Our health care systems are struggling to keep up with a growing, aging population and a rapidly changing world. The strain shows in physician burnout and suicide, long waiting times, increasingly unaffordable health care (for citizen and state), and the poor outcomes for our less privileged.
Part of this is due to poor use of resources and available information. The International Data Corporation projected that global health care data would amount to 2314 billion gigabytes by 2020.
Unfortunately, health care systems have gotten better at collecting data over the years, but not at using it. This is a task that’s easier said than done, considering the amount of data, and that it is spread across many systems that handle data differently.
This is where health care analytics comes in. Analytics includes collecting, checking, and structuring data from multiple sources, selecting bits that are relevant to a task, and running analytical models against the data.
After processing and analysis, information is presented in an understandable way such as a graph, or used to form predictions or recommendations. With enough data and resources, machine learning can also be used to create models that deliver faster and more accurate results, and are more flexible in the face of change.
Analytics has the potential to improve many aspects of health care. In terms of patient care, health care professionals can be provided with up-to-date information collected from multiple sources, including wearable devices, electronic health records, and even social media.
Machine learning algorithms could help detect risks that would remain hidden under standard operating procedures, and incorporate genetics and other data for highly personalised care.
Analytics could reduce health care workloads and costs as well. There is ongoing research on automating interpretation of laboratory tests and medical images using machine learning, either as an independent program or supporting a trained health care professional.
Another development, automated collection and analysis of information from medical devices, would reduce clerical work and allow for advances in telemedicine, thus reducing patient and health care system expenses. Automation of medical administrative work and replacing keyboards with voice interfaces is also a subject of research.
It could also facilitate research to improve scientific and clinical understanding, health care management, and health policy. This can replace some traditional laboratory studies, saving time and money, and can be highly effective – an example is drug discovery. As health care data grows, researchers will be able to find out more about our health and how to prevent diseases.
Health care analytics holds a lot of promise, but there are many barriers to overcome before it can deliver these benefits. These include issues with an important source of health care data, the electronic health record. It has a history of being difficult to implement and use, sometimes resulting in increased screen time, less patient contact, and higher rates of healthcare professional burnout.
Analytics software may also require extensive testing before it can provide significant assistance to health care professionals, because of the severe consequences of medical errors.
We should also question society’s readiness for these advances. Advances in scientific knowledge and technology do not directly translate to healthier behaviour and better outcomes. Steps should be taken to develop regulations on the use of health data and improve levels of health, data, and technological literacy. Existing inequities should also be taken into account because naively applied AI may end up worsening them.
Health care analytics is not a magic bullet. It is unlikely to fix toxic medical culture, improve emotional intelligence of health care professionals, or directly affect other systemic factors affecting health care. It is also not likely to replace doctors, according to Dr. Eric Topol, an American cardiologist, geneticist, and digital medicine researcher.
Analytics does, however, offer relief to a stressed health care system if used correctly. It has the possibility to give time to clinicians, empowerment to patients, and in the words of Doctor Topol, “restore the care in health care”.
Michael Khor is a graduate in Bachelor of Medical Sciences with an interest in health data analytics. He is a former intern with the Galen Centre.
- This is the personal opinion of the writer or publication and does not necessarily represent the views of Code Blue.
