Thought Leaders in Big Data: Interview with Lars Olrik, Group CEO of Verdande, and Jo Kinsella, CEO of Financial Services of Verdande (Part 2)

Sramana Mitra: In this use case you are primarily helping to manage the equipment used in drilling.

Lars Olrik: Yes. We are using vast volumes of data and we allow that data to be converted into actionable results that mitigate situations that can lead to a catastrophe.

SM: That is cool. Let’s have a look at the healthcare use case.

LO: In healthcare, as you can imagine, there is an enormous amount of data. Look at the data in a hospital with all the various machines, and then let’s say a doctor is going to conduct an operation. We are not using this data. We do not use the data to look at what kind of profile the patient has and what kind of corrective action [the doctor] may be able to take before the start of the operation. What we do is stream the data in real time and constantly search for things like blood avoidance, for example. If you are a bleeder, you typically stay in an ICU much longer than planned. That means that either the hospital or the insurance has to cover the gap, and typically you operate in a way that is not optimal for the patient.

Here, with the right data, we constantly search for similarities, and [the hospital] can start taking corrective actions so you avoid situations where the patient has to stay longer in an ICU. And, you make sure that the treatment is optimal to the individual. You can call that “days in ICU.” Days in ICU is cost.

SM: What are the data sources for the healthcare situation you are describing?

LO: Starting from heart monitors to various scanners, blood pump machines, and so on, all the information in an operating place is centralized and used. This is a vast amount of data. That data is reused in the context of how to take that data and look at what has happened in a similar situation to avoid the traumatic experience of having to undergo a different type of operation, for which the cost for the hospital or insurance increases.

SM: What is the state of the organization of all these systems – electronic medical records and all the imaging data, etc.? I know a lot of hospitals have made a lot of progress by having advanced their electronic medical record systems so you can sit on top of them and do this kind of analysis. But a large part of the ecosystem still doesn’t have those facilities. Is that observation correct?

LO: I would be bold and say that we are 10 years away from being able to use true electronic patient records. That means: “Lars is operated on in France, his records are automatically transmitted to Germany, where he is on his way to be operated because they can’t do it in France.” This really is the holy grail. We have a long way to go. But there are sub-packages that are being used. Certain hospitals are further downstream in the sense of electronic patient records. But we are a long way from having this globally integrated vision of patient records that you can transmit across the globe and which are easily transferable and easy to understand. We are not there yet. But to do this, you need data infrastructure, and you need the right data to do it. That is being worked on.

SM: How many clients do you have in the medical arena?

LO: Our healthcare business has partnered up with the Methodist Hospital System, which is the tenth biggest hospital group in the U.S. We are focusing on a specific area – cardiac surgery. That is our first product that we are in the process of developing together with Methodist Hospital.

SM: So, you almost have to develop custom heuristics for each of those specialized areas. For heart surgery you need a specific set of heuristics, and you need another set for brain surgery, and so on.

LO: I wouldn’t say that. What you need for each of these areas are agents that search for specific types of problem areas. A heart operation has different types of areas than a brain operation has. What you do is you have a fundamentally similar search platform in place, but you develop different types of agents that look for different types of problems within each of those categories. You have a 70%  to 80% scalable solution, and then you focus on building these search agents. If you had to recreate everything from scratch, you would never get there.

This segment is part 2 in the series : Thought Leaders in Big Data: Interview with Lars Olrik, Group CEO of Verdande, and Jo Kinsella, CEO of Financial Services of Verdande
1 2 3 4 5 6