Big Data and the IoT will allow doctors to provide better care.
Hu Yoshida, CTO of Hitachi Data Systems and prominent blogger and thought leader, spoke to CBR about the convergence of Big Data and the IoT and how they may come to transform healthcare.
CBR: What is the relationship between the Internet of Things and Big Data?
HY: The whole value of the IoT is being able to correlate, consolidate and integrate information from different sources. What we want to avoid is what’s happening in the medical field, which is an example of standards not going far enough.
In the medical field there’s a standard around medical imaging called DICOM. Vendors make a DICOM solution consisting of storage, server, network and applications and they differentiate by adding value. A number of people make these, and what they’ve done is go to the hospitals and go to where the money is: cardiology, radiology, etcetera.
They’ve bypassed IT, but [the different departments] don’t talk together. So the doctors treating the patient don’t get an integrated view across all of these archives. They’re now coming back to IT and asking for a vendor-neutral archive. So that’s why we have the Hitachi Clinical Repository (HCR), where we bring all of that data together, whether it is in Siemens or Fujitsu format, people will still be able to see across all those different verticals.
The whole point to me about Big Data is trying to integrate all of this information. If you keep it in siloes, you lose all the benefits of Big Data. That’s our approach to this whole problem, trying to provide common frameworks, common platforms.
What’s happened in the industry is that we’re more open to APIs, client providers and open source, which helps to do this integration. Those are very big movements making possible now what we couldn’t do five years ago.
CBR: Does the update to HCR aim to address this?
HY: You need some sort of middleware there that can integrate all of that information. So what [HCR] does is create metadata around the original data so that we can access things through metadata, but still serve it up to the original provider of that information or to somebody else who comes through the middleware. Recently we’ve expanded it out to mobile. It’s now an app that a doctor has on an iPad; when he talks to a patient he can have all of that information available.
We’re also trying to extend that out to the pharmacies. A doctor prescribes a prescription or a therapy and it’s up to the patient to do it. Next time he sees them he doesn’t know if the patient has done that so he can’t make any progress. People are fallible, they’re busy, they can be lazy, so sometimes they won’t do it. So it gives you that information.
We want to go beyond that, even further, to individual medicine. The IoT and all the sensors that we have now allow us to go much further in individualising medicine and therapy.
CBR: Do you expect that eventually every patient will have one account or platform to oversee all of their data?
HY: Yes, the doctor, patient and the insurance provider will be able to see certain things. For example, when treating a patient, diabetes is particularly troublesome and more and more people have it. Aside from insulin, there are now also pills that you take.
These pills are based on medical and clinical trials of thousands of people. This means that what you get is therapy that is the average of these thousands of people. Maybe this treats 10 percent. Then maybe you can go to gene therapy, which is maybe another 20 percent.
But the other 70 percent is really [based on] individual information: what you’re eating, are you healthy, are you sad, are you depressed, etcetera. That sort of information is very individualised. Now we can see it, with the IoT. You see people with FitBits, but you could also perhaps tattoo a sensor on your skin that monitors insulin levels.
You can then individualise the treatment, rather than give them the prescription for Metformin that works for the average person. There is tremendous potential in all of this. The technology is there, it’s just how do we put it together and how can we make it usable so that the average person can consume that information.
That’s what we like about Pentaho (a recent acquisition). You can download it on your laptop and play with it, and it gives you an easy way to visualise and correlate things and provides analytics. It’s a toolset, allowing you to integrate structured and unstructured data.
CBR: How is this being applied?
HY: For instance there’s a city in California where they have a welfare problem. It’s a farming community where the farms have now gone.
There wasn’t a regular relationship between a social worker and the un-wed mother, for instance. I’ve met some college kids that are integrating that information so that there can be a consistent relationship between the caregiver and the social worker. Are they computer scientists? No, they’re social workers. They’re trained in the social sciences and never had any computer science training except what they learned using their iPads. However, that’s enough to be able to correlate this information.
We want to make it usable so that it’s not just data scientists and engineers who can use it, but the people who are really solving these problems. Just pick it up and it works. That’s how the information has to be.