Hello, I am Pieter Van Gorp, Assistant Professor in the Information Systems group of Eindhoven University of Technology.My talk is positioned in the BIG DATA area, where it is clear that now and in the near future, data has big value.
Today, we have got used to using “free” services. Of course, these services are not really free, as we pay them by naively handing over our data.
For industrial HWB offerings, multiple vendors channel our data to their THEIR servers (which are often US based). As a result, app builders and researchers in health analytics fail to deliver added value across devices and servers of different ecosystems. The underlying barrier is that while the services tend to be offered for free, the posession of YOUR data is a strategic asset to the service providers. Here, the situation is even worse to Facebook and Google since for those we can at least liberate our data, which is not the case for typical sensor data!Also not that due to this chaos, there is a big gap between the Health ICT domain and the Healthcare ICT domain.
For healthcare data, the situation is fortunately different: hospitals are bound by many regulations and quite keen on keeping your data well protected behind their walls. Arguably however, in doing so they protect your data too much, since they also tend to keep it from you and they hardly share it between health organizations.
Indeed, it turns out that when moving across health organizations, your information typically only follows you by mail or FAX, after tedious procedures. How is this possible, you may wonder?
The problem is caused by a lack of incentives for the health organizations and their vendors to make their systems interoperable. As an exception to this, there are the clinical researchers, who wish to have as much and as rich as possible data. They however also are facing sad realities of inflexible IT service providers. Clearly, the current landscape requires a radical change since the value of the big data does not reach patients nor clinical researchers. Cynically, even the EMR vendors cannot reap the value of big data in HC since integration is essential for that.---- (too much) Also note that health organizations are burdened by complex data handling regulations, so ICT departments really are too burdened to even think about providing a rich, standard API to their EMR. Similarly they are too burdened to systematically consume data from ICTLabs HWB sensor innovations.Eventually, it is really frustrating to see that various hospitals partly use our money to finance the big players who refuse to integrate with SME innovations (e.g., from ICTLabs HWB).
Integrated PHR systems have been proposed as a remedy to overcome the problems for both your healthcare data and your health data. The general idea is that individuals exercise their right to their data, store the data in a private cloud and share it from there, with whomever needed. Also, they take control of their sensor data and share it similarly.
Promising PHR solutions have been deployed in industry.Unfortunately, manybarrierswereencountered, leadingtopoor adoption levels. The literaturerecognizes privacy barriers, barriers w.r.t. data entry and data use methods and business model concerns. The colored arrows on the slide also classify well known products but I do not have the time to explain the specific arrows here.In general, the business model barrier relates to the aforementioned poor incentives both from hospitals, from hospital IT providers and from commercial sensor vendors. The literature identifies privacy barriers for example in the context of insurers. At TU/e, we have interview results that indicate that with GPS and genomic data this will worsen. Data entry barriers relate to the observation that few people keep form-based records up-to-date and physicians mistrust patient-entered data in any case.
To tackle the business model barrier, we took inspiration from Apple. While Nokia etc. were monopolizing the development on their phones (like hospital IT vendors do), Apple opened up to the crowd and by now earns more in Store revenues than in hardware.
Second, let us first look into Privacy. In Health and HC, this is a key point due to the Open Store.It all starts from the assumption of having centralized European private data clouds, where citizens own their data. Citizens should be able to share, and companies should be able to operate on big anonymous data sets, but citizens should be in control of sharing (including the choice of anonymization algorithms). The cloud idea is opposed to current healthcare IT practice while the PRIVATE and CENTRAL cloud idea is opposed to current health IT practice.This concept is fully in line with the movement of EC VP Neelie KROES!Pictures and articles on movement towards EU cloud: - http://www.zdnet.com/cloud-for-europe-launches-as-sap-backs-eu-rules-on-data-privacy-7000023205/ - http://www.cee-cloudinsider.com/resources/Europe-begins-work-on-its-EU-wide-cloud-system
To combine Privacy withour Goal tocrowd-source Appdevelopment, we provideit at the platform level. In this way, even paranoid users can be sure that evil app builders cannot steal their data while ignorant users are protected in the first place.*** Essentially, we have built a safe App execution layer on top of the European data storage cloud.****
Regarding data entry effort, we aim at integrating as much as possiblewithexisting data sources (EMRs, sensor data providers, etc.). Technically, we do this by removing the key impediment to the transfer of data, being the endless pursuit of the “ideal data format”. In our approach, the platform can consume any type of file and the translation to a standard format is again moved away from hospital or vendor staff to the crowd of small software vendors and App builders. Note that this is also why we put forward MyPHRMachines as the right *technical* infrastructure for connecting health IT and healthcare IT.
We have demonstrated the effectiveness by showing how web access to ALL your radiology images can be supported JUST by an APP. ---(THE REST IS TOO MUCH)Note that in this demonstrator, we ride on the practical availability of DICOM CDs (no paternalism there!). In fact, we save radiologydepartments money iftheysimply have to push the data to the PHR platform ratherthanburning the CD (= manual task, involvingalsopatientwaiting). Soon after, we had our first externally contributed App for international REM around individual patients. We want to deploy that prototype in production and apply it to other use cases.
Summarizing, MyPHRMachines was built to have an open app-store with platform level privacy protection and a trivial data API.This resolves the privacy and data entry concerns and partially tackles a business model issue (i.e., by crowdsourcing app dev).In the following, we describe how in ICTLabs we wish to complete this from the business model point of view.
The proposition is to provide an alternative to the free use of HWB services and the expectation to get Healthcare data for free. By offering both device vendors as well as hospitals/clinics/GPs a fee (cf., the yellow arrows), they are reimbursed for providing access to your files/streams (cf., the blue arrows). Also, pharmaceutical companies and clinical researchers can now buy data sets from their ethical owners (the patients). An eBay style bidding platform should mediate between patients and research firms. The mediator provides a storage service (with provenance features), a dataset store (i.e., a shop for data), but also an app store with a safe execution service. The latter enables the research firms to offer a free app that tests whether a data set (or associated person) is of interest (cf., the green arrow). So besides buying data via the platform, a firm may buy participation to a clinical trial for personalized medicine.
The proposition is to provide an alternative to the free use of HWB services and the expectation to get Healthcare data for free. By offering both device vendors as well as hospitals/clinics/GPs a fee (cf., the yellow arrows), they are reimbursed for providing access to your files/streams (cf., the blue arrows). Also, pharmaceutical companies and clinical researchers can now buy data sets from their ethical owners (the patients). An eBay style bidding platform should mediate between patients and research firms. The mediator provides a storage service (with provenance features), a dataset store (i.e., a shop for data), but also an app store with a safe execution service. The latter enables the research firms to offer a free app that tests whether a data set (or associated person) is of interest (cf., the green arrow). So besides buying data via the platform, a firm may buy participation to a clinical trial for personalized medicine.
This business model is ethically correct, since (1) transactions are triggered by the ethical owner of the data, (2) patients can manage their health in unprecedented manners, (3) researchers get access to more data, and (4) App SMEs can enter a currently closed market.
Moreover, the approach is financially sustainable, since the platform owner can have a percentage on all transactions (like Apple but for health data, health software services, or physical services such as a trial participation or homecare for an elder). Also, insurers (and governments indirectly) can decide to reimburse some data purchases from patients. In this way, they invest in the use of a platform rather than in the mere construction/maintenance thereof.
We hereby warmly encourage you to contact us if you are either - involved in national or EU-level cloud storage, - developing sensors and looking for a future proof business model, - a hospital that is open to share data and leverage outside data