but evidence-based medicine only matters insofar as the evidence is complete and unbiased

The data is never complete, as medicine continues to evolve. Bias will also be an issue-- if it isn't a profit motive, it will be a career motive. It's just an important point for the lay public (I'm including you, Ezra) on what EBM does and doesn't provide-- clearer information on what works, and an expectation that physicians use the best information in making decisions. But it isn't complete, it isn't unbiased, it isn't infallible-- the medical debates will still continue.

Which leads to... while there are fair criticisms on the bias of data presented in industry-sponsored trials, the focus on drug companies by many demonstrates a tremendous bias against them. Compared to all other interventions in medicine, drugs currently have the best data available supporting their use-- as required by the FDA regulation. Medical devices have utilized what was supposed to be an uncommonly-used approval pathway, 510k, which requires significantly less data than we expect from drugs-- but instead has become the routine choice for devices. Therapeutic interventions, such as surgeries, have even worse data on average, than drugs or devices. So when we're speaking about what we need to do to improve data-- I'm pretty suspect of those that focus on drug companies. Its primarily from people that don't know the level of data currently available in all facets of health care, who have allowed the political demagoguing of the drug industry to cloud their judgment as to what's most important. Any reasonable assessment of existing evidence gaps would spend very little time focusing on drug companies right now. It isn't even close.

"Nemeroff's behavior is par for the course."

Isn't he pretty much a disgrace in the field now?

Another key point-- its important to distinguish between trials that lead to FDA approved uses, versus those that enable off-label use by physicians. I don't think we have a policy today that makes sense-- we require vast amounts of data for companies to get a drug on market for a specific indication-- trials that typically cost a couple hundred million dollars. But once on market, docs can use the drug for any disease that they see fit-- no matter how little information exists to support the use. How does that make sense?

Angell's criticism is focused on the off-label use, that's where the conflicts-of-interest issues she raises are most pressing. That's where you get multiple, smaller studies, with only certain ones published. The larger, primary studies for the approved indications aren't the issue, most of the time. So rather than jumping to a policy recommendation of nationalizing yet another part of health care (do Democrats have any solutions in health care besides government taking over things?), let's focus on the FDA approval process and perhaps think creatively about how drugs should be approved and used based on available data. There are some very good ideas out there, but it starts with not jumping to the conclusion that it isn't fixable and government needs to take it over and allow a more nuanced discussion to take place.

Thanks for the cogent comments, wisewon.

wisewon: how would you propose dealing with off-label use, given that it occurs in the context of classic, DTC-led prescription creep? Seeing Abilify advertised during SNL as a kind of turbo boost for an existing SSRI, having previously heard about it being pushed to children, reminded me of when I first saw the branded pens a few years back and was told that 'oh, you won't have heard of it -- it's for severe psychosis and schizophrenia'. Well, not any more.

If you institute restrictions, you'll just get diagnosis creep to ensure that the people who ask for their As Sold On TV pills get them. And as Angell suggests, the psych establishment is in the middle of rewriting the DSM to ensure that the next generation of happy pills have a new range of "disorders" to treat.

Fuck nuance: these are institutionalised drug-pushers.

I know somebody who used to work in a facility that did research on psychiatric drugs. My info is therefore second hand, but from somebody who did this for a living. As far as I know he's an honest guy. He sure as heck isn't rich.

There are two problems that come up when researching antidepressants. The first is "placebo responders". There is an entire group of people who will get better if you just tell them they are supposed to get better. It's hard to week them out of the research pool. And considering that many of these people are here because they can't afford meds without the research paying, is it ethical to turn them away?

The second problem is that for most people, just asking how things are going for them can make them feel better. Since any decent research of an antidepressant will attempt to measure how they are feeling, you end up with a Shroedinger's Cat problem. If you know how they feel now, you have just changed how they will feel in the future.

There are people whose lives are saved by this medication. The fact that these meds are over-prescribed and hard to research is not a reason to throw it all away.

These scientists have to make a living too.

Are you just jealous you aren't being paid millions to promote a corporate message?

I'm reminded of the huge battle that had to be undertaken to get the makers of Paxil to even begin to acknowledge that their product had some dangerous side-effects, both in use and in the cutting down and stopping stages.

wisewon is a shill for a number of things, but it always amuses me how he reserves the most enthusiasm for shilling for drug companies.

The picture on evidence is even worse than painted in the article. There was a journal article out of Cambridge UK - I'll try and dig it up later - about the choice of experimental conditions. Do you test against the existing drug for the disease, or against a placebo? Why might a company choose one or the other? And there are a whole host of other variables carefully manipulated.

And does this mean that various drugs don't work at all? No... but it very much affects the price that can be charged for them...

I'm working on a startup (CureHunter Inc), which works to quantify scientific research results. Our Research Interface product enables users to practice evidence based medicine and check the evidence in real-clinical-time.

Of course pharma funded, potentially biased, research gets blended in the results, but using timeline graphs you can see the focus of research and trends changing over time. Even drug company funded spikes in research disappear over time if they were not really scientifically viable targets.

My bigger concern is pathways that never get explored just because they aren't patentable or are prohibitively expensive to bring to market.

Hopefully technical advancements will make it easier and easier to develop new drugs. Right now though, it seems like we are still discovering more negative effects than positive.

Please give the system a try if you get the chance:
http://curehunter.com

(sorry for the blatant plug, but it seemed on topic)

Why not set up a system of "devil's advocates" - a formal part of the Catholic system for recognizing saints? Every study would be published with hostile criticism. At the moment, it's pretty random whether your research gets really dug into: nice guys finish first.

wisewon: how would you propose dealing with off-label use, given that it occurs in the context of classic, DTC-led prescription creep?

There are concepts around "staged approvals" based on levels of evidence available, and what would be allowed in terms of prescribing and promotion. This could allow us to get away from the on-label/off-label issue, and move towards the more helpful dialogue for physicians, patients and society-- the level of evidence a drug has for a certain disease compared to other therapeutic options. But you've directly touched on a key element of staged approvals-- DTC restrictions. A staged approach to approval may not allow any DTC promotion for an early approval. With a sufficient body of evidence, those restrictions could be loosened by the FDA. That said, there is a big need to reform DTC generally. But back to the on/off label point-- we should all be asking why physicians have the ability to prescribe any drug to any patient without any evidence burden. Its another example of the degree of autonomy that we give physicians that has been distorted over the past few decades.

PS Health IT is a key piece of the puzzle here as well. If we have linked medical records, the data that would be available on off-label use would grow dramatically, and we could make better decisions on the effectiveness of drugs across a number of diseases pretty quickly.

Of the 170 contributors to the most recent edition of the American Psychiatric Association's Diagnostic and Statistical Manual of Mental Disorders (DSM), ninety-five had financial ties to drug companies, including all of the contributors to the sections on mood disorders and schizophrenia.

It is much worse than even that. The contributors had to sign a non-disclosure agreement, that forbade them publicly discussing the behind-closed-doors deliberations about the contents of this document.

Completely unacceptable.

The DSM is, rightly or wrongly, one of the most influential documents in existence, and is claimed to be science based. There is no good reason for its deliberations to be covert, and every good reason for them to be completely transparent and open to public scrutiny and criticism.

@ShortWoman"There are two problems that come up when researching antidepressants. The first is "placebo responders". There is an entire group of people who will get better if you just tell them they are supposed to get better. It's hard to week them out of the research pool. And considering that many of these people are here because they can't afford meds without the research paying, is it ethical to turn them away?"

Why the heck would you want to weed these people out of the research pool? If there's a group of people who will get better by being given a placebo, let's find out who they are and give them placebos in real practice. And if the expensive, serious-side-effect-causing stuff works only slightly better than a placebo, then why are we giving it to people?

wisewon (as usual) makes a lot of the relevant points. I'd add this: psych drugs as a class are probably the big issue, for the reasons everyone cites... but that goes to the nature of psych and psych treatments more generally. Remember, prior to happy pills, we had great ideas like electric shock therapy. There's just a lot about treating depression, modd disorders, and schizophrenia that we've made up as we;ve gone along. People dealing in addiction therapy will tell you, outside of pushing people into 12 step programs... they don't have much. So some of this criticism, including criticism of the way DSM is compiled, really have to do with how much we're feeling our way around on mental health, even now. That some drugs provide some small measure of relief for a few people... let's not kick the idea out entirely. And let's keep in mind that it's much easier to quantify "reduces likelihood of cardiac events" than it is "reduces effects of depression," which is why other classes of drugs don't have quite the issues psych drugs have.

(These issues around psych, by the way, also have everything to do with the issues of getting proper mental health care treatment and, indeed, other rehab and long term care services.It's the very fact that results are so hard to quantify that makes insurers leery and likely to set absurd limits on metal health treatment and things like occupational therapy after a stroke. It's much easier to quantify heart surgery.)

Second, as wisewon explains, you will always have research biases, and you have to, at some point, accept that when good people say "I looked at this data objectively" that some of them did. I worked on pharma based medical education for a number of uears, and a couple of things were readily apparent: how to separate bad research from good, and how to separate the greedy docs from the talent. There are good doctors who recommend products because they believe in them, and beause they work. The only option most of us have, as an alternative is to start reading the studies ourselves. And frankly... that's not such a bad idea; I made my doc change my allergy med prescription when I could demonstrate, to him, that the side effect I described was in fact clearly indicated in the research literature.

There's a sense here that we expect medical information to be delivered by angels of purity and goodness, so that docs can be magic makers with perfect skill to solve every problem. That's not going to happen, now or ever, and it's a terribly passive approach for patients getting care. Terrible especially because what we need, most of all to reform health care, is getting patients far more active in how they think about and approach their own care. Drug companies have problems... but they also make life saving products that have changed medicine, often for the better. There has to be balance in looking at health care, because the good actors, as we see them, are not all good, and the bad actors are not all bad. But most of all, our unrealistic expectations can lull us into a passivity about health care being given to us. And that's a bad paradigm.

If we don't want health researchers taking money from drug companies to do research, then we are going to have to rethink the way we fund science research in this country. If you instantly disqualify everyone who has ever done any kind of drug research from taking some kind of leadership role, you are going to be left with a very small group of professionals who are going to be far less qualified. Drugs are a huge part of how disease is treated in the modern world. Unless you would prefer that the government take over all drug research and development (and that is certainly a possibility which comes with its own pluses and minuses), health professionals are going to have to work with drug companies to develop new and better treatments.

We don't want professionals being unduly influenced by drug companies, but marking people who have worked to create better treatments as "dirty" is not going to help matters either.

Wow, I just looked at that anti-depressant review. I'm not an expert on how the FDA conducts its analyses, but this:

A shortcoming in the FDA data is the absence in many of the reports of reported standard deviations

is shocking to me. How is it acceptable not to report standard deviations for outcome variables? That is like one of the first things you look at in an analysis. That is messed up.

Remember, prior to happy pills, we had great ideas like electric shock therapy.

ECT is still commonly used and is pretty safe and effective. Also, "happy pills" is an insulting way to describe medications that treat actual illness. There's enough stigma about mental health care already, knowledgeable people don't need to add to it.

I was anon @1:42 last night.

But back to the on/off label point-- we should all be asking why physicians have the ability to prescribe any drug to any patient without any evidence burden.

Do we really need to ask? In the US, physicians have that ability because people want the pill they saw on TV. If they can't get it from one doctor, they'll go to another, or use the friends-and-family route to find someone willing to write a scrip.

(Right now, there can be more hurdles if you want OTC pseudoephedrine than if you want an atypical antipsychotic for a child. It's a class thing.)

As weboy says, that's partly because the non-pharmaceutical provision for mental health is woeful, but it's also out of a general culture of quick-fix medicine.

So, who keeps watch? Pharmacies? Insurers? State licensing boards?

AMS: to my mind, marketing an anti-psychotic as a condiment for anti-depressants (SSRI's taste better with Abilify ketchup!) is far more insulting than the term "happy pills".

AMS: to my mind, marketing an anti-psychotic as a condiment for anti-depressants (SSRI's taste better with Abilify ketchup!) is far more insulting than the term "happy pills".

Why? I've never seen the Abilify ads, but using adjunctive therapy when monotherapy fails is common.

AMS: to my mind, marketing an anti-psychotic as a condiment for anti-depressants (SSRI's taste better with Abilify ketchup!) is far more insulting than the term "happy pills".

Why? I've never seen the Abilify ads, but using adjunctive therapy when monotherapy fails is common.

...yes, but as has been pointed out here by a number of folks, we don't actually have a lot of good evidence that Abilify as an adjunct really does all that much. Abilify's ads - which aren't, I think, as bad as they're made out here, but still - almost admit as much. It's a "they're kind of helpful, if you don't suffer some of the more dangerous side effects, and you think your current pill isn't really doing it." It's so full of qualifiers, I'd be hard pressed to know why it's needed... except that, for a depressed patient who feels things aren't getting better, it's a small raft to cling to.

Second, just to respond to a few folks: I don't use the term "happy pills" lightly. Aside from my general skepticism about antidepressants as a solution for depression, what we know is that Celexa, Prozac and several other drugs have curious efficacy data and for many patients, placebos would work as well; the SSRIs are prescribed, often, as a cheap, easy alternative to talk therapy and other forms of mental treatment that are hard, time consuming, and difficult, as I said, to quantify in hard results. Most of the drugs in question give a feeling of mild euphoria... and that's about it. And the drug companies have worked tirelessly to create - in that DSM book we've been mentioning - classifications of "diseases" that need their treatment: try to suss out what "social anxiety disorder" amounts to (and it's not agoraphobia, which has its own designation), and why it's needed except to provide more... um, "happy pills". We medicalize vague feelings of discomfort or anxiety... and then tell people that these pills are the only answer.

Look, I take mental illness very seriously - indeed, I think this country does a terrible job of stigmatizing and misunderstanding mental illness, and what we can do about it (which mainly has been to criminalize it and put treatment into the jail system - remember why the Virginia Tech massacre happened?). But having a serious concern for mental health in this country, it seems to me, would be challenging the hold psychotropics currently have on the way we approach treatment. So yes, I said "happy pills" and I meant it.

And PS, Med Student, I know EST is a valid therapy...as are lobotomies; I was mostly pointing out that when they were the only options, they were overused and misused, with disastrous consequences, mostly because practitioners were groping for solutions. We're still groping, even if we've been able to refine the use of shock therapy and brain surgery; it's still the case that a lot of how to treat mental illness is still a best guess, not necessarily clearly understood or proved. If we're going to address mental health issues better in this country, it just strikes me we need to face up to the complexities and challenges we face.

Ezra,
You make points that have been made many times before. Beyond just regurgitating facts about anti-depressant trials that you can find on Pharmalot or multiple other places, why don't you talk to the researchers, find out why trials sometimes turn out negative or sometimes show only a small difference between drug and placebo?

Especially with drugs that treat diseases with no objective measurements (such as depression or psychosis), placebo response is going to be a killer. What the nay-sayers don't tell you is that the survival curves over a long period of time show that patients on active drug do better. Placebo may work for six weeks while the patient is seeing the psych every week, but in the real world, where patients need meds for months and months and don't see their docs every week, placebo fails much more frequently than active drug.

Trials can be negative for many reasons--not always having to do with efficacy. Sometimes a company picks the wrong dose or the wrong length of time. Sometimes they get the patient population wrong. Just because a trial is "negative" doesn't mean a drug doesn't work.

The proof of the pudding is in the real world. Ask someone who has benefited from an SSRI and who has had the experience of NOT being on one. They'll tell you there's a difference between drug and placebo--and no trial will convince them otherwise.

Negative trials also give us important information about variations in the way different sets of people respond to drugs. We know from the SSRI trials that some people may be genetically predisposed to be responders--and some not. We don't have a mechanism to sort those people out on a wide scale yet, but at least we know now that there may be genetically-based differences.

And I don' think you want the government deciding if your disease is important enough to warrant research. I'd rather have a guy with a profit-motive working on the cure for my disease than a government bureaucrat using utilitarianism to decide my care.

And as for off-label use...

Drugs are chemicals with discernible mechanisms of action. Doctors are scientists who also know the chemical and electrical activity of the body. If a drug works by doing X and is indicated for disease Y, then a doctor, who knows that X also happens in disease W, may decide to try the drug in disease W.

Thus you get the atypical antipsychotics being used for refractory depression. These drugs bump serotonin and dopamine levels in different ways in different parts of the brain--primarily the frontal lobe and limbic system--which doctors know seems to help with depression. The drugs don't have terrible side effects when used for short times--why not try them for a patient who is miserably depressed and not responding fully to any other treatment.

Saying, "Oh that drug is for PSYCHOSIS," makes it sound like a nuclear bomb or something. It's just a chemical that causes chemical changes in the brain. Depressed brains and psychotic brains (and anxious brains and bipolar brains and OCD brains) all fall together in a sort of Venn-diagram like existence. Drugs that work for one may very well work for others. It's not malpractice to know that and try it out--it's good medicine and what we pay them a lot of money to figure out.

I don't know the answers to many of the questions posed.
As a cancer patient, I do know that off-label use of approved drugs extended my life. The first was in 2001, when I began using herceptin a breast cacner drug that DOES work for 10% of ovarian cancer patients. The second was in 2006, when I used avastin, a colon cancer drug, that had published studies showing good response in ovarian cancer women.
The third is a clinical trial drug (yes ,paid for by the drug manufacturer) that has worked for the last 8 months.

Oh course I want credible data, I can't afford to waste time trying ineffective drugs, but I'm out of options in the area of proven therapies, so I want to see more research, sooner!! 20,000 women die every year from ovarian cancer, most of which is advanced by the time its
found. Cancer patients LIVES depend on research expanding and being studied quickly.

One of the more interesting developments we should see over the next few years will be the role of genetic testing for both risk stratification and efficacy. Pharma resisted this at first because they wanted to sell to as many people as possible, but they now see it as crucial to ensuring that the millions they sink into a new MAB or cytotoxic drug have the best chance of profit.

While personalized medicine is still a long way off, even in the early stages we are seeing examples where patient populations who will never respond are being excluded. The math for the drug companies is becoming more compelling. It is better to identify genetic markers that correlate to efficacy and run trials in those patient populations. You get a better chance of success, potentially more significant differences with control group and you make insurers and government health plans more likely to justify the cost of your drug.

It's easy to paint pharmaceutical companies as bad guys, but they have to work within the system the government has created. I develop new drugs. I have seen a novel, very useful drug delayed for years due to trying to tiptoe throught the regulatory minefield. We watched as a competitor slipped up on a clinical study design which had a small flaw that ended up causing placebo to appear almost as effective as the tested drug, and causing the trial to be deemed a failure. Their funding dried up and they had to cancel the program. The company is on life support. We saw their mistakes and are able to learn and design a better trial, but who knows if we fail to factor in some critical tweak? You would not believe how small issues can sink promising drugs that can save tens of thousands of lives. Even the indication is a crap shoot. Due to regulatory burdens, you end up having to go after the small indication with the clearest and best chance of a positive outcome, and you design the trial for the best chance of success, not failure. Otherwise, you risk your whole company on a massively expensive and difficult clinical trial where anything less than complete success will kill the program. Don't blame the doctors for knowing what the true indication should be. We just didn't have the time and money to go after the bigger, more serious indication until we showed some data on a much smaller, less important indication, and received regulatory approval. We do no good to anyone by gambling ourselves out of business. Try not to look at everything in terms of good guy/bad guy. We are trying to develop safe and effective drugs. The whole system is so risk-averse that it's a wonder anything new ever comes out of it.

Two points. There is now a clinical trials registry which makes sure that the use of the FOIA will never be necessary again. The problem existed but it has been solved.

Second the statistical analysis of Kirsch et al is invalid -- just wrong. They did a crazy thing which reduced the estimated effect of anti-depressants and was roughly as important as publication bias (my last effort to comment included links and did not appear. Just google "kirsch" at my site and you will be able to click to the proof of these claims).

The level of "clinical significance" is completely arbitrary and has nothing to do with costs and benefits.

The publication bias and the gross methodological errors introduced by K et al. shifted the estimated effect from slightly above the arbitrary level to slightly below.

Kirsch et al included a pharmaceutical -- Nefazadone which is significantly less effective than the others and which is not approved in Europe anymore.

They presented no evidence of publication bias in the estimated effects of Prozac (another of the 4 drugs). The remaining 2 have estimated effects greater than the arbitrary "clinical significance" line based on their data.

The change of HRSD of placebo treated patients is not an estimate of the placebo effect -- it could be the normal course of untreated depression. The 80% figure is based on a complete misunderstanding of the point of controlled trials which is not and can not be to estimate the magnitude of the placebo effect.

Finally recall that Kirsch et all reported overwhelming rejection of the null that the 4 drugs they evaluated are no better than placebo.

In sum, there was a problem with publication bias which has been solved by mandatory publication. The rest is a mix of errors in statistical analysis included imposition of the assumption that 4 drugs are the same even though it was rejected by the data and an arbitrary definition of clinical significance.

Queen1: My point about atypical antipsychotics is that they have made the jump from use in heavily-controlled settings -- involuntary confinement or intensive outpatient treatment -- to the world of ad hoc off-label prescriptions and fuzzy DTC marketing in a relatively short timeframe.

There's a huge culture of self-medication in the US, and there's also a genuine blurring between 'ask your doctor about X' and enabled self-medication.

Doctors are scientists who also know the chemical and electrical activity of the body.

We can argue about whether doctors are really "scientists": they do scientific things, but doctoring has always been sui generis. Putting that aside, doctors are also businesspeople who know that refusing psychiatric meds can be bad business.

In short, I'm not so optimistic about the capacity of prescribing physicians

I'd rather have a guy with a profit-motive working on the cure for my disease than a government bureaucrat using utilitarianism to decide my care.

And I'd rather have a wider on changing the culture of American healthcare than engage with false dichotomies.

From Queen1:
And I don' think you want the government deciding if your disease is important enough to warrant research. I'd rather have a guy with a profit-motive working on the cure for my disease than a government bureaucrat using utilitarianism to decide my care.

So, just to clarify, you think it's just fine that an order of magnitude more research money is spent on hair loss (or erectile disfunction or whatever) drugs than is spent on research on, say, new classes of antibiotics, AIDS vaccines, et cetera?

And that more money is spent on advertising said drugs than is spent on researching them?

My God. If drug research were run by the US government with George W Bush in power, it could not possibly waste more resources and target its money more poorly than the system we have now.

-fred

Not credible to WHOM? To you? Why should anyone care if they're credible to YOU? You're not a doctor. Not credible to doctors, then? Based on what? Do doctors complain that they don't know what to believe? I don't hear that complaint, no matter what Marcia Angell says.

A review of seventy-four clinical trials of antidepressants found that thirty-seven of thirty-eight positive studies were published. Of the thirty-six negative studies, thirty-three were either not published or published in a form that implied a positive outcome

So your problem is more about not publishing bad data than it is about good data being corrupted? The literature is not the only source of information for doctors. They do work on these trials, you know...

I'm guessing you feel that the only way to get complete and unbiased data is for the government to run these trials? Right, like the government is unbiased. Additionally, you seem to clearly lack any practicial knowledge about just how much influence the FDA has in how trials are run, and how the information from these trials is used. That is to say: alot.

I work in this field as a biostatistician in an academic setting in partnership with the pharma industry.

We are frequently confronted with this dilemma of apparent or real conflict of interest.

One of the proposal that seem reasonable is to follow the European model and for a sort of consortium of industries to pool funds in a common account and then conduct research. Minor issues such as Credit still need to be sorted out.

This system is working pretty well in some European countries

Thanks, it is very good, I like it very much wow gold

So... meh. this blog post is the first I've heard about the issue, and based on it and reading the Politico article I'd say it's the kind of thing to ask questions about (if Calderone didn't already know that there are a lot of off-the-record list servs around DC like you say, which come to think of it would be a big thing to be ignorant of), but the answers aren't newsworthy. Smoke but no fire.
I just hate seeing "leftists" act like William F Buckley and the National fucking Review. This is why the left has gone to shit, too many damn rich kids and their summer-camp habits brought into the political arena, instead of actual organizing.

Laptop Batteries
Laptop Batteries discount laptop battery

http://www.llvgame.com/Maple-Story-Mesos/
http://www.mmo4s.com/product/Runes_of_Magic__Gold.html