Chance of FDA Approval

Long ago, one of my first blog postings described the "research funnel".  How a drug went from animal tests through three phases of clinical trials and was eventually approved (and how drugs could fail at each step in the process).  If you haven't read that posting, it contains a lot of useful background information on how drugs get approved in the USA:
http://cureresearch4type1diabetes.blogspot.com/2008/06/understanding-research-funnel.html

I used the best data I could find for that post, but I'm very happy to say that Biotechnology Industry Organization (an industry trade group) and BioMedTracker (a company which collects data on drugs in development) have published some much better data on success rates for US FDA approvals.  Their data set is much larger than what I had back then.  And it is interesting! At least to me :-)


The study was conducted from 2004 through 2010 reviewed more than 4,000 drugs from companies large and small and both publicly traded and private.  But NOT university or non-profit research.


Major findings about the process right now:

• The overall success rate (all the way through successful FDA approval) for drugs moving from early stage Phase I clinical trials is about 9%.  For phase-II it is 15%, for phase-III it is 44%.
• 63% of drugs in Phase I testing advanced to Phase II
• 33% of Phase II drugs made it to Phase III
• 55% of the drugs that made it to Phase III testing filed for approval applications.
• 80% of the drugs that filed for approval gained eventual approval (only about half were approved on their initial FDA review)
• Biologics had a 15% chance of going from Phase I through to FDA approval, compared with a 7% success rate for traditional small molecule chemical drugs. A biologic is more complex than a simple drug.  It is generally something derived from a living organism.  For example, a purified microorganism or specially treated or processed blood or tissue would be a biologic.  BCG, cord blood, and ATG are biologics that are being tested on type-1 diabetics.  
• The highest overall success rate from Phase 1 through likelihood of approval was for infectious diseases, such as hepatitis and HIV drugs, at 12%.  Next was endocrine system drugs, featuring [type-2] diabetes treatments, at 10.4%.  And then autoimmune diseases, such as rheumatoid arthritis, at 9.4%.
• Overall success rates from Phase I to FDA approval is nearly 9%. This number is comprised of lead and secondary indications. When separated, lead indications have close to a one in seven rate of approval and secondary indications have a rate of one in 30.  This was seen in all phases of clinical development as well as in all disease areas. 
• The study also shows that large molecule drugs are twice as successful in gaining approval than small molecule drugs.

Major findings about changes in the process:

• Overall success rate for drugs moving from early stage Phase I clinical trials to FDA approval is about 9%, down from one in five to one in six seen in reports involving earlier years.
• The 80% approval rate (for drugs that submitted applications), is down from 93% seen in early studies.

Commentary

Here are some very rough calculations:
There are currently about 15 phase-I trials with a 9% chance of eventual approval   (1.35 total)
There are currently about 10 phase-II trials with a 15% chance of eventual approval (1.50 total)
There are currently about 3 phase-III trials with a 44% chance of eventual approval (1.35 total)
So, a reasonable estimate is that out of the current crop,  4 drugs will eventually be approved.  It will be very interesting to see how this plays out over the next few years.

But with two serious limitations:

First, The study above only included commercial company's drugs.  it did not include drugs being researched at universities.  I would assume that university research is far less likely to eventually be approved than commercial research. (There are at least two reasons for this: first, universities should be more experimental and less worried about practicle applications, and second, they don't have the resources to push treatments and commercial companies do.)  Since several of the current clinical trials are university research, they are less likely to end up being approved.

Second, most of the current crop of treatments in clinical trials are not cures.  Indeed, only one (Burt) has kept people free of external insulin for 4 years.  That is a phase-I trial, so I think it is reasonable that we are going to need about 10 more treatments in phase-I that actually cure people (or get close) before such a cure will become generally available.

I know that many people want to spend a lot of time and resources investigating drugs that are already approved, for some other purpose, in the hopes that they will cure type-1.  This study shows the dangers of such a focus, since those are the secondary indications that have the much lower success rate.  It turns out that most treatments only work on one thing, so if you try to use them for something else, you chance of success goes way down.  (Even if it is cheaper and quicker.)

Also, it is interesting to me that more complex treatments have higher chances of eventual approval.  Biologics are more complex than drugs, but have a greater chance of approval.  More complex molecules are more likely to be approved than simple ones.  My guess is that because these are more expensive to develop, companies only push the very best of them.  Or maybe all the simple cures have already been productized, and only the more complex ones remain.

Handouts:
If you work in the pharma industry, or if it is important to you, I strongly suggest that you look at this PDF file.  It is thick with information:
http://insidebioia.files.wordpress.com/2011/02/bio-ceo-biomedtracker-bio-study-handout-final-2-15-2011.pdf
News reports:
http://news.yahoo.com/s/nm/20110214/hl_nm/us_pharmaceuticals_success;_ylt=AgsEciuYx.Hb5aUFo.oOX0gPLBIF;_ylu=X3oDMTJ2OGI2aGpzBGFzc2V0A25tLzIwMTEwMjE0L3VzX3BoYXJtYWNldXRpY2Fsc19zdWNjZXNzBHBvcwM4BHNlYwN5bl9hcnRpY2xlX3N1bW1hcnlfbGlzdARzbGsDc3VjY2Vzc3JhdGVz
http://www.bio.org/news/pressreleases/newsitem.asp?id=2011_0214_02


Joshua Levy
All the views expressed here are those of Joshua Levy, and nothing here is official JDRF or JDCA news, views, policies or opinions. 
Blog: http://cureresearch4type1diabetes.blogspot.com
To Get as Email Join here: http://groups.google.com/group/type-1-diabetes-clinical-trials-news
Web: http://joshualevy.pbworks.com/DiabetesCureReadyForHumanTrials

Why Fund J and J / Animas / DexCom Close Loop Trials?

Note: notes like this [d1] are footnotes to more discussion at the bottom of the posting.

Why Fund J and J / Animas / DexCom Closed Loop Trials?

About two weeks ago, JDRF made a big announcement that they were helping to fund research into a closed loop system.  Basically, they are working with Animas (insulin pump manufacturer and division of Johnson and Johnson) who work with DexCom (CGM manufacturer).  JDRF was putting in about 8 million dollars over about 3 years [d0], and was hoping for clinical trials within a year, and FDA approval of a device (although not a full Artificial Pancreas, or AP) within 4 years. 

So the first thing that popped into my head was "why bother?".  By which I meant, why should JDRF bother?  I don't have any problem with JDRF funding commercial companies [d1].  But Johnson and Johnson (especially) is a huge company which could easily fund it's own clinical trials. And it is clear to me (and many others) that closed loop systems were the next big thing in pump technology.

It so happens that last weekend I found myself talking to two guys who know more about corporate research into type-1 than I do, so I asked them this very question: Why should JDRF fund research that Animas was going to do anyway, when Johnson and Johnson had the money to do it?  Their reply was simple:

Animas would never have started the research into a closed loop trial, without the money and public support of JDRF behind it.

My initial reaction was disbelief.  I pointed out that Animas was in the pump business, and the obvious next step in pump development was a closed loop of some kind.  The first company to market a closed loop pump would have a huge advantage over all the other ones.  Of course they would do a closed loop clinical trial.  It had never occurred to me that any of the pump makers would not do a closed loop clinical trial.

But I was wrong.  As I quickly learned by searching through the FDA's clinical trial site.  In fact, Animas was not doing a closed loop trial; None of the big pump companies is currently doing a closed loop trial without JDRF's involvement. [d2]. There were two reasons why not.  The first reason was pretty simple: no pump company knows how to get a closed loop pump through the FDA approval process.  For one thing, no one's ever done it.  For another thing, a closed loop system would involve both a drug and a device.  In the world of the FDA, drugs have one set of rules (and are governed by one part of the agency) while devices have a completely different set of rules (and are governed by a completely different part of the agency) and of course the rules are not coordinated, and the two parts of the agency don't talk to each other much.  So there is a lot of uncertainty about exactly what sort of testing will be required [d3].  How long it will take, and so on.  And uncertainty can be fatal to big company research.  If J and J puts 10 million of their money into it, they will not know if it will take 3 years or 10 years to get through the regulatory process.  And that's a big difference when your stockholders want quarterly profits.

The second reason was simple: if no one does it, then no one has to do it.  This is the downside of capitalism when there are only a small number of companies.  There are only a few companies that market pumps in the US.  And they are all big and highly regulated, and they know pretty much what each other are doing.  So as long as they know that none of their (limited) competition is developing a closed loop system, they don't have to develop one [d4].

So here are some reasons why funding a closed loop clinical trial is a good thing for JDRF to do:

1. It (hopefully) gets one closed loop approved, so people can use it.
I'll put together another post on why closed loop is a good thing, even if it is not a cure.  It's too big a subject to put here, but as a quick summary: an AP is likely to lower many (if not all) of the long term complications of type-1 diabetes.  An AP is likely to be easier to use than a pump, and require much less work on the part of the diabetic.  An AP is likely to prevent the worst single complication of type-1 diabetes: "dead in bed". And an AP may well be required by the eventual cure, because excellent BG control from an AP may be required to regrow beta cells as part of an eventual cure.

2. It might motivate other big pump companies into starting their own clinical trials.
Capitalism is based on competition between companies.  So having JDRF work with one company is likely to motivate other companies to start, hurry, or continue their own research.  So putting money into Animas helps them directly, but also motivates Medtronic and Omnipod to put more of their money into similar research, or to work with JDRF themselves [d5].

3. Once one closed loop system is approved, it will be easier for the others.
One of the problems with the FDA approval process (especially for a combined drug/device), is that the first one through the gate pays a big price both in terms of extra work, but more so in terms of extra uncertainty.  So everyone wants to be second through the process.  Let someone else figure out the rules, and once they are known, then just go through.  Let someone else clear the path.  So if JDRF's money helps clear the path (ie. push Animas through the approval process first) it will end up benefiting Medtronics and Omnipod as well.  And in the end, all type-1 diabetics.

Put another way, consider this: If you have a problem, then you have a problem.  But if you can solve that problem with money, then you merely have an expense.  Companies don't like problems, but understand expenses very well.   Right now, getting FDA approval for an AP is a problem; no one knows the details of how it will be done.  However, if JDRF can spend some money and work with Animas to figure out how to solve that problem, then it will merely be an expense.  And companies know how to deal with expenses.  And an AP will be forthcoming sooner.

More Discussions

[d0] Although a lot of people focus on the money (as I do here), JDRF is bringing more to the table than just money.  They are bringing their previous research (especially previously tested algorithms), and well as a patient recruitment network, etc.

[d1] I think that the goal of JDRF should be to speed development of a cure for type-1 diabetes, and if they think the best way to do that is by funding a researcher at a University, great.  And if they think the best way is to fund a researcher at a company, great.  I don't care where the money goes; I care that it speeds development of a cure.

[d2] This surprised me, but it's true.  Using the FDA's clinical trials database:
There are three studies using Animas, but not a closed loop or feedback study.
There are four studies using Omnipod, but not a single closed loop or feedback study listed.
There are thirty one studies using Metronics, but only one testing a closed loop system and that one is a joint venture of Medtronic and JDRF: http://www.clinicaltrials.gov/ct2/show/NCT00831389
There are thirteen studies covering closed loop devices in diabetes.  One of them is the joint Medtronic/JDRF trial described above.  Of the remaining twelve, all are funded and sponsored by government agencies, hospitals, or universities.  Not one is funded by a company. 

So it is clear to me that the pump companies are not going to fund clinical trials aimed at getting a closed-loop system through the FDA approval process, unless JDRF catalyzes the process.  


[d3] There have been some drug/device combinations improved in the past. For example, drug covered stents.  However it does involve much more uncertainty than either a drug or a device alone.

[d4] This is not a matter of an evil conspiracy, more a case of no one spending money on something they don't have to.   They all think, since none of my competitors is working on a closed loop, then I'll spend my millions on something else.  If one of them does start, then I will start the year after, and still be pretty close to them.  Medical pumps are not a business where being a month (or even half a year) ahead of your competition matters that much, anyway.

[d5] And it is important to remember that JDRF is not playing favorites with companies.  They already have a similar agreement with Metronic, so this one with Animas will be their second, and it shows clearly that JDRF will partner with any company that is moving the research ahead.

This blog posting was unusual in that I asked several people to review it for me and provide suggestions and feedback.  I'd like to thank all of these reviewers.  

Joshua Levy
All the views expressed here are those of Joshua Levy, and nothing here is official JDRF news, views, policies or opinions.