LAST month I expressed concern about the imbalance that has arisen in health policy formulation, and last year I cautioned that as doctors, we too should have our privacy respected.

The latest “consultation” being undertaken by the government not only shines the spotlight on those two concerns but also portends worse.

Government is seeking opinions regarding the secondary uses of My Health Record systems data. If you want to have a say, you have until this Friday, 17 November 2017.

To summarise what the government wants to do is simple: it wants to share what is on My Health Record with the parties it deems suitable.

Open data are a problem that affects society at large, not only the health sector.

Earlier in 2017, Geoff Hughes wrote a wonderful opuscule on the topic that is worth reading. Hughes summarises the National Institutes of Health’s (United States) aims of data sharing:

  • open scientific inquiry;
  • diversity of analysis and opinion;
  • promotion of new research;
  • testing new or alternative hypotheses;
  • testing new or alternative methods of analysis;
  • education of new researchers;
  • exploration of topics not envisioned by the initial investigators; and
  • creation of new datasets by combining data from multiple sources.

The merits of data sharing, as outlined above, are easy to follow and appealing.

However, the flipside of sharing My Health Record data must be brought into this consultation.

My Health Record has had its fair share of growing pains. First, we had National E-Health Transition Authority, now we have the Australian Digital Health Agency. First, we had Personally Controlled Electronic Health Record and now we have My Health Record. The government has invested a lot of time and effort to get to where it is now and even so, only about 5.2 million Australians have registered. The uptake by practices has been slow and has necessitated incentives for GPs via the Practice Incentives Program. Understandably, politicians want to see a return on investment and practical benefits before My Health Record becomes a political hot potato like the National Broadband Network.

The slow uptake has been for three main reasons: privacy concerns, cumbersome registration and IT hurdles. To its credit, the government has worked hard at dealing with these issues, which are largely resolved.

Secondary use of My Health Record data did not feature in any of the government’s efforts to stimulate uptake; in contrast, reassurances about privacy were proffered, especially after the census debacle. Moreover, supposedly de-identified data can be relatively easily re-identified, as was seen when the 10% Medicare data set was published online and then had to be taken down.

Therefore, I opine that there is no social contract to justify the secondary use of My Health Record data. If this was the intention, it should have been declared at the outset.

Another concern is that the shared data will be too freely available. On page 8 of the consultation document, we find listed the potential recipients of data:

  • eligible researchers;
  • eligible government users;
  • accredited commissioning and data processing organisations; and
  • eligible external parties (private, non-government and foreign).

The way I read it, these categories are too open-ended, so any reassurances the government offers in terms of de-identification, documentation and checks are rendered useless. I fear that wolves in sheep’s clothing will be lining up immediately.

Those of us who work with medical software know the challenges of finding the right codes for diagnoses, tests and other data. So even though My Health Record data apply to about five million Australians, the reality is that the contents of most My Health Records are rudimentary, given it is still in its infancy. The IT principle of “garbage in, garbage out” therefore applies. Until data are more robust and better coded, the utility to research is low.

The Australian Government uses four examples it hopes will persuade us to back its proposal (pages 23-26 of the public consultation document):

  • the folate story – “using data from the Western Australia Birth Defects Registry researchers discovered a finding which reduced the number of births with neural tube defect, resulting in cost savings to the health and disability sectors and to families;”
  • ABC Catalyst story on statins in 2013 – “secondary use of My Health Record data will enable monitoring of the impact of policy changes, media reports and other events on the use of prescription medications in Australia, while also allowing analysis according to patient risk, and assessment of patient outcomes, using detailed information that is not currently captured in [Pharmaceutical Benefits Scheme] data;”
  • cost-effectiveness analysis of Indigenous patients in remote Northern Territory communities – “combining health databases even where those databases are limited in scope was used to improve access to primary care in remote communities for the management of diabetes, which results in net health benefits to patients and cost savings to government;” and
  • a Canadian study of implantable defibrillators – “demonstrates that the collection and use of the information is in the public good by furthering medical research”.

None of these are good examples, as none relied on a database comparable to My Health Record. They all relied on more well defined data.

If this proposal does get up, it will be the thin edge of the wedge, because it will embolden government to demand the use of My Health Record by all clinicians.

I foresee a transition from a government incentivising the use of My Health Record to a government penalising us for not using My Health Record. The subsequent step will be penalties for not using My Health Record with the codes and language the government wants. Rather than think for ourselves in a way that helps patients, we will become data punchers for the Department of Health. It reminds me of how accountancy has been so influenced by the data demands of the Australian Taxation Office.

Hughes refers to a speech made by Prime Minister Malcolm Turnbull, who had this to say about data:

“Governments hold an extraordinary amount of unique data, collected directly and indirectly in the course of doing our job. It is there. We have it. But there is no point in keeping all that data stored away. It needs to be accessed, analysed, understood, used and reused.”

Sadly for the patients and the genuine researchers, the risks of this proposal are far too serious to ignore or leave to blind faith in bureaucratic processes.

If that is Mr Turnbull’s attitude to the notes I keep about my patients, I prefer the government never to hold the data in the first place.

Dr Aniello Iannuzzi, FACRRM, FRACGP, FARGP, FAICD, is a GP practising in Coonabarabran, NSW, and a clinical associate professor at the University of Sydney.


To find a doctor, or a job, to use GP Desktop and Doctors Health, book and track your CPD, and buy textbooks and guidelines, visit doctorportal.




I support the secondary sharing of My Health Record data by the government
  • Strongly disagree (61%, 46 Votes)
  • Disagree (17%, 13 Votes)
  • Strongly agree (12%, 9 Votes)
  • Agree (8%, 6 Votes)
  • Neutral (1%, 1 Votes)

Total Voters: 75

Loading ... Loading ...

2 thoughts on “Secondary sharing of My Health Record data: no thanks

  1. Jesia Berry says:

    As a data analyst working in a university who works with linked data in early childhood development, I think there are so many benefits to be gained by data sharing and promoting the use of administrative data for secondary research. I agree the My Health Record data will be messy and incomplete, but having an end-of-use purpose through secondary use will change this over time. At the moment, so much important research cannot be done nationally in Australia because the data are not made available due to legislative hurdles or because the data are housed in diverse and stand-alone collections e.g. cerebral palsy/birth defect registers, Medicare’s pharmaceutical, immunisation registries, medical benefits schemes, and general practice visits. WA and NSW are the exceptions as leaders in data linkage in these new areas.
    Hospital, emergency, mortality, police crash records, and some child health data (e.g. perinatal, birth anomalies, education and welfare services) have been available for research since the 1970s starting in WA with other states joining via the Population Health Research Network (PHRN). So much public good has come from these datasets with wide-reaching government and policy implications e.g. recent SA research shows 1 in 4 children are notified to child protection services, an alarmingly high figure and their educational achievement is stimied.
    Accessing these data requires postgraduate training in biostatistics and epidemiology; these skills are essential to be able to tabulate and interrogate the anonymised data and researchers sign confidentiality agreements. The untrained person would not have a clue where to start. Access is restricted to employees (like me) vetted by the organisation approved for research access.
    The Scandinavian countries are world leaders by having a unique identifier for all people in their country and a commitment to public good research uses of administrative data. Shortly after the spurious Lancet article by Andrew Wakefield in 1998, The Danish published a whole-of-country child data linkage study which showed no causative link between measles-mumps-rubella vaccine and autism using psychiatry, hospital, general practice and immunisation records; we are grateful to them, as this study is not possible in Australia at the moment, unfortunately, due to our risk-adverse outlook and adminstrative data silos. Prospective cohort studies with opt-in consent to do the same take decades to finish and are comparatively extremely expensive! Meanwhile immunisation rates would plummet and reemergence of vaccine-preventable disease like measles would occur. I can understand doctors would be reluctant for their patient notes to become widely available, but the implications of business as usual are scenarios like the measles one above.

  2. Anonymous says:

    The rules and standards that govern GP access to Government data, should be the same rules and standards that govern Govt access to GP data

Leave a Reply

Your email address will not be published. Required fields are marked *