How We Handle Patient Data
Data analytics is booming in the United Sates. My daughter works as a senior director in a large, global professional services firm. She is definitely my “40 under 40” star in this emerging data and technology. Nearly everyone agrees that “data is king”. But, “garbage in, garbage out” is a real thing in medicine. So how will we handle our data responsibilities?
We’ll start with an example. Usually, when you see a primary care provider for a visit, a trained staff member takes your vital signs (heart rate, temperature, blood pressure, etc). Sometimes the staff take your blood pressure the old fashioned way by listening with a stethoscope. More recently, the staff may apply an automated blood pressure cuff. Taking your blood pressure manually means the staff must enter your blood pressure in the electronic medical record manually (first point for possible data error). If the staff member uses an automated blood pressure device linked to your electronic medical record (EMR), there is a second possible point of data error. Automated blood pressure cuff readings can vary with the position of the patient. application technique by the staff, or even anxiety. If the cuff reads your blood pressure incorrectly, and is attached to your EMR, the abnormal blood pressure reading is forever in your EMR, even if the staff believes it is incorrect. It can’t be erased based on a “Data Policy” (another potential point for data error). Each EMR and practice staff handle this situation differently, but that reading is not deleted. You’ll be explaining that one to your provider for the next several years. And don’t get me started on errors in coding your medical conditions…
So, what we want is data that is valid and enough data to determine that our medical device really does what we tell you it does. So, we are working with engineering partners on our hardware to make sure the measurements we are going to analyze are valid and reliable. We have run computer simulations. Our engineering partners have double-tested every sensor. We have planned on redundant data capture and redundant data storage. We have followed FDA recommendations for data integrity, HIPAA compliance and secure data storage for our upcoming clinical trials.
Electromagnetic field disturbance technology requires that we take millions of readings from an array of sensors during your study. There are lots of potential problems with recording that data correctly. Electric power in the study area can fluctuate. Sensors can malfunction, Electronic connections can come loose. The internet connection can go down. The list is long. The strength of our technology is in taking enough readings from your sensors that outliers can be identified and statistically removed during your data analysis. There is a principal in scientific research called “statistical significance”. Statistical significance means that there is less than a 5% chance that the outcome is random (device doesn’t work). It also means we have tested enough patients to be 90% sure that our data analysis is valid. Our clinical study designs are built around the principal of statistical significance so you can be confident that our cardiac testing is accurate.
The most exciting thing about our upcoming trials is that we will get to work with consultants who are at the forefront of machine learning and artificial intelligence (AI). Our data analysis will require 5 dimensions. Our measurements with be 3-dimensional over time and will have a 5th dimension of strength. This kind of reporting was unthinkable when I started my medical education in the 70’s. I admit I’ve had to brush up on data analytics to envision this. Thankfully, I have a really good consultant who is also the mother of 3 of my granddaughters.
Our Data Plan is to deliver information to the provider in multiple ways: facts, a 3-dimensional interactive CAD model and a movie of the heart beating while it is emitting energy. I hope that this multifunctional display will help providers identify deviations from normal that are statistically significant. That information should allow an earlier diagnosis of heart failure and an earlier opportunity for intervention. Follow our journey on our Web site as our clinical trials begin in 2 weeks.
