Big Names, Big Data and Stroke

Medical Malpractice

In the final installment of our exclusive series with ANZIIF: ‘How Advances in Medical Technology Are Impacting Insurance’, Karen Jones (Partner) and Dominic Flannery (Special Counsel)  predict the likely benefit of the historical data from stroke victims on patients and insurers.

You’d be hard pressed to find a reality where Sharon Stone, Richard Nixon, Joseph Stalin, and L. Ron Hubbard have much in common, so it may surprise you to know they’ve all suffered strokes, unfortunately some of which were fatal. Based on current statistics in Australia, anyone of us could suffer a stroke in any ten-minute period.

In 2015, stroke was the third leading cause of death in Australia[1], killing more women than breast cancer and more men than prostate cancer[2]. It is predicted that in 2017, there will be more than 55,000 strokes in Australia alone – that’s more than 1000 strokes per week, which translates to a stroke every 10 minutes[3]. While some affected go on to lead normal, healthy lives, many require a lifetime of care, which is costly to both the sufferers and insurers.

Over the past decade, instances of stroke continue to decline due to greater education and awareness of risk factors including tobacco usage, high body mass, diabetes, physical inactivity and high blood pressure[4]. Regardless, the financial ramifications of a stroke in Australia remain high, with the average cost in 2012 estimated to be $129,199 over the course of a lifetime[5].

With so many people reportedly experiencing strokes, medical practitioners are beginning to look at ways to harness population or ‘big’ data relating to stroke for both medical and monetary benefits, which will have significant implications for the insurance industry.

Big data and predictive modelling for stroke

Big data is exactly as it sounds: extremely large data sets analysed to identify patterns, trends, and associations. In the medical world, it comes via health records, clinical trials and genome data.

By using predictive analytics and computer aided diagnosis of health related big data, medical practitioners have the opportunity to develop a greater understanding of diseases and other medical conditions, as well as which treatments are most suitable and effective.

How this data is gathered, merged and analysed is key to the value of any insights and innovations that may emerge.[6]

How can big data support stroke prevention?

Research from the University of Nottingham has revealed that ‘self-teaching’ artificial intelligence systems analysing big data in the form of clinical records would be more efficient at predicting cardiovascular risk such as stroke than current standard risk models[7].

This is due to the fact that the versions currently used are too narrow and simplistic. In essence, detailed predictive modelling using big data from the population is likely to provide a more comprehensive and accurate overview of risks associated with stroke.

However, bridging the gap between the theoretical promise of using big data to manage stroke and its real world application remains an ongoing concern,[8] arising from two key areas[9]:

  1. Privacy – the ownership and sharing of the data itself including the usage, storage, interpretation, and dissemination of the data collected
  2. Quality and analysis – the diversity of data form and entry types comes from multiple sources. Merging this data is challenging both from a programming (databases are not connected and do not communicate with one another) and statistical perspective (once databases are merged, missing data from each database renders the connected links useless).

In order to overcome these concerns, there needs to be a standardised level of control over the data, at a national level to address privacy concerns specific to individual countries, and preferably at an international level to address the quality and analysis of the data.

The current process of data collection – a local perspective

Locally, the Australian Stroke Clinical Registry (ASCR) is the nation’s response to these concerns.

It is tasked with monitoring, promoting and improving the quality of acute stroke care by collecting data from participating hospitals across Australia[10]. In order to be able to access that data for research purposes, appropriate ethical, statistical and design criteria need to be provided.

While the data is available to researchers (subject to approval), the strict security and confidentiality policies at organisations like ASCR mean that, for the moment, raw access to this data by private companies such as insurers is unlikely.

However, this does not prevent an insurer from using the results of existing research, or conducting their own research.

Where to next?

The use of big data for the prevention of treatment of stroke opens up a world of possibilities.

Imagine if your local GP could upload your personal records (family history of stroke, medications, blood pressure, weight, co-morbidities etc.) into a database that compares it to thousands of other individuals who have suffered a stroke.

From there, the data could be analysed and a report prepared outlining the likelihood of a stroke occurring and when, the stressors that may increase that likelihood, medications or environments to take or avoid, as well as a plethora of other information that may help you avoid a stroke or minimise its effects when it occurs.

Impact on the insurance industry

While the predictive elements of this situation do not presently exist, the utilisation of big data in stroke healthcare is already seeing real world benefits for patients and medical practitioners alike from a retrospective perspective[11].

Historical data from thousands of strokes suffered is already helping medical practitioners understand stroke with the view to help prevent them in the future. The use of this data in the insurance industry is all but inevitable: a 2016 survey of Australian and NZ insurers found that 97% said big data initiatives were on their organisation’s agenda[12].

As long as the nature and scope of the use of big data is carefully regulated, particularly in privacy setting, historical data from stroke victims is likely to benefit patients and insurers alike.

 

[1] http://www.abs.gov.au/ausstats/abs@.nsf/Lookup/by%20Subject/3303.0~2015~Main%20Features~Australia’s%20leading%20causes%20of%20death,%202015~3

[2] Australian Institute of Health and Welfare, 2014. Australia’s Health 2014. – via https://strokefoundation.org.au/About-Stroke/Facts-and-figures-about-stroke

[3] https://strokefoundation.org.au/About-Stroke/Facts-and-figures-about-stroke

[4] Above 4.

[5] The Economic Impact of Stroke in Australia: National Stroke Foundation – Deloitte Access Economics (13 March 2013).

[6] Lidong Wang, Cheryl Ann Alexander. Stroke Care and the Role of Big Data in Healthcare and Stroke. Rehabilitation Sciences. Vol. 1, No. 1, 2016, pp. 16-24. doi: 10.11648/j.rs.20160101.13

[7] https://phys.org/news/2017-04-artificial-intelligence-accurately-future-heart.html

[8] Lidong Wang, Cheryl Ann Alexander. Stroke Care and the Role of Big Data in Healthcare and Stroke. Rehabilitation Sciences. Vol. 1, No. 1, 2016, pp. 16-24. doi: 10.11648/j.rs.20160101.13

[9] Electronic Health Records and Medical Big Data: Law and Policy: Sharona Hoffman – Cambridge University Press 2016

[10] http://www.auscr.com.au/

[11] https://www.ncbi.nlm.nih.gov/pubmed/27680330

[12] The Impact of Big Data on the Future of Insurance – Actuaries Institute

15/06/2017