Healthcare Big Data is Déjà vu All Over Again
My colleagues and I chuckle when we hear about all the ‘new benefits’ from Big Data. We chuckle because we have been using large datasets (what today is called Big Data) for years to help companies glean insights and intelligence from customer data. The goal has been to get the “right offer, to the right person, at the right time”. This means that we sought to understand the customer and their needs enough to be able to make the right product offer to them when they were most likely to accept or respond to the offer. To accomplish this, behavior analysis and predictive modeling was used to create more effective acquisition and retention programs and to build actionable customer segmentations. The segmentation schemes enabled the company to manage similar customers as a group with emphasis on keeping the top customers in their segments and motivating the other customers to ‘move up’ the segmentation schema.
What is old in most other industries is new in healthcare. Analysis and manipulation of large datasets has not been used in healthcare except by healthcare insurance providers, even then, the data has been used to manage insurance costs and not provide any real benefits to the customer. It has been used to analyze past transactions versus impacting current and future decisions.
The potential benefit of using Big Data in healthcare is very similar to how it has been used in other industries. Where it has been used to manage customer portfolios elsewhere, in healthcare, the data can be used to aid healthcare providers with their decision support. Instead of getting the ‘right offer, to the right person, at the right time’, healthcare can use Big Data to “get the right information/diagnosis, in the right (provider) hands, at the right time, to create the right treatment plan”. In other words, it is to help healthcare providers make the right diagnosis or treatment plan using data from thousands of similar cases.
For instance, a 43-year-old Hispanic male presents himself at the provider’s office with chronic high blood pressure. The doctor has several patients with high blood pressure and believes this should be a fairly easy case to treat. The doctor prescribes one of his preferred medications and sends the patient along his way. Two weeks later, the patient returns feeling very sluggish. The doctor may or may not have seen this in other patients. What the doctor may not have known was that the medication he/she prescribed has an added side effect for Hispanic patients with sluggishness being one of them. With access to a large dataset of similar cases at the time of diagnosis, instead of prescribing another medication for the patient, thus wasting the money the patient spent on the first medication, the doctor could have originally prescribed the medication proven to work best in this situation.
This story actually happened to someone I know. The patient’s regular physician was on vacation and he had to see one of his colleagues. When the patient went back to see the regular physician, the patient didn’t have to list off the negative side effects, his regular physician knew them already. He said, “I’ll bet you’re feeling sluggish, etc.” Neither physician was of the same ethnicity as the patient. The bottom line is that not all patients react the same to medications. There are variances in ethnicities, age, sex, etc. Having access to data to make better treatment plans at the time of the visit is what it’s all about.
Let me emphasize that this is not to take the decision out of the physician’s hands. The physician standing in front of the patient knows best which treatment plan to follow. This is in place to potentially help the physician make those decisions that will improve the overall patient outcome.