Since Crick and Watson figured out the structure of DNA in 1953, there has been a constant effort to make genomics meaningful to medicine in a broad scale way. Here we are, a mere 65 years later, and it seems that genomic sequencing has finally hit the big time. And by big time I don’t mean more scientific discovery – I mean broader reimbursement.
The ability to sequence the genome, or at least subsets of it, for the low, low cost of $99 – no wait! It’s now $79! – has actually been around for a few years, but the utility of doing so has been a mixed bag and most of what the public has heard about is what can be done to either identify heredity (hey, guess what? You’re 32% Neanderthal!) or to maximize anxiety about diseases you might just get but about which there is nothing you can do (well, you can get enough information to make you spend far more on anti-depressants!).
But clearly the real promise of genetic testing is about its use to make better medical decisions and the medical system has had a long road to making that precision medicine promise real. There are numerous reasons for this: early on, the costs of sequencing were too high to be practical, though that has clearly changed; until recently there have been few drugs that had a large genetic component to their use; additionally, most clinicians have not trained deeply enough in genetics to use the science meaningfully in practice and there is a vast under-supply of genetic specialty physicians and genetic counselors. Note there are something less than 1000 medical geneticists and 4000 genetic counselors in the whole country – in this article from 2007 , while slightly dated, it notes that there are more professional astronauts than geneticists in the U.S., which is kind of weird. If millions of people will need genetic testing and advice, we are going to have to retrain those astronauts. However, the number one killer of all new healthcare concepts has been a big part of what has held precision medicine back: there has been a dearth of payer reimbursement to encourage much broader use of finding genetic markers in order to guide treatment.
But it seems we have finally reached a double helix tipping point on the medical front, as this week CMS announced that it will now pay for diagnostics tests that tell us the genetic make-up of solid tumors in advanced cancer patients. CMS, the agency that oversees Medicare and Medicaid, “said it finalized plans to cover FDA-approved tests that scan tumors for a range of genetic mutations” and further agreed that CMS “will also immediately cover tests cleared by the FDA as companion diagnostics, as long as there are approved drugs on the market that clinicians can direct their patients to as a result.” This is a really big deal, as knowing the genetic make-up of a tumor can dramatically alter the potential treatment plan for how to kill it, provided the treating physician is up-to-date on this knowledge about this topic, which is not a given.
This change in reimbursement and the emerging agreement that the ROI is real is a pretty major new development. While the genetic sequencing tests are quite expensive (aka Foundation Medial expects to get over $3000 for it’s new Foundation One CDx test)), the cost of using an oncology drug, particularly if it has no positive impact, is far worse (the average of recently approved drugs is $171,000 per treatment). Unfortunately, it appears that a majority of drugs prescribed don’t have the desired impact. For instance, USA Today article reports on a November study published in JAMA Internal Medicine, where researcher Diana Zuckerman reviewed 18 approved cancer drugs that didn’t help patients live longer. Only one had clear data showing that it improved patients’ lives, such as by relieving pain or fatigue. Two of them appeared to make life worse for the patients. While some of these conditions will still exist, the hope of genetic sequencing of tumors is that the right patients will get matched with the right drugs more often, thus making the economics quite reasonable, all things considered. Since nearly 40% of people get diagnosed with cancer during their lifetime, we are talking big market opportunity and big clinical opportunity.
The import of genetic sequencing is much in the news right now, and consumers are becoming more and more aware of the opportunity this breakthrough is beginning to present. Once relevant only in fertility discussions and for some breast cancer patients, the utility of genomics to realize the promise of precision medicine is clearly on rise.
The risk we have, of course, is that the public doesn’t fully understand the relative value of genetic sequencing for medical reasons and doesn’t fundamentally distinguish it from consumer applications. There are companies using the genomics playing card for crazy pseudo-science that is no doubt causing consumer confusion. It’s one thing to want to know your ancestry or whether you’re a slow caffeine metabolizer; it’s quite another to sign yourself up to check your risk of serious cancers without knowing which tests are legitimate. There are over 70,000 genetic tests on the market and up to 10 new ones make an appearance every single day This may be a rate of proliferation that rivals the increased rate of use of the word “blockchain” in Silicon Valley. Yikes.
And worse, like in the early days of the internet, the charlatans are circling. There are numerous companies that claim to be able to customize a diet to fit your genomic profile, that can tell you what wine is best for your palate, and that can tell you if your kid is likely to be an athletic phenom, but the science here is highly suspect. Even more egregious is that hilarious entry that I read in the always entertaining CB Insights this last week, referencing a company called SpareRoom.com:
One Stanford professor got so fed up with it all that he started a parody company called Yes or No genomics which can tell you if you have genetic variants for $199, based on his adept use of a kaleidoscope.
I suspect that consumers are more likely to pay for the silly stuff than the serious stuff, as consumers are still largely allergic to self-pay for medical products. According to UBS Bank, only .04% has of the world’s population has taken any type of genetic test; if you limit it to the percentage of Americans, you double the take rate, so you can assume people are not wildly flocking to spit in a tube. But still, if people believe that you can actually pick a roommate or decide what to eat based on this stuff, and that is what you really care about, yikes. The estimates for the increase in genomics testing are pretty significant and one of the key reasons people state for wanting to take one is to learn about whether they have a tendency to heart disease or Alzheimer’s. I would submit that these aren’t the best reasons to spend your money; unless you have a particularly heinous cardiovascular condition (only about 5% of cardiovascular diseases have a genetic basis), I can prescribe for you right now what to do to keep your heart healthy: eat right, exercise, sleep well. End of prescription. If you find you might have a propensity to get Alzheimer’s, well good luck sleeping. Not much you can do about it except eat right, exercise. I’ll send you my bill.
Obviously there are massive differences between tumor sequencing when someone has advanced cancer and consumer-directed genetic tests. We should be grateful that the former is becoming commonplace, as it will likely create massive benefit. As for the latter, well….it remains to be seen, except for the entertainment value; or finding a roommate. One of the potential values of the consumer-paid testing is that the volume may lead to enough data to create medical value from the findings. But that assumes that companies like 23andMe and others will sell their data to pharmaceutical companies (oh wait, they do that) and that consumers will be ok with that. Some are, some aren’t. Maybe those whose data are used for studies should get their $99 back? Maybe those whose data are used should get free access to any medicines created from their genome if they ever need it? The whole business model around this has yet to be written in its entirety.
I was teaching my Berkeley class last week and was lucky to have Lisa Alderson as a guest speaker. She is founder and CEO of Genome Medical (a GE Ventures investment), and veteran of two other highly successful genomics-related companies (and a recent Tech Tonics podcast guest). Lisa made two predictions:
- Within 5 years every cancer patient will receive genomic testing
- Within 20 years, the vast majority of the US population will be sequenced
I’m definitely with her on the first prediction, and the FDA, CMS and many health systems and others seem to be heading there as well. As for the second prediction? We shall see. To date only about 1 million people have been sequenced. Will start sequencing people at birth? Perhaps. But what about everyone else who is already born? 20 years isn’t really that long considering how long it has taken to get where we are now (65 years post discovery of DNA). And the cost implications of widespread sequencing are pretty significant, particularly if we still lack treatments to address many conditions uncovered by the sequencing. A small study in Boston found that sequencing a group of 50 people with no discernible need for it resulted in an average of a $350 increase in healthcare costs for those people and virtually no change in health outcome. Yes, it’s too small a study to be generalized to the masses, but I’m sure my readers get the wallet problem if you extrapolate to 300 million.
More precision medicine will come to the fore over the next 20 years for sure (and thank goodness for that), but it also appears we are on a track to fewer insured Americans, at least for now, so the forces are for and against sequencing. Will the healthcare system voluntarily step up to add $350 or even $100 per human if the ROI isn’t there? Doubtful. Will consumers spend their own money for full blown sequencing when they aren’t looking for a roommate? Time will tell.
Nicole Boramanand says
Thoughtful & spot on, as always. At the nexus of broader proliferation of consumer based sequencing (ie outside of oncology use cases) sit privacy and utility. The former is a clear concern, attempting to be tackled by companies like Nebula Genomics and LunaDNA – time will tell how secure blockchain really is, and how engaged consumers are in managing the use of their genomic data. Utility is best realized when accompanied by clinical data — most valuable: longitudinal data over time. This part is much tougher, both in terms of access and fidelity of data given the state of EHRs and how easily usable data can be harvested. Oncology is farther ahead in this respect, in part because the economic goal posts are clearer … the incentives are there for all stakeholders, and while complex, the problem is still reasonably clearly defined. Boiling the ocean of the human genome and distilling it into personal, relevant & clinically actionable insights (beyond entertainment value) is a different proposition, and expesnsive in terms of time and cash. It’s likely reimbursement won’t come until that utility is clearly demonstrated.
Lisa Suennen says
Nicole, yes a very complex landscape indeed. When incentives align the complexity seems to disappear. Lisa
Cassie Wallender says
Very interesting research you’ve done here; I’m glad I subscribed!
I was an early 23andme adopter and it told me to get my act together or my heart would almost certainly kill me. Though that was the kick in the pants I needed to get going and drop 40 pounds, you’re absolutely right that the prescription was universal: eat right, exercise, and sleep well.
Lisa Suennen says
Cassie, glad you’re on a better path to health! Lisa
Harsh Vathsangam says
Seems like we are a few decades off. As such, both positive and negative confirmation studies for arbitrary testing should be taken with a grain of RNA before more robust meta-analyses are available.
On a lighter note, I wonder if the Indian arranged marriage market has caught on to this sequencing-based roommate matching trend yet.
Akmal Jameel says
Was at a lecture by Dr. Leroy Hood who also predicted that everybody will have a health cloud with genomic and other data. I would also add that an additional area where genomics are going to be important is infectious diseases. Currently, in a number of diseases doctors are not able to find the exact pathogen between 10-70% of the cases. They prescribe a best guess antibiotic, then another and another until the patient is cured or dies! Companies like Karius, Pathoquest or my own Fusion Genomics have solutions. Ours is the best of course 🙂
Lisa Suennen says
Hi Akmal, yes that is a very interesting application for sure! I’d be curious to see how yours develops. Lisa