Article Text

Download PDFPDF

Public good, personal privacy: a citizens' deliberation about using medical information for pharmacoepidemiological research
  1. Lianne Parkin,
  2. Charlotte Paul
  1. Department of Preventive and Social Medicine, University of Otago, Dunedin, New Zealand
  1. Correspondence to Dr Lianne Parkin, Department of Preventive and Social Medicine, University of Otago, PO Box 913, Dunedin 9054, New Zealand; lianne.parkin{at}


Background Epidemiologists have long argued for access to personal medical information in order to undertake research in the public interest, while legislation and ethical guidelines have increasingly focused on protecting the privacy of individuals. How the public weighs up these public interest and privacy arguments is unclear.

Methods A citizens' jury was held to explore public views about the use of medical information for the post-marketing surveillance of medicine safety. A steering group of stakeholders oversaw the process, including the framing of the question for the jury and the selection of expert witnesses and jurors. An independent chair and facilitator managed the 3-day hearing.

Results The jury unanimously concluded that researchers contracted by a public body should be permitted to use medical information about identifiable people, without their consent, to study the safety of medicines—providing existing ethical guidelines and relevant laws are followed. Jurors who were initially uncomfortable about the use of information without their express consent all became more comfortable after obtaining further information and deliberating with other jurors.

Conclusions This outcome suggests that an informed public does not place personal privacy above societal benefits in the particular circumstance of medicines' safety research, given appropriate privacy safeguards. The specificity of the example, the framing of the public interest and privacy considerations—in the context of scientific, legal, ethical, clinical and consumer input—and the opportunity to deliberate, may explain why the conclusions of the jury differ from public opinion surveys about secondary uses of medical information.

  • Citizens' deliberation
  • drug safety
  • epidemiology
  • ethics in epidemiology
  • privacy
  • research ethics

Statistics from

Epidemiologists have long argued for access to personal health data to monitor, investigate, and improve the public health. Doll1 declared in 1974 that research workers would continue to need access to personal information ‘if the public is to be protected against the hazards of living in a complex industrialised society’. At the same time in the USA, Gordis and colleagues2 gave examples of landmark findings that would not have been obtainable if strict privacy principles had been in force. Twenty years later, privacy legislation in Europe threatened to bring epidemiological research to a standstill. Vandenbroucke3 called for epidemiologists to combat these developments by proposing alternative basic principles: ‘the ethics of serving the community first’. More recently, when professional guidelines restrictive of access to records were introduced, Doll and Peto4 reacted by suggesting that the right to medical care should generally include a responsibility to allow information gained in the course of such care to be used to benefit others.

Nevertheless, legislation and ethical guidelines have increasingly tended to be framed in terms of the privacy rights of individuals overriding the public interest in the results of research. For example, in the UK the legally constituted Patient Information Advisory Group and research ethics committees made such secondary use of medical records without consent almost impossible,5 and the situation may not be improving.6 Secondary use can often be of anonymised data, but there are times when identifiable information is required to link datasets, maximise value and allow quality and completeness to be validated.7 Moreover, obtaining consent is often not practicable and a requirement to do so may bias the study results.8

In New Zealand, the Privacy Act and Health Information Privacy Code are less restrictive. Health information may be disclosed for research purposes when it is not desirable or practicable to obtain consent and when no identifiable information will be published.9 10 Ethical guidance takes a similar approach, which entails weighing public benefits and potential harms.11 Nevertheless, research ethics committees have sometimes been restrictive in their judgements,12 perhaps reflecting members' perceptions of a higher value put on privacy by the public.

Missing from the debate has been a clear view about how the public weighs the public interest and privacy arguments. A Nuffield Trust report in 2002 on the secondary use of data recommended a dialogue with the public about the arguments for use without consent and the appropriate safeguards.13 In 2006, the UK Academy of Medical Sciences noted that evidence of public attitudes towards the use of medical information in research was still largely absent, and the investigations that had been undertaken asked undifferentiated questions that were not adequate to assess attitudes towards different types of research conducted by different groups for different purposes.14 Moreover, strong support for research from patient groups consulted by the Academy suggested that engagement with the public was key to a more informed view of the worth of, and safeguards for, the use of personal medical information for research.

A dialogue with the public entails informing people about the issues at stake for particular types of research and from all relevant points of view. As well as information, the opportunity for thought and involvement matter.15 One such approach has been referred to as ‘deliberative democracy’; a way of reinvigorating politics by encouraging citizens to take the public good seriously rather than private or factional interests.16 Deliberation can involve both internal reflection and public discussion.17

We have used the aspects of information giving, deliberative discussion, and a detailed proposal, to investigate what the public wants. In particular, we have used a citizens' jury approach to explore the views of the public about the use of personal medical information about identifiable people, without their consent, for the purposes of post-marketing surveillance of medicine safety (including both monitoring and research) in New Zealand.


A citizens' jury brings together a group of non-experts to be informed and to deliberate on a matter of public importance.18–20 The key elements are: a steering group of relevant stakeholders; participatory representativeness; expert evidence; the opportunity to question the experts; and the delivery of recommendations to those who have the power to act on them.

The steering group

A steering group composed of relevant stakeholders was established to guide the overall process and to ensure independence and credibility. The key stakeholders represented included the Office of the Health and Disability Commissioner, the Office of the Privacy Commissioner (as an independent observer), the Ministry of Health and researchers on medicine safety. The two commissioners have statutory roles to promote and protect the rights of health and disability consumers and the privacy of individuals, respectively.

The steering group guided the framing of the question for the jury; the selection of expert witnesses, jurors, chair and facilitator; the choice of a venue, and the development of the agenda. The group also identified various authorities and organisations to which the jury's recommendations should be reported.

The question for the jury was in two parts (box 1). Post-marketing surveillance of medicine safety includes spontaneous reporting of suspected adverse reactions, systematic collection of information about adverse events in users of selected newly marketed medicines (prescription-event monitoring), examination of routine statistics and analytical epidemiological studies.21 These approaches can be seen as a continuum—although, in practice, they are usually subdivided into two distinct categories of activity: ‘monitoring or audit’ and ‘research’. However, because these terms can be understood differently,22 the question avoided their use and simply described particular activities.

Box 1 Question for the jury

  • Should researchers contracted by a public body1 be permitted to use medical information2 about identifiable people, without their consent, for the following purposes:

    • i. Routine analysis to identify potential adverse effects of medicines that are newly introduced into New Zealand?

    • ii. Investigation of emerging concerns about the adverse effects of medicines currently being used by New Zealanders?

  • If so, under what circumstances and with what safeguards, if any?

  • If not, why not?

  • Explanatory notes:

  1. Public body refers to a publicly funded body such as the Health Research Council of New Zealand (which is responsible for administering government funding of health research) or Medsafe (the unit of the Ministry of Health, which is responsible for the regulation of medicines and medical devices in New Zealand)

  2. Medical information refers to two types of data:

    • a. Health-related data collected by the New Zealand Health Information Service (about medicine use, hospital admissions, cancer diagnoses and death) for statistical and research purposes

    • b. Limited medical information from, for example:

      • i. Hospital discharge letters or general practice records, which was primarily recorded for treatment purposes, or

      • ii. Death records

Selection of expert witnesses, jurors, chair and facilitator

The steering group recommended that the expert witnesses should include a specialist in privacy law, an ethicist, two health consumer representatives (with potentially differing views on research and the use of medical information), a medically qualified pharmacoepidemiologist and a medical practitioner. Those people proposed by the steering group agreed to take part. Experts were asked to speak for 20–30 min, to provide a short written summary of their main points and to answer questions from the jury. Decisions about the specific content of their presentations were deliberately left to the experts.

A random sample of 96 people, stratified by age and sex, was taken from the local electoral rolls in the expectation that this would provide 12 jurors,23 and an invitation letter was sent to each person. A sum of NZ$150 was offered to cover expenses such as travel and childcare.

An independent chair and facilitator were appointed to manage the hearing and to support the jury, respectively. The chair also chaired a university ethics committee, while the facilitator was an experienced community worker.

As this was a consultation process, ethical review was not required.

The hearing

A neutral and accessible public institution was chosen as the venue. Legal and ethical aspects were addressed on day 1, and on day 2 the jury heard from the other experts. Day 3 was set aside for juror deliberation and for reporting back their decision and recommendations.

During the introductory session on day 1, the chair explained the citizens' jury process and introduced the question. Jurors were encouraged to consider the question for all people in the community, and were told that a unanimous decision was not essential as we were interested in people's reasons for their decisions as much as the decisions themselves. The jury was also provided with a list of potential issues to consider (based on ethical guidelines)11 in coming to a decision (box 2).

Box 2 Potential issues for jurors to consider

  1. Are there potential benefits of undertaking this work? If so, what are they? Are they for the individual or the public? Are there potential harms in not undertaking this work?

  2. Would a requirement to obtain consent cause unnecessary anxiety?

  3. Would a requirement to obtain consent affect the scientific value of the study?

  4. Would obtaining individual/next of kin consent be possible in practice?

  5. Would not obtaining consent cause disadvantage to the individuals or their relatives or any of the collectivities involved?

  6. Does the public interest in investigating the safety of medicines outweigh the public interest in privacy?

  7. What safeguards for confidentiality should be required?

  8. Is there anything else that matters in answering these questions—for example whether the public knows about monitoring the safety of medicines or this use of medical information?

Following the chair's introduction, one of us (LP) provided the jury with background information about the development of medicines, approaches to studying medicine safety, the data required for such research, the sources of those data, and a proposed national programme of post-marketing surveillance of medicine safety. The content of this background material was approved by the steering group.

The jurors completed pre and post-hearing questionnaires. The first collected demographic data and included questions from a national telephone survey about public attitudes to sharing medical information.24 These questions were repeated in the post-hearing questionnaire. The second questionnaire also asked jurors whether their views about the main question had changed since the beginning of the hearing and, if so, what had led to the change. Feedback on the process was also elicited.

With the agreement of the steering group and the jurors, two independent observers attended some of the sessions.


The process

Thirteen people agreed to take part in the jury; the extra juror was included to allow for a drop-out. Unfortunately, three women were forced to withdraw at the last minute because of sudden illness and injury, and one man who attended on day 1 withdrew. The demographic characteristics of those who took part in the deliberation included both sexes (six men, three women), a wide range of ages (18–65+ years, although seven were 45 years or older), a good range of educational qualifications and occupational groups (a clerical worker, kitchen hand, manager, nurse, plumber, storeman, taxi driver and two retired people—a marine engineer and a nurse), and people of New Zealand European and Māori ethnicity.

Of the 62 non-participants who replied or were contacted by phone, most were unable to attend because of work commitments, ill health or disability, childcare issues, or because they would be away on the relevant dates. Only three of the 62 directly reported a lack of interest in participating. Nineteen people could not be contacted and thus nothing is known about their reasons for not participating.

All jurors had questions for the expert witnesses; the main questions are shown in box 3. At their request, jurors were provided with additional written information about the identifiability of medical information (as classified in ethical guidelines)11 and copies of the national ethics application form25 26 and the Code of Health and Disability Services Consumers' Rights.27 To clarify what was meant by the word ‘identifiable’ in the question for the jury, one of us (LP) explained that normally the information would be key-coded (ie, potentially identifiable), but on rare occasions it might be necessary to obtain the names of individuals.

Box 3 Main questions from the jurors to expert witnesses and organisers

Meaning and scope of the question

  • What is ‘identifiable information’? What is meant by ‘researchers’? Are they independent of the pharmaceutical industry? Why are there two parts to the question? Why are both types of activity necessary?

Legal and ethical protections of health data

  • Where is existing health information held? Who owns medical records (health services or patients)? Is there a limitation on how long medical information and privacy must be protected? What is the relevance to research of the Health and Disability Code of Rights? Legally, are individuals able to stipulate that their existing summary health data (held by the New Zealand Health Information Service) are not used in research?

  • Are all ethics committee members appointed by the Minister of Health? Who are the professionals on the committees? Are pharmaceutical companies represented on ethics committees? Who do ethics committees report to? Are the committees publicly or privately funded? How long does it take to get ethical approval for a study? What happens if there is an urgent drug safety issue and people might die while researchers are waiting for approval? Can ethical approval be transferred from one group to another if, for example, funding arrangements for monitoring are changed? If something unexpected is found in a study, can researchers obtain further data to investigate this under the original ethical approval?

Current post-marketing surveillance

  • Who is currently monitoring the safety of medicines in New Zealand? Which medicines are being studied? How? What are the arrangements for ethical oversight and protection of privacy?

Proposed national post-marketing surveillance programme

  • Is there anything new about the proposed programme with regards to legal and ethical issues? Would the programme be permitted within the current legal and ethical frameworks? How else could drug safety be studied?

  • Why is linkage of health data being proposed now? Is it less expensive than previously? Who would have access to the linked data? Why is it (currently) not possible to use anonymised data at all times? Would it be necessary to obtain the names of individual patients? Would identifying or potentially identifying data (named or key-coded) be required for both of the activities described in the question?

  • Why not make use of advances in information technology, such as data mining, to detect adverse drug reactions rather than relying solely on health practitioners' suspicions and voluntary reporting? Is this not being undertaken because of concerns about privacy?

Individual consent to use health information

  • How many people would be too many for it to be practicable to seek individual consent? Do you think that informed consent is really informed?

Specific questions for the clinician

  • How much do doctors worry about prescribing treatments for which the information about benefits and harms is incomplete? What information is available—sources, volume, trustworthiness? Would it be helpful if the information was filtered and a trustworthy party provided a summary? If a study suggested that a particular drug should not be used by certain people, would not that help your current patients? How do you deal with patients who do their own research on the internet?

The pharmaceutical industry

  • Are pharmaceutical companies required to have ethical approval in order to undertake research? Is it necessary to develop a system to ensure that incomplete linked health data are not misused by pharmaceutical companies? Why not levy the pharmaceutical industry to fund independent research? How is it that pharmaceutical companies and independent researchers sometimes get different answers in studies of drug safety?

Regulatory environment

  • Why are some drugs allowed into the country without going through the usual regulatory approval processes? Would the proposed programme be vulnerable to political and pharmaceutical industry pressures? What if Medsafe did not support further research on an important safety issue?

Public awareness of research

  • In addition to being able to read the minutes of ethics committee meetings on-line, should the public be informed through other media about research currently being undertaken? Are researchers required to inform the public about the results of their research? How is this done?

From the outset, jurors engaged in lively discussions with expert witnesses and with each other. On day 3, the chair, independent observers and organisers withdrew to allow the jury to deliberate in private with the facilitator. These deliberations continued for more than 4 h. On two occasions the organisers were called back to answer questions about medicine safety research and the proposed national programme of post-marketing surveillance.

The jury's recommendations

The jury reported back to the organisers orally, and in writing, on the afternoon of day 3. Their decision, which was unanimous, was that researchers contracted by a public body should be permitted to use medical information about identifiable people, without their consent, for the two purposes stated in the question. The decision was qualified with specific circumstances and safeguards (box 4).

Box 4 The jury's decision and recommendations


  • Researchers contracted by a public body should be permitted to use medical information about identifiable people, without their consent, for the two purposes stated in the question.

Under the following circumstances

  • The data should be key-coded (eg, using an identifier such as the national health index number), although identifiers such as names, dates of birth and addresses could be used if necessary.

  • There should be ethics committee oversight.

  • The relevant legal provisions should be adhered to.

Recommended safeguards

  • The ethical review process should be retained.

  • There needs to be more public awareness about research approved by ethics committees.

  • The public should be informed about the outcomes of such research.

  • Health professionals should be informed about the outcomes of such research.

  • Medicines coming into New Zealand without a formal approval process should be monitored in the same way as approved medicines.

Six jurors reported that their views on the question had changed during the jury hearing. The reasons cited were the information provided by the expert witnesses, learning that there were systems (legal and ethical) in place to oversee the use of medical information, and having the opportunity to discuss these matters with their fellow jurors. Two others reported feeling more confident about their original views (both supported the use of medical information).

Changes in responses to national telephone survey questions

Most jury members were comfortable with the use of their medical information (in which they were identified solely by a medical record number) for research ‘to improve the health of New Zealanders’ at the outset (mean level of comfort 9.0, range 7–10, in which the value 10 represents ‘very comfortable’), although several felt less comfortable if that information was to be used without asking them (mean level of comfort 6.7, range 2–10). Following the hearing, all of these people had moved towards being much more comfortable about their information being used without their consent (mean level of comfort 9.0, range 7–10). Two jurors did not answer the questions in the baseline questionnaire.


A group of nine citizens, selected at random from the electoral roll, took part in a deliberative discussion over 3 days. They unanimously concluded that researchers contracted by a public body should be permitted, in particular circumstances and with certain safeguards, to use medical information about identifiable individuals, without their consent, for investigating the safety of medicines. Six jurors reported that their views about the question had changed over the course of the hearing; they attributed this to being more informed and being able to deliberate with others. Changes were also observed in the responses to a question about levels of comfort with researchers accessing medical information without explicit consent.

This outcome suggests that an informed public does not place personal privacy above societal benefits in the specific circumstance of post-marketing surveillance of medicine safety. Furthermore, it suggests that the public would not see the small loss of privacy entailed by the use of potentially identifiable medical information as in conflict with societal benefits, so long as researchers adhere to existing legal and ethical protections for confidentiality. Nevertheless, the public expects that benefits be realised in practice: the jury recommended that the public and the profession be made more aware of the outcomes of such research.

We are not suggesting that these empirical findings trump normative claims about what ought to be the case. This investigation has shown how a group of people weighs normative claims. Such findings, although not decisive, are relevant to policy making.

Can these results be relied on? Potential defects are that the choice or instruction of experts could bias the results,28 that the criteria used by the jury organisers will dictate and thus limit the moral framework in which the jury makes decisions,29 and that self-selected subsamples of random samples are unrepresentative. Each of these warrants consideration.

First, to ensure independence, all stages in the planning and conduct of the jury processes were overseen by a steering group that represented the potential interests. Because of unforeseen circumstances on the day, the ethics presentation had to be given by one of us (CP). However, the material presented was approved by the steering group and followed closely the National Ethics Advisory Committee guidance on observational studies,11 and the original and stand-in presenters were both members of the National Ethics Advisory Committee. Although the balance of the presenters was carefully considered beforehand by the steering group, two jurors commented that they would have liked to have heard more arguments against permitting the use of personal information. Both of these jurors were very comfortable (before and after the hearing) with their information being used for research in the public interest. In fact, one of the consumer representatives gave a strong privacy view and all the other presenters provided some balance between the public interest in privacy and the public interest in medicines' safety. The expert pharmacoepidemiologist, as expected, put the public good argument and one of us (LP) described the practical and scientific reasons for not seeking consent, in the context of maximising the value of the research. Although not anticipated, the medical practitioner and the other consumer representative put the public good argument strongly. The medical practitioner raised the question of responsibility to others, and described how information provided by studies of previous generations of patients informed her practice and thus benefited current patients. The consumer representative argued that allowing such research would result in substantial benefits and minimal harm.

Second, although the jury members were given criteria for decision-making, in the form of questions that they were encouraged to use in deliberations (box 2), these questions were drawn from guidelines that are used by ethics committees and thus are part of the existing moral framework. The jury was also free to explore other criteria.

Third, because participation involved a substantial time commitment and a potential loss of income, many people were unable to take part. The final group was unrepresentative in being older and with a predominance of men, although the group included a mix of educational qualifications, occupational groups and ethnicity. Most of those who declined to take part did so because of other commitments rather than a lack of interest. The term ‘participatory representativeness’ is sometimes used in these settings to denote the subset of the general population that is prepared to participate in civic endeavours.

The strengths of deliberative democracy: involvement, thought and information were all present in the jury process.15 One of the main reasons for recommending more deliberative styles of public engagement is that, when it has been studied, the public's political knowledge is very limited.16 This also applies to knowledge of scientific, legal and ethical processes. The questions posed to the presenters (box 3) demonstrate that the jury members sought information across a wide range of issues and they commented on their gains in knowledge in many of these areas in their evaluation responses. The length of the final deliberation by the jury (more than 4 h) shows that involvement, reflection and discussion were vital features of this jury process.

How do these results relate to what is already known? Population surveys have, on the whole, shown that many people have little knowledge about the purposes of research and regard research uses as peripheral in the sharing of medical information.30–36 They are also much more inclined to want to have control over who has access to their records for research purposes than for clinical care. However, one survey in the context of information giving about a specific example, cancer registration, showed that most people were happy about the collection and research uses of this information and did not regard it as an invasion of privacy.37

In relation to deliberative discussions,38–44 once again a consultation that used the most specific example—national confidential enquiries into suicide and homicide, maternal and child deaths and patient deaths—showed the most acceptance for the use of data without consent.43

As well as the specificity of the example, the framing of the issue appears to be important. One qualitative study in the context of the use of medical information in primary care for research framed three specific scenarios in terms of increasingly more extensive ‘breaches of confidentiality’.44 The scenario that was reported as demonstrating the most extensive ‘breach of confidentiality’ was the transfer of patient data to an external disease register. The study participants reported ‘much anxiety’ about access to data for population disease registers and concern even about the use of aggregated anonymous data. The discussion of ethical issues in the paper proposed that moral wrongs may be committed even when the ‘wronged’ are unaware of their occurrence. This is clearly going beyond the cases in which anonymous data may identify a specific ethnic or social group in a way that unjustly discriminates to a theoretical wrong. The authors concluded that public acceptability of the use of medical records for research cannot be assumed.

Those findings are in stark contrast to the lack of concern about invasion of privacy in the survey findings of Barrett and colleagues37 in the context of questions about the cancer registry, in the National Institute for Health and Clinical Excellence deliberation on confidential enquiries into deaths43 and in the findings reported here. We suggest that the crucial difference is in the framing of the problem to the public.

The usual frame for these sorts of investigations is simply to find out about the attitudes and views of the public. An alternative frame is to find out their views in the context of information about the current legal and ethical framework and the public good arguments for research. Does the public accept these arguments and the current protections, or are they inadequate and is more required? Deliberative democracy is about encouraging people to take the public good seriously.16 As such, it is not completely impartial between private and public goods. The frame for the jury described here was to find out people's views in the context of scientific, legal, ethical, clinical and consumer input. The aim was to ensure that the public good arguments were presented along with more commonly heard individual privacy concerns.

As well as the frame of the question, the particular topic—the safety of medicines—is one in which people can easily appreciate the need for good information given that most people take medicines and many have experienced adverse effects. The New Zealand national health index number, a unique identifier, also provides privacy safeguards; most linkage can be done at arms length from the researchers, through the Health Information Service of the Ministry of Health.


A citizens' deliberation, framed to include both the public good and privacy arguments in relation to the use and linkage of medical information for research on medicines' safety, concluded that such use was warranted given the existing protections and the minimum of identifiable information. Both the framing of the issue and the specificity of the example may explain why the jury was comfortable with the small loss of privacy.

Internationally, the implication of our results is that epidemiologists should continue to make the public good argument for the secondary use of medical information—an informed public is likely to be receptive and to agree that the public good might sometimes take priority over the demands of privacy, given that appropriate safeguards are created and maintained about their medical information. Alongside this is the need to develop ways to use and link data that allow a minimum of identifiable information to be disclosed.

What is already known on this subject

Access to personal medical information about identifiable individuals, without individual consent, is sometimes necessary in order to undertake epidemiological research in the public interest. Legislation and ethical guidelines have increasingly been framed in terms of the privacy rights of individuals overriding the public interest in the results of research. However, there has been no clear view about how the public weighs up these public interest and privacy arguments.

What this study adds

The findings of this citizens' deliberation suggest that an informed public does not place personal privacy above the public good in the particular instance of post-marketing surveillance of medicine safety, given appropriate privacy safeguards. Unlike many population surveys that have been less positive about secondary uses of medical information without consent, the participants in this consultation process considered a specific example, in the context of information about both the current legal and ethical frameworks and the public good arguments for research. It is noteworthy that the public expects that societal benefits be realised in practice and researchers have a responsibility to inform the public of their results.


The authors would like to thank the jurors, the chair and facilitator of the jury, the expert witnesses and the members of the steering group for their participation in this project. They are grateful to Inga Hunter for providing copies of questions used in a telephone survey about public attitudes to sharing medical information. They also thank Patricia Priest, Peter Herbison and Tom Cunningham for their comments on an earlier draft of the paper.



  • Funding This work was supported by a Joint Initiative of the Health Research Council of New Zealand and the New Zealand Ministry of Health. The funders had no role in the study design, in the collection, analysis and interpretation of data, in the writing of the report or in the decision to submit the paper for publication.

  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.