Research-Based argument essay

58 journal of law, medicine & ethics

Ethics, Information Technology, and Public Health: New Challenges for the Clinician- Patient Relationship Kenneth W. Goodman

O ne of the largest, oldest, and most interest- ing challenges in health care is the balanc- ing act in which clinicians have generally

uncontroversial duties both to individual patients and to communities. Physicians and nurses must — so we teach them — put patients first, and at the same time recognize that individuals are members of com- munities. Individuals affect the health of communi- ties, and communities affect the health of individuals. Thus, the moral and professional duties that result are sometimes in conflict.

Moreover, the traditional, prosaic clinical encoun- ter is evolving in an environment increasingly shaped (albeit too slowly according to some) by electronic health records, personal health records, pharmacog- enomics and vast networks of data collection and stor- age for public health surveillance, human subjects research, health services evaluation, and comparative effectiveness research. Health information technology is changing everything. It would be perverse otherwise: imagine large amounts of data and information either ignored, missed, or collected and then ignored. The foundations of evidence-based practice and evidence- based public health undergird bold-faced duties both to learn more about the world and to communicate this knowledge to those in clinical and public health prac- tice.1 Even those who criticize evidence-based medi- cine for its failure to celebrate the wonders of clinical judgment would never argue that data from clinical encounters not be collected, analyzed, and shared.

The challenges of making sense of all this clinical data are amplified now that there are tools adequate, more or less, to do these tasks. We can suggest with comparatively large granularity a first-approximation typology of these challenges:

professional education;•  clinicians’ duties to collect and share patient •  data; and patients’, clinicians’, and society’s duties to con-•  tribute to public health.

In what follows we survey these challenges and suggest the features of an ethically optimized stance toward the use of information technology for pub- lic health with special regard to the clinician-patient relationship.

Kenneth W. Goodman, Ph.D., is a Professor of Medicine and Philosophy and Director of the Bioethics Program at the University of Miami.

Professional Education We have long known that physicians and nurses have duties to people who are not their patients. These are interesting and distinctive professional obligations because they are owed to people one has not met, does not know, and with whom one has not entered into any kind of professional or social relationship. For physi- cians or nurses, however, this is just what we mean by saying they have duties to the health of the community they serve.

From Hippocrates and Avicenna through Parcelsus and John Snow, there has been clear recognition of the relationship, of some relationship at least, between the health of individual patients and the health of collec- tives. The birth of formal epidemiology is customarily traced to Snow, who in London in 1854 famously plot- ted clusters of cholera deaths in Soho and identified a well at Broad Street as the source of contaminated water. For Snow, “information technology” was a map, a pen, and the addresses of people who died. The insight has come to be a common, if perhaps less than appreciated, component of the training of physicians, especially regarding disease prevention.2 Information technology as such enjoys somewhat less emphasis in the core training of clinicians. The Liaison Commit- tee on Medical Education hints at the importance of health information technology, but emphasizes it in terms of libraries and access to literature: “The medi- cal school must have access to well-maintained library and information facilities, sufficient in size, breadth of holdings, and information technology to support its education and other missions.”3 The growth of medi- cal and nursing informatics as disciplines in their own right and with commensurate pedagogic standards has been documented,4 but precious few medicine and nursing postgraduate programs include information technology as a core component of the curriculum.

There is unfortunately no formal pedagogic link between public health and information technology in the training of clinical health professionals. Indeed, the idea that crowded professional training curricula

might add content in public health informatics, when they barely if at all address these disciplines sepa- rately, is optimistic. Nevertheless, if we take seriously the inferential chain that clinicians have duties to individuals and populations; that both relationships are information-intensive; that the clinician-patient relationship entails duties to public health; that clini- cal and hospital information systems are expanding; and that this provides a mechanism by which to meet such obligations — then it should be uncontroversial

to suggest that professional training in nurs- ing and medicine should include content on knowledge, skills, and attitudes related to public health information processing.

What gives such a recommendation moral force is the hypothesis, equally uncontrover- sial, that such training will improve the health of populations. Indeed, there are increasingly good reasons to believe that health IT educa- tion is needed to foster the development of clinicians,5 in part because of evidence that such training improves health outcomes. The point can be made in the opposite direction:

the failure to provide adequate training in health infor- mation technology will impede the evolution of HIT as a public health resource — a congenial way of saying that such a failure will prevent populations from real- izing certain health benefits, that is, will allow people to be harmed.

Along with general clinical practice and clinical information technology, such a pedagogic initiative as advocated here will raise and must address correlate ethical issues.

Clinicians’ Duties to Collect and Share Patient Data It has long been established that clinicians have moral and/or legal obligations to report or disclose certain maladies, injuries, adverse events, errors, and other events. These obligations generally fall under several labels: public safety, duty to warn, public health sur- veillance. While these are related in many instances, we here confine our attention to the last. The origin of such mandatory reporting lies in the discovery of the value of the collection of vital statistics: “The impetus for a truly effective system came from the realization by some very astute statisticians and physicians… that records of births and deaths, particularly records of deaths by cause, were needed for the control of epi- demics and the conservation of human life through sanitary reform.”6 The history of vital statistics is commonly traced to the work of John Graunt of Lon- don, who published in 1662 his ‘‘Natural and Political Observations Mentioned in a Following Index, and

the effects of health information technology on the physician-patient relationship • spring 2010 59

Kenneth W. G oodman

The failure to provide adequate training in health information technology will impede the evolution of HIT as a public health resource — a congenial way of saying that such a failure will prevent populations from realizing certain health benefits.

60 journal of law, medicine & ethics

SYMPOSIUM

Made Upon the Bills of Mortality,” a masterpiece of untutored demography. Some 350 years later, we have achieved broad agreement on the duties of clinicians to report to authorities or otherwise disclose informa- tion about child and elder abuse, gunshot wounds, medical error, nosocomial infections, various infec- tious diseases, and so forth. In some practices, clini- cians contribute to tumor registries and the like.

There is a great deal of evidence documenting the privacy-friendly public health benefits realized from analysis of data acquired during clinician-patient encounters. Practical examples come from domains as varied as infectious disease,7 maternal-child health,8 vaccination efficacy,9 and syndromic surveillance.10

Once it has been established that an action can reduce suffering, improve health, or impede untimely mortality, there generally follows a corresponding obligation to repeat the action and, as appropriate, coordinate similar actions and communicate about and analyze them. It is only when such an obligation is in conflict with (an)other obligation(s) that we might consider modifying it or even setting it aside.

The first analyses of ethical and legal issues in health information technology included the important observation that it might be blameworthy not to use a computational tool if there were reason to believe the tool would improve patient care.11 We are in a position now to extend this duty to apply to public health con- tributions by clinicians using health information tech- nology. In some respects, this is a call for physicians and nurses to become fully fledged members of the public health community. Such a change could lead to vast increases in the amount of personal patient data shared with public health officials. We consequently need to attend to the warrant and scope of such shar- ing, and we need to address any special issues raised by the use of information technology itself. These tasks can be accomplished under three headings: valid con- sent, privacy, and human subjects research.

Its historically special if not unique properties cast the doctor-patient relationship as a dyad: knowledge- able professional and vulnerable patient, the former owing an ensemble of duties to the latter. Acknowl- edging cultural and historical variance regarding the extent to which a physician controls, oversees, or man- ages the relationship, there has nevertheless evolved international recognition of the core importance of valid consent for clinical care and human subjects research. Valid consent limits or at least modulates physician authority. If clinicians are to become more reliable partners in the health of communities, it is reasonable then to inquire how much of a say patients have, and should have, in the data sharing that under- girds these partnerships.

Let us imagine the following state of affairs. The current and unprecedented initiative to foster univer- sal use of electronic health records (EHR) is a success. It includes widespread adoption of personal health records, including portable and online resources for patients to hold and track their medical informa- tion in isomorphs to the official electronic health record. Every physician and nurse now uses an EHR to acquire and store patient information, and inform diagnosis and treatment. In conjunction with a hypo- thetical National Initiative for Public Health Technol- ogy Infrastructure (NIPHTI), health data of use or interest to public health officials and epidemiologists are shared continuously in real time, and retained in a very large national data base. (This mechanism is not unlike the current national system for poison control data collection and sharing.) Our NIPHTI also therefore constitutes a comprehensive and ubiquitous early-warning system for public health emergencies, including pandemic influenza, bioterrorism, and so forth.12

Following are questions NIPHTI raises, questions which are orthogonal to those customarily raised by the professional-patient relationship:

Is patient consent required for such data shar-•  ing? If so, why — given that current disclo- sures of patient information to public health authorities are not preceded by consent? How fine grained would this consent need to be? That is, how much detail about the data, the data base, and their use must be disclosed to patients? Should doctors and nurses introduce an expanded consent process into the clinical relationship, or are there other ways to obtain consent without interfering with or burdening the clinical encounter and without introducing selection bias into the process? What are the consequences of including or excluding patient- controlled and patient-augmented personal health records? Thoughtful analyses of the pri- vacy-consent nexus repeatedly demonstrate that clumsy and vigorous invocations of privacy rights can impede public health and research without improving privacy protections.13

Independent of any putative consent require-•  ment, do patients regard such data sharing as a violation of confidentiality? Indeed, do they regard current data reporting as a violation? If there is a difference, is it by virtue of magnitude or kind? Do standards for data de- and re-iden- tification matter to patients? If they understood the intent and value of public health data report- ing as mediated by their doctor or nurse, and if

the effects of health information technology on the physician-patient relationship • spring 2010 61

Kenneth W. G oodman

their data were adequately anonymized, would their views change? Preliminary evidence sug- gests that patients are in fact willing to share their information and, indeed, that privacy con- cerns do not necessarily pose the kinds of con- straints and inhibitions customarily invoked to limit information sharing.14

Does such data collection and analysis constitute •  human subjects research? One of the conceptu- ally most interesting questions raised by public health data collection and analysis is closely related to the relationship between research and surveillance. Attempts to demarcate the two are complicated, at least in the United States, by the definition of “research” in the Code of Federal Regulations: “a systematic investigation, includ- ing research development, testing and evaluation, designed to develop or con- tribute to generalizable knowledge.”15 Epistemologists rightfully cringe at the very idea of something called “general- izable knowledge.” Because much that does not look like, and in fact is not, human sub- jects research falls under this definition, no small effort has been spent trying to clarify it. Surely the efforts of John Graunt, the Centers for Dis- ease Control and Prevention, and similar others all count as research under this definition, even as systematic public health data collection also is commonly described and understood as “surveil- lance” and not research.16 Clearly we do not want to regard or render all clinical encounters under NIPHTI as human subjects research.

This delicious and ancient tension, between “patient- focused health care” and “population health” (to use current terms of art) suggests an ensemble of opportu- nities for empirical and conceptual research: we know too little about patient preferences regarding consent for public health and secondary use of anonymized data, and we require more and better tools for estab- lishing distinctions between human subjects research and the garden-variety missions of public health researchers and epidemiologists. Indeed, the effects of population health-driven uses of information technol- ogy on the clinician-patient relationship is itself a rich opportunity for future empirical research.

Patients’, Clinicians’, and Societies’ Duties to Public Health Clinical epidemiology is the application of epidemio- logic findings, methods, and inferential strategies to the care of individual patients.17 Individual patients are,

sooner or later, the source of evidence for these find- ings. The cycle feeds on information, which becomes evidence, which improves the cycle. That is, individual patients and their populations grow healthier, in prin- ciple, as more information produces more and better evidence. This should be uncontroversial.

Because we rightly value the health of individuals and populations, it follows that efforts to improve the process are generally praiseworthy, and efforts to impede the process are generally blameworthy. More- over, it is, as above, widely and correctly agreed that physicians, nurses, and other clinicians have duties both to individual patients and to groups of patients,

and some of the duties to the latter involve using infor- mation about the former. Relatedly, civil society has bold-faced obligations to foster individual and pop- ulation health. To the extent this is an information- intensive undertaking, debates about privacy and con- fidentiality are, in this respect, largely about how to balance these duties. Too often overlooked in debates about the duties of clinicians and societies are the duties of patients.

Consider a hypothetical physician who refused to incorporate years of experience when making future clinical judgments because to do so would mean using information about prior patients without their explicit permission. We would regard such a physician as unfit for practice — the very idea that one would decline to learn from experience (and cite patient privacy prefer- ences in the process) would be medically and morally fatuous and kooky. Moreover, any patient who insisted her physician or nurse not apply insights from her case to future patients would be no less peculiar. It would be equally and ethically bizarre if the patient cited a right to privacy, that is, suggested that such a subse- quent use of her personal information would consti- tute a violation of her confidentiality. If, on the other hand, the clinician were to tell the subsequent patient that her case was just like that of Ms. Crabtree of 123 Elm Street, then such a disclosure would constitute an unacceptable violation.

The point is obviously that patients, too, have obli- gations to contribute to public health, and that a broader adoption of clinical information technology

Indeed, the effects of population health- driven uses of information technology on the clinician-patient relationship is itself a rich opportunity for future empirical research.

62 journal of law, medicine & ethics

SYMPOSIUM

(perhaps especially including personal health records) provides a means for this in the context of the doctor- and nurse-patient relationship — as long as certain conditions are met and certain safeguards are in place. Such conditions and safeguards are simply and pre- cisely the stuff of contemporary debates about second- ary use of health information technology: data de- and re-identification policies and technologies; the need for and extent and role of patient control of personal data; the relationship between individual consent and “community consent”;18 and so forth.

There is much to learn about how such conditions and safeguards will affect the clinician-patient rela- tionship in an information-intense ecosystem. Tradi- tional conceptions of the relationship as trust-based make clear the importance of bi-directional commu- nication. Information technology, whether for clinical care or public health, will doubtless affect this rela- tionship — although there is no reason to suppose that public health component of this will add anything to the effects of the clinical computing component. That is, if clinicians do in fact widely adopt health informa- tion technology for clinical purposes, the public health uses of data generated during clinical encounters are unlikely to add in any significant way to any effects on the relationship caused by the adoption.

There might however be one exception, namely the fear that if clinical data are shared and analyzed for public health purposes, then such sharing and analy- sis might somehow increase the risk of social bias or insurance discrimination. We have said little so far about the extent to which personal identifiers would be attached to a large-scale public health informatics system, but it should be clear that the needs of pub- lic health are traditionally and would continue to be served by data which have been adequately anony- mized. No doubt “adequately anonymized” requires additional analysis and explication, and we need to be mindful of the fact that disparate anonymized data can in some cases be concatenated or reanalyzed to elicit information about identifiable individuals. But this is a well-known ethical, legal, and security issue,19 and it would be hyperbolic and mischievous to allow this issue alone to derail or impede efforts to realize a new generation of public health benefits. More- over, to the extent that fears of discrimination focus on access to health care, we have identified another reason to support a system of universal coverage. The idea — the very idea — that a patient would identify a self-interest in withholding anonymized health data from public health analysis is an ethico-socio-political tragedy.

Patients, Clinicians, and Health Information Technology People are known for making decisions which in many cases subvert their interests. Consumer psychology research has for instance shown that people happily accept “default” positions such that if one asks if some- one wants or values something, the number of respon- dents who agree is smaller than if respondents are asked to “opt out” or decline to disagree.20 This insight, were it applied more widely in organ procurement, for example, which most people find inoffensive, would save thousands of lives.21 What Max H. Bazerman and colleagues have called the “irrational preference for harms of omission over harms of action”22 has yet to be explored in the context of privacy and secondary use preferences regarding the use of personal data for public health. Hypotheses therefore to be tested explicitly, as suggested above, include that ordinary patients, all things being equal:

1. would not object to the use of their information for public health;

2. assume such use is already occurring; and 3. agree they have a moral obligation to permit such

use.

There is much to be learned about the effects of health information technology on the clinician-patient rela- tionship. Since adoption of this technology is moving forward with good reason anyway for clinical pur- poses, it would be a missed opportunity not to use the same technology for public health. It has been our the- sis here that this technology bids fair to improve pub- lic health in nontrivial ways, an empirical claim; and that clinicians and patients therefore share a moral obligation to support such improvement, an ethical claim. Fortunately, there is plenty and good evidence in support of the former. What remains is to demon- strate that a fully fledged public health information technology system that incorporates quotidian clini- cal and patient-entered data will support and not degrade the clinician-patient relationship; indeed, we have suggested, there is reason to believe it will have no effect at all. This means there is no good reason to oppose it.

Acknowledgement Work on this article was supported in part with funding from the Robert Wood Johnson Foundation’s Pioneer Portfolio.

References 1. K. W. Goodman, Ethics and Evidence-Based Medicine: Fal-

libility and Responsibility in Clinical Science (New York: Cambridge University Press, 2003).

the effects of health information technology on the physician-patient relationship • spring 2010 63

Kenneth W. G oodman

2. D. R. Garr, D. T. Lackland, and D. B. Wilson, “Prevention Education and Evaluation in U.S. Medical Schools: A Status Report,” Academic Medicine 75, no. 7 (2000): S14-S21.

3. Liaison Committee on Medical Education, Functions and Structure of a Medical School: Standards for Accreditation of Medical Education Programs Leading to the M.D. Degree, Chicago, 2008, at 26, available at <http://www.lcme.org/ functions2008jun.pdf> (last visited November 30, 2009).

4. C. P. Friedman, R. B. Altman, I. S. Kohane, K. A. McCormick, P. L. Miller, J. G. Ozbolt, E. H. Shortliffe, G. D. Stormo, M. C. Szczepaniak, D. Tuck, and J. Williamson, “Training the Next Generation of Informaticians: The Impact of ‘BISTI’ and Bio- informatics – A Report from the American College of Medi- cal Informatics,” Journal of the American Medical Informatics Association 11, no. 3 (2004): 167-172.

5. K. S. Desjardins, S. S. Cook, M. Jenkins, and S. Bakken, “Effect of an Informatics for Evidence-Based Practice Curriculum on Nursing Informatics Competencies,” International Journal of Medical Informatics 74, nos. 11-12 (2005): 1012-1020; R. M. Gardner, J. M. Overhage, E. B. Steen, B. S. Munger, J. H. Holmes, J. J. Williamson, D. E. Detmer, and AMIA Board of Directors, “Core Content for the Subspecialty of Clinical Informatics,” Journal of the American Medical Informat- ics Association 16, no. 2 (2009): 153-157; C. Safran, M. M. Shabot, B. S. Munger, J. H. Holmes, E. B. Steen, J. R. Lump- kin, D. E. Detmer, and AMIA Board of Directors, “Program Requirements for Fellowship Education in the Subspecialty of Clinical Informatics,” Journal of the American Medical Infor- matics Association 16, no. 2 (2009): 158-166.

6. M. A. Freedman and J. A. Weed, “The National Vital Statis- tics System,” in P. W. O’Carroll, W. A. Yasnoff, M. E. Ward, L. H. Ripp, and E. L. Martin, eds., Public Health Informatics and Information Systems (New York: Springer, 2003): 269- 285, at 271.

7. F. T. Bourgeois, S. C. Porter, C. Valim, T. Jackson, E. F. Cook, and K. D. Mandl, “The Value of Patient Self-report for Dis- ease Surveillance,” Journal of the American Medical Infor- matics Association 14, no. 6 (2007): 765-771.

8. J. R. Lumpkin, “History and Significance of Information Sys- tems and Public Health,” in P. W. O’Carroll, W. A. Yasnoff, M. E. Ward, L. H. Ripp, and E. L. Martin, eds., Public Health Informatics and Information Systems (New York: Springer, 2003): at 16-38.

9. W. W. Lowrance, Learning from Experience: Privacy and the Secondary Use of Data in Health Research (London: The Nuf- field Trust, 2002).

10. G. Hripcsak, N. D. Soulakis, L. Li, F. P. Morrison, A. M. Lai, C. Friedman, N. S. Calman, and F. Mostashari, “Syndromic Surveillance Using Ambulatory Electronic Health Records,” Journal of the American Medical Informatics Association 16, no. 3 (2009): 354-361.

11. R. A. Miller, K. F. Schaffner, and A. Meisel, “Ethical and Legal Issues Related to the Use of Computer Programs in Clinical Medicine,” Annals of Internal Medicine 102, no. 4 (1985): 529-536. Compare R. A. Miller and K. W. Goodman, “Ethi- cal Challenges in the Use of Decision-Support Software in Clinical Practice,” in K. W. Goodman, ed., Ethics, Computing and Medicine: Informatics and the Transformation of Health Care (Cambridge and New York: Cambridge University Press, 1998): at 102-115; and K. W. Goodman and R. A. Miller, “Eth- ics and Health Informatics: Users, Standards, and Outcomes,” in E. H. Shortliffe, ed., Biomedical Informatics: Computer Applications in Health Care and Biomedicine, 3rd ed. (New York: Springer, 2006): at 379-402.

12. M. C. Szczepaniak, K. W. Goodman, M. W. Wagner, J. Hut- man, and S. Daswani, “Advancing Organizational Integration: Negotiation, Data Use Agreements, Law, and Ethics,” in M. W. Wagner, A. W. Moore, and R. M. Aryel, eds., Handbook of Biosurveillance (Boston: Academic Press, 2006): at 465- 480.

13. L. O. Gostin, “Privacy: Rethinking Health Information Tech- nology and Informed Consent,” in M. Crowley, ed., Connect- ing American Values with Health Reform (Garrison, NY: The Hastings Center, 2009): at 15-17; N. Lurie and A. Fremond, “Building Bridges between Medical Care and Public Health,” JAMA 302, no. 1 (2009): 84-86; M. A. Rodwin, “The Case for Public Ownership of Patient Data,” JAMA 302, no. 1 (2009): 86-88.

14. J. L. Marquard and P. F. Brennan, “Crying Wolf: Consumers May Be More Willing to Share Medication Information Than Policymakers Think,” Journal of Health Information Manage- ment 23, no. 2 (2009): 26-32.

15. “Public Welfare,” Title 45 Code of Federal Regulations, Pt. 46. 2005 ed., available at <http://www.hhs.gov/ohrp/humansub- jects/guidance/45cfr46.htm#46.102> (last visited November 30, 2009). Compare E. M. Meslin, “Shifting Paradigms in Health Services Research Ethics: Consent, Privacy, and the Challenges for IRBs,” Journal of General Internal Medicine 21, no. 3 (2006): 279-280.

16. The CDC adopts a definition of “surveillance” as “the ongoing, systematic collection, analysis, and interpretation of outcome- specific data, closely integrated with the timely dissemination of these data to those responsible for preventing and control- ling disease or injury.” See Centers for Disease Control and Prevention, Guidelines for Defining Public Health Research and Public Health Non-Research, 1999, available at <http:// www.cdc.gov/od/science/regs/hrpp/researchdefinition.htm> (last visited December 15, 2009); citing S. B. Thacker and R. L. Berkelman, “Public Health Surveillance in the United States,” Epidemiologic Review 10 (1988): 164-190.

17. “…[B]ecause your diagnosis and ‘subsequent clinical acts’ have their rational basis in our prior, collective experience with groups of patients, it follows that the strategies and tac- tics of understanding the distribution and determinants of health and disease in groups (i.e., epidemiology) can be use- ful to you as a physician.” See D. L. Sackett, R. B. Haynes, G. H. Guyatt, and P. Tugwell, Clinical Epidemiology: A Basic Science for Clinical Medicine, 2nd ed. (Boston: Little, Brown and Co., 1991): at 4, emphasis in original.

18. W. A. Yasnoff, “Privacy, Confidentiality, and Security of Public Health Information,” in P. W. O’Carroll, W. A. Yasnoff, M. E. Ward, L. H. Ripp, and E. L. Martin, eds., Public Health Infor- matics and Information Systems (New York: Springer, 2003): at 199-212.

19. K. W. Goodman, “Ethics, Information Technology and Public Health: Duties and Challenges in Computational Epidemiol- ogy,” in P. W. O’Carroll, W. A. Yasnoff, M. E. Ward, L. H. Ripp, and E. L. Martin, eds., Public Health Informatics and Infor- mation Systems (New York: Springer, 2003): at 251-266.

20. J. Baron, Judgment Misguided: Intuition and Error in Public Decision Making (New York: Oxford University Press, 1998).

21. E. J. Johnson and D. Goldstein, “Do Defaults Save Lives?” Sci- ence 302, no. 5649 (2003): 1338-1339.

22. D. A. Moore, P. E. Tetlock, L. Tanlu, and M. H. Bazerman, “Conflict of Interest and the Case of Auditor Independence: Moral Seduction and Strategic Issue Cycling,” Academy of Management Review 31, no. 1 (2006): 10-29.

Copyright of Journal of Law, Medicine & Ethics is the property of Wiley-Blackwell and its content may not be

copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written

permission. However, users may print, download, or email articles for individual use.