Artificial intelligence & machine learning – what does it mean for healthcare?

| February 12, 2018

Is the onset of artificial intelligence to be feared or embraced? Healthcare consultant Megan Kennedy considers the revolutionary potential AI presents.


As the artificial intelligence and machine learning momentum grows, and begins to make a mark on healthcare, there is  no shortage of hype – but what does it really mean?

Firstly, what is Artificial Intelligence? It’s basically about getting computers to behave in an intelligent manner. This can be achieved through what we call curated knowledge or machine learning.

Curated knowledge is about using existing knowledge and embedding it into a system. For example, ‘if a temperature reading is above 38 degrees, the patient has a temperature.’ There is basic curated knowledge in the EMR I currently use, as if I enter a temperature >38 degrees, the system advises me the parameter is outside normal values and prompts me to review.

With machine learning, a computer derives knowledge from the data, beyond just the basic notion described above, to uncover new insights. For example, it is being used to help doctors detect intracranial bleeding, resulting from either head trauma or stroke. An algorithm has been developed that uses deep learning, machine vision, patient data and clinical insights to automatically highlight for a doctor, areas that might indicate the potential presence of a cerebral bleed on a scan.

The most obvious application of artificial intelligence in healthcare is data management. The first step in revolutionising existing healthcare systems is to collect it, store it, normalise it, and trace its lineage. You can then mine the data of medical records to provide better and faster health services. For example, each scan and test result contains crucial information about whether a patient is at risk of a serious condition, and what needs to be done. Being able to interpret that information quickly and accurately is essential for hospitals to provide appropriate treatment. Right now, there are some patients that suffer some form of avoidable harm because this information isn’t interpreted and acted on in time.

Artificial intelligence is also starting to provide clinicians with evidence-based treatment options for their patients, especially in the area of Oncology. Analysing the meaning and context of both structured and unstructured data in clinical notes, reports and external research data, is assisting with the selection of appropriate treatment pathways.

Artificial intelligence will also have a huge impact on genetics and genomics. Deep Genomics aims to identify patterns in huge data sets of genetic information and medical records, looking for mutations and linkages to disease.

Most importantly, we need to start with ‘what problems are we solving?’. Many people think that artificial intelligence is herculean. It’s not going to replace clinicians/healthcare professionals, it’s a set of tools to help them do their jobs, and when appropriately implemented, will revolutionise healthcare as we know it today.

As a healthcare professional wearing two hats – emergency clinician and healthcare consultant – I look forward to all that artificial intelligence is going to be able to do, to support my vocations.

SHARE WITH:
Megan Kennedy

Megan has passion for improving health outcomes for people, whether they be from a business or clinical perspective. This is supported by many years of experience in the areas of consulting, strategy, business development, solution sales, marketing, product management, education, clinical care and clinical management. Megan has participated in numerous eHealth engagements and the development of innovative strategies to
transform hospital and health systems.

One Comment

  1. Alan Stevenson

    Alan Stevenson

    February 28, 2018 at 9:37 am

    Computers are logical machines – they react to predetermined stimuli. For this reason one cannot equate AI with morals or ethics. I am not a medical professional, but would assume that at times both moral and ethical considerations would be the ideal model i.e. if a patient could be given a few extra months of life but at the expense of continued pain or suffering. Psychological considerations might also enter the mix. By all means use AI to help the practitioner, but I would be wary of letting the computer go unsupervised.