Opting out of My Health Record
I remember one warm evening in 2003, when I was still a medical student based in the emergency department of a major metropolitan hospital. An elderly woman walked in to the waiting room and collapsed. The team transferred her to a resuscitation bay, ensuring she was stable. Then, my supervisor turned to me and asked, “How do we work out what is wrong?”
My head spun. The different causes of collapse, which I had diligently memorised, relate to every single organ system in the body – heart attacks, lung infections, stroke, low blood sugar, low blood pressure are only the most common causes. I had been taught that a good history from the patient provides the best localising information. But the patient could not speak, her previous hospital medical records from an admission two years ago were in the basement and her GP practice was closed.
Without the clues a history can provide, my supervisor got on with treating the patient using all the other diagnostic tools at her disposal. Within minutes, she had established what was wrong and started appropriate care. As a medical student, this dark magic seemed impossible. I remember wishing in those days for the time medical records would be stored online. But now, 15 years on, I find myself a surgeon with serious reservations about such a system – not only as a clinician but also as a patient.
The suggestion that My Health Record (MHR) can provide a more complete picture of health or avoid a full consultation is one I believe to be disingenuous. A complete picture of someone’s health comes from speaking to them, not just a list of diagnoses or drugs; even repeat prescriptions obligate a full consultation for medical safety. For privacy reasons, this is a system specifically designed to allow patients to edit their own record, but that also means no clinician can ever fully trust how comprehensive or accurate the record might be.
Doctors and privacy experts have long-voiced these concerns, particularly since the Australian Digital Health Agency (ADHA) announced in May 2017 that all Australians would be given a national centralised electronic medical record unless they opted out. In response to expert fears about privacy and data use, and the potential for unforeseen harm, especially to vulnerable people, the Australian government passed legislation in November 2018 and set a new opt-out date of January 31, 2019. As this deadline has come and gone, millions of Australians continue to grapple with whether they should delete or maintain their record.
From my perspective, these changes were vital – blocking the release of medical records to insurers, employers or law enforcement unless authorised; increasing privacy for teenagers and victims of family violence; allowing the permanent deletion of records at any time; and legislating against future privatisation. Researchers will still be allowed access to de-identified data – although they of course face the usual potential for legislative amendment by this or future governments.
While there’s momentary reassurance in the fact these legislative protections have addressed some of the most glaring flaws in the system, many clinicians and other experts remain concerned about My Health Record, which has cost $2 billion to establish, and will cost $500 million a year to run. These issues relate to the system’s ongoing security and safety, its viability in day-to-day practice, its ability to deliver the outcomes it promises, the ongoing risk to specific groups of patients and whether it even represents good value for money. After weighing these concerns, I chose to opt myself and my family out.
Sometimes I consider the astronomical cost of MHR and think about the other, proven, health interventions that could have been instituted for $2 billion. ADHA proposes some immediate savings will offset the cost of the MHR program, by reducing the rate of “avoidable” pathology and radiology tests. But even if all 6.5 per cent of pathology tests and 4.4 per cent of diagnostic images deemed “avoidable” were eliminated, this is a cost saving of only $300 million based on 2016-17 expenditure – well short of MHR’s yearly running cost.
Electronic medical records have been around for more than a decade. Many doctors, including myself, use these systems in public hospitals and choose to use them in our practices. Dissent should not be dismissed as a fear of technology.
There is good evidence electronic medical records in hospitals very modestly improve health outcomes, but there is no evidence that national systems save lives or improve health. There are too many confounders to compare countries with each other, or to themselves at different time points. To suggest that MHR specifically, or a national health record generally, will improve health is merely an assertion and one that isn’t yet backed up by the data.
Australia already has a heavily digitised and connected medical environment. Virtually all of our public hospitals and many clinics run on electronic records. This means that even if health outcomes are improved by electronic health records, many of these benefits may have already been realised. In a move that strikes me as bizarre, MHR has not been engineered to be compatible with these existing medical programs. This dilemma is likely to cost clinicians time – which could be otherwise spent providing substantive healthcare – and also affect the accuracy of uploaded records.
The information the government has provided the public about MHR largely extols the virtues of having a record, without any real acknowledgement of concerns. Health information will travel with patients, will negate the need to provide a medical history to a new doctor, will provide a “more complete picture of health” and the ability to replace a lost prescription quickly “rather than have a full consultation”. For the elderly, end-of-life decisions are easily documented and available and, for everyone, MHR promises better care in an emergency where you may be unconscious and alone. Although there’s mention that de-identified data may be used in future research, the details are vague.
On its website, the ADHA explains its broader population level goals for My Health Record. These include decreasing adverse drug reactions, increased adherence to evidence-based medical care and cost savings from avoiding duplicated tests. All seem reasonable. A greater ability for patients to manage their own health and secondary uses of data by researchers raise some questions. There is a promise to clinicians that MHR will save us time.
At a population level, de-identified patient data, collected with explicit informed consent, can improve clinical practice. Analysis of such data aids our understanding of health outcomes from different treatments, predicts future healthcare use and enables governments to provide better value for money. While a nationalised medical record has the theoretical potential to serve this purpose, as it stands MHR would require laborious manual interrogation to extract any meaningful data, even if enough people agreed to maintain a comprehensive record.
Regardless, health researchers remain excited about the possibility of access to population-level health data. Those interested in the integrity of electronic systems and patient privacy are less excited. Associate Professor Vanessa Teague, a cybersecurity expert at the University of Melbourne, found that both doctors and patients could easily and confidently be identified from a de-identified dataset published online by the Australian Department of Health in 2016. Only a few points of information, such as year of birth and dates of surgery, were required to retrieve all their Medicare bills and PBS prescriptions for some years. There does not appear to be any way to truly de-identify health data, unless it is altered so much that it becomes useless to researchers.
Teague says that while patients can maintain a health record while opting out of any secondary uses of their health data, she feels Australians are unable to make an informed choice about this because ADHA continues to inaccurately assure the public de-identified data is secure.
For researchers, although My Health Record has obvious appeal, the combination of poor uptake, patient editing and incomplete data entry by time-poor clinicians using non-compatible computer systems may result in inadequate data to be genuinely useful. It seems particularly galling that so much taxpayer money has been spent on a system that cannot improve health unless it is a data source, only to have it fundamentally fail at accurately collecting data in a comprehensive and usable way.
For clinicians, it is of little use, as the technologically literate may heavily edit their record while those among us who struggle with technology will likely find it difficult to generate the personal health summaries that are central to the system.
For patients, I also have privacy concerns, particularly within families. While others having access to a record is optional after the age of 14, some teenagers and some adults in controlling relationships may struggle to maintain independent medical care if others compel them to have, use and share their MHR.
Offline, the limitations of the system come into even sharper relief. If you get hit by a bus? It is unlikely your emergency physician will collate your Medicare card number, date of birth, surname and gender before seeking emergency access to your MHR. Instead, they will probably just get on with the job of saving your life, which they are already excellent at doing.
Overall, in my opinion as a medical professional, My Health Record is not fit for purpose. The tension between individual loss of privacy and participation in data collection for the greater good is a genuine moral question. But a system that fails both at protecting privacy and at being an effective research and clinical tool has no justification.
This article was first published in the print edition of The Saturday Paper on Feb 2, 2019 as "Not what the doctor ordered". Subscribe here.