An official website of the United States government
Here's how you know
Official websites use .gov
A
.gov website belongs to an official
government organization in the United States.
Secure .gov websites use HTTPS
A lock (
) or https:// means you've safely
connected to the .gov website. Share sensitive
information only on official, secure websites.
As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with,
the contents by NLM or the National Institutes of Health.
Learn more:
PMC Disclaimer
|
PMC Copyright Notice
Editor—The editorial by Al-Shahi and Warlow is timely and well argued.1 The implications of a strict requirement for written consent from patients in observational and epidemiological research are extremely serious. Research in histopathology would be particularly severely affected. It is essential that patients' consent is obtained in interventional studies or where information, or, in the case of pathology, tissue, is collected solely for the purpose of research. However, applying this requirement for every study of archived biopsy, surgical or necropsy tissue blocks, or review of patient records, where diagnosis has been made and where patient management will not be affected, is impracticable.
Studies of archival material have been the core of research in histopathology. A glance at the methods section of the papers of any pathology journal reveals that most have used archival tissue with no indication that specific patient consent has been obtained. For recent issues of the Journal of Pathology, Histopathology, and the Journal of Clinical Pathology the proportions of archival, non-consented research were 9/16, 9/9, and 11/13 papers, respectively.
Until now, this form of research has been simple and cheap to undertake, and has provided valuable information on the causes and pathology of many human diseases. It is unlikely that any of the large studies of gastric biopsies, cancers, etc would have been conducted had every patient or their relatives had to be contacted to obtain consent. We would be left with only anecdotal information on which to base prognosis and treatment for many common cancers and we would know much less about many common diseases. Despite considerable problems in obtaining funding for research in histopathology, authors in the United Kingdom continue to produce papers of high quality, but it is doubtful that this will continue if individual consent for each study becomes a legal requirement.
Either a general consent for use of residual tissues for research should be allowed or alternatively the NHS needs to fund the staff necessary to obtain the consent from central or regional research and development budgets, which could be accessed by researchers wishing to undertake archival studies. This would, however, be both cumbersome and costly. If observational research in pathology or other subjects and epidemiological studies are to be outlawed, the public interest will be damaged and patients in the future will be the losers.
References
1.Al-Shahi R, Warlow C. Using patient identifiable data for observational research and audit. BMJ. 2000;321:1031–1032. doi: 10.1136/bmj.321.7268.1031. . (28 October.) [DOI] [PMC free article] [PubMed] [Google Scholar]
BMJ. 2001 Apr 7;322(7290):858.
Argument for consent may invalidate research and stigmatise some patients
Editor—We would like to congratulate Al-Shahi and Warlow on their editorial.1-1 A blanket requirement for anonymisation of data and informed consent from all participants to use identifiable data about them would jeopardise the methodological integrity of research and audit. This point has been highlighted during our recent work.
Patients with schizophrenia suffer from increased physical ill health and excess mortality.1-2,1-3 As such patients increasingly rely on primary care for their physical health care, it is important to audit that the care they receive is comparable to that of patients who are not mentally ill. Our study, addressing this issue, was a case-matched retrospective review of primary care records. Local research ethics committees originally requested that patients provide consent. But systematic bias could invalidate the findings of observational studies if people were excluded because they did not consent.1-1 Obtaining consent heightened this risk in our study for two reasons. Firstly, we are examining care given to patients—those feeling strongly about this (having had very good or poor care) may be more likely to consent. Secondly, study patients have schizophrenia and characteristics of the disease themselves—for example, paranoia—may reduce the likelihood of unwell patients consenting. Therefore, consenting patients could be less ill and better approximated to the control population. The need for consent may minimise the effect observed, potentially invalidating the results.
Production of substandard flawed research is less ethical than the use of anonymised data by professional researchers. Mental health research poses unique problems, and adopting the argument for consent may further stigmatise an already stigmatised group by the production of low quality research. After discussion, the three local research ethics committees we approached agreed with this. We were also confused by the Data Protection Act 1998 while trying to clarify whether our research fell under the umbrella “necessary for medical purposes.” Practice recruitment was hindered by our inability to give firm guarantees about the work not breaching the Data Protection Act. We agree that the law needs clarification, to protect both researchers and the public. It may be appropriate for a committee to be appointed to review protocols prospectively in a similar manner to research ethics committees. We aim to research issues of clinical importance and attempt to select the most robust methods. It is inappropriate for researchers to have to choose between adopting a weaker and biased method or risk breaching the act.
References
1-1.Al-Shahi R, Warlow C. Using patient identifiable data for observational research and audit. BMJ. 2000;321:1031–1032. doi: 10.1136/bmj.321.7268.1031. . (28 October.) [DOI] [PMC free article] [PubMed] [Google Scholar]
1-2.Kendrick T. Cardiovascular and respiratory risk factors and symptoms among general practice patients with long-term mental illness. Br J Psychiatry. 1996;169:733–739. doi: 10.1192/bjp.169.6.733. [DOI] [PubMed] [Google Scholar]
1-3.Brown S. Excess mortality of schizophrenia: a meta-analysis. Br J Psychiatry. 1997;171:502–508. doi: 10.1192/bjp.171.6.502. [DOI] [PubMed] [Google Scholar]
BMJ. 2001 Apr 7;322(7290):858.
Integrity of communicable disease surveillance is important patient care
Editor—As Al-Shahi and Warlow point out in their article there are conflicting goals: the drive toward greater patient autonomy and the wish to maintain an evidence base for medical practice by collecting and interpreting observational data.2-1 Both are important, but the former should not needlessly jeopardise the latter.
The General Medical Council's guidance acknowledges the legal requirement for local statutory notification of certain infectious diseases without explicit consent (paragraph 43).2-2 But non-statutory laboratory and clinical reporting systems have added appreciably to our understanding of communicable diseases. There is great concern that bias will arise in national reporting of incidents of infectious disease, either as a consequence of patients not agreeing to their infections being reported or because clinicians do not have the time or the opportunity to obtain explicit consent. Both will diminish, and potentially destroy, the surveillance of communicable diseases. In the short term, there might be failures to recognise or mount timely responses to adverse reactions to vaccines or to outbreaks of infectious disease, especially those that do not form local clusters but occur as a series of apparently sporadic cases countrywide. Only rigorous national surveillance and national collation of data will identify these events as coming from a common source.2-3
At the surveillance centres we rely on the goodwill of health professionals for prompt reports of communicable diseases. We are aware that all data relating to individual patients must be secure. Some within the healthcare professions, and some patients, however, feel that patients' right to privacy overrides the need to maintain surveillance. In response to such concerns we are continuing to seek ways of reinforcing the security of data entrusted to us.
The consequences if the surveillance of infectious diseases unravels include reduced quality or even loss of trend data; absence of monitoring residual incidence of infections preventable through immunisation programmes and inability to monitor effectiveness and safety of vaccines; delayed response to outbreaks; incomplete national data on HIV and other infectious diseases with serious long term consequences; possible failure to recognise imported and nosocomial infections and zoonoses; and inability to monitor effectiveness of interventions such as infection screening.
Just as detrimental in the long term would be loss of familiarity among health professionals with the epidemiology of infection, resulting in inferior care of patients and their contacts. The integrity of surveillance of communicable diseases is as important to the care of individual patients as to the health of the community as a whole. Strict precautions are already in place to protect confidentiality. Insisting on explicit individual consent as a prerequisite to the reporting of infectious diseases will serve neither patients' interests nor the wider public health.
References
2-1.Al-Shahi R, Warlow C. Using patient identifiable data for observational research and audit. BMJ. 2000;321:1031–1032. doi: 10.1136/bmj.321.7268.1031. . (28 October.) [DOI] [PMC free article] [PubMed] [Google Scholar]
2-2.General Medical Council. Confidentiality: protecting and providing information. London: GMC Publications, September; 2000. [Google Scholar]
2-3.Rushdy AA, Stuart JM, Ward LR, Bruce J, Threlfall EJ, Punia P, et al. National outbreak of Salmonella senftenberg associated with infant food. Epidemiol Infect. 1998;120:125–128. doi: 10.1017/s0950268897008546. [DOI] [PMC free article] [PubMed] [Google Scholar]
Editor—We are delighted by the supportive responses to our editorial. It is surely testament to the far-reaching implications of strict confidentiality and data protection guidance that authors from so many other disciplines have echoed our concerns: pathology, public health laboratory sciences, primary care and general practice, and unpublished electronic responses from oncology and clinical audit.3-1 This further reinforces the need for urgent action to protect the wide range of activities carried out by bona fide professionals with a utilitarian desire to improve the public health.
It is absolutely essential that policymakers and the public are fully aware of the detrimental effects of excluding people—who cannot, or do not, consent—from ethically approved and peer reviewed observational studies (authorisation bias). As Cox, Roberts and Wilson, and Evans and Ramsay persuasively argue, studies subject to such bias may generate the wrong conclusions, thereby discriminating against current and future patients. The profound impact of authorisation bias does not seem to be as clear to policymakers as it does to researchers, perhaps due to the sparse literature on the subject.3-2 We encourage anyone concerned about this issue to explore their own datasets for authorisation bias and publish their results. Moreover, this concern needs to be explained more to the public both in the national press3-3 and by doctors when seeking their patients' consent.
Legislation and professional guidance on confidentiality and data protection have both moved on since our editorial. Clause 59 of the Health and Social Care Bill now enables the health secretary in England to sanction the use of patient-identifiable information for medical purposes in the interests of improving patient care or in the public interest.3-4 The General Medical Council has given cancer registries a breathing space until October 2001 in which to put appropriate mechanisms in place for seeking and recording consent according to their guidance on confidentiality. But surely registries of all important diseases, and not just cancer, should be included? Nevertheless, any observational study will struggle to avoid authorisation bias under the GMC's terms.
We would welcome a return to pragmatic guidance on the conduct of observational studies that both protected patients and promoted medical progress.3-5 But for now, will the medical community just stand back and allow people, with the best of intentions, to compromise the greater good of research, disease surveillance, audit, and clinical governance?
3-2.Jacobsen SJ, Xia Z, Campion ME, Darby CH, Plevak MF, Seltman KD, Melton LJ. Potential effect of authorisation bias on medical record research. Mayo Clin Proc. 1999;74:330–338. doi: 10.4065/74.4.330. [DOI] [PubMed] [Google Scholar]
3-3.Ballantyne A. Silence that could be a death sentence. Times 2000 Oct 27:22.
3-4.Dyer C. Bill gives government power to breach patient confidentiality. BMJ. 2001;322:256. . (3 February.) [PMC free article] [PubMed] [Google Scholar]
3-5.Royal College of Physicians. Independent ethical review of studies involving personal medical records. London: RCP; 1994. [Google Scholar]
