Recent Posts



How to Improve Patient Outcomes with Artificial Intelligence

Many of you may have heard the "AI" buzzword, but how do we actually use it to solve problems?

One tool that we use to solve problems is Machine Learning. Machine Learning allows us to train the computer how to learn. Teaching a computer to learn is not that different than teaching a child to learn from a high level. When the child does something wrong, we apply negative reinforcement to explain to the child that what they did is wrong. When the child does something good, we tell the child that they have succeeded.

In data science, we can use machine learning to solve the answer to the following five questions:

Classification (is this high blood pressure or low blood pressure?)

Anomaly Detection (this blood pressure reading seems out of the normal range)

Regression (how many times have we recorded readings like this?)

Clustering (how are blood pressure readings organized? Is it linear? Or are their clusters where groupings tend to be in a particular range, then the next cluster jumps to some reading above that)

Reinforcement Learning (Now that we know all of this information, what should we do next?)

By leveraging large data sets, we can create very accurate models to analyze any kind of biometric data such as blood pressure by using the Machine Learning tool.

Consider the new clinical experience, you arrive at the clinic, you speak into a voice device such as Alexa or Cortana, recording all of your symptoms. Your vital signs are taken and sent to an AI system which passes the data to the model where it will respond back to you in a series of follow up questions again using the voice technology.

Voice System: "Does your family have a history of high blood pressure?"

Patient: "Yes, my mother"

Voice System: "Does she or did she suffer from any diseases related to her high blood pressure?"

As you can see, AI will have an enormous impact on how we receive medical care. There will be a significant shift from relying on a doctor or nurse's knowledge to relying on the machine. The hope is that this will increase the quality of patient outcomes.


©2020 Copyright Dynamics Intelligence Inc