Michael Nierenberg, Medical Director of Navigenics comments on “Absence of “High Penetrance” in SNP Genomic Services” regarding my statement “existing SNP genomic services just aren’t that useful.” (comment appended below)
Rather than entertain an “ongoing discussion” about “preferences” and confidence, this debate about Navigenic’s efficacy can be immediately resolved with a simple concrete response to the following challenge:
Dr. Nierenberg, as a physician, what action would you almost certainly prescribe given a Navigenics test result which you would almost certainly not prescribe without that information?
Assume no bound on cost. Please remember that you must also defend why you would almost certainly not prescribe the given action without the genomic information if cost were no issue. For example, any non-invasive screening test would be a bad example because one could reasonably prescribe the test to all cost-insensitive patients without genomic testing.
Additionally, you may respond with an economic argument. As of 2008, state an approximate dollar amount savings for what healthcare spending and by what savings mechanism (e.g. insurance reimbursement, state provider). State as an expected net present value savings per average American greater than the cost of the net present value cost of a Navigenics test ($2500 + $250 each year for life). If you respond with an economic argument, you must respond with a falsifiable savings hypothesis specific enough to be tested by real patients.
Please respond within one month, though sooner would be better considering Navigenics has been for sale to consumers for several months as a cost-insensitive service for improved healthcare efficacy, and I thus hope Navigenics has a long list of excellent responses to this efficacy challenge already prepared.
To be fair, I will meet this challenge immediately for BRCA testing.
Cost-Insensitive Efficacy Challenge: BRCA
Test: BRCA 1&2 genomic test by Myriad
Physician Advocate: Dr. Barbara Ward, surgical oncologist at Yale-New Haven Action: prophylactic bilateral mastectomy (the removal of both breasts before evidence of cancer) Positive Clinical Precedence: Jessica Queller. In fact, she wrote a book about it which she discusses at Helix: Jessica Queller and “Pretty is What Changes” [audio] Physician Testimonial:
Dr. Barbara Ward: “In terms of managing hereditary [breast cancer], I have to say as surgeon, I believe very much in prophylactic surgery. I look forward to the day that we’ll look back and say ‘can you actually believe that we did mastectomies for [BRCA] gene carriers?’ but at this point in time it is the best that we can offer.” Source at Helix [audio]
Why Action is not prescribed without genomic information: Action is removing a woman’s healthy breasts Other relevant qualified action: testing for family, oophorectomy, fertility counseling
I am unable to make an additional economic argument that meets my challenge criteria.
Thank you for your comment, Dr. Nierenberg, and we anticipate your response. I understand you and your team at Navigenics are genomic enthusiasts, but Navigenics markets itself as a genomic test held to a high medical standards of efficacy that other genomic tests are not. I and others simply are yet unconvinced regarding that advertised usefulness. Please help better inform us.
Appendix
Dr. Michael Nierenberg’s original comment [source]:
Mr. Andrew Yates offers the following comment in Absence of “High Penetrance” in SNP Genomic Services about espousing the uselessness of genetic testing by companies such as Navigenics:
“The reality is that except for a few enthusiasts (like anyone in the DNA Network, including myself), existing SNP genomic services just aren’t that useful. It’s not useful to know that one has a 7% lifetime risk rather than a 5%. I need to know risks of about 40% and higher.”
In response to Mr. Andrew Yates I offer the following information about lifetime risk assessments offered by Navigenics (http://www.navigenics.com).
On the Navigenics HealthCompass six conditions have a lifetime risk potential that exceeds 35% (Five conditions exceed 40%). Since some conditions are less common in the population, it is useful to also consider relative risk, as opposed to absolute risk.
Half of all 23 conditions reported have a relative risk (compared to the average genetic risk in the general population) greater than 2; seven conditions, greater than 3; and three conditions greater than 4. For comparison, this level of relative risk is in line with other well-accepted, commonly used risk factors such as positive family history and environmental risk factors (high lipids, age at menopause etc.)
I hope this clarifies things for Mr. Yates who can now feel better about being an enthusiast.
I would also ask Mr. Yates to reconsider his statement about why Navigenics is not doing BRCA testing. Patent considerations aside, it is of import only in about 5% of breast cancer cases (those with “extreme” positive family histories and early age of onset). The contribution of BRCA1/2 mutations in sporadic breast cancers has not been determined conclusively. Our preference is truly to offer a test that is applicable to the majority of individuals. We are cognizant that in a minority of people BRCA testing is important and have addressed this issue on our web site in several places. We realize that no test is complete in and of itself; otherwise people would have heart surgery based only on a treadmill. So, I would ask Mr. Yates to please give us credit for more than trying to get around patent considerations. We are enthusiasts as well.
Incremental Agglomeration, Creative Destruction, and the Impending Forest Fire of the American Medical Establishment
Part 1: Introduction
The great promise of every modern medical initiative has been “cost savings.” Indeed, what magnanimous medical aspiration in print or powerpoint could preclude a fast-fact stat about the alarming overspending in American healthcare? Yet, as America continues to realize these health initiatives —ostensibly in the spirit of “cost savings” —yet, health spending inexorably grows despite typically marginal improvements in care.
Electronic records, evidence-based medicine, patient-centric care, personalized medicine, universal coverage, and genomics —sure, they all sound good, but can America afford another revolutionary “cost saving” medical industry initiative? No, because revolution is destructive, and what institution willfully destroys itself? Not institutions that still exist. So like the non-directive flaws of biological group selection theory, the drastic improvements promised by revolution are beyond the incremental and self-sustaining mechanisms of institutionalism —as intellectually unaesthetic as that may seem.
Every new medical initiative purports revolution by some combination of better education, savings, standards by implementing some new discovery, invention, or idea. But there’s no grand medical conspiracy inhibiting the potential of these ideas; the mundane simplicity is that change is hard work, people are busy, and everyone stakes in the status quo. Only outside competitors are willing and able to metabolize the medical establishment, and while these well-minded initiatives may snuff immediate needs, rather than sparking significant change, they continue to agglomerate like a thicket. Thus the longer revolutionary fire is procrastinated, the more dead wood accumulates, and the greater the inevitable conflagration in healthcare will be.
But today, two decades of information technology has made healthcare revolution imaginable. First, medicine has matured from scholarly practice to information science, and science needs no priesthood. Next, decades of accumulated initiatives choke American healthcare in waste information, a problem information technology solves well… too well. From problems like “pager tag” to communicate simplistic information like blood pressure and on-calls to “governance by bureaucratic harassment” like insurance coding and liability threats outside strict procedure, the revolutionary solution destroys the existing systems and their vested interests. But, destroying useful systems is locally suboptimal, and institutions can act for their own immediate growth and survival. Thus, revolution is fundamentally unachievable by insider initiatives despite theoretical feasibility and best intentions. So finally, a willing and able outsider is necessary to imagine revolution, and the elite groups behind information technology giants like Google have announced their revolutionary intentions with ventures like 23andMe and Google Health. However, revolution is unpredictable, and by the power of the Internet, the eventual leading health revolutionaries could be anybody.
So manifests the quintessential American pragmatic hypocrisy: all tout the wealth-building virtues of creative destruction until the torch of change is under you. Then, change is “unethical.” Only law is more artificially sustained by its own ethical sophistry and willfully abstruse erudition as is the gross establishment of American health. America hates its healthcare, revolution is coming, and no insider medical initiative can stop it forever.
This is a four part serialized essay which will discuss what a revolution in health will and will not look like, why revolution is possible today, and why direct-to-consumer genomics is sparking our imaginations. Subscribe to Think Gene to be notified when new sections are published.
Tip to Dr. Steven Murphy for bringing this article to my attention. He notably accompanied his email with much more profanity than I will use in this post.)
But because Jason’s mother had a history of strokes, Doneen wanted to dig deeper. She used deCODEme’s genetic test and searched Jason’s genome for a gene that medical studies have related to A-fibrillation. Jason was positive for the A-fib gene, so despite the fact that his heart-monitor test was negative Doneen put him on a heart monitor, but this time for a month in order to be more accurate. Sure enough, the month-long test showed that Jason was going into a-fib. Doneen immediately started treating Jason with a different course of medicine.
‘The test changed his course of treatment,’ she says enthusiastically.
Doneen cautions that just because a patient has an abnormal gene doesn’t mean that patient should be treated with medicine. In fact, she says there is no evidence to treat patients based on an abnormal gene. However, knowing that Jason had this abnormal gene, and had a family history of stroke, Doneen decided to take action, and put Jason on a heart monitor, which fundamentally altered his clinical course.
Why not just put everyone on a heart monitor for a month, instead of a week, and therefore make sure patients don’t have a false negative test?
‘Putting everyone on a monitor for a month is not feasible. But putting someone who has an abnormal gene on a monitor, that’s feasible.’
In other words, while putting everyone on a monitor for a month would catch more heart attacks, it’s not a “feasible” allocation of resources. So, the deCODEme test, which as stated has no evidence, is being used like a random number generator to justify picking some patients for additional care and not others. Why? They tested positive for the A-fib gene. Magic.
So, since there is no evidence, only “decision to take action,” you could replace the deCODEme test with a homeopathic test that produces the same ratio of positive results for the same clinical effect. I understand that people sometimes feel they need a special reason to pursue healthcare from which they could always hypothetically benefit, but there is no rational necessity to produce this impetuous prior to treatment other than the psychological comfort of rationalization.
I would like to see a study where one person gets the genetic test, and one person gets a random jibberish report recommending the same thing, and see if there is a clinical difference. Random jibberish is much cheaper to produce. Why not save your money and get a random jibberish test to help you “decide to take action?” I will put it in a leather case and have a doctor sign sign it on real parchment and read it to you in a very authoritative tone and use big words. You will be 238.8% as decided to take action as if I posted the result on your genomic Web 2.0 deCODEme profile. (there is no evidence to back this claim)
Non-casual risk testing at low-mid penetrance like deCODEme has great potential someday, but only to more optimally ration scare health resources for better net efficiacy that otherwise should be provided to everyone. However, there is no shortage of “eat healthier, reduce stress, exercise, and see a doctor if you have a problem” which is the best medical recommendation that can be produced by today’s DTC testing.
On an individual scope, one must ration personal wealth to pay for healthcare. It’s feasible that low-mid penetrance can aid in making more rational decisions. However, it’s much more likely that personal decisions to pursue which healthcare strategies are in no way rational or scientific and that “optimization” is a delusion. You can’t perform risk benefit assessment arithmetic on “your feelings.” 20% + you’re scared isn’t a number with a dollar sign.
On a social scope, low-mid penetrance risk assement can help better allocate a fixed amount of resources per person to purchase the most personally effective healthcare. However, this is assuming that the healthcare industry isn’t corrupt and would use the tests to justify whatever medically-arbitrary decision they’ve already made to benefit themselves. I don’t think that assumption is justified until the medical benefits of these tests are so obvious as to prevent such bureaucratic abuse excused as an “ongoing discussion.”
But, by all means, buy a deCODEme test. It’s the American way. But do it because you want it, not because of some magic you wish was medicine because you’re afraid of illness and death.
Adriano Squecco of DNA-forums.org is collecting Y Chromosome information from DTC consumers and compiling an open community ancestry database: Y-DNA-FORUMS [excel spreadsheet] (my results are under “Yates” on the 23andMe v1 page)
Interestingly, he is advertising his project on the 23andMe user boards, which is how I found Adriano and his project. Since users can download their 23andMe (or deCODEme) genomic test results as a text file, any 3rd-party service in a secondary genomic market can provide interpretation and collaboration services without ever handling biological samples over the internet.
This is the future.
The sequencing of the genome itself isn’t as interesting because we know what the end is: a cheap, accurate, complete genomic sequence. We may think of genomic testing today as an obscure luxury service, but the technology will bottom out and sequencing companies will consolidate, all offering the same undifferentiated sequence, all competing on price, economy of scale, and marketing.
Instead, what is interesting, what offers an unlimited potential for differentiation, competition, and innovation —is the secondary genomic market, services which accept biological test results and produce research and interpretation. And while genomic testing is the obvious start for a secondary medical information market, why not any medical test? We already send away blood and cultures to a lab and get back data at medical facilities, why not cut out the middle and do it directly through the mail and the Web?
Que Dr. Steven Murphy freaking out about involving doctors and irresponsibility… except, now we got smart. We’re going to beat the “hackers” at their own game, right Steve?
Introducing HelixGene
PS: obviously, this is just a quick proof of execution. Give us at least more than one night to build a prototype of a radical new model of health care, ok?
Introducing: HelixGene Foundation for Better Genomic Medicine. You email us your medical questions about genomics, our licensed medical genomic doctors email answers. This announcement is our soft-launch: we are accepting and answering emails and paid subscriptions, but we are still building the service and have not officially launched. Our genomic specialist physicians include:
Steven A. R. Murphy, MD (also known as The Gene Sherpa)
Clinical Genetics Fellow, Yale School of Medicine;
Managing Partner, Helix Health, PLLC
Adam J. Messenger, MD
Pharmacogenomics Specialist, Department of Pharmacology, New York
Medical College and Graduate School of Basic Medical Sciences
Matthew B. Lubin, MD
Jennifer Ibrahim, MD
Joy Samanich, MD
Our general subscription is $14.95 per month (FREE first month alpha accounts) to post medical genomic questions to HelixGene’s public forum by email. Private emails are $145 each and will be answered by a HelixGene health professional in 48 to 72 hours.
We DO discuss the medical validity and implications of all DTC genomic tests in a medical setting including: 23andMe, deCODEme, and Navigenics.
For the press, we provide expert opinion services for journalists and sell expert review subscriptions for mass media publications including medical genomic information. To enforce accountability we feel is lacking in medical reporting in the mass media today, HelixGene publishes medical genomic report cards for publications. Pre-submitting your publications and consulting with our experts helps us help you accurately and honestly report medical genomic information to the general public. Ask our about media medical consulting plans including those for bloggers, independent journalists, and major publications. Email us for pricing.
We also host a private, invitation-only genomic specialist forum which is free. If you are qualified, ask for an invitation to our expert forum. This is a forum for doctors and scientists to discuss academic issues, particularly to discuss the merits of new genomic services, academic issues, and troubling medical genomic reporting in the popular media. Email us for an invitation.
Statistics and disease risks are invented for this exercise!
1% of people will have Hubbub Disease by age 70. Up to 80% of people with Disease will get positive HEBOT mutation tests. 9.6% of people without Hubbub Disease will also get positive HEBOT mutation tests. A patient had a positive HEBOT mutation test in genomic screening. What is the probability that patient will have Hubbub Disease by age 70?
The correct answer is about 68.1%. If you got this, good job, you’re qualified, click here.
If you got a different answer or need an explanation, click here.
note: updated to be even more nonsensical to avoid confusion. Thanks, neandrothal.
Josh: For those interested, I wrote a paper for a class 2 years ago applying Bayes Theorum to medicine. It’s moderately related to the post.
I feel rationally obligated to distrust anything self-promoted as “Web 2.0″ in the same way I distrust statistics tacked with some vague quality like “40% more kick!” It’s as if those tacks are nailed right in to the body of Science itself, and I feel its pain. I feel it. Nonetheless, as I can be trusted to regularly grease the world economy $3.29 USD every late night caffeine sortie at the 7/11 across the street, there’s nothing physically impairing me from both believing that savant-like super powers wrought at the heights of all human achievement will give me more energy and help my money work for me while simultaneously loathing the idiocy that same electric blue sans-serif salespitch will sap my attention and wallet if only they could brand their goods with some impressive looking number. You know, like science and stuff. It’s as if behind the decimal point lies a secret realm where mere digits morph into runes of ancient magic, drawing the true, terrible power of tenth decimal place into a furious ball of psychogenic witchcraft burned by marketers into every web service and sugary softdrink for their duplicitous intent of short-circuiting my brain —not impairing rational thought itself, merely its ability to keep my money in my pocket and my faith to myself. It’s a near optimal function to illicit self-disgust.
Hey, I have an idea for a radical new financial instrument, it’s called: the hundredth decimal place. With marketing that tight, I may as well be selling soma to cashiers, right? It’s a brave new world out here on the interwebs, and I’ve got the confidence interval for you right anterior to my magic symbol “%.”
What was I talking about? Oh right.
Bertalan Meskó of Science Roll has officially launched Webicina, a service that sells education material and consulting to doctors about using the Internet and helps them launch a website. Regarding doctor blogs: Dr. Steven Murphy said today about his doctor blog, Gene Sherpas: “all they good things that happened to me this year have come from my blog.” Webicina is a valuable service that I endorse that will help your professional career as a medical professional —and it’s fun, too.
Here’s a funny picture of a dancing little boy wearing decimal point on his shirt. I want you think of it every time somebody uses a decimal point or any other statistic without justifying their significant digits or measurement confidence.
PS: I never want to see the expression “Web 2.0″ ever again, and God help you if I see anybody try to peddle *shutter* “Web 3.0.” To a lesser extent, that goes for that “59.5%” penetrance estimate for g2019s LRRK2 report at 23andMe, too.
Story: genetic councilors are people you pay to console you about the results of a genetic test report to make you feel better and help you make rational decisions about your health when your reasoning may be distorted by emotional distress. If that’s a service you want, buy a couple hours from a genetic councilor.
Computers can never provide human consolation, no matter how excellent and rational their reports. Doctors and scientists feel that councilor work is beneath them, nor are they typically any good at it, nor do they realistically have the capacity to do it. However, scientists tend to be much better at analyzing complex information, and medical doctors can use reports to better synthesize a general understanding of your health to recommend medical action not already prescribed by obvious standard practice.
For example, I would probably make an excellent systems biologist, but the world’s worst genetic councilor. “Beep-bop-boop, you have Huntington’s. Don’t have children.” See? A that’s perfectly rational statement, but it’s illustrative why genetic councilors are important.
But how many people in health care today can read the following expression? Or write it? Or tell me what it means not just clinically, but biologically?
Steve at Gene Sherpas reports systemic medical insurance fraud in many top institutions providing genetic health care. Clinics are billing as if doctors are seeing patients, but only genetic councilors ever see the patient. “But I have a solution,” Steve exclaims. “The answer: Nurse Geneticists.” He continues, “Call your insurer. Demand to be seen by a physician or physician extender… Why should you demand this service? It will put stress on the broken system. To repair that system, we must first rebuild the foundation.”
Be careful what you wish for, Steve. Systemic abuses like this that risk high liability for meager reimbursements suggest a deeper problem than petty greed. Either the system has collectively concluded that genetic councilors are sufficient to perform the service and that they are following procedure “in spirit,” or they are desperate to keep unsustainably high margins and must resort to abuse to protect them. I suspect some of both.
The problem genetic councilors is political. Genetics used to be called “eugenics,” literally, “the science of the well born.” After World War 2 and the civil rights movement, it became taboo to socially direct population growth and the practice of human genetics was legally and institutionally castrated. Today’s “non-directive” genetic councilors are the progeny of this purposefully impotent profession.
However, scientific progressed. We discovered DNA. We learned molecular biology. We sequenced the human genome. Suddenly, the members of this small, marginalized profession were both keepers of the hereditary taboo and keepers of the code of life.
So, that’s what genetic councilors are: specialized nurses who aren’t supposed to touch you or tell you to do anything, but who have become the hands, eyes, voices, and smiles of this newest, most vital paradigm of medicine. So long as people need a warm consult and an authoritative opinion, there will be a place for you.
That’s probably the nicest thing I’ve ever said on this website.
Old medicine was dead the day you adopted “evidence-based medicine.”
There, I feel better.
What medicine? Scientific medicine. Mathematical medicine. Mechanized medicine. Terabytes of cited hyper-linked studies compiled into statistically weighted results medicine. Input symptoms and test results, output diagnosis and CPT medical billing code medicine.
Beep-bop-boop you’re dead, have a nice day. (the industry, the not patient, who of this ethnicity and phenotypic profile can expect 55.2% efficacy and 82.1% efficiency over control who received traditional care)
And no other medicinal discipline is as scientific mechanical as genetics. How mechanized? Watch, I’ll make a free clinical diagnosis machine prototype right here on my blog. Yes, not a risk report, a real diagnosis. Just input your genome sequence service account login, and my machine download your genome securely from your sequence service and practice medicine.
Would you like to subscribe to our genetic counseling service? Now, only $89.95 a month!
Don’t have a sequence? Sequence with us! Now only $999.
Enter your credit card and you gmail account and password below.
The real question is: are you dead like “mama bell” AT&T, the century-old institution that suddenly crumbled and consolidated. Or, are you dead like the newspaper, a slow and humiliating decline of consolation and standards of with only a few prestigious survivors?
So, OK, some genetic councilors will go back to school for additional nursing or physician assistant credentials, and some genetic centers may begin staffing appropriately if they can afford to despite the gross waste in health care already. But realistically, I’m supposed to believe that genetically-trained nurses are the solution? A new, more expensive kind of medical profession for which academic and institutional support will take decades to mature? When America is straining under grossly inflated health care costs during an economic depression? Meanwhile, I can hypothetically write a Python script to practice medicine and hire a perky and well-educated Indian genetic councilor to answer phones now for cheap?
Have fun breaking the system, Dr. Murphy. It needs to be broken, sure, but I’ll meet you down here at the bottom.
Think Gene at Technorati
