By Dr. Heather Packer, Research fellow, the University of Southampton.
After attending both the International Semantic Web Conference (ISWC) and the Joint International Semantic Technology Conference (JIST), I considered what this meant for research and development in this exciting area.
Some of the most interesting work presented at ISWC and JIST were in the areas of cultural heritage, such as the recording of personal stories, and integrating historic maps with new timelines. This was highlighted in the opening keynote at JIST, which was given by Eero Hyvönen of Aalto University in Finland, presenting its work on using Semantic Web technologies to preserve Finnish cultural heritage.
These included the preservation of ancient shoemaking methods, through the digitisation and documentation using Semantic Web metadata with multimedia, interviews and written sources. Secondly, the complete transcription of the Finnish national epic, the Kalevala, into a Semantic narrative, and the process by which topics are linked from their own ontology portal, so that topical connections in the Kalevala narrative can be made with other Finnish cultural heritage work.
Aalto University’s work shows that Semantic Web technologies, such as its Cultural Sampo ontology portal, allow cultural heritage artefacts and narratives from multiple sources to be brought together and automatically cross-referenced. Examples were shown where the cross-linking between sources has already benefitted researchers, with technical barriers easily overcome.
In the future, I am aiming to work towards a Semantic Web which will allow narratives to share workflows and stories about companies, as opposed to more traditional methods like statistics calculated from databases. These narratives can be used to explain things based on people’s past experiences and their interests (taken from their actions on the web), to make them both more useful and engaging.
One problem, however, arises from where is it acceptable to gather and use data. Many of the people I have spoken to in academia and industry have said that information taken from their emails is too intrusive and people as a whole are unwilling to use such a system. However, people are more willing to adopt systems that use information from social networks where they can freely censor information about themselves.
Yet in my experience the most useful information is often to be found precisely in private online places such as email and calendars. In the future I would like the Semantic Web to allow me to attend a conference in another country, and automatically (with optional and minimal input) handle my flights, hotels, conference registrations and restaurant recommendations based on preferences that I had made in the past, such as price range and hotel recommendations and amenities.
In addition to academic research, the Semantic Web also has applications for business and handling personal data. The latter, in particular, has recently seen its research spurred on by a number of initiatives, including the midata initiative from the UK government’s Department for Business, Innovation and Skills (BIS). The initiative, which is due to start in 2013, mandates that companies must supply data they hold about a person back to that person in a machine readable format, and under an open licence.
The Semantic Web provides an obvious framework for enabling this at low cost to businesses – there are already numerous examples of marking up personal data under appropriate licenses such as the Open Government license used on data.gov.uk. Semantic Web technologies would therefore enable businesses to comply with new data protection legislation in a cost-effective manner. End-users that receive their data will also benefit, because there are numerous analysis, visualisation and storage mechanisms which already work with Semantic Web data.
The need for storing, managing, using and sharing personal data continues to grow. In response, numerous business startups which focus on providing such services have been launched. Meanwhile groups such as the W3C Read Write Web community group are discussing approaches to using Semantic Web techniques for publishing, receiving and sharing private data. For users this means that it will be easier to make their data work for them, including sites that use your data to help you save money, such as Bill Monitor, which analyses your mobile phone bill to find out how much you can save by getting a new phone contract. It is very likely that similar services will exist in the future for other utilities, such as electricity, gas, and broadband.
The future of the Semantic Web is making it easier to access increasingly richer presentations of our history and heritage, and also publish, and thus increase the amount of cultural heritage material being preserved and made available online. The future of personal data is also one which is expanding rapidly, towards the goal of helping people to make more financially beneficial purchases, and to better manage their private data.
Commentary: Back to the interoperability future
The recommendations for Stage 3 of meaningful use are now out for comment. Coincidentally or not, there is a new degree of pessimism about when health IT interoperability will ever be achieved. The issue of progress, or relative lack thereof, on interoperability surfaced just before the election with members from both houses of Congress questioning whether HITECH funding of electronic medical records should be continued without interoperability standards or more rigorous meaningful use requirements in place.
Some dismissed the questions from Congress as election season rhetoric, but at the same time, many industry professionals have again resigned themselves to a long, slow road ahead. Recent, non-political, Congressional testimony suggested interoperability is still another decade away. And there are enough renewed discussion threads of potential “interoperability solutions” by newbies and statements of dispirited resignation by old hands to substantiate a serious trajectory problem.
HIT’s déjà vu all over again
Information exchange and interoperability have long been seen by people involved in health IT as being central to achieving meaningful outcomes with technology. Health IT professionals certainly recognized that providers needed to adopt health IT to start, but they also have long held that the data needed to be mobile and not stuck in a particular IT system or organization for many of the benefits of health IT to develop. Given how vague and ill-defined interoperability can be, and given the sparse empirical evidence for some of these assumptions, it is a little surprising how resolute the professionals are with these conclusions. Perhaps it comes from the practical challenges of trying to support continuity of care, or of working to aggregate data for quality, efficiency, public health, and research purposes, or from simply battling the obstacles to making disparate hospital systems work together.
For many of these same professionals, HITECH and its billions of dollars for health IT were thought to be a potential interoperability game changer. It was the first major national investment in health IT. And with the magnitude of the funding, the “hook” of the meaningful use mechanism, a second generation certification program, and the mission of supporting the needs of health reform, many felt we should have turned the corner on interoperability. Yet for those in the trenches, interoperability is still an uphill grind if not largely elusive. They just aren’t seeing many health IT systems that can easily process information that other systems provide. While information exchange is advancing in some ways, without broader exchange and the interoperability needed to process “foreign” information, health IT can actually act to increase the unnecessary information that a provider has to review rather than help make the provider more efficient.
The corner seems to have been turned on the adoption of electronic medical records, but many remain worried that the same is not true for interoperability. The interoperability trajectory is not proportionally steep, the HITECH tools to accelerate to this goal seem to be receding, and the newly installed base of largely non-interoperable EHRs may now be yet another obstacle to achieving a fully interoperable health IT infrastructure.
LOWTECH?
How, with all of HITECH, did we arrive back at a distant future for interoperability? Perhaps there was not enough money to overcome the entrenched interoperability problems of healthcare? Perhaps money alone is not enough? Opinions here vary as much as they do with potential solutions. It should be noted, however, that much of the political leverage of HITECH has been focused on quality measurement rather than interoperability in the early stages of meaningful use. There has only been so much pressure to apply and the lion’s share of it seems to have gone into process and quality measures. The theory has been that health reform and measuring quality will pull together the meaningful health IT that is needed to achieve it, rather than the competing approach of engineering the health IT needed to improve the quality of healthcare that can be performed.
Some health IT people point to the interoperability testing components of the certification process as also being problematic. One issue, perhaps inherent to how HITECH is structured, is that certification is only applied to EHRs. For testing in other industries and even testing of successful health IT systems like e-prescribing, all or most of the participant systems and organizations are certified – not just one piece. There have also been many comments about the overall lack of specificity of the interoperability related criteria of Meaningful Use. Of course, this lack of specificity can be traced upstream to the general lack of well-specified, unambiguous implementation guidance for most of the health transactions that need to occur.
Finally, many health IT professionals continue to point to the lack of a business case for the transactions that are being advanced to move health information around. Health IT implementation guidance includes standards for terminologies (for which there has been substantial progress in Meaningful Use) and standards for messages (groupings of data to support an activity) and technical transactions. The most prominent transactions of downloading data to a patient (Blue Button) and pushing data to another provider (Direct) do not align with compelling business cases for healthcare information exchange or requisite interoperability. Blue Button may align with efforts to make healthcare more patient-centric, but patient mediated electronic information exchange, such as that which would be fostered by non-portal Personal Health Records, is almost non-existent. And stated bluntly, what does a provider get out of pushing data (via Direct) to another provider? It may be the right thing to do, and it may be itself pushed by meaningful use measurement, but it is not a compelling need for the provider who must initiate the action.
Measure at least once
Certainly one part of forward movement on interoperability will be to move from what is traditionally a gross, subjective, and qualitative assessment of its state to a fine-grained, objective, and quantitative one. Health IT interoperability is now almost always talked about in subjective and coarse terms. “Semantic interoperability” and “plug-and-play” are bandied about as interoperability goals, but are characteristically referred to as either existing or, mostly, not. Interoperability, however, is both multifaceted and nuanced. There are data, transactional, business, cost, privacy, policy, and incentive elements of interoperability. And almost all of these elements have shadings and nuances of impact.
Effective interoperability measurement will help with decisions about what else can be done to move forward faster. Measurement, at least, will substantiate a need to double down or change course on activities to bring interoperability back to being a possibility for the present.
John W. Loonsk is chief medical information officer of CGI Federal.
Other commentaries by John Loonsk:
Health IT orthodoxy 2.0 after the Supreme Court ruling
Why stage 2 MU transactions need more than SMTP
Comments for the meaningful use commenters: Specificty matters
View the original article here