III. Attitudes of General Practitioners About Genetic Screening for Common Diseases Statement
Believe in involvement for diseases like Alzheimer's disease, breast cancer, and cardiovascular disease Feel sufficiently prepared to take family histories and draw pedigrees Feel sufficiently prepared to counsel about genetic test results Given the necessary training and information, are willing to take family histories and draw pedigrees Given the necessary training and information, are willing to counsel about genetic test results Want to learn more about genetic screening for susceptibility to common diseases
Strongly agree
Agree
Neutral
10 (32) |50 (165)| 24 (80) 4 (14) I 25 (83) 1 (3)
14 (47)
Disagree
Strongly disagree
(32)
6 (19)
10
22 (72) | 40 (132)| 21 (69)
50 (166)
9 (29) 14 (47)
9 (30) |54 (178)| 19 (62)
14
(47)
4 (13)
9 (30) |55 (182)| 17 (55) I
14
(46)
6 (19)
19 (61)
15 (49) 59(195)|
5 (16)
2 (8)
Notes: Values are Percentages (Numbers) of General Practitioners, Bold and normal boxes highlight median and interquartile ranges, respectively.
55-65%), agreed or strongly agreed they should be involved with genetic screening for common diseases, and 78% (95% CI, 73-82%) wanted to learn more about genetic screening. However, only 29% (95% CI, 24-34%) felt sufficiently prepared to take family histories and draw pedigrees, and only 15% (95% CI, 11-19%) felt sufficiently prepared to counsel patients about their genetic test results. General practitioners who had referred or counseled patients were only slightly more prepared. Of those respondents who had referred to a medical genetics unit in the last year, only 79/246 (32%) felt prepared to take family histories, and of those who had counseled about genetics, only 33/160 (21%) felt prepared to do so, indicating that exposure alone does not sufficiently increase preparedness. Given the necessary training and information, 63% (95% CI, 58-68%) and 64% (95% CI, 59-69%) of the whole sample were willing to provide these services.
Willingness to Provide Genetic Services Willingness to take family histories and draw pedigrees and to counsel patients about their genetic test results was greater among those general practitioners who reported having counseled in the last year or who reported higher than average demand from patients concerned about their risk of a common disease because of a diagnosis in a family member (Table IV). Having taken a course in medical genetics increased willingness to counsel but did not significantly increase willingness to take family histories and draw pedigrees. There was no association between willingness and year of qualification or gender.
GPs Providing Genetic Services
307
308
Suchard, Yudkin, and Sinsheimer
Of the responding general practitioners, 163/339 (48%) agreed to participate in further research. Those who agreed to participate were more willing to provide genetic services; for example, 132/162 (81%) agreed or strongly agreed that they were willing to take family histories and draw pedigrees, compared to 76/168 (45%) respondents who did not want to participate. Also, 133/163 (82%) agreed or strongly agreed that they were willing to counsel about genetics, compared to 79/169 (47%) not wanting to participate. Of those not wanting to participate in further research, 158/176 (90%) gave "Not enough time" as a reason. Open-Ended Comments A substantial minority of respondents (20%, 68/339) provided additional comments. From the 68 comments received, 38 (56%) mentioned a need for more time and funding to provide genetic services. Eight (12%) identified a lack of referral/information guidelines and the need for further training for general practitioners in genetics. One general practitioner wrote, "I commonly see women with strong family histories of 'premenopausal' breast cancer and am not sure where to refer them." A substantial group of general practitioners were concerned about the ethical implications of genetic screening. Twenty-one (31%) general practitioners raised ethical issues about genetic screening for common diseases, including increased anxiety among patients, screening for untreatable diseases, confidentiality, insurance discrimination, and disclosing genetic information among family members. A general practitioner commented, "We frequently have to write medical reports on our patients, and an increased risk of a disease may increase premiums or even make them uninsurable/unable to have a mortgage." DISCUSSION The general practitioner respondents to the survey reported that approximately 4 patients per 1,000 inquire about their own risk of a common disease because of a family history. In other words, roughly every 2 weeks, a general practitioner is forced to consider the genetics of common diseases. General practitioners who reported a higher than average demand were more willing than others to take family histories and to counsel their patients about their genetic test results, suggesting that general practitioners will be responsive to an increasing demand. Unlike single-gene disorders like Huntington's disease with a population prevalence of 7.6 per 100,000 in South Wales (Walker et al., 1981), at least one common disease such as Alzheimer's disease, breast cancer, and cardiovascular disease is likely to impact on most patients seen in general practice. The prevalence of first-degree relatives affected with cancer was investigated during a randomized trial of health checks in general practice (Imperial Cancer Research Fund
GPs Providing Genetic Services
309
OXCHECK Study Group, 1994). Johnson et al. (1995) estimate that a typical general practitioner would have 18 women aged 35-64 with a history of breast cancer in a first-degree relative under the age of 70, putting them at a higher risk of the disease. Women aged 20-44 with an affected first-degree relative are approximately twice as likely to develop breast cancer as women without a family history (Claus et al., 1991; Thompson, 1994). Given the prevalence of family histories for common diseases, growing media attention to genetics is likely to increase patient demand for genetic services, forcing general practitioners to take a greater interest in genetic screening and other related services. This is the United Kingdom's first study focusing on general practitioners' attitudes to genetic screening for common diseases. We achieved a remarkably high response (94%). Surveys of general practitioners generally achieve poor responses rates (McAvoy and Kaner, 1996), some as low as 18% (McDonald, 1993). A survey of general practitioners regarding presymptomatic testing for Huntington's disease in Scotland achieved a response of 74% (Mennie et al., 1990), but in the United States, only 62% of family physicians responded to a survey about cystic fibrosis screening (Tambor et al., 1993). The short length of our survey and the inclusion of an "opt-out" question may be responsible for some of the improved response over these surveys; but Lydeard (1996) maintains that the most important factor in determining response is the general practitioners' perception of the value and relevance of the project to their patients. If true, genetic screening for common diseases is an important and topical issue for general practice. The most disturbing results of our survey conscerned general practitioners' self-reported preparedness for genetic counseling. Many general practitioners are compelled to provide genetic counseling but feel they are not well prepared to do so. While 50% of the respondents indicated they counseled patients about genetics in the last year, even in this group of general practitioners only 21% felt sufficiently prepared to counsel. Similar findings have been reported previously. For example, from a study of physicians' use of commercial APC gene test results in the United States, as many as 32% misinterpret the results for their patients (Giardiello et al., 1997). There is clearly a need to improve the skills of general practitioners in this area; however, there are at least two potential barriers. Conventional methods of training are of limited effectiveness (Davis et al., 1995). Moreover, in our study, a substantial group of general practitioners (20%) indicated that they would not be willing to take family histories, or to counsel about genetics even given training and information; their responses indicated that lack of time may be a major deterrent. "Practice-enabling" strategies (Davis et al., 1992), such as computerized aids to help determine familial risk, may improve performance without demanding too much time from general practitioners. The ethical issues raised by genetics cannot be eliminated as a strong deterrent to general practitioners. While our survey did not specifically address ethical issues, respondents expressed many unsolicited comments addressing concerns such as increased anxiety among patients, screening for untreatable diseases,
310
Suchard, Yudkin, and Sinsheimer
insurance discrimination, and disclosing genetic information among family members. These responses are consistent with the attitudes of obstetrician-gynecologists in the United States toward genetic breast cancer screening (Rowley and Loader, 1996). The long-term relationship between the general practitioner and the patient/ family makes primary care a valuable but difficult place to provide genetic services, especially when considering general practitioners' workloads, their lack of preparedness, nondirectiveness in genetic counseling, and the ethical issues raised. All of these concerns must be addressed if genetic services are to be provided in general practice. ACKNOWLEDGMENTS We thank all the general practitioners who kindly replied to our questionnaire. We also thank Drs. Godfrey Fowler, Peter Rose, Paul Salkovskis, Mike Murphy, and Tim Lancaster for their critical comments and support, and Thiru Thangarajah and Hannan Lamlum for technical assistance. MAS is supported by a predoctoral fellowship from the Howard Hughes Medical Institute. JSS is partially supported by NCI grant CA16042 and AIMH grant 28697.
REFERENCES Bell J (1998) The new genetics: The new genetics in clinical practice. BMJ 316:618-620. Claus EB, Risch N, Thompson WD (1991) Genetic analysis of breast cancer in the Cancer and Steroid Hormone Study. Am J Hum Genet 48:232-242. Davis DA, Thomson MA, Oxman AD, Haynes B (1992) Evidence for the effectiveness of CME: A review of 40 randomized controlled trials. JAMA 268:1111-1117. Davis DA, Thomson MA, Oxman AD, Haynes B (1995) Changing physician performance: A systematic review of the effect of continuing medical education strategies. JAMA 274:700-705. Genetics Research Advisory Group (1995) The genetics of common diseases: A second report to the NHS Central Research and Development committee on the new genetics. London: Department of Health. Giardiello MF, Jill D, Brensinger JD, Gloria M, Petersen GM, Luce MC, Hyline LM, et al. (1997) The use and interpretation of commercial APC gene testing for familial adenomatous polyposis. N Eng J Med 336:823-827. Harper PS (1995) Genetic testing, common diseases, and health service provision. Lancet 346: 1645-1646. Harris R, Harris H (1996) Genetics in primary care. J Med Genet 33:346-348. Hoskins KF, Stopfer JE, Calzone KA, Merajver SD, Rebbeck TR, Garber JE, Weber BL (1995) Assessment and counseling for women with a family history of breast cancer: A guide for clinicians. JAMA 273:577-585. Imperial Cancer Research Fund OXCHECK Study Group (1994) Effectiveness of health checks conducted by nurses in primary care: Results of the OXCHECK study after one year. BMJ 308:308-312. Johnson N, Lancaster T, Fuller A, Hodgson SV (1995) The prevalence of a family history of cancer in general practice. Fain Pract 12:287-289. Kinmonth AL, Reinhard J, Bobrow M, Pauker S (1998) The new genetics: Implications for clinical services in Britain and the United States. BMJ 316:767-770.
GPs Providing Genetic Services
311
Landis JR (1977) The measurement of observer agreement for categorical data. Biometrics 33:159-174. Lydeard S (1996) Avoid surveys masquerading as research [commentary]. BMJ 313:733-734. McAvoy BR, Kaner EPS (1996) General practice postal surveys: a questionnaire too far? BMJ 313:732-733. McDonald P (1993) Response rates in general practice studies [letter]. Br J Gen Pract 43:484. Mennie ME, Holloway SM, Brock DJH (1990) Attitudes of general practitioners to presymptomatic testing for Huntington's disease. J Med Genet 27:224-227. Qureshi N, Raeburn JA (1993) Clinical genetics meets primary care: Patients should benefit. BMJ 307:816-817. Rowley PT, Loader S (1996) Attitudes of obstetrician-gynecologists toward DNA testing for a genetic susceptibility to breast cancer. Obstet Gynerol 88:611-615. SPSS Inc. (1994) SPSS 6.1 User's Guide. Chicago: SPSS Inc. Tambor ES, Chase GA, Faden RR, Geller G, Hofman KJ, Holtzman NA (1993) Improving response rates through incentive and follow-up: The effect on a survey of physicians' knowledge of genetics. Am J Public Health 83:1599-1603. Thompson WD (1994) Genetic epidemiology of breast cancer. Cancer 74:279-287. Walker DA, Harper PS, Wells CEC, Tyler A, Davies K, Newcombe RG (1981) Huntington's chorea in South Wales: A genetic and epidemiological study, Clin Genet 19:213-221. Welsh Health Planning Forum (1995) Genomics: The Impact of the New Genetics on the NHS. Cardiff.
