Off Label Prescribing as Good as On Label

Once a drug has been approved for some use it may be legally prescribed for any use. New uses for old drugs are discovered quite often so off-label uses can be very different from FDA approved uses. Mitomycin, for example, was approved to treat stomach and pancreatic cancer but is used off-label in laser-eye surgery. Drugs prescribed off-label have not been through FDA-approved efficacy trials for the off-label use. In Assessing the FDA via the Anomaly of Off-Label Drug Prescribing I pointed out that off-label prescribing, therefore, gives us a window onto a world with much less FDA regulation.

Since off-label prescribing is common and in rapidly progressing areas of medicine often the gold-standard, I argued that the behavior of physicians validated off-label prescribing and demonstrated that physicians were willing and able to draw upon non-FDA sources of information to make rational prescribing decisions. Dan Klein and I also showed that physicians are supportive of off-label prescribing saying, for example, that it would be “crazy” to require FDA approval for off-label uses.

The support of physicians for off-label prescribing is telling but not dispositive. Perhaps physicians make hubristic mistakes in prescribing off-label. A new paper by Ladanie et al. (including John Ioannidis) provides important information. The authors search the literature for all the RCTs when an off-label drug was pitted against an on-label drug. They conclude:

Our meta-epidemiological analysis of 25 different treatment indications for off-label drug use
provides no empirical evidence supporting any assumption of generally inferior treatment
effects associated with off-label use. On the contrary, the summary effect estimates across all
indications would even be compatible with more favorable effects, on average, of the off-label
treatment. However, the heterogeneity is substantial and the on-label comparators are not
necessarily the best approved treatment option in all 25 topics. While some off-label
treatments are clearly better, others are clearly not.

The finding is especially impressive because although off-label treatments are sometimes the gold standard they are also often used when standard treatments have failed. Thus, in an RCT, off-label treatments could be worse on average and yet still provide a very useful weapon in the medical armory.

One might argue that if off-label treatments are as good as FDA-approved treatments then the FDA should have higher standards. FDA required clinical trials, however, already cost hundreds of millions of dollars and years of effort, creating drug lag and drug loss. Rather than condemning the FDA, what these results indicate is that the medical system–physicians, hospitals, insurers, scientists–does a good job at evaluating new uses for old drugs. As Dan Klein and I noted in our precis on off-label prescribing:

The off-label experience testifies to the fact that much knowledge
about efficacy and safety is produced outside the FDA regulatory
apparatus. The Pharmacopoeia’s recognition of off-label
indications years ahead of the FDA demonstrates that physicians
and scientists have certified thousands of drug indications quite
independently of the FDA, even when those indications are not
very closely related to the original indications. In addition to the
Pharmacopoeia, there are several other forms of professional certification,
including the American Hospital Formulary Service Drug
Information, HMO formularies, and a wide
array of specialist professional periodicals
and information services. NIH studies,
clinical results and determinations
from other countries, and other professional,
science-based judgments are
examples of nongovernmental, non-mandatory

Hat tip: Michelle Dawson.


'Drugs prescribed off-label have not been through FDA-approved efficacy trials for the off-label use.'

Which just might explain why a drug manufacturer advertising an off label use is considered to have broken a regulatory framework in place for more than a century.

'One might argue that if off-label treatments are as good as FDA-approved treatments then the FDA should have higher standards. '

But not Prof. Tabarrok, Bartley J. Madden Chair in Economics at the Mercatus Center, one can safely assume. One could however argue that drug companies are more interested in profit than anything else, and being able to get around the current framework would definitely pay off for the bottom line. As pointed out in the next sentence, if with more credulity - or through a carefully nurtured perspective - by Prof. Tabarrok.

And when will we finally be able to compare laetrile with those other, presumably much less effective in the eyes of those involved in promoting laetrile, cancer treatments?

Umm, but letrile is not used "off label" because it's never been recognize as a safe and effective treatment for anything.

One can surely see that the supplements business has become a medicine show, but that's different than off-label as the use of supplements is not gated by physicians.

The question is, can we trust physicians to use consistently good judgement when prescribing off-label? In some cases the answer is obviously "yes," but in others physicians (due perhaps to laziness and/or greed) seem unduly influenced by pharma salespeople, and under-influenced by published medical science.

Off-label use can be a threat as well as a boon to pharmaceutical companies. For example, Revatio is approved to treat cardiovascular disease, yet it is chemically identical to Viagra. Since it sells at a much lower price, prescribing Revatio off-label can threatens profit earned by Viagra.

Albigensian writes, "The question is, can we trust physicians to use consistently good judgement when prescribing off-label? In some cases the answer is obviously “yes,” but in others physicians (due perhaps to laziness and/or greed) seem unduly influenced by pharma salespeople, and under-influenced by published medical science."

We can't even trust physicians to use good judgement when prescribing on-label; witness the huge opioid crisis.

Yup. And it's not just the physician's fault. It's that we truly don't know much at all about most off-label uses. Typically the only research that most docs will see on a new off-label use is selectively published from a study funded by a company that stands to make money from it.

And usually the research is being shoved in the doc's face by a drug rep bearing sandwiches for the whole office staff (which is a nice perk for those folks that docs don't want to take away) and reminding the physician that he saw his prescribing of Company XXX's drug was down a bit last month and is there anything he can help clear up for him?

The second paragraph here is patently wrong.

There are strong (and enforced) regulations that prohibit Sales Reps from speaking at all about Off-Label Use, much less promoting it to doctors. Doctors CAN request data from the company related to Off-Label Use, including speaking to a Medical Science professional about the data, but this is strictly done by request of the physician (unsolicited).

The two functions must remain separate (Commercial/Sales and Medical Science) and the latter can have no relationship whatsoever with prescriber volumes, even to the level of being aware of them.

Finally, the "sandwiches" are also largely a thing of the past. The "Sunshine Act" required doctors offices to report any "gifts" of any level of value, whether it's a reprint of a published article or a club sandwich for the office admin.

What you describe would have been the case in the era of Lipitor and Viagra, in the early 2000s, but snce then regulations and compliance have actually gone a long way towards limiting the worst of it.

Yes, Erik, are you aware of how many billions in settlements there have been because of systemic off-label drug promotion by the largest drug makers? It is a cost of doing business and it definitely happens.

Drug reps actually are permitted to share off-label info and it is perfectly legal. Here is the FDA guidance on exactly how drug reps can share information on unapproved uses that have had no public health agency vetting whatsoever: The drug rep just technically has to be asked, which is basically impossible for regulators to track.

Your "Medical Science Professionals" are simply drug reps, no matter what name the industry adopts for them.

Meals are most definitely not a thing of the past. The Sunshine Act required disclosure but in no way limits meals to docs or their staff. In fact, companies made $2.8 billion in "general payments" to docs last year, most of which are meals--totally separate from research payments.

In fact, evidence from the Sunshine Act database points to the strong link between accepting meals from drug companies and prescribing practices.

So, unfortunately, this is the current landscape. Not much has changed.

This doesn't say anything about FDA regulation for safety though, as these drugs have already been tested to show that they are safe.

Safe only at the dosage and indication that the drug was approved for. Doctors can prescribe such drugs for new indications (off label) and at dosages either above or below what is recommended in the FDA approved drug label. Safety for these "new" indications is unknown until sufficient use has been observed and adverse drug reactions monitored.

'Safe only at the dosage and indication that the drug was approved for. '

From a certain perspective, sure, the studies have limits. But there is absolutely no actual regulation saying what the legally mandated dosage is - that is what the term 'malpractice' and its legal framework is designed to cover to a large degree.

It's hard to imagine that doctors would prescribe uses far outside the proven-safe dosage. That's one reason they prescribe the drugs for off-label uses - they know the safe dosage limits.

LOL, there are lots of examples of drugs with very narrow therapueutic and safety ranges having caused adverse drug reactions when prescribed outside the FDA approved dosage.

@clockwork_prior - the FDA approved drug label that has regulatory authority does say what the dosage is. Whether companies can be sued for adverse reactions when drugs are prescribed outside of this range is another question entirely. Case law usually has sided with the pharma company in such cases.

I thought a physician had every right to double an approved dosage - such as an antibiotic - when confronted with a situation where the alternative of maybe having a patient die was definitely having a patient die.

Of course, there are also medications which are lethal at higher than prescribed dosages, such as paracetamol (and by clearly saying that, drug companies avoid liability when some ignorant person destroys their liver). My thought was more in the antibiotic example than in the paracetamol one - however, whether an antibiotic or paracetamol, it is not the drug company prescribing the medicine, but the doctor, and it is the patient that does or does not follow instructions. Obviously, there is an interplay between various actors.

And from wikipedia - 'Paracetamol toxicity is the foremost cause of acute liver failure in the Western world and accounts for most drug overdoses in the United States, the United Kingdom, Australia, and New Zealand.' Something that the FDA actually tried to deal with to an extent -

Sorta. There are a lot of well-established off-label uses for drugs that most physicians are aware of and many make use of regularly.

There's also a lot of doctors who *think* they know how certain patients will react to certain drugs because they had a patient (or heard about a patient from someone they trust) that reacted in a certain way. Human everywhere tend to over-generalize from small samples, and physicians are no different. In fact, the self-confidence (arrogance?) that's almost required to be a physician who makes life-or-death decisions daily probably makes them more susceptible to this type of reasoning. (IMO.)

These are competing challenges that aren't easy to address with one-dimensional solutions like, "The FDA should be more lax when approving off-label usage of drugs!" What's needed (IMO) are better data collection methods and proper analysis to show which types of off-label uses work well and which don't. The paper Alex cites is trying to do this, though they're taking a fairly broad view and pointing out the need to better study off-label drug usage given how successful some types of off-label prescription have been.

Freedom always works better than government. All uses should be off-label and the FDA should be abolished. It's a corrupt agency that primarily regulates speech and exists only because the judiciary doesn't enforce the First Amendment.

Just like clockwork, yet another blog post by Alex on off label prescribing. FDA already has a streamlined process for getting new indications on the drug label. If Alex is so sure that off label prescribing is so good why do we still continue to see companies sanctioned? Maybe Neurontin is actually good for all the indications that Pfizer was fined for promoting (sarcasm; and of course it wasn't that it was Pfizer who were responsible for the over promotion but Warner-Lambert who Pfizer acquired).

so exactly "why do we need the FDA" & and why do FDA bureaucrats have so much arbitrary power over American citizens and medical practice ?

So that people such as you are not at risk when they take unapproved drugs.

so all "unapproved" drugs are an automatic risk to humans on this planet... and only FDA bureaucrats possess the knowledge, skills, and motivation to properly evaluate all drugs and eliminate such risk ?

Would I trust you or a scientist whose job it is to protect me from unsafe drugs.

When you use the word bureaucrat you are just using the word scientist, investigator, statistician; so quit using loaded words and begin thinking.

That's my unapproved prescription.

FDA employs over 15,000 personnel, with the largest percentage of them working in the Office of Regulatory Affairs. Most in FDA are paper-pushers and watchers, not serious scientific doers.

And drug companies tend to employ a lot more sales reps than researchers. Your point being?


Policeman are watchers too. So are FBI agents.

And, you need to get your idiotic and unthinking buzzwords up to date.

It is not paper pushers anymore.

It is electron pushers.

'so all “unapproved” drugs are an automatic risk to humans on this planet'

Of course not. But laetrile being sold, even if unapproved, is the sort of thing that people who intend to make money selling laetrile definitely agree on. And to be honest, they have zero concern whether you die from cancer as long as they get their money up front.

Dennis C misleadingly writes, "FDA employs over 15,000 personnel, with the largest percentage of them working in the Office of Regulatory Affairs. Most in FDA are paper-pushers and watchers, not serious scientific doers." I don't know what the current employment in ORA is (been retired from the pharma industry for 7 years now) but their activities range over a lot of areas and are not focused solely on enforcing off label prescribing issues. They have responsibility for assuring that data submitted to the FDA in registration filings is compliant and also they monitor manufacturing facilities for Good Manufacturing Practices (GMP) both pre- and post-approval. I'm sure you want assurance that a sterile drug product is actually sterile prior to it being injected. These folks are not just paper pushers but have sophisticated training.

The historical perspective here is that although FDA has existed since 1906, its authority was significantly expanded after thalidomide was prescribed to pregnant women and the subsequent birth defects horrified the public.

So, a historically-informed answer might be, "The FDA has vast regulatory power because the public wishes to avoid another thalidomide disaster."

Yet a rational person recognizes that practically everything has costs as well as benefits, and that the natural tendency of bureaucrats is to avoid risk, yet risk avoidance inevitably creates opportunity costs.

A helpful summary, Mr A.

Milton Friedman....something...something...Beta Blocker...mumbles...

The FDA has grown much larger than it was after the thalidomide disaster. It wasn't regulating speech in the 1960's as it is today.

I work in this field (Medical Affairs Strategy). There is actually a spectrum between "FDA approved" and "Off-Label Use". (Not from a regulatory and compliance perspective, of course, but in relation to the course of medical science).

Leading physicians in their fields often want to pursue independent research, and they may apply to pharma companies for free product, funding, or both. The protocols for these studies are extensively reviewed, including by the FDA and EMA.

Promising results from one of these small studies spread, and on the back of that scientific evidence, more physicians may begin following the studied use when there are no other options. Others might apply for a follow-up study with a higher patient population or with a more specific protocol to examine the strengths and weaknesses exposed through the first pilot study.

Eventually there may be enough "real world data" from off-label use to compile a cohesive analysis at scale which may then lead the company into investing into a full-scale clinical trial to pursue FDA approval.

All in all, medical science is a staircase, and off-label use, although it should never be PROMOTED by the company, is an important piece of that.

As an aside, almost all products initially gain only very limited regulatory approval. Only for a specific subset of patients, or in "moderate to severe" X, or only after first-line treatment has failed. Off-label use often pushes the boundaries of these tight restrictions and again, over time, a case can be built using real world evidence as a guide, and in fact sometimes the FDA will now even accept real world evidence as long as the populations are correct and the analysis is sound enough. This is how, over time, the approved indication may expand.

Off-label uses should always be promoted when they're the best option. The FDA should never regulate speech. Trump should drain the swamp but he's not bright enough to understand what's happening.

"[I]f off-label treatments are as good as FDA-approved treatments then" the physicians prescribing the drugs may be guessing as to both the patients' condition and the best treatment for that condition. But at least it's only a drug, rather than surgery or radiation:

The title of this post is absurd. This is a conclusion based off 25 off-labels usages that were chosen specifically to be included in meta-analyses. Why were they chosen from the population of all off-label prescription?

Also, "There was substantial statistical heterogeneity across comparisons (I2=43%)," so maybe you shouldn't try to generalize but instead focus on the particular off-label usages?

I generally think that it should be easier to get off-label usages approved (especially when you start talking about subgroups and so-called "personalized medicine"), but misleading posts like this just muddy the waters.

Note that I'm just objecting to Alex's framing. The study itself looks interesting.

You say the title of the post is absurd and you object to "Alex's framing" yet the title of the article under discussion is

"Off-label treatments were not consistently better or worse than approved drug treatments in randomized trials"

i.e. off label as good as on-label, so Alex just shortened the title.

"not consistently" is key. Some are better, some are worse. Alex's title is categorical.

Just because two groups are on average equivalent (or one group is on average better or worse than the other) doesn't mean they're element-wise equivalent. It's a faulty thought process that people engage in all the time. They consider some summary of two groups and do a comparison of those summaries. Based on those conclusions, they generalize to every element in the group. It's a convenient short cut and often a useful heuristic when there isn't time to compare things element-wise, but when it comes to prescribing drugs (and lots of other things in our modern world), there's no reason to be so hasty or rely on generalizations.

AlexT has a habit of setting up strawmen and knocking them down, so I am suspect of this post. AlexT cites somebody who says: "The off-label experience testifies to the fact that much knowledge about efficacy and safety is produced outside the FDA regulatory apparatus".

But, if you deconstruct this sentence, it assumes the FDA regulatory apparatus produces knowledge about efficacy and safety, when in fact all the FDA does is review such efficacy and safety. In other words, it's not the FDA's mandate to produce knowledge about efficacy and safety, but to evaluate such knowledge. Indeed, since most FDA material is not public, but trade secret information between the FDA applicant and the FDA, it's probably true that the FDA produces very little public knowledge on efficacy and safety except ex post approval of a new drug, when it doesn't do much good at all. Which leads to the inquiry: what exactly is AlexT's point? Except perhaps to highlight the FDA is a simple delaying or stalling organization set up by the US government to prevent another thalidomide disaster, mainly by delaying drug approval until human guinea pigs outside the USA can use new drugs and alert the USA as to their danger. Presumably AlexT wants US medical patients to elect to become human guinea pigs as well, a reasonable libertarian stance I guess.

Bonus trivia: Thalidomide has off-label uses in fighting Hansen's disease (leprosy).

Given that there are still instances of harmful drugs being approved and used even *with* the full FDA clinical trial rigmarole, it seems hard to swallow (as it were) that the use of drugs *without* these procedures would be somehow better-informed. My immediate (perhaps cynical) guess is that while off-label uses don't benefit from FDA trials, their patents have often already expired, and therefore don't "benefit" from pharma salespeople pushing doctors to prescribe them for all possible cases. If this were true, the FDA is just a sober counterbalance to "Big Pharma" influence over doctors' judgement (and Big Pharma came about because of the prodigious muscle required to get a new drug through FDA testing- a bad equilibrium!) Looking at the differences in effectiveness between off-label use of patented and post-patent drugs would indicate if this were true.

Another possibility is that off-label uses only typically get discovered if the drug is just *wonderfully* effective for this use. Perhaps also, that it's good for a condition that (for reasons of perceived potential market or otherwise) no direct novel drug research is devoted to. Think of the story of Viagra, which was initially a drug against heart disease; it only got recognized as a erectile dysfunction drug because it had an effect which immediately stood out (as it were) to patients. I would imagine that in the pre-Viagra days, very little research was done on sexual dysfunction drugs- how could that be more popular than a drug that saves people from heart attacks?- so it likewise would have been unlikely to be discovered on purpose. This one I'm not sure how or whether it could be falsified.

While I generally support the continued utilization of off-label prescribing of drugs, analysis and referenced study really do not support this:
1) There is massive selection bias in the study. Not all off-label uses are studied and even fewer generate enough RCTs to be included in systematic reviews. The Ladanie paper only evaluated systematic reviews. It's been well established that there are publication biases toward studying and publishing positive results. And, as Erik indicates, much of this research is supported financially by pharmaceutical companies. My own experience is that even when the study design and execution are truly independent from the sponsor, industry sponsors only want to sponsor studies likely to produce favorable results. The total effect is that a literature review of systematic reviews of off-label vs. on-label drugs is virtually assured to result in showing good performance for off-label drugs. It's likely that the total health impact of off-label prescribing of drugs not included in the Ladanie paper is less positive, perhaps even harmful.
2) there are plenty of counterexamples. To pick just one topical example, fast-acting, sublingual (absorbed under the tongue) opioids are only indicated for breakthrough cancer pain, but have been wildly overprescribed for off-label uses with disastrous health and economic effect.
3) Saying off-label prescribing is safe is not the same as saying that FDA regulation should be loosened. Adam correctly points out that FDA has already proven these drugs to be safe at a specified dosage. It's not clear that, if drugs were approved without the same scrutiny, that clinicians would substitute rigorous safety analysis. My gut is that clinicians are more likely to correctly evaluate the efficacy of an approved off-label use of a medication than to evaluate the safety of an unapproved or more laxly approved drug.

The critical point is that a drug found to be safe and effective for one use is unlikely to be unsafe for another use, whether or not it is effective in another use.

Allowing prescriptions of drugs found to be safe is enough to trust licensed medical practitioners with them, even though a lack of proven effectiveness is a good reason to keep drugs out of the OTC market.

1. It's not regulation but clinical studies that are the expensive. They take time to set up, you have to do them carefully to keep the data from being contaminated, you have to wait a long time for results.

2. "demonstrated that physicians were willing and able to draw upon non-FDA sources of information to make rational prescribing decisions"

Few things to unpack here:
First, approved medications *have* been safety vetted by the FDA. So using them at a different dose for an off-label condition is not quite the same as using a drug that has had no previous study or evaluation.

Second, non-FDA sources of information? What exactly is that? The FDA doesn't do the studies so they are not the source. So what is really meant here is 'non-clinical trial information'. Ok so what is that? Anecdotes, hunches, theories etc. There's nothing wrong with that but it's good to keep in mind those things are nowhere near the same as a well designed and executed clinical trial.

So here's the thing. If a drug really works well for something the clinical trial angle is not going to be a huge problem. There are plenty of examples of clinical trials that were stopped early and all patients switched to the test drug because the results were so good. Sadly, though, that only happens occasionally. Most of the time a good drug is only able to slightly make it's value known above the random noise after a complete statistical analysis is completed.

So in terms of a wonder drug, the 'non-clinical studies form of information' will often work pretty well. If a drug cures 90% of patients with no side effects, it will generate a lot of anecdotes very quickly. A drug that causes 50% of patients to live 6 more months versus only 3 more months using the standard treatment, is not going to be so easy to see and there is where the well done clinical trial is going to yield insights that will be missed by individual cases of success or failure.

So in a world absent the FDA or the requirement for clinical trials what is the incentive for them to be undertaken? Marketing wise the inspiring story or anecdote is usually more effective than the dry clinical study. Yet at the margin it's the dry clinical studies that lay the ground for incremental innovation in health. A company with a good but not amazing drug would seem to be handicapped. Funding a dry study would not yield a great marketing tool *but* the competing drugs could dilute their merit by flooding the market with anecdotal based marketing.

I think perhaps we have a 'sweet spot' where the expensive clinical studies are needed to get to market and to make actual claims but off label is an area doctors can explore potentially new places a drug could help.

What happens, though, when someone has a decent drug that actually works modestly but has to compete in a 'snake oil' world against less reliable drugs making extravagant claims indirectly?

I think we are mixing up exactly what product we want to maximize here. It isn't drugs so much as the controlled studies that actually tells us what works and what doesn't.

Comments for this post are closed