Hyperledger Foundation Hyperledger Global Forum 2021, 11 Jun 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Fast, Secure, and Accurate Evaluation of Prescription Drug Reimbursement... - Alevtina Dubovitskaya

Description

Fast, Secure, and Accurate Evaluation of Prescription Drug Reimbursement Claims using Hyperledger Fabric - Alevtina Dubovitskaya, Swisscom AG

For the reimbursement of some new medicines, health authorities authorize a pre-approval process based on clinical criteria by the health insurance provider. The traditional pre-approval process is paper-based, lacks interoperability and trust, and is often time-consuming, costly, and error-prone. A consortium consisting of a pharmaceutical company, insurance companies, two large teaching hospitals, and a technology provider, created a blockchain-based system using Hyperledger Fabric to optimise the pre-approval process and improve patient treatment. The system was tested with clinical patient data in the setting and compared with the traditional paper-based process regarding. Alevtina will present the system design and implementation, comparative results, and the advantages and barriers for the adoption of blockchain technology for medication cost approval and reimbursement management

Sessions Details & Slides (if available)- https://sched.co/j3dE
Access All Sessions: https://events.linuxfoundation.org/hyperledger-global-forum/program/schedule/

----
Hyperledger – Open Source Blockchain Technologies

Hyperledger is an open source community focused on developing a suite of stable frameworks, tools and libraries for enterprise-grade blockchain deployments.
It serves as a neutral home for various distributed ledger frameworks including Hyperledger Besu, Fabric, Sawtooth, Indy, as well as tools like Hyperledger Avalon, Cactus and libraries like Hyperledger Aries, Ursa. Learn more about Hyperledger projects: https://www.hyperledger.org/use

Case Studies: https://www.hyperledger.org/learn/case-studies
Training & Certification: https://www.hyperledger.org/learn/training
Tutorials: https://www.hyperledger.org/use/tutorials
Webinars: https://www.hyperledger.org/learn/webinars
Events: https://www.hyperledger.org/events
Vendor Directory: https://www.hyperledger.org/use/vendor-directory

Subscribe to the Hyperledger Newsletter: https://www.hyperledger.org/newsletter

Follow-us on Twitter @Hyperledger

Learn about Hyperledger Membership: https://www.hyperledger.org/about/join

#Hyperledger #HyperledgerGlobalForum2021 #Blockchain

A

Hi, my name is, and I'm happy to present you today fast secure and accurate evaluation of prescription drug reimbursement claims using hyperledger fabric.

A

This work has been done by a value-based partnership consortium that consists of two insurance companies, css sanitas, universitat hospital zurich, orbiting, geneve, two university hospitals and amgen. Switzerland agreed pharmaceutical companies in collaboration with swiss commerce as a technology provider.

A

So for the reimbursement of different medications, uh there exists a sort of pre-approval process that is based on different clinical criteria, so this traditional process is paper-based, lacks interoperability and trust, and often time-consuming, costly and error-prone.

A

So the aforementioned consortium um created a blockchain based system uh for the fast uh prior for to facilitate using a hyperledger fabric uh digitalization of the process and as well as its realization, using uh smart contracts on blockchain allows for optimization of this process, and we have tested our system with real-time patient data and also compare it with the traditional paper-based process.

A

I, on this slide, I would just like to compare the traditional process that currently exists and to get and with our blockchain based uh process that we developed, so traditional, um just uh maybe to check is everything good okay, um so the traditional cost demand process uh consists of the following steps.

A

So first patient comes to the medical practitioner and then medical practitioner has to submit a paper-based form that is sent by post to the health insurance, where health insurance involves a trusted physician that has to evaluate the form and decide whether a patient is eligible for reimbursement of prescribed medication.

A

The problem that all this process takes about two to eight weeks, because it can also involve back and forth exchanges between medical practitioner, health insurance and also involve a trusted physician that can take time to evaluate the the request.

A

uh Also, this is the error prone, because we're talking about the paper-based form and also sharing of the patient data and also sending the data over the post, which can also can impose some privacy threats. So in our blockchain based process, we propose to evaluate those forms automatically and also in the digital manner.

A

So basically, uh the patient would count the same to the medical practitioner and then, instead of filling the paper-based form, the practitioner will use web interface in order to insert patient data, and then the smart contract will evaluate whether the person is uh the patient is eligible for this treatment for the reimbursement of the treatment. So this lower the cost, because now we don't need to involve trusted physician.

A

It also allows to collect data to collect the outcomes of such process and also no additional visits is required for the prescription by the patient, because then the patient can receive from the medical practitioner answer immediately whether he is eligible for this treatment and or not.

A

So, as I already mentioned, this work has been done uh um in collaboration with a pharmaceutical company with two different insurances in switzerland and two uh different uh university hospitals in switzerland, in zurich and in geneva, and in collaboration with the swisscom that provided the technical solution.

A

So we started from those um forms that exist in this uh usual paper-based paper-based process where so. These are the forms that the medical doctor would need to fill in and then send by posts to the insurance company for the mission. So what we did is from those forms- and uh here you see a part of the form and also the description uh which was by the way happening in different languages, because we have uh some of some part is in the one part in the french farther season.

A

There are french language and also in the german, uh there is german used, so we had to basically uh develop the interface for two languages and then uh to create those rules in order to then um implement them in the form of a smart contract. So basically we had the form we had the information on how to interpret the form field in by the doctor, and so we created the rules that then become became a part of a smart contract.

A

So from those indications in the questionnaire, the doctor would feel on the paper we created those rules and those rules became the logic that was deployed on all the peers of our blockchain network, where we had two organizations, one for hospitals. uh We had two hospitals, uh testing the system and also two insurance companies, and we had also created the front end for uh and web interface for, each of the type of the user, either hospital or insurance company.

A

So this web interface looks like this and I will demonstrate it to you in the demo, so I prepared the video in case of a demo effect, but I will just try to to demonstrate you in real life how system works.

A

So let me just another.

A

A

So now we see on the web interface, and I also I would like to to show them.

A

The logs of our system, so uh now we actually uh connected um to our blockchain network and we can see uh that there are peers from the hospital organization and the peers from the insurance organization.

A

So now we're going to enter uh the the logs of one of the peers here and basically we can see that uh this pier joined one uh channel and we will see that we have the the height of the chain is like this. So now we go back to our web interface, um where we see how the system works for the doctor and for the insurance company so for now and lagin as uh one of the doctors.

A

So I will again as a doctor from our university hospital of geneva, so here we can see different requests that has already been created by this doctor. We can see different patients a time of the creation of the request and two different companies that has received those requests. We see that some requests were approved and then also there is a reason uh right now.

A

You see them in french and also uh some request has been rejected and again there is an explanation why why this happened so now, uh like I'm a doctor, I would create another request uh that corresponds for this medication and I'll first need to provide the information uh for the patient information about the patient, uh which is going to be jane doe with the entrance number with the davis first and some address.

A

Information now I can choose uh to which in company I would like to send this request, and the information is already previewed automatically about this insurance company. uh Then I have to um put information about the patient medical history, where uh I'm selecting the those information based on our imaginary example, where, for example, patient has diabetes, that is controlled, um diet doesn't smoke, but his cholesterol level is unfortunately, quite high and without this medication.

A

So we can also list here the information about some medications that patients received previously, because here we are filling in the form for the um basically for the offensive medication and therefore the request to be provided information about the previous treatment of the patient and now uh the most important part. We have to select the indication. So basically, uh we, the doctor, would like to prescribe this medication, because the patient has a specific disease and his cholesterol level is pretty high.

A

So now we can either save the uh this request or submit it directly and we're going to submit it.

A

And now we see in the list of the requests, we see that the information from this doctor about the specific patient is sent and the status is committed. So then, almost immediately we see that this request was approved and with the reason why this request was approved. So now we can. um We can log in as an insurance company this one and see whether we receive this request.

A

So now I'm looking at the admissions.

A

Provider and basically.

A

I can see this request that we just submitted that it is approved and as an interest company, I can only see without this data without notification and now coming back, while still have a couple of minutes. I would like to get back to uh our logs of the pier, and then we see so the previous height of the block was 36, and now we're gonna have the height of the block increased by two, because the each we basically have two blocks and each block has just one transaction. That is uh so.

A

Each block is generated 32 seconds and one first transaction is just the data sent by the doctor and then the second transaction is actually execution of this logic of our smart contract.

A

So now just I would like to end the presentation with this.

A

Last slide uh with with the results, and so the biggest result is that we actually were able to test our framework with the real patient cases. Of course, after the consent was provided by every patient for each case. So we test first our system with about 26 cases from insurance companies that was created by insurance providers based on some realistic cases that they had before and also with real patient cases.

A

We had six patients from two different hospitals from two different insurance companies, and so basically, uh the the very important result is that, instead of the average process that takes about 11 days in this specific six patients, uh we had the results that we received in just two seconds and basically also. This approach allows to also identify and help the correct input errors and gives a great potential for optimization of such decision making process.

A

So, thank you very much. I don't know if we have time for any questions I'll be happy to to.

A

A

I see the the question from brian was the hardest part about making this all work. So actually it took us quite a bit of time and I think that the most difficult part was really not on the blockchain side. Well, first, I would say there are two most difficult tasks, so one is creating this logic, from the description and from the paper-based form, because but not even creating the logic but verifying it with every stakeholder, because this is what we had to do we had to when.

A

When we created those business rules, we had to really make sure that they are correct before even implementing that, and then we had multiple reviews with the people from insurance companies from the doctors and really to make sure that we captured any single every um single possible case for different patients.

A

That could that could come and another part was actually web interface and also really making this really trying to make make system more easily adoptable and really um implementing those forms uh in in such a way, so that it's easy for physicians, so that it captures the the missing data, but doesn't try to mess with the logic, doesn't propose, doesn't suggest to add something or to change, something that would then influence the the give more data and influence the execution of the smart contract and yeah. We will try to yes exactly.

A

We will we're trying we're actually writing a paper right now and we're also talking with the we are going to talk with the office of public health in order to in switzerland in order to uh get their opinion and hopefully uh bring the spock to the next level and hopefully have it uh really used by by doctors and insurance companies.

A

A

A

Oh, I'm opening the q a.

A

Oh, yes, that's a great question uh about the mechanism for ensuring that no private data is shared with unauthorized parties, so uh we actually so the smart contract is running uh on the on swiss comp premises. However, it can also be running all the peers could also be running on site on the insurance companies or the hospitals.

A

What we have implemented in order to ensure the the privacy of the patient is that all the patient data is encrypted with the key shared between a doctor that is submitting the data and the insurance company, so all the patient, uh all the uh patient identifiable information.

A

All this information about the the patient like address and age and date of birth it's encrypted and therefore only the designated uh insurance company can access can can get the information from from the ledger about the the patients so only about the patients from this insurance company and also as as a service provider. Suscom cannot see any information about um patient data.

A

So like this, we have really uh point-to-point encryption between um doctor and the interest company to have access to the patient data, and no other participant cannot cannot see any information.

A

So another question: uh do you have a process for deleting editing information? The doctors have inputted incorrectly. So uh yes,.

A

So if this is spotted by the web interface, so, for example, if some important information is missing like indication, then um I can actually share this real quick just to to demonstrate.

A

Like if I would be the doctor.

A

And I would try to create the form, so I will start filling the information and then I would try to submit the form that is um completely not um filled. So, first of all I can save. If I start filling the form, I can save some information.

A

But if I try to submit the form, then uh the I will responsification, that will say, please feel all the necessary um all the necessary fields which are fields such as uh information about private information, about the patient insurance company and, of course, the indication.

A

So, basically uh already on the stage of feeling the information we can spot. If there are any information missing. Also, if doctor made a mistake and maybe forgot to add some information about patient history, maybe about the previous indications.

A

Then and the request is rejected well, the doctor can see the reason of the reject and uh see if there was, uh if maybe there was by mistake or if maybe something suddenly happened,.

A

So another question about uh multiple chain have access to the data it is encrypted, so we have just one blockchain one channel uh where different organizations and different users can either write or query the data. So, first of all, the mechanism is that the the insurance company can, when there is a query to the uh to the blockchain, so only the data that is that is written for this insurance company can be accessed by this insurance company.

A

uh This is this: is the mechanism in the in the query of the uh of the chain code and another part of the question authorized party share the key to decrypt the data, so here we have basically, we use diffie-hellman in order to have the keys generated for each of the pairs.

A

So we have two doctors, so it means that each doctor, when sending uh information, if it if information, goes to uh just one so information gonna go to only one of the insurance companies and then the all the uh personal identifiable data is then encrypted with this key.

A

So it means that only designated insurance company can decrypt this information, while uh the data such as facts like whether this patient uh ex would have have or not have the diabetes or whether some facts from the history they stay unencrypted but because the personal identifiable information is encrypted, then this is how we make sure that even the technology provider, even the nodes, that even the entity that is running the nodes, cannot really get or even get. It cannot get any information about the patient.

A

Thank you for your questions, so I'm just going back to the chat. Yes, I think I'm a little over time. So thanks a lot. Thank you very much.