Family Practice Vol. 18, No. 5, 487-490
© Oxford University Press 2001
Genetics |
Referral of patients with a family history of breast/ovarian cancer—GPs' knowledge and expectations
Department of Primary Health Care, University of Oxford and
a Churchill Hospital, Oxford, UK.
Peter W Rose, ICRF General Practice Research Group, Department of Primary Health Care, University of Oxford, Institute of Health Sciences, Old Road, Headington, Oxford OX3 7LF, UK.
Rose PW, Watson E, Yudkin P, Emery J, Murphy M, Fuller A and Lucassen A. Referral of patients with a family history of breast/ovarian cancer—GPs' knowledge and expectations. Family Practice 2001; 18: 487–490.
 |
Abstract |
|---|
 Top Abstract BackgroundMethodResultsDiscussionAppendixReferences |
|---|
Background. Primary care is expected to play a significant role in the management of patients with genetic problems. Currently, this particularly involves patients with a family history of cancer. If GPs are to exercise their gatekeeper role efficiently in this area, they must be able to evaluate genetic risk and make appropriate referral decisions.
Objectives. The aim of this study was to examine GPs' ability to assess risk and to make appropriate referral decisions for women with a family history of breast/ovarian cancer, and to determine their expectations of a referral to secondary care.
Method. A questionnaire survey was carried out of the 282 GP principals working within Bedfordshire Health Authority. GPs were asked to make decisions for six simulated cases of women presenting with a family history of breast or ovarian cancer.
Results. A total of 164 (58%) GPs returned completed questionnaires. Across the six family histories, the percentage of GPs making an appropriate risk assessment ranged from 21% [95% confidence interval (CI) 14–27%] to 63% (95% CI 56–71%), and an appropriate referral decision ranged from 40% (95% CI 32–48%) to 80% (95% CI 73–86%). Regardless of their accuracy of risk assessment, most GPs were consistent in deciding not to refer low risk women and to refer moderate and high risk women (range 71–85% of GPs for the six family histories). Only 43 (26%, 95% CI 20–33%) of GPs knew the three most important criteria for risk assessment.
Conclusions. GPs require more help and education to enable them to perform their gatekeeper role satisfactorily when assessing patients with a family history of breast/ovarian cancer.
Keywords. Breast cancer, GPs, ovarian cancer, referral, risk assessment.
|
Background |
|---|
|
TopAbstractBackgroundMethodResultsDiscussionAppendixReferences |
|---|
Primary care is expected to play a significant role in caring for patients with genetic problems,1 and there is already evidence about the ways in which GPs are taking on this role.2 Patients already present to their GP to enquire about their family history of disease, and these enquiries frequently concern inherited cancers. Approximately one patient per month will present to a GP with a list size of 2000 patients with a concern about their family history of breast cancer.3 There is also evidence of increasing GP involvement from the number of referrals to secondary care. In Oxford, referrals to the Family Cancer Clinic have risen from 152 in 1992 to 960 in 1999, and account for most of the rise in total referrals from 820 to 1650 to the Regional Genetics Service over the same period (Anneke Lucassen, personal communication).
If GPs are to exercise their gatekeeper role efficiently in this area, it is important that they can evaluate a patient's risk of inherited disease accurately. They can then reassure those with lower risk and refer those with higher risk. Guidelines have been produced locally and nationally to help GPs make appropriate referrals to Clinical Genetics departments,4,5 but at the time of this study there was a lack of consensus between the various guidelines.
Preliminary work for this study using case vignettes at educational meetings indicated that GP principals and registrars had little difficulty in identifying lower risk patients when there were only one or two relatives affected with breast cancer, especially if the age of onset was late. However, they had difficulty in evaluating more complicated family histories. A case vignette approach has been used successfully in other studies, including research into genetic risk assessment and prescribing by GPs,6–8 but only one study6 used this method to assess genetic risk assessment and referral.
The aims of this study were to assess GPs' performance when assessing risk and making referral decisions for women with a family history of breast or ovarian cancer. In addition, we sought to find out what criteria GPs used to assess risk in such women and what they expected to happen when these women were referred to a clinical geneticist. Answers to these questions will inform the development of educational strategies for GPs in this area.
|
Method |
|---|
|
TopAbstractBackgroundMethodResultsDiscussionAppendixReferences |
|---|
A questionnaire was sent to the 282 GP principals working within Bedfordshire Health Authority in December 1998, with two reminders at 3 and 4 months for non-responders. The questionnaire asked the GPs to categorize six written family histories (see Appendix) into low, moderate or high risk and say whether they would refer the case to secondary care. These risks equate approximately to the risks of a woman carrying a dominantly inherited breast cancer gene of 0–10% (low), 10–25% (moderate) and >25% (high) as calculated with the Claus model.9 In the questionnaire, the GPs were told that the three risk categories equated to a low, moderate or high risk that the patient would develop an inherited form of breast cancer. As there is a lack of complete agreement of risk categorization between the Claus model, consensus statements4,5 and the many local guidelines, we chose family histories that gave the same result by all methods of calculation. Local guidelines had not been distributed systematically to these GPs at the time of the study. A clinical geneticist (AL) determined the risk category for each family history and ensured that they provided an appropriate range of complexity. The assessment of results followed the recommendation that lower risk patients should be dealt with in primary care and moderate to higher risk patients should be referred.4,5 Subjects were also asked to record, using a 5-point Likert scale, how likely they thought that specific topics would be discussed in a consultation with a clinical geneticist.
It was calculated that with a sample size of 150 respondents, a percentage of 50% would (with 95% confidence) be estimated to within 8% and a percentage of 20% would be estimated to within 6%. Proportions were compared using the chi-squared test. Confidence intervals were calculated using CIA (Confidence Interval Analysis).10
|
Results |
|---|
|
TopAbstractBackgroundMethodResultsDiscussionAppendixReferences |
|---|
A total of 282 questionnaires were mailed and 176 (62%) were returned. Of these, 12 were not completed, giving a response rate of 58%. Fifty-two [32%; 95% confidence interval (CI) 25–39%] of the 164 respondents were aware of referral guidelines and 42 (26%; 95% CI 19–32%) of these said they used the guidelines in clinical practice.
There was no difference between responders and non-responders in terms of their sex, but responders were more likely to hold the MRCGP and have qualified more recently (Table 1
)
|
Table 2
shows the risk evaluation and referral decision for each family history. Correct risk assessment ranged from 21 to 63%. Correct referral decision ranged from 40 to 80%. However, apart from Family history 4 (very low risk) and Family history 2 (high risk), fewer than one-third of the GPs made both a correct risk assessment and referral decision for each family history. The problem that the GPs found with the assessment of low risk cases is highlighted by Family history 3 and Family history 6: a low proportion assessed them correctly as low risk (21 and 21%, respectively) and the majority thought incorrectly that the cases should be referred (60 and 58%, respectively).
|
Regardless of the accuracy of their decisions, most GPs said that they would not refer those cases they assessed as low risk and would refer cases assessed as moderate or high risk. The percentage of GPs showing this consistency ranged from 71% (Family history 1) to 85% (Family history 5).
GPs' expectations of the content of a consultation with a clinical geneticist by a woman with a family history of breast or ovarian cancer were generally accurate. There were three areas where the GP views were at variance with standard practice: 74/161 (46%; 95% CI 38–54%) thought that it was likely or very likely that a moderate risk woman would be offered a gene test (currently only high risk women are offered a gene test and only if blood is available from an affected relative); 67/156 (43%; 95% CI 35–51%) thought that it was unlikely or very unlikely that a moderate risk woman would be offered mammography outside the National Screening Programme (the consensus meeting4 recommended annual mammography from 35 years but there is variation in local service provision in Bedfordshire); 100/155 (65%; 95% CI 57–72%) thought that high risk women and 43/157 (27%; 95% CI 20–34%) thought that moderate risk women were likely or very likely to be offered ovarian cancer screening (screening for ovarian cancer is usually only offered as part of a research programme).
In response to the question about criteria that would prompt them to make a referral, 97 out of 164 GPs (59%; 95% CI 52–67%) stated the age at diagnosis in the affected relative, 113 (69%; 95% CI 62–76%) the relationship of the affected relative to the proband and 89 (54%; 95% CI 47–62%) the number of affected relatives. However, only 43 (26%; 95% CI 20–33%) correctly stated all three.
|
Discussion |
|---|
|
TopAbstractBackgroundMethodResultsDiscussionAppendixReferences |
|---|
This survey demonstrates some of the problems that doctors in primary care may experience when dealing with patients with a family history of cancer. Although doctors may not always behave the same in surveys and real life,11 it is not surprising that GPs have difficulty with assessing patients in a rapidly expanding field such as genetic medicine. Many GPs currently practising will not have had any formal education in genetics. The relatively low response rate (60%) in this survey may mean that our findings overestimate GPs' performance, especially as non-responders were less likely to hold the MRCGP and had qualified longer ago than responders. Over the six family histories, the percentage of doctors getting both assessment and referral correct ranged from 17 to 61%. We found that GPs have significant difficulty with risk assessment unless the risk is very low or high, confirming our original impression from the pilot. For example, their assessment was much better for the lowest risk family history (Family history 4) than for the other two low risk cases in this survey. Any woman presenting to a GP with more than one or two affected relatives is unusual and their risk is likely to be overestimated. These results confirm our original impressions and the findings of Fry et al.6 that GPs tend to overestimate the risk of lower risk cases but are consistent in their decision to refer those cases they assess to be at higher risk.
One limitation of this study is the lack of complete agreement on risk categorization between the various guidelines available. The vignettes were chosen so that they produced the same risk assessment by all available methods, and fewer than a third of the GPs were aware of guidelines anyway. Although there are limitations to the use of guidelines,12 there is evidence from a recent study13 that distribution of guidelines to the GPs in this survey might have improved performance.
GPs are willing to undertake work in the field of genetics but recognize their need for education,14 and our survey confirms this. Only 26% of GP respondents were able to name the three most important criteria used for risk assessment. Although GPs were well informed about the process of care in a genetics clinic, some GPs stated unrealistic expectations about interventions that would be offered to their patients if they are referred. These findings highlight the need to find ways to help GPs with the evaluation of risk, either through guidelines or with the development of computer programs that can be used in the consulting room.7 Educational programmes should be developed to help GPs acquire knowledge that enables them to gain the confidence to reassure low risk women and ensure that patients who are referred have realistic expectations.
|
Appendix |
|---|
|
TopAbstractBackgroundMethodResultsDiscussionAppendixReferences |
|---|
Family history 1
The patient is a healthy 35-year-old woman.
She has one sister who is well, age 37.
The patient's mother developed breast cancer age 49 and is still alive.
The patient has two maternal aunts. One developed breast cancer age 47, the other is now age 50 with no history of cancer.
The patient's maternal grandmother is alive and well age 82.
Family history 2
The patient is a healthy 34-year-old woman. She has one sister, age 36 who is well.
The patient's mother developed breast cancer at the age of 37. When she was 50 she developed ovarian cancer and died 2 years later.
The patient has one maternal aunt, age 57, who is well.
The patient's maternal grandmother died of ischaemic heart disease age 82.
Family history 3
The patient is a healthy 44-year-old woman.
Her mother developed breast cancer age 66. She is still alive and well.
The patient's maternal grandmother developed breast cancer age 59 and died 2 years later.
The patient's paternal aunt developed breast cancer age 49.
The patient's paternal grandmother died of lung cancer age 71.
Family history 4
The patient is a woman, age 37 who is well.
The patient's mother developed breast cancer age 60 and recently died of the disease when she was 63.
The patient's grandmother died aged 84 from a stroke.
The patient's paternal aunt is 64 and is well.
The patient's paternal grandmother developed breast cancer age 62 and died from the disease.
Family history 5
The patient is a healthy 31-year-old woman.
She has a mother aged 54 and a maternal aunt age 49 who are both well.
Her maternal grandmother died aged 68 from congestive cardiac failure.
The patient has two paternal aunts who both developed breast cancer, one diagnosed age 46 and the other diagnosed age 48.
The patient's paternal grandmother also had breast cancer, diagnosed at age 51.
Family history 6
The patient is a 43-year-old woman with no relevant medical history.
She has two sisters. One is age 45 and is well. The other recently developed breast cancer age 53.
Her mother developed breast cancer aged 68. She is now aged 79.
Her maternal grandmother died of a heart attack, aged 82 years.
|
Acknowledgments |
|---|
We thank Sophia Nightingale for help with administration of the survey, all GPs completing the questionnaire, and Oxfordshire Health Authority for funding the study leave for PWR during which this work was undertaken.
|
References |
|---|
|
TopAbstractBackgroundMethodResultsDiscussionAppendixReferences |
|---|
1 Population Needs and Genetic Services. London: HMSO, 1993.
2
Emery J, Watson E, Rose P, Andermann A. A systematic review of the literature exploring the role of primary care in genetic services. Fam Pract 1999; 16: 426–445.
3
de Bock GH, Perk DC, Oosterwijk JC, Hageman GC, Kievit J, Springer MP. Women worried about their familial breast cancer risk—a study on genetic advice in general practice. Fam Pract 1997; 14: 40–43.
4 Wellcome Trust. Report of the Consensus Meeting on the Management of Women with a Family History of Breast Cancer. London: Wellcome Trust, 1998.
5
Eccles DM, Evans DGR, Mackay J on behalf of the UK cancer Family Study Group. Guidelines for a genetic risk based approach to advising women with a family history of breast cancer. J Med Genet 2000; 37: 203–209.
6
Fry A, Campbell H, Gudmundsdottir H et al. GPs' views on their role in cancer genetic services and current practice. Fam Pract 1999; 16: 468–474.
7
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. Br Med J 2000; 321: 28–32.
8
Walton RJ, Gierl C, Yudkin P, Mistry H, Vessey MP, Fox J. Evaluation of computer support for prescribing (CAPSULE) using simulated cases. Br Med J 1997; 315: 791–795.
9 Claus E, Risch N, Thompson D. Genetic analysis of breast cancer in the Cancer and Steroid Hormone Study. Am J Hum Genet 1991; 48: 232–242.[Web of Science][Medline]
10 Gardner SB, Winter PD, Gardner MJ. Confidence Interval Analysis. London: BMJ Publishing, 1989.
11 Jones TV, Gerrity MS, Earp J. Written case simulations: do they predict physicians' behaviour? J Clin Epidemiol 1990; 43: 805–815.[Web of Science][Medline]
12
Cabana MA, Rand CS, Powe NR et al. Why don't physicians follow clinical practice guidelines? A framework for improvement. J Am Med Assoc 1999; 282: 1458–1465.
13
Lucassen A, Watson E, Harcourt J, Rose P, O'Grady J. Guidelines for referral to a regional genetics service: GPs respond by referring more appropriate cases. Fam Pract 2001; 18: 135–140.
14
Watson EK, Shickle D, Qureshi N, Emery J, Austoker J. The ‘new genetics’ and primary care: general practitioners' views on their role and educational needs. Fam Pract 1999; 16: 420–425.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  M. Jiwa, J. Reid, C. Handley, J. Grimwood, S. Ward, K. Turner, M. Ibbotson, and N. Thorman Less haste more speed: factors that prolong the interval from presentation to diagnosis in some cancers Fam. Pract., June 1, 2004; 21(3): 299 - 303. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Gramling, J. Nash, K. Siren, C. Eaton, and L. Culpepper Family Physician Self-Efficacy With Screening for Inherited Cancer Risk Ann. Fam. Med, March 1, 2004; 2(2): 130 - 132. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Knottnerus Community genetics and community medicine Fam. Pract., October 1, 2003; 20(5): 601 - 606. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||