Between 2007 and 2013, Australian physician David Henry led a not-for-profit Canadian research institute tasked with analysing the vast array of health data generated by Ontario’s 13.5 million residents.
Statisticians, clinicians, epidemiologists and analysts at the Institute for Clinical Evaluative Sciences in Toronto have been examining anonymous patient records and clinical and administrative databases for 25 years. These data include electronic medical files held by general practitioners, administrative records from hospitals, clinics and practices, and records of diagnostic and imaging tests, as well as data regarding income, schooling and social service needs gathered by the welfare, education and taxation departments. The records are collected during routine care and service delivery and are linked and de-identified before being made available for research and policy analysis. The institute is accredited by the Information and Privacy Commissioner in Ontario.
The institute “can link a whole lot of data sets at the level of the individual, but it is all done anonymously,” Henry explains. “The analyses can address a wide range of questions. For instance, they can look at the education levels of children and ask ‘what are the impacts of childhood asthma or diabetes on educational achievement?’ ” One study discovered that the incidence of diabetes was 20 to 30 per cent lower in suburbs that were deemed ‘walkable’ – a finding that has influenced urban design in Toronto and beyond.”
Government agencies in Ontario routinely consult the institute when designing policies, Henry continues. For example, they might ask “can we work out after six months whether a new policy designed to improve the care of patients with heart attacks is working”.
“The power of these records is extraordinary,” says Henry, who now lives in northern NSW. But Ontario’s use of health data is a generation ahead of what is possible in this country. “Australia is still debating under what circumstances a person’s data should be linked,” he says.
The power of these records is extraordinary.
Henry backs the University of Melbourne’s plans to train local researchers in the skills required to work productively with large and diverse health data sets. It is important that a professional cohort is ready for when the digital infrastructure is finally in place, he says. This will include epidemiologists, statisticians and analysts who can apply their skills to “very large population data sets”, and data scientists who are able to act as stewards, linking the data and making sure the information is secure, and suitable for high quality research.
Henry says digitising and linking medical records is expensive, but it can ultimately deliver significant savings; the process becomes efficient once regular feeds of updated data are being received. Also, individual study costs are much lower than for primary data collection. Henry hopes that a portal to Ontario records will soon be available for University of Melbourne researchers. This will help the Faculty prepare for a time of ready access to linked de-identified Australian data – a development that Henry describes as “inevitable”.