The PROOF (Prevention of Organ Failure) Centre at St. Paul’s Hospital, hosted by the University of British Columbia in Vancouver, is one of those leading-edge research organizations aiming to move us a little closer to the world of precision medicine, with the help of semantic technology. In the process, it also hopes to have a positive impact on the high costs of health care.
That’s a problem not just in the U.S., but also in Canada where provincial governments bear the burden of rising health care costs, which make up 45 percent of the budget for British Columbia alone.
PROOF has underway health economic impact analysis studies that show the promise of improving the accuracy of diagnostic blood tests so that physicians could better manage patients with diseases associated with organ failures and transplants. Results of its initial models for patients with heart problems that required transplants show that the savings over 5 years in the U.S. could be around $40,000 per patient, according to PROOF Centre chief operating officer Janet Wilson-McManus.
The computer models explored the protein and gene markers derived from plasma and white blood cell tests of patients who reject heart organ transplants vs. those who did not. The goal is understanding what markers are most influential in outcomes, and so which will be the best ones physicians should test for in the blood to better manage patients. These tests could also be used to help determine the right amount of therapy for these patients. Some 38,000 molecules are measured across hundreds of patients in different cohorts to discover which markers to use in the final blood test.
Heart failure takes its toll on health care system costs, but the most frequent contributor to ER and hospital costs in the U.S. and Canada is COPD (chronic obstructive pulmonary disease). Some COPD patients seem to need more treatment than others, and physicians could improve care quality and potentially reduce repeat ER visits and hospital stays if blood could be tested for markers that correlate to a patient’s chances of being more or less stable. The former group might not need the most rigorous treatment, while the latter group could benefit from a more intensive regimen that keeps them out of the hospital as often, and therefore also could positively impact health care costs.
When it comes to COPD, in Canada, “if we were able to implement tests so doctors could help prevent people from coming to the hospital so often, we could save $40 billion over 25 years,” Wilson-McManus says.
Where semantic technology comes in, or will come in, is around the data integration and visualization work on the research side, to better understand factors that may be at play for particular patients that could affect how their organ disease treatment should progress and where risks may lie. There’s a lot that has to be understood about different patient phenotypes, for example, to help identify what is considered normal for patients from different countries or ethnicities when it comes to things like blood pressure or creatinine levels, she says.
Getting It Together
“So the semantic piece is bringing that information together to help us make those improved diagnosis so that we can then improve the treatment and care regimens,” Wilson-McManus says. “All the information we gather eventually can be re-mined for new types of blood tests or for new therapeutic targets.”
PROOF is working with partners including IO Informatics on this, and also hopes to use the vendor’s Sentient products to further build out connections among its gene and protein models.
“The problem is people create all this data and then it kind of just gets lost. So by making sure this data is available, and by using semantic technologies that are now out there, you can actually mine everybody’s data and do this all more quickly ,” Wilson-McManus says. “In the transplant world there are lots of people doing similar studies but then everybody kind of keeps in their own little space. Once we mine our data and get our specific markers that we have questions for we are going to put our work out to the world for use.
So with IO Informatics and semantics, someone can come along, combine our large data set with another large data set, ask different question and more rapidly do discovery and validation without having to enroll more patients because the data is already there …. So it will allow for future studies to be done much more rapidly and much less expensively.”
You need to be able to “visualize relationships between [data sources] and identify patterns from the data, regarding, for example, what is unique to a patient at risk of organ transplant rejection,” Robert Stanley, IO Informatics president and CEO, recently told The Semantic Web Blog. “That’s a real use case.” See this article for a recent discussion with Stanley about some of its latest technology. The company also recently announced a strategic partnership with computational proteomics vendor Sage-N Research, Inc. to discover potential biomarkers and a microbial detection system; future applications of the work could enable automated screening for biological threats, to characterize origin and type of disease, and to develop preventive measures (drugs or vaccines) effective for several classes of microorganism.”
PROOF is a not-for-profit, so its work has to be focused on creating blood tests that can be done at lower costs, both to help the health care system but also to attract commercialization partners for clinical lab apps. ”We have to look at whether we provide enough value to partners who want to do that, and we also must look at the regulatory piece,” Wilson-McManus says, as the FDA and/or Health Canada sometimes have requirements around clinical labs’ blood tests and trials also will have to be conducted to ensure tests meet agency requirements.
PROOF is furthest along with its transplant work, as it’s been working on it since 2004, and it has hopes of getting its blood tests to the clinical lab space sometime this year. That foundation work, she thinks, will make future blood tests it develops more readily available, in something like a three-year window.
- AgilOne Introduces Self-Learning Email Optimization Tool to Democratize Big Data for All Marketers
- What The NSA Can Do With All That Data
- FIBO Summit Opening Remarks by EDMC Managing Director Mike Atkin
- Expert System Develops Semantic Search Engine for Wolters Kluwer Italy