In our latest article, we started digging into NLP. To learn more, I talked to Veniamin Tzingizis – data manager at Capio Private hospital – who has a solid background in engineering, management, and business analytics. With his skills, he is an important player in the team around Peter Fredriksen and the project where Capio works with NLP in finding chargeable diagnosis codes.
We are all aware that we live in a data society, where the amount of data increases almost exponentially. But for many, maybe especially health carers, it might seem like “it’s too much!” It’s hard to sort, hard to grasp, hard to use, almost like it’s too much of a good thing.
Hi Veniamin! Comments on that?
Understandable. Luckily, NLP is a great tool for giving structure to unstructured data! If you have a text, a computer can actually learn how to understand it. I think many companies find NLP important because they want to take new steps in understanding and using their data.
I know you are doing some interesting things in Denmark. Tell me more.
When it comes to optimizing workflows and procedures, you need to start by rethinking and redesigning the whole flow. A common mistake is to put new tech into old flows and hope it will be enough to accelerate. It’s not.
But it’s equally important to use current tech - not just pre-historic. Our aim is to be at the forefront of testing innovative stuff, looking for anything that can benefit our transformation. We have contacts with universities, for example, DTU Denmark, and together with some professors in healthcare, we are exploring different techs. And we have a good understanding of working on data.
For example, we are working on ways to involve NLP in finding chargeable codes in EMR texts. (We will share some more project insights into the coding project in the next LinkedIn article.)
Since data is the foundation before machine learning can start working its magic, the first challenge is to get the data into the right format for analysis, to be understandable.
For the analysis of the data, you need to train the algorithm. This is a big challenge: You need to come up with a way to link the performance of the algorithm to an actual business value. When you have done that, it’s time to find the correct measurements. You need to have more than one that defines the combination of performance. The complexity of this gives opportunities to me and others to contribute.
We have a very good team in Capio Denmark, supported by the culture in Ramsay Santé. If the high management is not keen on transformation, if goal and vision are lacking, there will be no investments and no successful outcome. But I feel that the whole organization is moving toward this and we’re just getting started, we have a lot of ideas and ambitions.
It gives a purpose to my competence when we can use technology to benefit faster results and better healthcare for our patients. I’m excited about the future!”