Data for doctors?

If there was one topic in the medical curriculum that you wish you could remove what would that topic be? I came across this question not too long ago. A very common answer was (unsurprisingly) – biostatistics.  Many believe that this time could be used to learn other useful things. I can see where that angst is coming from. The medical curriculum is crowded. We do miss out on teaching several things. From that perspective is biostatistics or even epidemiology really needed in the undergraduate curriculum. Not everybody becomes a researcher anyway?

Data Use in the larger context

Lets look at the “non medical” world. The value being given to “data” seems to be rising every day. The Economist published a famous piece declaring “Data the new oil”. Billion-dollar companies like Facebook, Amazon, Instagram make their money with their prized commodity – Data. Massive advances are being in the fields of Machine Learning and Artificial Intelligence (The more celebrated older siblings of old school data analytics). “Data experts” seem to believe that AI replace a lot of what doctors are doing. Many doctors seem to detest technology . The middle ground seems elusive.

Using Data in Medicine? Is it new?

In medicine, we have been using data long before data science became a thing. That is in fact the foundation of “Evidence-based medicine”. How do we know that 140 /90 is the cut off for hypertension? Why do we use CURB65 scores to decide management in Pneumonia? Why do we use the GCS Score in Trauma? We conducted studies, looked at the data and predicted that people with prolonged elevated BP / High CURB65 Scores / Low GCS / High Respiratory Rates tended to have poorer outcomes if they were not treated immediately. Pediatricians have been using “Z Scores” for years to classify nutrition status. (A Z Score refers to standard deviations which is a direct statistics term)

It is said that “Medicine is a form of art built on the foundation of biostatistics”. The idea that Data Science (including Machine Learning and Artificial Intelligence) and Medical Science are at odds with each other is a faulty assumption that many physicians hold. The heart of data science is problem solving. When a clinician pieces together several sheets of prescriptions and investigations to understand the past history of the patient I would call it a type of data science! Each prescription is a piece of raw data. When put together with what the physician knows it can provide valuable information. For large populations you need computers, software and programming but those are only tools.

Decision Making in medicine

In medicine we have the difficult job of taking decisions and giving advice. Decisions that are extremely hard while making them. Do I operate or manage conservatively? Do I admit or send home?   Start Antibiotics or wait and watch? After the outcome it is ridiculously easy to turn back and judge. What genuinely seemed right at the time can seem so wrong in hindsight. Tools like CURB65, Glasgow Coma Scale etc. help make difficult decisions. They are by no means the only thing we base decisions off . Spotting a sick child in a busy casualty is definitely more than just counting a respiratory rate. Clinical experience matters. Counting the respiratory rate definitely helps the process. Spotting a sick child with a normal respiratory rate takes experience but relying on the respiratory count will prevent us from discharging a baby breathing at 66/min!

Research for everybody?

Doctors may not necessarily be creators of research papers but ALL doctors will be consumers of research. Learning the foundations of bio-statistics and epidemiology and being able to interpret the basics of a study is important now more than ever. In this day and age when there is a relentless onslaught of fake news, propaganda and psuedo-science this becomes important not just in medicine but in life as well. We need to build a culture of promoting scientific temper. Let us work with data science, not as a magic bullet to solve all problems , but as one of many effective tools to guide clinical decision making. The heart of medicine will always remain the doctor – patient relationship and rest assured nothing can replace that.

Leave a Reply