Data Science- Healthcare woes 3

In continuation of the series, I’d recommend you read up the blog post here.

I am going to address the most crucial issue with the lack of adoption of AI and ML in healthcare. The value of the data pipe is tough to measure- data doesn’t translate into value. I have been grappling with this problem too (and often my failure) to convert the “digital medicine” into a concrete deliverable.

Frankly, I don’t wear blinkers in front of my eyes that screams the “value of access and disadvantaged communities”. Those are social and political problems, and I am ill-equipped to handle them. I can take the horse to water, but I can’t make it drink.

Most digital medicine roles on Twitter appear to be dubious with “influencers” and crowing about “tweet impressions/engagement”. How does it translate into real life? What would be the actual value to the organisation? Does it bring about a behavioural change? Do more people become proactive to seek help, for example, when a woman discovers lumps?

I think it boils down to accessibility and “predictive analytics”. Coming back to the original question on how data science can add value to healthcare would be the ability to demonstrate the impact in simple quantifiable terms.

For one, Data Scientists are often in “support” roles. Most organizations make the majority of their decisions on intuition that stems from past readings and personal experiences – not from a Data Scientist’s analyses….Another challenge being usually a “support” role in a company is quantifying your impact. A common data science task is to help a Product Manager answer a question about some recent activity in the data. You can also issue a product recommendation based on your insights. So what? How do you measure whether this work of yours was impactful?

Does this mean that healthcare organisations would be willing to hire someone in that role? I think, and it is a long term view, that doctors should understand the value of structured data and proceed accordingly too.