Analytics in Healthcare Industry is an Irony, its needed, there is a market of consumers, there are suppliers of the same, but at the same time from a resource perspective its very challenging. When i talk of Health care there are various stake holders , prominent among them being practitioners like Doctors, Pharma Industry, Device manufacturers, Diagnostic Industry. While all of them use analytics through some source, when it comes to induction of a analytical resource , its very challenging. The reason for this is data is not public. A doctor will never publish his patients medical data on a public platform for new budding professionals to create analytical models on those, neither would the patient be comfortable having his data published for others to scrutinize and do research. A pharma company , which as a process does a lot of data recording right from dossier preparation, testing , submission, approvals , field trials and so on would never share this data , as this is their Intellectual property and they have spent substantial years, money and efforts to bring a drug from inception to conception making it fit for public usage. The same story goes for Device manufacturers and Diagnostics Industry.

So, how would you and i as a person who knows model building or wants to learn it with respect to pharma or healthcare or medical IOT , create analytical models. Top this up with the challenges like lack of domain knowledge specific to the vertical you are dealing, it becomes very frustrating when you are unable to connect the dots between the model, the subject matter experts inference and the data. 

The first time, I started working on a data challenge, it was a new world for me. But , I being an experience data scientist had the leverage of talking to the best Diabetologists, market leaders in Wearable devices, API developers in the segment, Data from AIIMS experimental study and so on. So funtionally , using the human body for self as we operate day in and day out might seem pretty ordinary, but when you start analyzing the data for analytical model building you wake up to the fact that sleep itself has roughly five stages , wake, light awake, light sleep , deep sleep and REM sleep. And they again have sleep cycles which vary in length and pattern with different individuals. Now , how on earth , not having studied human anatomy and that too specifically related to sleep would we be able to make sense of tonnes of sensor data shared to us on a drive.

Now , this is just one aspect, it needs a doctors and researches view to help us understand that heart rate at rest and heart rate at activity are different for different people and although there are bench marks , its subjective.  Variables like Exercise intensity, activity distance,duration, calories burned, REM Sleep, times woken, resting heart rate and so on can help us predict a possible heart attack , but we need access to a lot of data , the underlying functional knowledge and technical capability .

All of this also points out to one fact, that if you are some where in the healthcare industry and have been working on data, although not core analytics, you are in a better position to leverage this situation and learn analytics and apply it .So there is a considerable scope for people in pharma industry to jump into analytics.
 

Let me now talk of some typical applications of Analytics in Health-care Industry

1. This one is where I have personally contributed, There were various stake holders , AIIMS geriatric ward and Apollo Hospitals was the custodian of historic data, Fitbit, Garmin and Inbody were the wearable device and an eminent team of doctors on SicureMis advisory council were the guiding factors. The result was a model owned by Sicuremi which connects to any wearable  devices IOT data and uses syndicated data as well as instream data to predict the heart condition, sleep apnea and many other health conditions which could save a patient from aggravated medical trauma.
 

2. 
Pharma industry uses analytics for analyzing  data from field trials as well as for new drug discovery . The API research teams use a lot of simulations, test and analytics to come out with novel drug discovery as well as product innovations.

3. 
Diagnostic industry has practically made it easy for a patient to receive the dash board of their health card on their proprietary  apps. Patients can then discuss their health with the doctor in a much more informed way. Although on a lighter side , doctors are not amused about it, we have four doctors in our family and they say one thing , google can give you the meaning of anything , but its only a doctor who can infer things, so don’t try to be a doctor. Well they are actually true on that part.

Authored : Mohan Rai