Information technology, ‘big data’ underpin efforts to move personalized medicine to the mainstream.
Healthcare information technology, data mining and data analytics are driving rapid advances in personalized medicine across the country and around the world. The work being done today on this front is but a thin slice of what is yet to come, experts say, and many predict vast advances in the next four to five years.
The paradigm is changing, assert the clinicians, geneticists, researchers and technology makers working to make the shift from generic to genomic. Targeted, insightful, translational are some of the words they use to describe medicine of the future.
At Moffitt Cancer Center in Tampa, Fla, researchers are synthesizing data from numerous clinical sources, a state-of-the-art biobank, and patient-reported information. The Istituto Nazionale dei Tumori in Milan, Italy is testing a new analytics platform with the end goal of helping physicians provide personalized care.
The work these two institutions are undertaking signal the vast changes to come for patient care – not only with the efforts seen on the wellness front, but also with how physicians treat sick patients. How they target disease will increasingly depend on the patient’s makeup – his or her DNA.
Perhaps that’s what inspired Republican presidential candidate Newt Gingrich, a longtime advocate of health IT, to liken personalized medicine to makeup the other day when talking to a group of high school girls in North Carolina.
“If you’re going to church you probably wear one level of makeup, if you’re going out on a date, you may wear a different level of makeup,” Gingrich said. “If you were going to be in a play up here, you may wear a different level of makeup and it would be literally unique to you – to each one of you. We’re going to be able to have very personalized medicine, just the way we have personalized makeup.”
Mark Hulse, RN, VP of information services and CIO at Moffitt, leaves analogies aside to describe the personalized medicine work that is under way at the cancer center where researchers and clinicians are employing a health and research informatics platform developed by Oracle to capture and analyze clinical data.
“The personalized medicine piece comes in because of the molecular data,” Hulse says. “Typically the way it works in healthcare today, physicians go through a diagnostic process, and they treat the patient based on that particular diagnosis. Where the personalized medicine piece comes in is at the level of a particular person based on their genetic profile – molecular profile. We can actually target therapy very specifically.”
Across the pond, in Milan, something similar is happening in a pilot project the Istituto Nazionale dei Tumori is undertaking with IBM, using IBM’s biomedical analytics platform called Clinical Genomics (Cli-G). The platform can integrate and analyze all available clinical knowledge and guidelines, and correlate it with available patient data to create evidence that supports a specific course of treatment for each patient.
"By providing our physicians with vital input on what worked best for patients with similar clinical characteristics, we can help improve treatment effectiveness and the final patient outcome,” says Marco A. Pierotti, MD, scientific director at the Istituto.
Moffitt Cancer Center
The more data there is the better, says Moffitt’s Hulse. Not only is Moffitt capturing its own data into its HRI platform, but it is also collecting data from 17 other hospitals that are members of a consortium founded for this purpose.
At Moffitt, as in all academic medical centers, research and clinical trials inform treatment. The HRI platform has pushed up workflow efficiency.
“What the researchers have said,” Hulse reports, is that it helps them reduce what could have been weeks – oftentimes months – of querying around many different databases. They can look and literally determine within a few minutes whether enough patients exist that fit the inclusion or exclusion criteria for a particular research study, or even clinical trial.”
In the year since Moffitt rolled out the Oracle platform, Moffitt has also seen some advances on the clinical front.
“We’ve been in the process of developing over the past year, clinical pathways – essentially algorithms – to guide our clinicians based on evidence, in terms of what are the right diagnostic and treatment options for patients with a particular disease and a particular staging, for example, within cancer,” he says.
The analytics adds a critical layer. What helps Moffitt stand out are the linkages across the clinical data and the tissue data and linkages to the molecular data.
Other hospitals that have established clinical data warehouses, and some of them are starting to link to molecular data to study what works and what doesn’t at the detailed personalized medicine level, Hulse says. “But I’m not sure there are many that have linked together both the clinical data and the tissue-level data in the bio repository. That’s a particularly important piece particularly in cancer because so much is determined by how advanced the disease is, and that’s often done as part of the pathology piece around the tissue, and then, of course, linking that to the molecular tissue.
“A lot of folks who have tried to do this type of integrated analytics in the past have spent a lot of money and not really gotten what they wanted, says Kris Joshi, VP of product strategy for Oracle. “This time around I guess what’s different is that the technology has matured. There’s much more of a product-driven approach rather than consulting and services-driven, one-off approaches.
As Joshi sees it, Moffitt is one of the leaders in the move toward personalized medicine. He commends Moffitt for “evolved data processes and data governance and integrative care centered around the patient.”
For Joshi and the Oracle team, it’s been easy to work with Moffitt.
“When we went to Moffitt and we discussed this vision with them, it was a meeting of the minds,” Joshi says. “We found ourselves finishing their sentences for them.”
Istituto Nazionale dei Tumori
In Milan, physicians at the Istituto Nazionale dei Tumori, are piloting a new analytics platform to personalize treatment based on automated interpretation of pathology guidelines and intelligence from a number of past clinical cases, documented in the hospital information system.
Developed at IBM Research in Haifa, Israel, the new prototype Cli-G platform works by investigating the patient's personal makeup and disease profile, and combines this with insight from the analysis of past cases and clinical guidelines to help clinicians choose more effective treatment options.
In addition to supporting decision-making about treatment, it can provide administrators with an aggregated view of patient care, enabling them to evaluate performances and using this knowledge to streamline processes for maximum safety, says Chalapathy Neti, director of life sciences at IBM. For example, hospital administrators can drill down into the data to better understand what the guidelines were for insights, what succeeded, and whether treatment quality improved.
"Our clinical genomics solution may enable caregivers to personalize treatment and increase chances of success," explains Haim Nelken, senior manager of integration technologies at IBM Research - Haifa. "The solution is designed to provide physicians with recommendations that go beyond the results of clinical trials. It may allow them to go deeper into the data and more accurately follow the reasoning that led to choices previously made on the basis of subjective memory, intuition, or clinical trial results."
“The more we prove through pilots like this that actually having an understanding of the genomic underpinning can give you more effective and personalized decisions will make the gathering of such information much more prevalent,” says Neti. “The challenge today is not everybody collects this kind of genetic data in everyday practice. It’s an additional test and a lot of times the value of that is not known. So, these types of pilots are valuable.”
Personal across medicine
While much of the research in personalized medicine today is being done in academic medical centers like Moffitt and Istituto Nazionale, and it seems that much of the focus is on cancer care or HIV, personalized medicine crosses specialties. “We know the HIV effort was really about optimizing the drug cocktails,” says Neti. As for cancer, it “clearly has a strong underpinning in genetic mutation.”
That’s not to say, though, that personalized treatment would not be a boon for other conditions, Neti notes.
“Almost every disease, the effectiveness of the drug – the treatment – has a strong underpinning in your genetic makeup in the sense that each of us will metabolize drugs very differently from the other, and that has a genetic underpinning,” he says.
Hulse agrees. “There are many diseases, diabetes, for example, and certain types of cardiac disease where there’s a strong link to genetics. Even just in general when you think about the effect of medication, patients are going to respond differently regardless of what the disease is, just based on protein metabolism. There are small variations from one human to the next. By understanding and being able to look at these variations at the molecular level you can better target therapy. It’s important, too, to understand biomarkers – how can you get an earlier handle on disease and treat it while it’s still very treatable.
Next for Moffit, Hulse says is expanding, both the number of patients bringing the consortium patient data in as well as expanding the data elements.
“The next piece for us is to bring in an even richer set of clinical data, for example, from electronic medical records,” he says. The idea is to treat patients based on their molecular makeup, follow them closely, and record outcomes data.
“In a way, every patient becomes a part of this ongoing clinical study.” Hulse says. “That’s the whole idea that’s being talked about around a learning healthcare system, where you really have a continuous data flow between the point of care – the clinical side – and the research side, and one is continuously informing the other. So, you’re continuously improving care.”
IBM’s Neti advocates for more pilots like the one at Istituto Nazionale dei Tumori. Creating more pilots – more proof points – will help drive the move toward personalized medicine, he says. “The challenge today is not everybody collects this kind of genetic data in everyday practice. It’s an additional test and a lot of times the value of that is not known. So, these types of pilots are valuable.”
Bringing down the cost of genomic sequencing from the $50,000-$60,000 it is today to less than $1,000 also will be a huge factor, in Neti’s view. IBM is working with Roche to achieve a quick, easy – and cheap genome sequencing ($100-$1,000) using what IBM calls “DNA Transistor” technology.
At the Center for Connected Health, part of Partners HealthCare in Boston, center director, Joseph Kvedar, MD, has been pondering the link between genomics and connected health (see his commentary on P. 17) and he has an idea for creating “Personalized Prevention.” With the $1,000 genome sequencing nearing reality, he writes, “there are all sorts of implications, but the most mind-bending is the idea that we will eventually be able to create diagnoses that are unique to you and therapeutic responses that are equally unique.”