Many have mused that we are all just players in one big simulation. That every bit about us can be codified into simple data pieces. Including but not exclusive to our height, weight, and gender, there isn't an aspect about us that can't be documented in some meaningful way. Where I am attempting to add a bit of levity with the Sims link above, and where most if not all of the programming language we rely on in using our everyday phones/computers/cars are derived from a binomial 1's and 0's programming foundation, we really are just a roaming band of 1's and 0's. And this is significant in the healthcare world. One type of machine learning could be described as a feedback loop of data aggregation. Tesla has "Full Self Driving" technology powered by a supercomputer aptly named "Dojo." Dojo does little more than review millions of images to differentiate a stop light from a yield sign rendering two decidedly different outcomes. These images are also paired against responses of human drivers, and simulated drivers in a consistent quality check of "when this happens, I should do that." Where Tesla technology is based on pictures, what makes our own individual programming code is indeed pictures with a side of something extra. Some of you may have a deeper understanding of what Tesla technology does, and that's great. Though digging too deep into my comparison defeats the purpose. I'm only looking to show an example, not debug the latest build. The more data from the more people accumulated in a central data warehouse location, the better. This may sound a bit too Orwell/1984'ish but work with me here since it gets to the point of framing the case study you just read. We need data today to pave the way for a healthier life tomorrow and machine learning is certainly nothing to be shrugged off to make that happen. The more data we can collect from men with prostate cancer, or women with breast cancer, or boys and girls with conditions that could have been identified and addressed in utero, the better. The more data points you give to a machine learning system, the more capable it becomes of picking out the outliers that may have the disposition that can be treated if not eliminated and this is what hierarchical coding does. Now we're getting to the good part ... Initial health assessments and annual wellness visits are each big deals in their own right. An initial health assessment is the primary care provider's first chance to get a detailed view of the patient in front of them. This interaction is intentionally broad and generally includes an opportunity for the patient to share a little bit about themselves, and I would encourage you to check out that link for a sense of what I mean. The annual wellness visit is a chance to update what was learned from the initial health assessment. New biometrics in concert with an update on the patient's life who by then is a bit older with characteristics that are just as likely to change as remain the same. Where data is gathered every healthcare encounter including the two types identified above, a provider isn't gathering any data if the patient isn't bothering to schedule or to show up. Lack of patient participation has the downstream effect of reducing the efficacy of what a machine learning apparatus could do. Question: How would you incentivize the completion of an initial health assessment evaluation, including the Stay Healthy Assessments above?
Many have mused that we are all just players in one big simulation. That every bit about us can be codified into simple data pieces. Including but not exclusive to our height, weight, and gender, there isn't an aspect about us that can't be documented in some meaningful way. Where I am attempting to add a bit of levity with the Sims link above, and where most if not all of the programming language we rely on in using our everyday phones/computers/cars are derived from a binomial 1's and 0's programming foundation, we really are just a roaming band of 1's and 0's. And this is significant in the healthcare world.
One type of machine learning could be described as a feedback loop of data aggregation. Tesla has "Full Self Driving" technology powered by a supercomputer aptly named "Dojo." Dojo does little more than review millions of images to differentiate a stop light from a yield sign rendering two decidedly different outcomes. These images are also paired against responses of human drivers, and simulated drivers in a consistent quality check of "when this happens, I should do that." Where Tesla technology is based on pictures, what makes our own individual programming code is indeed pictures with a side of something extra. Some of you may have a deeper understanding of what Tesla technology does, and that's great. Though digging too deep into my comparison defeats the purpose. I'm only looking to show an example, not debug the latest build.
The more data from the more people accumulated in a central data warehouse location, the better. This may sound a bit too Orwell/1984'ish but work with me here since it gets to the point of framing the case study you just read. We need data today to pave the way for a healthier life tomorrow and machine learning is certainly nothing to be shrugged off to make that happen. The more data we can collect from men with prostate cancer, or women with breast cancer, or boys and girls with conditions that could have been identified and addressed in utero, the better. The more data points you give to a machine learning system, the more capable it becomes of picking out the outliers that may have the disposition that can be treated if not eliminated and this is what hierarchical coding does.
Now we're getting to the good part ...
Initial health assessments and annual wellness visits are each big deals in their own right. An initial health assessment is the primary care provider's first chance to get a detailed view of the patient in front of them. This interaction is intentionally broad and generally includes an opportunity for the patient to share a little bit about themselves, and I would encourage you to check out that link for a sense of what I mean.
The annual wellness visit is a chance to update what was learned from the initial health assessment. New biometrics in concert with an update on the patient's life who by then is a bit older with characteristics that are just as likely to change as remain the same.
Where data is gathered every healthcare encounter including the two types identified above, a provider isn't gathering any data if the patient isn't bothering to schedule or to show up. Lack of patient participation has the downstream effect of reducing the efficacy of what a machine learning apparatus could do.
Question:
How would you incentivize the completion of an initial health assessment evaluation, including the Stay Healthy Assessments above?
Step by step
Solved in 2 steps
