The hype (and over-hype) of AI in home-based care

In this blog, we discuss how AI can help the home health and hospice industry — and the AI tools that are over-hyped.

What is AI (and what is it not)?

AI looks at broad sets of information and derives population-level insights. Care professionals then need to personalize the treatment plans based on their knowledge of the situation and the patient’s preferences, issues and history.

While AI filters through large samples of data and identifies “signals” that are strong predictors that an event has or will occur, it’s not a panacea. Its only as good as the models created and the information that we feed into it.

AI is not there to diagnose patient issues, but rather to raise insights and awareness to information and trends that healthcare workers can use to diagnose and treat patients. Great care will continue to rely on skilled, dedicated and informed home health professionals.

Examples of AI in home-based care

Fall risk

One example of AI in home-based care is fall risk, where large samples of data combined with accurate tracking of actual events point back to key objective and subjective information that can be easily collected. This data includes changes in medications, not taking prescribed medications, dehydration, changes in gait or walking style, and changes in activity level.

Machine learning in healthcare also filters through body-worn sensors, such as personal safety and alert devices, to detect with high accuracy when an actual fall has occurred versus someone just sitting down rather abruptly.

Behavioral health

Another example of AI in home-based care is behavioral health. Similar to fall risk, subjective information (e.g., how did the patient feel today?) combined with objective information (e.g., did the patient take their medications as prescribed?) can yield remarkably accurate assessments based on population comparisons. This can then be tailored and refined at the individual level over time.

Where does AI fall short in home-based care?

AI fails to handle fringe or “exceptions.” Personal metrics can vary significantly and can be temporal in nature. An AI system may identify a possible cause, but it really comes down to the healthcare professional to make decisions based on their understanding of the patient, the longitudinal patient record and their own unique history.

Maximize AI for your organization by using it to intelligently analyze data, helping you gain insight into your patient population and make more informed decisions.