| Potential impact of guidelines discouraging prostate cancer screening on general practitioners' perceptions of medicolegal risk |
| REVISED IN RESPONSE TO REVIEWERS'
COMMENTS (2498 words) This is the first revised version of this article. Click here for the current version or click here for the original submission |
| Needs Assessment & Health Outcomes Unit |
| Seham Girgis MBChB MPH, Public Health Officer |
| Jeanette E Ward PhD FAFPHM, Director |
| Australian Institute of Health, Law and Ethics |
| Colin JH Thomson BA LLM, Executive Officer |
| Abstract |
| Objective: |
| To ascertain GPs’ perceptions of medicolegal risk when screening for prostate cancer and explore the potential impact of three national guidelines on perceptions and test-ordering. | par 7 |
| Design: |
| Postal survey in August 1997 | par 8 |
| Participants: |
| par 9 |
| 219 randomly selected GPs in NSW (65% response rate) |
| Main outcome measures: Self-reported response to case scenarios; perceptions of medicolegal risk and protection afforded by national guidelines before and after reading extracts of three national guidelines; ratings of current and potential strategies. | par 10 |
| Results: |
| 198 (90%) would screen an asymptomatic male patient and 170 (78%) of the respondents would advise that DRE & PSA in combination was the best available test to screen for prostate cancer. 133 (61%) indicated GPs would be at risk if they did not screen. Fifty-three 53 (24%) were aware of all three guidelines about prostate cancer screening. While significant changes in responses were found after respondents' had read guideline extracts, 101 (46%) respondents continued to perceive medicolegal risk if screening was not performed. Two thirds (n=145, 66%) supported a clear statement about the legal status of guidelines in a court of law. | par 11 |
| Conclusions: |
| Even when made aware of national evidence-based guidelines against prostate cancer screening, GPs in our survey perceived limited hypothetical medicolegal protection . Their lack of confidence in a legal defence derived from national evidence-based guidelines is troubling . Perhaps those strategies suggested by GPs as important in increasing their sense of medicolegal protection should be considered as a matter of some urgency. | par 12 |
| Introduction | par 13 |
| In Australia, the risk of medical litigation has increased. For general practitioners (GPs), the risk of being sued has doubled from 1:160 in 1990 to 1:84 in 1994 1. While many are concerned about the increase 2, not everyone agrees there is a crisis 3. If doctors believe these risks are substantial however, this belief alone may influence their clinical behaviour 4-7. | par 14 |
| By synthesising scientific evidence, evidence-based guidelines might reduce medicolegal risk when clinical practice is within such guidelines 8. Prostate cancer screening represents a 'test-case’ for such guidelines 9. Although there is no evidence that premature mortality from prostate cancer will reduce with mass screening, men report high rates of testing 10,11. GPs use prostate specific antigen (PSA) assays for screening purposes, either alone or combined with digital rectal examination (DRE) 12-14. Neither approach yet meets accepted criteria for screening 15 . At best, prostate cancer will be present in about one-third of men in whom a PSA test is considered raised and, at worst, in less than one in ten. Current tests cannot distinguish innocuous from aggressive malignancy. Treatment options currently available are 'unnecessary for some and insufficient for others' 16. On current evidence, men’s quality of life may be deleteriously affected by anxiety, unnecessary treatment and adverse complications if screening is recommended 15. | par 15 |
| Unsurprisingly, 39% of GPs surveyed in 1995 indicated that prostate cancer screening guidelines would be 'extremely’ or 'very’ useful 17. Three sets of guidelines ensued, each one recommending against screening 18-20. Yet Pinnock et al recently reported 'anecdotal evidence that general practitioners are concerned that if a PSA test is not offered, and prostate cancer is later diagnosed, they may be seen as negligent' 11. Those American physicians who perceive that a patient who develops prostate cancer would be successful in sueing his physician if he had not been previously screened are more likely to perform PSA tests than those who perceive that the patient would not be successful in such a suit 21. No Australian research has quantified the influence of medicolegal concerns on prostate cancer screening. We designed this study to ascertain GPs’ perceptions of medicolegal risk when asked to screen for prostate cancer and, by using case scenarios, to assess the potential impact of national guidelines on their views. | par 16 |
| Method | par 17 |
| Questionnaire development and content | par 18 |
| The first section of our self-administered questionnaire commenced with this case scenario: 'Mr Smith, a 58-year-old employed repairman, presents to his regular GP after prompting by his wife to have a test for prostate cancer. He has no urinary symptoms, no family history of prostate cancer and has not had a vasectomy’ . GPs were asked when they last had a similar request; what should be done; whether the GP would be at a risk medico- legally if s/he either did or did not perform either or both tests for screening purposes; what they would do if they were the GP and which test(s) they would advise as being the best available. We also randomised respondents to receive one of two versions of the case scenario which were identical except that, in one, the patient was an architect and, in the other, he was a repairman. | par 19 |
| We next provided verbatim the policies of the ACS 18, NHMRC 19 and RACGP 20 about screening for prostate cancer and asked if GPs were already aware of each guideline and, having read excerpts from all three, whether respondents would change their answers to the initial case scenario. We then repeated the first scenario and questions. | par 20 |
| Our second case scenario read as follows: 'Imagine the following scenario with the earlier patient, Mr Smith. Having presented for a PSA test, you discouraged him from having it. Imagine that six months later, he is diagnosed with prostate cancer after a blood test organised by a locum. He then proceeds with a formal complaint against you because you did not do the test when he first requested it’ . We asked respondents to indicate if the three guidelines would protect them in the event of such a complaint; the extent to which such a possibility influenced their practice and their perceptions of the likelihood of such a scenario. | par 21 |
| We next asked respondents to indicate their opinion of each of seven statements about prostate cancer screening, using a five-point Likert scale. Respondents then ranked five current and fourteen potential strategies to increase GPs' sense of medicolegal protection in this aspect of clinical practice. The questionnaire concluded with six socio-demographic questions. | par 22 |
| Survey administration and analysis | par 23 |
| par 24 |
| We purchased a list of all NSW general practitioners from a commercial company and randomly selected 400 names. Questionnaires and reply-paid envelopes were mailed in mid 1997 after an advance telephone prompt 22. At Day 16, non-responders received a reminder letter. At Day 35, a second questionnaire was posted to remaining non-responders. Two weeks after the second mail-out, a research assistant telephoned non-responders. | par 25 |
| The initial sample size was calculated to yield at least 200 questionnaires for analysis, thereby permitting independent and paired univariate analyses. Descriptive statistics were performed, using SPSS version 6.0. We used chi-squared to examine differences in GPs’ responses to either repairman or architect scenarios. Univariate analysis using chi-squared also was performed to determine significant associations between respondents’ beliefs and five nominated outcome variables. For these analyses, categories were collapsed into dichotomous variables. McNemars chi-squared was used to determine differences in GPs' responses to the case scenarios before and after reading the three guidelines. For all tables, where rows or columns do not add to 100%, data were missing. | par 26 |
| Results | par 27 |
| GPs' response to a request from an asymptomatic 58-year-old request for prostate cancer screening before and after guidelines | par 28 |
| Of the total GPs participating, 116 (53%) had received the repairman version of the case scenario and 103 (47%), the architect. There was no significant difference in GPs’ answers to management of patient’s request for screening before reading the guidelines (chi-squared=0.81, 1df, p=0.4) or after reading the guidelines (chi-squared=0.49, 1df, p=0.5). Responses to all scenarios were combined, therefore, irrespective of patient occupation. | par 29 |
| More than two thirds (69%) of the participating GPs had had a request for prostate cancer screening from an asymptomatic man within the previous week (n=48, 22%) or the previous month (n=102, 47%). Of the total sample, 115 (53%) GPs mentioned that PSA and DRE in combination should be done if a man requested a screening test and 121 (55%) GPs would perform the two tests in combination if they had such a request (Table 1). For men insisting on having the 'best available’ screening test, 170 (78%) of the respondents would advise both DRE and PSA in combination; 25 (11%) DRE alone and nine (4%) PSA alone although nine (4%) still advised neither test. The proportion of GPs perceiving that the GP would be at medicolegal risk if s/he did not screen for prostate cancer (61%) was significantly higher than that perceiving a risk if s/he did screen (15%) (chi-squared=13.83, 1df, p<0.01) (Table 2). | par 30 |
| Less than a quarter of the sample was aware of all three guidelines extracted in our survey (n=53, 24%). After reading the guidelines, significantly fewer respondents indicated that the GP in the case scenario was at risk medicolegally if s/he did not perform either or both tests for screening (Table 2). Further, significantly fewer perceived a medicolegal risk if the GP did screen, after reading the guidelines. Sixty-one (28%) indicated that, given the three sets of guidelines, the GP would be protected medico-legally if screening tests were not performed (Table 2). | par 31 |
| GPs’ reaction to a threat of legal action by a patient with prostate cancerGPs’ reaction to a threat of legal action by a patient with prostate cancer | par 32 |
| In the second case scenario (patient diagnosed with prostate cancer initiating a formal action against his doctor for not previously ordering a screening test), few (n=32)(15%) thought that the three guidelines 'most likely’ would protect them in the event of such a complaint. Respondents who thought such a scenario would 'most likely’ or 'likely’ influence their test-ordering (n=151, 69%) were significantly more likely than those indicating it 'unlikely’ (n=52, 24%) to consider such a scenario could happen (chi-squared=10.7, 1df, p=0.001). | par 33 |
| Respondents who stated, after reading the three Australian guidelines, that they would not screen an asymptomatic man were significantly more likely than those who would to indicate that guidelines would protect them medicolegally (chi-squared=43.81, 1df, p<0.01). Respondents who indicated that the GP would be at risk medicolegally if s/he did not perform any screening test were significantly less likely to consider that the guidelines would protect them in the event of formal complaint compared to those who did not consider the GP at risk (chi-squared=45.40, 1df, p<0.01). | par 34 |
| Respondents who self reported they would change their response to the case scenario after reading the guidelines presented in the survey were significantly more likely to think that the guidelines would protect them compared with those who would not (chi-squared=8.3, 1df, p=0.003). Likewise, respondents who, having read the guidelines, considered the GP was protected medicolegally if s/he did not perform any test also were significantly more likely to perceive that guidelines would protect themselves legally compared with those who consider the GP unprotected (chi-squared=43.12, 1df, p<0.01). | par 35 |
| GPs' views about influences on prostate cancer screening | par 36 |
| Table 3 summarises responses to seven statements about prostate cancer screening in general practice. Of the sample, 191 GPs (87%) 'strongly agreed' or 'agreed’ that patients' decisions to be screened should be based on full disclosure of what is known about diagnosis and treatment of early prostate cancer. While 189 (86%) of the sample either 'strongly agreed’ or 'agreed’ a positive PSA test may cause unnecessary anxiety for patients, 130 (59%) also indicated that a GP could be sued for not ordering a PSA test for someone who subsequently develops prostate cancer (Table 3). Respondents were significantly more likely to agree that a GP could be sued by a patient who subsequently developed prostate cancer for not performing a PSA test than by a patient who subsequently experienced adverse consequences of treatment resulting from a PSA test which had been performed (chi-squared=6.6, 1df, p=0.01). | par 37 |
| GPs' views of guidelines and other current or potential strategies to increase their sense of medicolegal protection | par 38 |
| Of five current strategies already in place to increase their sense of medicolegal protection, 110 (50%) of respondents considered guidelines based on systematic reviews of the evidence as 'very important'; followed by NHMRC endorsement (n=96)(44%); quality of patient information (n=96)(44%) and availability of patient information about prostate cancer screening (n=84)(38%). More than 10% (n=24, 11%) were unsure of the impact of the current adversarial court system however. With regard to potential strategies, there was strong support for written advice about the legal status of guidelines and inviting GP peers as well as specialists as expert witnesses to court (Table 4). | par 39 |
| Discussion | par 40 |
| Evidence-based guidelines appear to have little influence on prostate cancer screening by GPs. Requests for such testing appear to be frequent. While there was no discrimination in the provision of a screening test on the basis of a patient's occuption, at least 90% of respondents would proceed with at least one screening test (DRE, PSA or both in combination). Furthermore, a significant majority of respondents considered the GP in the first scenario would be 'at risk' medicolegally if s/he did not screen and moreso than if s/he did screen. Awareness of national guidelines was low. Surveys repeatedly demonstrate poor penetration of guidelines into general practice 24, generating interest in more deliberate implementation 9,25,26. | par 41 |
| Further, our proxy for assessing guidelines impact on test-ordering augurs poorly for their influence in reducing perceptions of risk of suit . Although (having read the guidelines) significantly fewer considered the GP was at risk if s/he did not screen, only 28% indicated the GP was protected medicolegally by three sets of evidence-based guidelines, even when complying with their recommendations against screening. Even fewer (15%) considered protection would be afforded them if a man, having been diagnosed elsewhere with prostate cancer, then proceeded to initiate a legal action for previous failure to screen. Those who perceived such a situation was likely were also more influenced by such a situation in their test-ordering. In contrast, those who would not screen were more likely to feel protected by guidelines. Those who adopted this view point after reading the guidelines appeared to place greater faith in the medicolegal protection afforded by them. | par 42 |
| Respondents' views differ from those of participants at a forum held in 1997 to examine legal implications of guidelines 27. While a majority of respondents supported a clear statement about the legal status of guidelines in a court of law, it was not recommended by the forum. To ensure that guidelines ' are best used within the current legal framework' 27, an increase in the use of GP experts, reference to evidence-based guidelines whenever possible and attention to public information especially in languages other than English will be needed . | par 43 |
| Our results also confirm the observation of Pinnock et al that prostate cancer screening can be motivated by medicolegal concerns 11. In the US, physicians self-reported PSA testing practices correlate with their views on a medical malpractice case vignette 21. Our findings suggest this to be the case also in Australia, prompting us to agree that ' if physicians perform PSA tests at least to some extent out of fear of malpractice, then the standard of care may eventually be set by these defensive practices' 21 . GPs nonetheless support 'informed decision-making ' by men themselves. Specifically, 88% of our respondents 'agreed' or 'strongly agreed' that men's decisions to be screened should be based on 'full disclosure' of what is known about the diagnosis and treatment of early prostate cancer. Information necessary to make such a decision has recently been clarified through expert survey 28. | par 44 |
| Two methodological caveats are recognised. First, the validity of our use of scenarios before and after exposure to guidelines has not been formally tested against actual behaviour. We know of no other feasible method to examine the influence of medicolegal risk on test-ordering however. Second, our response rate was lower than we had hoped (65%) but the professional and demographic characteristics of our sample were comparable with the NSW GP profile . | par 45 |
| In summary, GPs in our survey perceived limited medicolegal protection from national evidence-based guidelines. Their lack of confidence in a legal defence derived from national evidence-based guidelines is troubling . Perhaps those strategies suggested by GPs as important in increasing their sense of medicolegal protection could be considered further by health departments, medical defense organisations and professional opinion leaders. | par 46 |
| Acknowledgments |
| References |
| 1. Keaney M. Is there a medical litigation crisis? Individual viewpoints on the perceived medical litigation crisis. Is litigation increasing? Med J Aust 1996;164:178-182 |
| 2. Nissel P. Is there a medical litigation crisis? Individual viewpoints on the perceived medical litigation crisis. Is there a crisis? Med J Aust 1996;164:178-182 |
| 3. Dunn I. Is there a medical litigation crisis? Individual viewpoints on the perceived medical litigation crisis. Crisis or beat- up? Med J Aust 1996;164:178-182 |
| 4. Weisman C, Morlock L, Teitelbaum M, Klassen A, Celentand D. Practice changes in response to maplractice litigation climate. Results of a Maryland physician survey. Med Care 1989; 27: 16-24 |
| 5. Kessler D, McClellan. Do doctors practice defensive medicine? National Bureau of Economic Research working paper series, Cambridge, 1996 |
| 6. Summerton N. Positive and negative factors in defensive medicine: a questionnaire study of GPs. Brit Med J 1995; 6971: 27-9 |
| 7. Hancock L. Defensive medicine and informed consent: a research paper. Review of professional indemnity arrangements for health care professionals. Canberra. AGPS. 1993 |
| 8. Barratt A, Bates P. O tell me the truth about evidence. Aust NZ J Pub Hlth 1997; 21: 441-4 |
| 9. Ward J. Prostate cancer screening: too much, too soon? Cancer Forum 1998; 22: 18-23 |
| 10. Ward J, Hughes AM, Hirst G, Winchester L. Men’s estimates of prostate cancer risk and self-reported rates of screening. Med J Aust 1997; 167: 250-3 |
| 11. Pinnock C, Weller D, Marshall V. Self-reported prevalence of prostate-specific antigen testing in South Australia: a community study. Med J Aust 1998; 169: 25-8 |
| 12. Sladden M, Dickinson J. General practitioners’ attitudes to screening for prostate and testicular cancer. Med J Aust 1995; 162: 410-3 |
| 13. Ward J, Young J, Sladden M. General practitioners' views and use of tests to detect early prostate cancer. Aust NZ J Pub Hlth 1998; 22: 374-80 |
| 14. Ward J, Gupta L, Taylor N. Do general practitioners use prostate-specific antigen as a screening test for early prostate cancer? Med J Aust 1998; 169: 29-31 |
| 15. Hirst G, Ward J, Del Mar C. Prostate cancer screening: the case against. Med J Aust 1996; 164: 285-287 |
| 16. Whitmore W. Management of clinically localised prostate cancer; an unresolved problem. JAMA 1993; 269:2676-7 |
| 18. Australian Cancer Society. Prostate cancer screening: guidelines for health professionals. Cancer Forum 1995; 19: 47-50 |
| 19. Australian Health Technology Advisory Committee. Prostate Cancer Screening. AGPS, Canberra: 1996. |
| 20. Royal Australian College of General Practitioners. Guidelines for preventive activities in general practice . 4th ed. Sydney: RACGP, 1996:26. |
| 21. Collins M, Fowler F, Roberts R, Oesterling J, Annas G, Barry M. Medical malpractice implications of PSA testing for early detection of prostate cancer. J Law, Med & Ethics 1997; 25: 234-42 |
| 22. Young JM, Ward JE. Improving survey response rates: a meta-analysis of the effectiveness of an advance telephone prompt from a medical peer. Med J Aust 1999 (in press) |
| 23. Commonwealth Department of Health and Family Services. General practice in Australia: supplementary tables 1997 . Canberra, GP Branch, 1997: 19 (Publication no. 1838) |
| 24. Gupta L, Ward J, Hayward R. Clinical practice guidelines in general practice: a national survey of recall, attitudes and impact. Med J Aust 1997; 166: 69-72 |
| 25. Hirst G. Clinical practice guidelines - to what end? (letter). Med JAust 1997; 167: 288 |
| 27. Pelly J, Newby L, Tito F, Redman S, Adrian A. Clinical practice guidelines before the law: sword or shield? Med J Aust 1998; 169: 330-3 |
| 28. Ward J, Girgis S. Making an 'informed decision' about prostate cancer screening: what information is required? Hlth Prom J Aust 1998; 8: 173-6 | par 0 |
| par 1 | ||||||||||||||||||||||||||||||
| par 2 | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| par 3 | ||||||||||||||||||||||||||||||||||||||||||||||||||||||
| par 4 |
| par 5 | ||||||||||||||||||||||||||||||||||||||||||||||||||
This is the first revised version of this article. Click here for the current version or click here for the original submission | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||