Health care Information and Analytics

Business Intelligence evolved in 80s and 90s and became mainstream in 2000s. I didn’t see what was happening before that, but I can imagine that marketing and salespeople were themselves responsible for all data gathering, analysis as well as making sense out of that. Nowadays even small and medium sized companies have a dedicated IT teams focusing on data gathering, analysis and reporting. There’re businesses which are dedicated to data collection and are going to extreme lengths to get as much data as possible.

Health care industry is past due for this transformation. It’s certainly a lot more complex to categorize and analyze it. However, for various reasons, we don’t collect enough data and we don’t even record it properly. Simple example, if someone has cancer, we don’t ask them ‘Where all have you lived?’. Are we leaving it to Erin Brockovichs of the world to figure it out? We leave it to mostly to doctors to make sense out of mountains of information. We don’t have processes to collect information properly when doctors talk to patients. Many times, such information stays with that doctor only and if you had to change your doctor for some reason, you can be certain that not all information will be transferred. The process of transferring information is also in archaic — you can be handed a bunch of printed sheets or at best a folder full of printed sheets. Then the receiving doctor’s office, at best, will scan those sheets into their system!

Human health is a lot more complex than anything else we deal with. We don’t know all the factors which affect our health — from genetics to food to environmental factors and everything in between. It’s extremely important that we get better at it, and we’ll get better at it only after we start collecting and managing the data better. Couple of things from my observation as a consumer of health care services –

1. Recording calls between patients and health care professionals. Then feed that to AI engines to pick up relevant information and add to the case if it’s not there already pending approval by related professionals.
2. In one instance, I came to know that people working on analyzing health data are called ‘Bio Statisticians’ which points to the fact that medicine relies heavily on statistics. It’s understandable that we’ll like to relate cause and effect when it comes to health information. However, we already have many instances where we know certain drugs work in certain cases, but we don’t fully understand how and why.
   Here again there’s huge opportunity to use AI. Neural networks can be employed to deep dive into data and present findings to health care professionals with Confidence Score. As with any ML system, the results will need to be vetted by professionals. One prerequisite for this is collection of as much information as possible.
3. One can make a case regarding privacy of such information. I think its easily achievable and similar to other private information, patient can have full say on what information can be shared with who.
4. Anonymized data can be freely shared across research institutions which will expedite progress in our understanding of health problems and possible solutions.