How is artificial intelligence relevant to your healthcare?

Artificial intelligence (AI) refers to a broad set of computing technologies that are able to perform tasks such as reasoning, sensory understanding, and learning. AI is not one technology – it encompasses many approaches and techniques. For a more in-depth understanding of the different types of artificial intelligence, please read this primer.

There are a number of ways in which your health care can be directly impacted by devices or applications that are powered by AI technologies, for example:

  • For a patient with a chronic condition like type II diabetes: with the help of AI driven applications, a doctor or nurse practitioner can give you information on how to manage your diet and medication remotely. It can also coach you on lifestyle habits and provide other feedback to improve your health. Through sensors and wearables that report changes in your conditions, healthcare providers can know if your health is deteriorating and get in touch with you.
  • For a patient with Parkinson’s disease: AI driven surveys and phone sensors can better track your symptoms. This can help your doctors and caregivers better understand changes over time and the impact of exercise or medication on your condition.
  • As a patient: you might often find that you are interacting with an AI driven chatbot to schedule a doctor’s appointment, to find a clinic in a specific area, for instructions on how to use certain medication, or to accomplish many other tasks.

Why is data relevant to artificial intelligence?

AI and machine learning (ML) processes are driven by algorithms that are tested through large datasets. ML is the most common category of AI techniques in the health care context. It refers to statistical techniques which attempt to ‘fit’ models to data and train the models using large datasets. These techniques use large datasets to test and train the models. For example, doctors and researchers can use large datasets of patient records to understand which treatment has worked for other patients with similar features or underlying conditions.

For AI and ML to drive successful outcomes, they need high-quality, clean and accurate data. Researchers emphasize the need for large, multifaceted datasets that allow ML processes to incorporate as many factors as possible into analysis. Unfortunately, in the United States, healthcare institutions experience a number of issues while implementing electronic health record systems which makes the datasets they produce difficult to work with. They require robust data infrastructure, technical standards and expertise, and are often not compatible across different institutions. These challenges ultimately impact patients and their ability to access their own health data.

How are providers benefiting from applying artificial intelligence?

Researchers and doctors have used and continue to use machine learning techniques to better understand how particular types of treatment will affect patients with certain attributes.

AI can improve a doctor’s ability to diagnose certain medical conditions. For example, more sophisticated types of AI, like deep learning, can help doctors more accurately detect cancer based on scans and images. On their own, doctors may misdiagnose such conditions because cancerous lesions are not always observable to the human eye.  Similarly, researchers at Google use AI to analyze images of retinas and diagnose diabetic retinopathy with over 90 percent accuracy. AI has also shown potential of predicting acute kidney injury two days before it actually occurs.

Physicians and payers also use chatbots for certain health management tasks. Chatbots are computer programs that are able to have a conversation with you through text message, applications, or instant messaging. Providers use AI driven chatbots to help with administrative tasks and reduce costs. In this way, AI can improve access to treatment in resource constrained environments by freeing up time that would otherwise be spent by health care professionals on administrative tasks.

In what other ways is the healthcare industry leveraging AI?

AI and ML techniques can lower the cost of drugs by helping researchers in several ways. Drug development is a costly and time-consuming process. AI can help improve drug development through its lifecycle, from identifying gaps in current therapeutics to bringing new products to market.

Pharmaceutical researchers can use AI to sort through huge numbers of research papers and patents, as well as comprehensive lists of chemical compounds and their properties, to suggest opportunities for drug development. By analyzing the growing databases of biomarker data, they can then work to target different treatments to different types of patients. When drugs or other treatments reach the clinical trial stage, AI can help match ideal patients to the right trials.