HIStalk Interviews Don Rule, Founder, Translational Software

Don Rule is founder of Translational Software of Bellevue, WA.

Tell me about yourself and the company.

My background is in software, first at Dun & Bradstreet and then Microsoft. It occurred to me while I was at Microsoft that the ability to digitize biology through sequencing is something that’s going to be very important to healthcare. I spent a lot of time thinking about it there. When I left Microsoft in 2008, I spent a year in a genetic testing lab and realized that just about every lab is going to be interested in genetic testing, but the ability to understand the implications of those tests is not readily apparent.

You’ve heard of the $1,000 genome and the $100,000 interpretation. Getting the cost of that interpretation down is critically important. That’s what we’re focused on. Having looked at all the different shiny objects we could follow, we focused very much on pharmacogenetics because we feel pretty strongly that that’s going to be the first and most pervasive use of precision medicine.

How often do genetic test results change a physician’s mind about prescribing a given drug?

Something came out from Mayo Clinic recently that said if you look over all the potential mutations that there are, the vast majority of people have some mutation that will be actionable at some point in their life. In terms of a specific individual, it’s a little bit skewed because often they don’t get tested unless there’s a suspicion of a problem, so we know we have a sampling error here. But I would say at least 60 percent of the time there’s something that’s actionable.

That patient’s genetic predisposition could mean that a given drug might be entirely inappropriate, or it could be that the dose that would otherwise be chosen might be too high or too low, correct?

That’s correct. For example, 20 percent of the population doesn’t metabolize Plavix well. But if you put together a collection of drugs — and it’s not uncommon that people are taking anywhere from five to 15 drugs — across that collection, it’s pretty common that there is something that you would either adjust the dose or you might look for an alternative on the basis of the person’s metabolism and other factors.

Can you correlate a patient’s new genetic testing results against their old medical history to learn something new, like why treatments have failed or that doses were inappropriate?

Forensically, looking at somebody’s metabolism is not uncommon in trying to understand the cause of adverse drug effects. The most famous case was in Toronto. A woman who had just delivered was given codeine for pain. Four days later, her baby died. It turns out she had multiple copies of the gene that metabolizes codeine into its active form, which is morphine. She instantly processed that codeine into morphine, it was expressed in her breast milk and the baby died. It was only through that sort of forensic analysis that they understood what was going on there.

Are drug companies going back to look for genetic reasons their products may not always work well?

Absolutely. In fact, even some of the development pathways they’ve taken have mitigated away from the cytochromes that they know are variable in different people, or at least mitigated toward different cytochromes. From the CYP2D6 or CYP2C19 that they know are altered in many people in the population, they’ve moved to drugs that are CYP3A4 and CYP3A5 and potentially killed some drugs that would be very beneficial if you could understand who in the population would benefit from them.

Can they determine that genetic influence in the lab while developing the drug or do they have to wait until the drug is rolled out to a broad population to see what happens?

That’s one of the reasons we think pharmacogenetics is going to be so compelling. There is a lot of good data about how drugs that have been approved are metabolized. The FDA, for a very long time, has required studies that show exactly what genes are in effect at the time it’s metabolized to get an idea of what pathways clear it and, to a lesser extent, what pathways are affected by the drug.

As a company, will you stick to pharmacogenetics or expand into other areas of personalized medicine?

There certainly will be others. We look at ourselves as more a platform for genetic analysis. Pharmacogenetics, again, we think there are hundreds of millions of people that could benefit from it and the data is well understood because of the FDA and other studies. But we have begun to broaden. We have a cystic fibrosis panel that’s coming out. We have some other infectious disease that we’re looking at for later in this year, as well as some licensing around functional medicine. There are lots of areas that it’s applicable to. But again, we see pharmacogenetics as well proven, very important to the clinical process, and readily available.

Does the decreasing cost of genetic testing justify having it done just to guide drug therapy decisions?

One of the transitions that the industry will go through in the next couple of years is from reactive to proactive. Right now, it’s common to get a genetic test when you think you’re going to be prescribing Plavix. You’ll see what happens, what is the viability of Plavix, because there are other alternatives, but they’re much more expensive.

What we see happening over time is beginning at hospitals like Inova, where they get the test early in life and keep it in the medical record. From that point on, for the rest of your life, anything you get prescribed, you can at least check it to see if there are genetic determinants of the efficacy or toxicity of the drug. You can make decisions on that basis. The real key there is building that into your clinical decision support in such a way that the physician can make use of that test throughout the future.

Is only one lifetime test required for a given patient to determine not just the pharmacogenetic influences that have already been documented by research, but also those that might be discovered in the future?

There is one broadly relevant test that would be relevant to, say, 180 drugs. There are a few a little more specific. For example, specific drugs for HIV,  there might be a gene that’s fairly difficult to test that would be relevant to that, so you might do a reflex test if you’re considering Abacavir for a particular patient. Certainly there are panels now that cover the vast majority of the drugs that are known to have important genetic effects.

Other than the patient, the beneficiary would seem to be insurance companies that can avoid the cost of ineffective therapy or the treatment of genetically driven therapy complications. Are they willing to pay for the testing?

They are willing. There’s a big challenge right now, though, in reimbursement. If you’re a pharmaceutical company going in to get a new drug approved, you can afford to spend for a gold standard clinical trial for it. In the world of a diagnostic, where the drug may be off patent for 20 years, diagnostic companies don’t have the same returns as drug companies. Even once they’ve produced the evidence, they can’t necessarily patent that evidence, so it might be available to all their competitors. The evidence creation has lagged behind.

In fact, there’s a really challenging dichotomy now between NIH and FDA. They are pushing forward in precision medicine and CMS is pushing back. That’s a difficult place where the industry is in right now. We really haven’t figured out how to get beyond that.

What is especially interesting about that, though, is that we’re beginning to see some forward-thinking payers who are willing to run tests themselves, who are willing to run trials themselves, to see what they could potentially save by putting pharmacogenetics in place. They look at it as a competitive advantage to lower their costs relative to their competitors.

What information from your system do Inova’s clinicians see in Epic?

At this moment in time, what they see is a static report. The evolution that we see in the future is that we can provide, in that static report, the information that’s relevant to the physician at the time they’re ordering the test, but then make the rest of the data available in the EMR as clinical decision support for other decisions in the future. That is certainly a vision that we all share. We’re early on in the implementation of that.

First Databank is distributing your knowledge in their reference content that drives order guidance and alerts from vendor clinical systems. Will that make your information more easily used and widely available?

That’s exactly the approach we’re taking. We’re working on providing what we have, making it available available through a standards-based API so that anyone — whether it’s a pharmacy system, an EMR, an application in an EMR, First Databank, or someone who works with the payers — can plug into our system and say, "Should this person be tested on the basis of the drugs that they have? Where should I order the test from? Once I have the results of that, can I go back and re-query it on the basis of some new set of drugs or some prescription change that I’m doing in the future?"

Where do you precision medicine going in the next five years?

There are a couple fields to look at. Cancer is pretty well along now. There’s a lot of work going on and that will be pervasive in the next five years.

It  takes more parties to put pharmacogenetics into place, so I think in the next five years, we will see the majority of forward-thinking organizations incorporating pharmacogenetics into the prescribing decision factor.

For things like heritable disease, the interpretation and the understanding will be so readily available that for many of the things that are diagnostic odysseys now and many of the things that are rare diseases that are heritable, those will be much, much easier to find in the future, much easier to understand.