I spent years working for a company which performed genetic testing for a variety of conditions, germline and somatic (IE, cancer).
Internally, the results we accumulated were only used for providing diagnoses, estimating population frequencies, and writing papers.
Other companies, however, like Guardant Health and Foundation One, initially provided inexpensive or free tests in order to build enormous patient data stores and were then bought by large pharmaceutical companies primarily for this data.
In the end, the patients were the product, and the cheap/discounted tests were overhead cost in taking advantage of these people. I do not know where these companies are now, but my distaste for the industry was a large part of why I've chosen a different field of research.
Im curious why do you find that distasteful assuming there was a chain of consent somewhere in there. Data security is super important as are chains of unbroken informed consent, for using someones individual data in a study. But pharmas getting to see patient data is not something I have a problem with. If you want new cures for diseases how could you possibly have an issue with that? There is no law really stopping pharmas directly interacting with health providers, instead they choose these third party intermediaries because everyone has this unrealised "distaste" in their minds. What's distasteful about the institutions that produce cures having a clear view of the disease? Long term genetics is not even private data, once its cheap enough and portable enough there will be companies sequencing all humans from just the environment through metagenomics. What if you have a rare variant and I can see it at the local brothel as well as at a work place and at a home in large copy number. The future is Google/Facebook style of genetics regardless of what you think its going to be like. The sort of super strong blanket regulation required to block that would destroy any positive value of the technology. Your super secret data is only super secret right now because of a limit in technology and scope of the market. In 50 years genetic data will be as valued by people as their home address, especially as high value products become available for presenting companies with their data. I really dont have a problem with this. Smart watches are a larger existential threat to say insurance premiums for all of us.
I think the company disguising what business it is actually in is fraudulent and defeats the informed consent. If the transaction is actually to collect my genetic data for a database that will be sold for millions of dollars and eventually used to make billions of dollars selling drugs, then why am I paying a fee at all? I should be paid several hundred dollars for contributing my sample, with the cheap genetic testing thrown in as a bonus.
There's a company out of MIT working on this and using a blockchain as the medium for sharing your genome and medical information. Basically it would allow you to get paid for your DNA and choose to whom you want to sell it. I think there's another company that wants to treat your genome similarly and allow other companies to write apps that use it (same idea but no blockchain). It is kind of a chicken and egg problem though. You want the test to find out if you have something actionable/drug-able so you agree to give up your sample and related IP rights to get the result. The real issue though is that while the technology is developing and we are gathering a lot of information, there is nothing actionable except in a few limited cases.
> The future is Google/Facebook style of genetics regardless of what you think its going to be like.
I would rather die at 35 of a genetic disease than live my entire life in the dystopia you describe. If we need a Butlerian Jihad, so be it. The fact that so many in this thread take the opposite attitude is terrifying to me. As far as I'm concerned, genetic technology represents an existential threat.
Yes, it would be fantastic if we could take the good without the evil, but it doesn't look like we're capable of doing so. Nevertheless the attitude of many is that technology (and not just genetic tech) will march on, and damn the consequences.
Historically, IQ tests were frequently used to justify racism and anti immigrant sentiment. IQ tests are strongly influenced by things like culture, so outgroups typically perform less well on them.
You need to be very skeptical as to the mental models being applied to something you cannot currently change, like your genes.
I have a genetic disorder. I was diagnosed late in life. So I had been managing my condition a long time without a diagnosis and getting a diagnosis empowered me to get a lot healthier. I have been endlessly crapped on for talking about that online because people are sure it cannot possibly be true. This is where identifying someone's genes gets very dangerous because our current mental models frame genes as inescapable destiny and such widespread belief tends to become self fulfilling prophecy by cutting you off from other options.
One night years ago, I dreamed of someone with chains around their neck that they could not remove from that end. The solution in the dream was to break the chains at the other end in order to get free, though it would not entirely free them of the burden of forever carrying these chains around. They at least we're no longer trapped.
I interpreted this dream to mean that the chains are my DNA, which I cannot escape. But I can change my diet and lifestyle. It isn't a cure, but it beats the hell out of what conventional medicine has to offer me.
It is the equivalent of saying "No, thanks, doc. I will keep my nuts and go get bigger undies." And I live in fear of the possibility that someday I shall be told that getting bigger undies is not a valid choice and I must submit to castration for my own good and the good of the larger community. (metaphorically speaking since I don't actually have any nuts)
Flatiron Health was just acquired by Roche for 2 billion dollars. A few years ago Flatiron purchased a company called Altos Systems and as a net result acquired an EMR and patient data for 20%+ of cancer patients in the US. A major part of Flatiron's business was de-identifying patient data and flipping it to pharma companies. They do contribute to some FDA studies and submit papers here-and-there, but the largest sources of income are the EMR and pharma, pharma, pharma.
Are you really comfortable with one company owning so many parts of patient care? I'm not.
> assuming there was a chain of consent somewhere in there
What if a genetic disorder is determined by screening close family (parents, siblings, first cousins)? Could an insurance company reasonably determine that the trait exists within the family, thus block coverage as a "likely" pre-existing condition? Or worse, force a person to be tested if they want coverage?
What are the odds those consent forms will have blanket permissions? Allowing for fishing expeditions.
I understand the motivation to leave a field if the practices are dishonest. I've been there, too.
However, did you note whether or not these efforts may have been positive contributions to the healthcare of humans? I would imagine gathering health data from a large population has the potential to yield remarkable results.
Bingo! You just described 23andme, Anchestry and some others.
Fun fact: you can't make any reasonable profit off of sequencing human DNA at $49 per pop. Its just won't happen. Maybe with technology in iPhone250 in year 3560, but sorry not today.
Back in a day when I was a Noogler, I had interesing conversation on data scraping and purity of infornation in DNA sequencing database. Basically I was explain it will eventually be sold to insurance companies so that they could estimate your premiums better based on probability of your future diseases. As you can imagine that's a gold mine for any large or mid health insurance provider out there. I was also explained how data verofication makes sure of consistency; for example, if male order a kit (via credit card [prepaids dont work bc they cant know your identity]) but result came as female, such result would not be marked as purity and never sold. There are many other checks and balances, but you get the point.
Bottom line: if you want to get the kit and test your DNA out of fun, at least make sure your identity is not obvious.
23andMe costs $199, not $49. I used to do research based on DNA sequencing, and doing DNA extraction and high-volume sequencing costs way under $200 per sample. That being said, I don't know how many replicates 23andMe does per customer. If its a lot, or their prep process is more involved than I'd imagine, its possible their costs are over $200 a sample, but I doubt it.
23andme does not perform sequencing. They do array genotyping, which is far cheaper. They recently abandoned their experimental exome sequencing product, which carried a far higher price.
They do now. They did have a facility for developing sequencing assays, but eventually they gave up. I don’t know how long it will be until they try again. (In fact, they poached a number of our employees for that initiative.)
Which is not really an ironclad guarantee it won't happen (I'm thinking in terms of the law being modified or repealed, not in terms of companies ignoring it).
But the solution to the insurance problem is pretty obviously at the societal level. Once the testing is cheap enough, if they are allowed to, companies will just refuse carriage to people that won't submit the DNA (well, assuming that DNA profiling provides meaningful pricing information).
I don't think repeal of GINA is remotely likely, just look at the political discussion about pre-existing conditions on both sides of the aisle, there is no appetite for having companies accurately price insurance.
I think the fact that health insurance premiums are the #1 thing brought up in privacy discussions is basically evidence of the fact that we all know that insurance is a bad model for healthcare, but we can't bear to socialize it yet.
Should we be socializing the cost of Amazon Prime and life insurance?
Because that's what removing information from the pricing of premiums does, it shifts costs from higher risk individuals to lower risk individuals.
For life insurance, something like the US already does with Social Security (directly providing support payments to minors) probably makes more sense than forcing private insurers to omit information from their pricing. I'm a little less certain about disability and LTC, but those are also already somewhat socialized (Medicaid is the payer of last resort in both situations).
But for life insurance? There's nothing pernicious about someone paying the correct actuarial price for a multi-million dollar contract.
This is a good point and I suppose it’s for society to determine.
Since your genes are unchangeable (currently) it means some people in society will not be at a significant disadvantage if they can’t buy disability insurance or long term care insurance. This has ramifications throughout society.
There’s a lot to be said for paying the correct actuarial price. But I suppose you could also charge different races different prices for their actuarial differences.
So society needs to determine what things they distribute for justice. Social security is one form of this. But it seems unfair that one person with Alzheimer’s will suffer and die in a Medicaid nursing home because they can’t afford long term care insurance during their asymptomatic years while another with Alzheimer’s will use their long term care for appropriate end of life care.
There’s a separate issue of why the poor have bad end of life care. But today the current example takes place because gene testing excludes some from insurance.
>if you want to get the kit and test your DNA out of fun, at least make sure your identity is not obvious.
DNA can be used to identify the person it came from!!!
If you send off your DNA with a different name, there are many ways that they can use to determine that a different name was used. Traits estimated by the DNA could, at least in theory, be used to figure out who you might be. And if your relative is already in the system, your identity may be trivial to uncover.
I worked in insurance. I have a pretty jaded view of it.
I wouldn't hold my breath counting on every insurance company to do right by their policyholders. If the world were so filled with paragons of virtue, we wouldn't need to make laws about it and detail what the consequences are for disobeying them.
> if you want to get the kit and test your DNA out of fun, at least make sure your identity is not obvious.
Once a relative uses the service with their real identity, even if you don't use the service yourself, wouldn't the company know your family (and thus you) have a high chance for X disease or traits?
Short answer is yes, but it depends on the relationship and the disease. For example, if a parent takes the test and is positive for a breast cancer risk genetic variant, then all offspring will have at least a 50% chance of the same result. If, instead of a parent, a grandparent is tested, that percentage drops to 25%.
The cost to sequence has been dropping very rapidly, though. Illumina's 'desktop' scanner claims to be as cheap as $25/sample[1] (and that would be high compared to doing higher volumes on the bigger machines). I would expect there to be hard limit for the reagents etc. but it wouldn't blow my mind if it isn't in the $5-$10 range.
I've never done one of these genetic tests. Partly because it's scary to me to learn what diseases I'm likely to get as I get older, but also I just don't want that information in some else's database. We're very fearful of what Google and Facebook already knows about us now imagine they buy 23andMe and then join all their data about you with your DNA (which you've already willingly given up and allowed them to "profile".
Google/Facebook user behavioral data + their users' DNA is about as scary invasive as it gets.
On the other side of this, if massive databases of DNA and medical histories were accumulated, it's possible a great amount of human suffering and harm could be solved by aggressively mining this data to find out the connection between DNA and various diseases.
I mean, besides being denied health insurance because of what someone might find in your DNA, why are you so afraid of something even having your DNA, because it might result in GATACA-like society? If so, we need strong legal protections against discrimination by DNA, but we can't throw out the baby with the bathwater.
There's too much good that can be done to human wealth and welfare by eradicating disease and paranoia about theoretical danger should be challenged into laws to protect against discrimination.
What do you think Facebook + DNA is going to do to harm you? Try to sell you a pharmaceutical or medical treatment? A cure for balding or erectile disfunction?
I can imagine "look at me, my genes are superb!" having a slightly different effect on potential dating partners' level of interest to the one intended...
Great. Maybe we will have less cystic fibrosis in the world. Not sure many people would argue they would be keen for their children to inherit that. There is a big difference between knowing you naturally are a carrier of a disease, and acting on that info, and then manipulating radically and arbitrarily the genetics of humanity. Regulation internationally is fairly clear, sifting out incredibly debilitating disease is going to be ok. Making your children more aryan looking, not so much. Now if only they posted a recent STI panel result too.
You'd be fine with someone not finding a partner because they have a chance that their child would carry a disease? Something like "I cannot find a husband because my child will have a 5% chance of cystic fibrosis, so nobody will go for me".
That isn't remotely how that works. There is no circumstance whatsoever where your child has a 5 percent chance of having CF.
CF is a homozygous recessive disorder. Both parents have to contribute a defective gene in order for the child to have CF. If one parent is a carrier and the other is not, there is zero chance of the child having CF, barring random mutations which have been known to happen. If both parents are carriers, there is a 50 percent chance the child will be a carrier, a 25 percent chance the child will have no defective genes and a 25 percent chance the child will have CF.
If you actually have CF, you have two defective genes and you can only produce children who are either carriers or who have CF, depending on the genes contributed by the other parent. If the other parent is not a carrier, the children cannot have CF.
As someone who both has CF and has raised a child with it, I absolutely do not want more children with CF. It's a horrible burden for the parents and only a deeply sick sadist would wish such a thing upon their child.
I have never lacked for male interest. I don't know why the hell that is because I'm quite open about my medical situation. I am now menopausal, so it is a moot point. But I absolutely spent some years agonizing over how to screen out CF carriers from the dating pool. Having another child with CF is one of my worst nightmares. I would rather be taken out and shot.
Uh, no. All other heritable diseases are either homozygous recessive or they are passed via a single dominant gene. In neither case is there a 5 percent chance of passing it on. All recessive disorders follow the exact same math outlined above.
And I am guessing that's not a case where they have identified specific genes. Please correct me if I am wrong. But I am not aware of bipolar genes being a thing.
There is a difference between hand wavy "runs in the family, but we don't really know how it works" and a bona fide genetic disorder whose alleles have been mapped to some degree or another.
Edit: To be perfectly clear, if it is not genetic and merely tends to run in families, it may be due to a variety of nongenetic factors, such as diet and lifestyle. Smoking also tends to run in families. Smoking is not a genetic disorder, though it significantly impacts health.
If it is not genetic, it's heritability is not going to show up on your 23 and me profile. Fretting that it will basically means you don't know what you are talking about.
There is a disease that was thought to be genetic, but was ultimately traced to ritual cannibalism spreading an infection. Relatives would eat the dead and what body part you got depended upon how closely related you were. This caused the disease to very much mimic a genetic disorder in its pattern of dispersal throughout the family.
"Runs in families" needs to be very clearly distinguished from genetic disorders per se. The former suggests correlation with an as yet unidentified cause or causes, as there may be various contributing factors. The latter is a case of identified cause.
Easy enough to confuse if your own life does not depend upon it, but absolutely not the same thing.
No, but you are being intentionally insulting in violation of HN guidelines and also moving the goal posts. I already stated above my diagnosis appears to be no barrier whatsoever to getting romantically involved with someone. Except from my end, as I have chosen to be celibate for medical reasons for a lot of years.
I am 100 percent okay with someone like me being able to screen out carriers. Because I have CF, it is entirely up to the father's genetic contribution as to the outcome. I am also 100 percent okay with having some means to have a potentially awkward conversation up front.
There are, in fact, worse things in life than sleeping alone.
I think the OP you responded to is being intentionally obtuse, you have made your point clear repeatedly. If you are aware you have CF gene and are also aware your partner does too, then you will not mate with them and instead find another mate. Some diseases can be eliminated like this.
I also have a condition similar to you and there is no way I'm letting another child suffer with that, so I'm not going to ever mate with a woman who carries that gene.
I think the value in knowing upfront is you just avoid getting attached to begin with and you go bond with someone who doesn't introduce such big scary questions to the relationship. It is much easier to walk away before you get attached, no hard feelings. It is only when you fall in love first, have already invested substantial time into the relationship and then find out you both carry genes for something horrifying that it winds up being an agonizing question.
I understand the negatives of having everyone compare genes can get kinda "master race"-ish but at the same time it is still heart breaking for kids to be born with stuff that could be prevented.
For people really heart-set on having children, I think it would be important to know your odds ahead of time before you get to the point of "ok lets have a kid together".
I was 26 or thereabouts (it's been a while), in good health, but had one "touch base" appointment in my prior 2 year medical history (what the prospective insurer wanted, at that time). The appointment had no ongoing medical/treatment implications; it was simply a follow-up with a prior provider. Fortunately, a friend of a friend ran my application "unofficially" by the insurer's underwriters; if it had been a formal application, the result would have ended up in a cross-insurer database of decisions made whether to insure. Yes, there actually is -- was, certainly, at that time -- one, and if you landed in it with a "deny", getting anyone else to insure you became an order of magnitude more difficult. So I was told by this industry professional -- again, as a favor to my professional contact and friend.
More recently, I've had a minor condition that an eminent surgeon refused to operate on. Cost/benefit didn't merit it; his recommendation was simply to monitor.
When I changed from employer provided insurance to individually insured, I was denied. There was nothing I could do, and the condition did not hinder me in any way. Nonetheless.
Again, through the fortune of having a (different) professional contact and friend in the industry, with an allied company, I was able -- via a flaw in their processing of the application -- to get a re-evaluation. And a policy written, albeit with a rider excluding that condition.
But their initial reaction, before I "pulled strings"? They didn't offer me a policy with an exclusion rider. Nope, just "deny".
SO, many people who are paranoid about this kind of thing and "their health data getting around", are for damned good reason.
I agree, the knowledge and health benefits of sharing this data could be enormous. Think of what it could do for epidemiology. Best practices. Etc. Etc.
But, you want me to share it? You want me to sign on to a public program of same. You're going to have to guarantee that the data won't be used against participants.
As in, you do so use it, and you personally go to jail with a felony conviction and a multi-year sentence. And your company gets severely penalized and, if the behavior is widespread and with significant negative outcomes for those discriminated against, its own "death penalty".
You want the data? Demonstrate your good will. In ironclad terms.
I could also live with a system like Germany's, where insurers are private but the state requires equal treatment and effective systems of cost control.
No system's perfect, and every system needs to be effectively managed.
I've long since come to the point of wondering how much the U.S. behavior actually negatively impacts other countries' health care systems.
In many ways more indirectly that this following, but also explicitly in Canada, they keep pushing for increased privatization. Want that marketshare, and to be the middleman between patients and services.
In Germany very few insurers are private. Most are state-nonprofit, private is split into most being private-nonprofit and the tiniest part being private-forprofit.
I think the good part is that we have a very good state healthcare, it's non-optional for basically 99.9% of the population.
The sad thing is that our actual healthcare system (doctors and hospitals) are in need of workers and doctors which leads to them preferring the privately insured patients.
The policy I got was with a good company (I mean, once you actually had a policy), for a reasonable rate. They then exited the individual market in my state. I ended up being shunted to a lower benefit, higher cost policy with a "second tier" company (if that). My premium costs about tripled in the span of a couple of years.
The Affordable Care Act's full enactment came just in time, for me.
Unfortunately, all legislation consists of two essential components: 1) The law, itself; and 2) funding (the budget process).
Republicans baldly stated, after Obama's election, that their primary, number one objective was to make him a one-term president. Before all else.
Despite its coming out of what was essentially a Republican design, then Governor Romney's health care insurance program in Massachusetts, Republican's went after the Affordable Care Act -- after labeling it "Obamacare" -- with a vengeance.
A primary way they crippled it, was by not providing the funding that was written into the law.
Insurers were given a 2 year time frame in which they could recoup their losses, until new participants' "deferred care" was taken care of and the companies had better demographics and cost projections (actuarial studies) upon which to carry forward.
When they went to the government to be made whole, under this provision, they were paid about 15 cents on the dollar.
So of course, premium costs increased dramatically, plan options became more limited, and companies started to exit the ACA marketplace.
Meanwhile, the Republicans made endless attempts to "repeal" the ACA in one fashion or another. Claiming they had a "better plan", that would provide better coverage at lower cost.
Now, they've had both houses of Congress and the executive for over a year. Still no sign of an actual "plan". Just continuing efforts to destroy the ACA.
So, going forward? No, I don't think this problem of pre-conditions can be considered solved. The party in power keeps trying to reintroduce it.
And I, for one, have no more trust in our society. Maybe I'll find another position that includes an employer-provided group insurance plan. Or a group plan through a professional organization.
If I didn't have various things tying me here right now, including elderly parents, and I had the opportunity, I'd leave the country.
I've always lived modestly. And it seems that is even part of my problem. The U.S., always somewhat hypocritical, is becoming more and more a land of "winners" and "losers".
Just look at our "fearless" leader, using the "L" word -- quite derogatorily -- left and right. If we're supposed to take our cue from our leadership, well then, f-ck this place.
Republicans used the ACA as a wedge issue to motivate their base. Driving up premium costs was them hitting the hammer. The press failed to hold them to account. There was a big backlash though in some town halls where they would get called out so much so that they stopped attending town halls! Insurers want rescission and pre-existing conditions because it is a very profitable business when you don't have to pay out claims. I agree that we are lacking in trust as a society, we only trust the almighty dollar and maybe our local communities. The real problem is that people keep voting republican because they refuse to think. But that excludes the massive voter suppression they engage in in their red states as well. You can see that if people vote most states will turn blue. We need more engagement.
>Despite its coming out of what was essentially a Republican design, then Governor Romney's health care insurance program in Massachusetts,
Calling anything that comes out of MA "republican design" is a little less than truthful.
Regardless of how you feel about the merit of the MA system or the federal system there are no republicans in MA, just democrats who will say what they need to say to run against incumbent Democrats.
We already have so much data in the healthcare system that people are unable to use or hamstrung by regulation to use (and to be fair it is probably very difficult to make sense of..). One doctor has no idea what another doctor wrote in my chart. Hospitals don't share data most data with others or with scientists for research.
When medical research even begins to use the massive amount of data they already have on patients from billions of routine visits, tests and hospitalizations productively, we can think about giving them more. We are so laughably far from that point that adding more data to the mountain doesn't make sense.
Why couldn't we have our cake and eat it too? Have anonymised data used for research that could never be traced back to you. A government agency could be in charge for that (like they are for censuses for instance). I'd sooner give that type of information to my democratically elected government than some corporation overseas.
> What do you think Facebook + DNA is going to do to harm you?
If you're actually looking for a pie in the sky answer, what happens when one if these entities gets popped and someone (state actor?) takes regional data to produce a very effective targeted bioweapon?
But the real answer is: you tell us when you find out the hard way. In the meantime, my personal info, of any kind, will continue to be protected, to the very best of my ability.
While I agree that throwing babies or bath water out is wasteful...
The primary goal of business is to make money. Or, in other words, to isolate us and make us feel inadequate so we buy more things with money we don’t have to impress friends we don’t have.
I mean, if I was being cynical. I’m not, but if I was...
I did 23andMe many years ago around when it first came out (and I therefore have access to certain results others can’t get since this was pre-FDA settlement).
At the time, I ignored the fact that it said I was 4x more likely to develop colon cancer (I was in early 20s). My sister was diagnosed with colon cancer 3 years ago. When you are a woman diagnosed in your early 30s, apparently it is due to genetics. She discovered it by going to three different doctors after insisting she wasn’t feeling well and that it felt “deeply internal”. The doctor who suggested a colonoscopy did so because he had an internship for a year at a cancer clinic in Boston. The others thought it was in her head. Well they found a malicious tumor and the prognosis from the nations leading cancer clinic was to remove her colon, her uterus and ovaries and large intestines (in their entirety). You can imagine this is a lot to ask of anyone. We asked for data that supported the conclusion that if we don’t remove all of these parts the cancer would spread to her ovaries and kill her and there was literally none to be had. It was 100% based on the doctors personal experience (again we were at one of the top clinics in the world for this).
There is something structurally off with the modern medical system. I believe it relies too heavily on anecdotal experience and memory. I don’t know about you but I can’t remember what I ate yesterday. I think services like 23andMe are on to something, if only as a data point to empower patients to self advocate. We did nothing with the data but he moment I got the phone call I couldn’t help but wonder if we could have found this earlier.
> We asked for data that supported the conclusion. . . and there was literally none to be had. It was 100% based on the doctor's personal experience
On a semi-related note, I've had a few conversations with physicians about the utility of AI in medicine, and all of them seem to be adamantly against any technology that displaces the individual doctor as dictator/hero of the medical hierarchy.
In one memorable exchange with a surgeon (and friend), I pointed out that doctors, like most other educated humans, are pretty terrible at applying statistics to any kind of complex decision-making - and I brought up your observation that doctors are limited to the knowledge/anecdotes/heuristics that fit in their own heads. Why not employ some kind of machine intelligence to at least aid in diagnosis, double-check treatment plans, prevent medical errors, etc? My friend abruptly changed the subject and began waxing poetic about how his decades of experience allowed him to glean enormous amounts of information from the mere tactile sensations of a loop of bowel in his hand during surgery. How would a machine replicate that, he asked. The conversation wasn't even about robotics.
I don't have a lot of faith that doctors - outside of a few pioneers - will begin using medical expert systems until they are forced to.
You bring up a great point. I think a lot of people, not just surgeons, would trust their own experience and intuition over AI (at least initially). Not to mention the whole human characteristic of resisting pretty much any sort of change (even if it's for the better).
I think AI in medicine can be great, especially if you start using it to augment a physician by keeping them up to date with the most recent research/papers/literature on a given topic.
I've had this conversation with doctors quite a bit. The missing link, in my opinion, is that the mind is an incredible tool only made better with good data. I don't think you or anyone is advocating to replace doctors with a TI-86, we just want data driven / supported decision making and I just can't figure out why we can't bridge that little gap.
I have a spreadsheet that I actively log the total amount of time I spend with my doctors and their staff when I see them in person. All I am going to say is god bless nurse practitioners.
Before, lots of data was based on billing codes, which are a total mess. With electronic medical records, now they can finally collate the data in a meaningful way. Dr. Atul Butte is doing that at UCSF for the entire UC system, but it just didn't exist before a year ago. I just saw a presentation of his a few weeks ago, and he's digging into stuff like "why are there hundreds of treatment paths in the first 12 months for patients diagnosed with T2D at UC hospitals?"
Thank you for sharing these details. It's helpful to know details about aggressive therapeutic regimens doctors might suggest.
Excuse me if I am intruding and do please disregard my question if you prefer not to answer.
I'm presuming your sister did not follow the therapeutic regimen recommended to her. If this is the case, would you mind sharing what treatment she did pursue and whether she is currently in remission?
In any case, I hope your sister is well as she can be. I hope also that you and the rest of your family remain healthy despite your genetic predisposition.
I appreciate the kind words and don't mind chatting about it at all. My sister taught me that it is much better to talk about these things as it could save someone's life someday. She did take the treatment and the cancer is in remission.
I am happy to chat about this and she is as well to the extent you wish to go straight to the source. After her recovery, she organized a group of young mom's that meet regularly to help each other cope with cancer (and what she considers to be the more onerous issue - the thought of not seeing your kids grow up).
I’ve thought since consumer-level genetic testing first came out that there would be a place for a company to do it fully, truly anonymized using tor+crypto or something similar. It’s a bit surprising to me that it doesn’t exist yet; if I had the right background I’d probably go for it myself. There has to be a decent number of people that want to know, but don’t want anyone else to.
Unfortunately, there is limited interest in genome privacy. There is also a BIG interest in getting people to surrender their genomes.
""But its for the good of humanity, so don't worry about any possible downside.""
Personally, I am trying to develop molecular cryptography, so that genetic data can be protected even if, theoretically, we cannot trust our computers.
It would have to be something to the effect of getting the kit in a store and sending it in with a public key and no return address. They check that the associated account has paid, run whatever analysis and add the encrypted results to the account. The major bottleneck is probably getting the kits to the user; workarounds I can think of that don’t involve having your own kit available in stores would be either having a list of generic kits that would work or piggybacking off of 23andme’s distribution and accepting their kits (though I’m sure they wouldn’t be happy about that, to say the least).
The Alice company would either put DNA collection kits for sale on Amazon or tell people where they can buy a kit that meets the required criteria for collection.
Then people pay the company in bitcoin (how to separate your bitcoins from your identity is left as a exercise for the user) and once they've paid they're provided an input field to enter a public key they've generated.
The website would then generate a barcode you'd use to label your sample before mailing it in, or it could be a QR code of your public key.
When a sample arrives at the lab they scan the barcode, and check to ensure that there is a public key tied to that bar code and that they've paid.
They then run the sample and publish the results online, accessible to all, but also encrypted with that users public key.
The user then checks the website every day for week after they've sent their sample in. Once they find their results by searching all results for the one labeled with their public key they download the results and then use their private key to unencrypt them.
I did 23andme back before they had to stop doing some of the medical stuff. They discovered a serious and potentially harmful mutation that I had. I showed my GP who then confirmed it with other tests. It's something that could kill me, and knowing about it allows me to avoid situations where it could.
The ancestry stuff is really really cool too. I'm 2.7% neanderthal! My whole family discovered aspects of our history that we didn't expect.
It really is scary. They claim to anonymise the data, but to me it feels like removing your name and address from your DNA is like removing the make and model from a car schematic (where only one model was produced).
I wanted to know. It turns out I’m a carrier for cystic fibrosis, and my wife is not. If we were both carriers, we might have chosen to adopt children rather than have our own. So there is more to learn than just your future health. It’s also interesting seeing who my distant cousins are.
It's worth noting that's not a selling point for many Americans.
From incomprehensible billing and having to fight for coverage with health insurance, to dealing with antiquated offices (fax? really?), never mind the hassle of scheduling an appointment over the phone, going in to an office, and seeing a doctor who's rushing you out the door, "normal healthcare avenues" is cumbersome, and persists only due to lack of real competition in the marketplace. (The doctor, of which there are really truly awesome ones, is a cog in a far larger machine.)
Still, the technology that allows us to screen a fetus at 12-weeks of pregnancy is quite impressive.
Of special concern would be entities like FB and LinkedIn having thins kind of data _and_ availing it to private health insurance as well as potential future private sector employers.
The insurance argument for genetics is incredibly over blown. Very few people have some large quantifiable risk with high certainty found from their genetics. Something like a smart watch mobile phone fitness app that records you are clinically obese and dont move around is far more of a problem than your genetics will ever be. Just because you have the genetic potential for higher or lower risk is nothing compared to measuring the actual state of your person right this second and over your lifetime. I work for a company that records and uses both simultaneously. Trust me when I say your smart watch with a heart rate monitor is probably about 1000x the existential threat to your insurance premium. The genetics is helpful for parameterising models and adjusting you from say an "average" human. But there is nothing like actually measuring daily who you are. The power of that data is actually incredible.
Aren't these PCR tests against known sequences of base pairs? It's not like they're sequencing your DNA or anything, they are just testing for the presence of selected markers. I don't think the data they have is meaningful outside the context of the tests.
I feel I'm in an interesting scenario in that both of my parents have done 23andme, but I have all the concerns you have.
You're not just making this choice for yourself, you're making it (to some degree) for everyone you are or will ever be related to. That has profound implications that people aren't entirely considering when they spend the $50 (or whatever) and spit in a cup.
This doesn't test the full gene sequence, only a few SNPs. Say someone does the test, and it comes back positive. They then call all their relatives. Those relatives have the choice of getting the same test done with 23andMe, or seeing their doctor. Usually proper BRCA gene tests are only reimbursed under certain circumstances, so mostly the relative will have to pay at least $500 to get tested properly. Most of them will just go to 23andMe. The growth in testing because of this effect is going to be significant. All the hard work of doing this kind of testing eg counselling, detailed family pedigrees, contacting family members, discussion of pros/cons of testing etc is being put back onto the individual and their healthcare providers. Good for 23andMe I guess, especially since their value seems tied to how much data they have. Whether this is net beneficial to society I'm not so sure.
"23andMe demonstrated a high level of accuracy (greater than 99% concordance to Sanger sequencing) and precision (demonstrated by studies yielding greater than 99% reproducibility and repeatability)"
Sanger sequencing is pretty accurate, on the order of 99.9%. So if 23andMe is matching those at over 99% accuracy and precision, a full sequence may not be necessary. The FDA took a long time to approve this, so I'd imagine it passes their standards as well.
If you are thinking of macro-societal impact... I think you have to ignore the specifics of this particular test and think of the impact of these sorts of test on medicine, culturally.
For example, there's a sort of Principe in medicine that you shouldn't order tests unless you will action treatments based on results. For example, if you detect a risk of breast cancer that can only be avoided by mamosectamy, only do the test if you're willing to to the procedure. Otherwise, all you have is a more stressed out patient. Data gathering is totally different in medicine for such reasons.
The big, relevant exception is cancer, where a statistical, risk estimating approach is commonplace.
In any case, genetic testing (and other ambient data gathering) may give us good reasons to change this approach. It's a transition and taking a step from toy to medicine (this is what 23andme is doing) is en route to such a change. What could emerge downstream is personalized cancer screening protocols. Maybe you should start getting colonoscopies at age 20 and I should start at 60.
There is definitely a science of designing and performing useful diagnostic tests. I have seen many cases of illogical or misinformed testing that has harmed patients physically, psychologically and financially. I ultimately believe that more information is better, however, so far the transition to more direct consumer testing is failing miserably (eg Theranos, or all the companies claiming SNPs can help design dverything ftom skin care to exercise regimens). Then there are just crackpots who coerce vulnerable patients into expensive and totally unproven, never-to-be validated tests that cost thousands of dollars.
Nearby variants are highly correlated due to their spatial proximity, a phenomenon known as linkage disequilibrium. Therefore a dense array of SNP genotypes is often sufficient to make statistically accurate inferences about the portion of the genome you do not directly assay. I myself have used this strategy to estimate my risk for possessing the ApoE variants from my 23andme callset. Is this 100% accurate? Of course not; no medical diagnostic is. But it is accurate enough to be used as a screening procedure in a lot of cases.
If you start off with an inaccurate screening test, you waste a lot of resources with follow up testing, not to mention the psychological impact with BRCA tests.
...for the few SNPs that they actually test. A negative 23 test does not exclude BRCA mutation. To apply link disequilibrium as you brought up doesn't change that.
Your 1st point is something like "these tests are useful for screening, you can get information about unassayed regions using statistical genetics approaches". My point is that this is not useful, since screening tests need to have certain properties not provided by SNP tests. The biggest issue is that 23andMe do not test the whole gene. This is like only doing a mammogram on one tenth of only the left breast, obviously a futile screening endeavour even if the FNR for that 10th is gold standard level. But most of all, unlike other screening tests which are designed to help the patient as much as possible (within societal resource constraints), 23andMe are offering a very limited test because that is what suits them, that is all that is possible with the SNP arrays around which they have built their business model. My original post was highlighting that unlike what is suggested on the gushing press release, this is not actually a great moment for direct to consumer genetic testing. If they were really interested in helping screen for BRCA mutations, 23andMe could swap out some of the other genes which tell you trivial stuff like how fast you metabolise caffeine, or some rare drug metabolism effect and include more BRCA SNPs. But they don't, because it is all about how many tests they sell and how useful the information is to pharmaceutical companies.
I haven't read through the T&Cs at 23andMe, but can this data be cross referred to (bought by) by insurance companies to quote me an insurance price depending on the outcome of 23andMe analysis ?
Or even by mortgage providers for that matter ? Or potential employers ?
They probably will partner up with insurance company’s and will provide discounts if you take this test as part of your physical. And increase it if you don’t or start at a higher price for all products that the client doesn’t take the test. Yeah it’s gonna come to that
> They probably will partner up with insurance company’s and will provide discounts if you take this test as part of your physical. And increase it if you don’t or start at a higher price for all products that the client doesn’t take the test. Yeah it’s gonna come to that
Even if they could do that [0], once you (and presumably they) get the results they can't discriminate based on them due to the Genetic Information Nondiscrimination Act of 2008, so there is very little incentive for them to pressure you to take the test.
[0] They can't, AFAICT, also because of the GNA, though employer-sponsored insurance might be able to back door something similar through a voluntary employer-based wellness program nominally separate from insurance.
Insurance companies already _can_ do this today, but the practice is uncommon or nonexistent. Why do you think that 23andMe being granted FDA authorization would be the final tipping point to begin this behavior?
If the current FCC approved something, I wouldn't assume that it was approved on its merits. Is the FDA better? An honest question; I don't know much about its current performance or composition, but it's under the same political control as the FCC.
I don't know much about the FCC other than reading about net neutrality, but my answer would be probably yes.
For all its faults, FDA is pretty good all things considered, especially considering the lack of resources they have to do their job. The reviewers are often helpful and the conversations are very scientifically focused and objective. There are many controls within FDA to prevent corruption or outsiders to influence policy.
That said, ive been involved with some decisions where FDA has clearly made questionable decisions that have set back innovation for no good reason. There are also many lobbyists pleading their case to FDA, but congress has very limited power to influence in most cases other than by writing laws (which i think is as it should be)
Many companies that complain about fDA simply havent done their homework or are pushing shoddy science
While I respect the FDAs intentions and notable positive actions, I would like to present several articles from reputable sources describing the FDA as a "revolving door" for various industries - an obvious inlet for corruption, dubious research, delayed actions, etc. As I understand it, this is a well known issue.
I believe it is important to be vigilant and critical when it comes to three letter acronyms.
That is fair and skepticism is warranted. A lot of those links point to one study by vinay prasad, who has a reputation as something of an alarmist. Just want to make sure people don't take that study to definitively claim a trend where none may exisr
While you may be right about Vinay Prasad being an alarmist (I have no idea), the second article is actually based on a bill written by 7 senators with (seemingly) no relation to Vinay's work.
The FDA is pretty excellent. You’ll find all sorts of criticism online, but they are doing a remarkable job. Just remember they routinely have to stand in the way between seriously ill patients and the supposed “miracle cure” when you read anything online.
Unfortunately, it seems that you're out of luck if you're outside of the US, but you do get the raw data regardless.
Is there a tool which can help analyze this raw data? If not, I think it would be beneficial to create one, but I imagine this might be hard. Does anyone know how difficult it is to programatically test the genome, as 23andMe provides it, for various genetic disease risks? Is research on this easily available? Would a person need a significant know-how in the field, or can this be done by someone after a reasonable amount of reading on the subject?
geneticgenie.org analyzes your 23andme data. You can upload raw data or just link your 23andme with oauth. It might be more narrow in scope than what you had in mind since it is primarily used to generate a methylation and detox profile. I used it primarily to determine if I had a MTFHR mutation but it was helpful since it showed other potential issues with my methylation system.
I don't know how hard it would be to programmatically do, but if geneticgenie is any proof, it can be done.
I wonder how this stacks up vs other BRCA tests in terms of performance and coverage of common variants. And also whether it is cheaper to use their DTC test rather than going to a doctor and getting insurance to cover the test? I assume this is not a sequencing based test?
It's a SNP chip and only covers a few common mutations.
To anyone who wants to know for sure whether you're at increased risk of breast cancer, you'll want to get a test that sequences all the known mutations, along with another half-dozen or so genes that can have essentially the same effect as a damaging mutation in BRCA. I'd also urge you to find a genetic counselor, who can help explain the diagnosis in detail, as well as the limitations of the each kind of test.
Additionally, many harmful mutations are essentially "private" within a family and do not inform any of the GWASs used to create these SNAP chips. Regional sequencing is still important for persons with a family history of illness.
What I did a while back is download the raw data and run it through https://promethease.com/. Will tell you a lot more about risks, drug interactions. What it does is it compares your data to that of all the journal papers and research around that gene. Interestingly I have a SNP that is correlated with lack of empathy and another one that is the "warrior" gene. The BRCA SNPs would be identified.
I'd urge you to be cautious when interpreting those data. The effect size of most of these SNPs is tiny, often explaining (at most) only very small amounts of the variability observed between groups. This is especially true for things like personality traits, which have a huge environmental component.
So, if finding out the empathy thing is fun, then by all means, keep poking around in your data! Promethease is probably the best of the bunch, and I've run my data through it. Just be mindful of the evidence level and effect size!
Yes, this is a good note for anyone interested in doing this. Other things to keep in mind, the paper source, while may be peer reviewed, most likely was not reproduced (a problem in the industry in general), so would be prudent to filter on # of publications supporting the SNP correlation. Also, the report will tell you relative risk, not absolute (e.g. 1.5x increased risk of getting breast cancer). In other words, your absolute risk may be 0.00001, so it can easily be 8x and it won't really matter. The most actionable items I've found for me personally were drug interactions due to genetics. Thanks for bringing up the caveats.
l'll second that, drug drug interactions and genetics is an under learned topic in the UK, I know that because my GP borrowed my book, 'Clinical Manual of Drug Interaction Principles for Medical Practice' :)
Hmm. How long before health insurers use this to, uh, "personalize" rates? Note that many have wellness programs that incent employees to present blood samples.
Internally, the results we accumulated were only used for providing diagnoses, estimating population frequencies, and writing papers.
Other companies, however, like Guardant Health and Foundation One, initially provided inexpensive or free tests in order to build enormous patient data stores and were then bought by large pharmaceutical companies primarily for this data.
In the end, the patients were the product, and the cheap/discounted tests were overhead cost in taking advantage of these people. I do not know where these companies are now, but my distaste for the industry was a large part of why I've chosen a different field of research.