Transcript Pharmaceutical Regulation: A Matter of Life and Death

Pharmaceutical Regulation: A Matter of
Life and Death
Sam Peltzman
Booth School of Business
University of Chicago
Overview
• Regulation of the Introduction of New Drugs
• Brief Early History
• Costs and Benefits
– In principle
– Evidence
• Political Response
• The Future
Historical Background
• New prescription drugs must be deemed ‘safe’ and
‘effective’ by Food and Drug Administration (FDA) before
introduction to the market
• Proof of Safety: FD&C Act of 1938
– Mfr conducts tests
– Submits to FDA
– Fixed time period. Default is “safe”
• 1935-1960: Industry enters a “golden age”
– Sulfa drugs, antibiotics, early psychotropics
– Enormous growth
– Political scrutiny increases
1962 Amendments
• Two concerns come together
1.
Consumers are wasting money on drugs that don’t work (better)
•
2.
Rush to market compromises safety
•
•
•
“me too” , mixtures, etc.
Wonder drugs are also strong medicine
Thalidomide: a close call
1962 Act
1.
Adds proof of efficacy requirement
•
2.
3.
Seller can only claim effects on approved drug label
Removes fixed time period. Default is “no, not yet”
Regulates testing process
•
•
Approval of all phases of testing protocol
In effect, requires double-blind randomized trials v placebo
A Brief Sketch of Drug Testing & Approval
• Pre-Clinical discovery
• IND: Regulated testing begins
• Phase I
– Mostly healthy subjects
– Primary concern: safety
• Phase II
– Patients with targeted disease
– Safety & also efficacy
• Phase III
– Large scale, double blind test v placebo
– Efficacy & also safety
Main Effects of the Regulation
• Drug development takes more time
– Today: 10 years (average) from IND to market
• 1 to 6 years more in pre-clinical: animal testing etc
– How much is due to regulation?
• Only 1.5 years directly in approval process
• But all previous testing affected by need to prove efficacy
– Rough guess: 3 to 5 extra years
• Drug development is more costly
– Around $1 billion/new drug
– Probably more than half due to extra costs of regulation
• Most affected part of process (phase iii) has 3x monthly cost of
earlier phases
And Fewer New Drugs
Average Annual Number of NCEs. US. 1951-2013
These Effects Can Be…
• Good
– More unsafe, ineffective drugs screened out by longer,
more expensive testing
Or
• Bad
– Some safe, effective drugs not developed because of the
higher R&D costs
– The good drugs are more costly
– The good drugs come to market later
• So potential benefits are missed
So, Do Costs Exceed Benefits, or Vice Versa?
• Most studies suggest that costs exceed benefits
– Caveat: I said so many years ago
• For both safety and efficacy
• Will focus on safety
Limits on Net Benefits of Efficacy Regulation
1. Ineffective drugs lose sales rapidly
•
So direct benefits (saving wasteful expenditures) are
limited
2. Growth of “off label” prescribing lost benefits
• Seller cannot claim new benefit for existing drug without FDA
approval
• But doctor can prescribe the drug for any indication
– Even if the official label doesn’t mention it
• Up to half of all Rx’s “off label”
• Suggests testing costs are restricting benefits
– From restricted information
– Or similar, but better targeted, drugs that are not developed
Safety Regulation: Outline
• What is the plausible safety benefit of extra testing?
• Compare it to just 1 cost
– Delay in benefits from extra time in testing
• Ignore loss of benefits from fewer new drugs
– Or extra R&D costs
• Use recent history to give you a sense of
magnitudes – why benefits are much less than costs
Safety Benefits of FDA Testing
• Pre-1962 : Unsafe drugs are rare
• 1 or 2 per decade with significant mortality risks
• Approx 1000 deaths before withdrawal
• Not entirely eliminated - e.g., “fen Phen”
– Diet pill marketed for 1 yr 1996-97
– Withdrawn after adverse safety effects
• Pulmonary hypertension
– Estimated 1800 excess deaths (by opponent of drug’s introduction)
• Low incidence effects hard to detect in clinical trials
– Need very large samples
So What Are the Actual Safety Benefits of Extra
Testing?
• Hard to tell
– How many lives lost per year today
– Compared to how many would be lost with less testing
• And greater reliance on market forces & tort law
– We do not observe the counterfactual
• But we can get a sense . . .
– Of why benefits are less than costs
– By exaggerating the benefits
• i.e., the lives lost in the counterfactual world
Say The Extra Testing Saves 500 Lives Per Year
• Less testing = 4 fen-phen’s per decade
– 3 more than we observe in today’s world
• Probably exaggerated, because…
– Observed only 1 or 2 per decade before 1962
– Tort law has gotten tougher
• fen-phen cost Wyeth $34B
• But saving 3 fen-phen’s would save ~5000
lives/decade
What are the Safety Costs of the Extra Testing?
• Look only at 1 cost
• The extra time it takes to get every good new drug
to market
– Which means some benefits are missed
• Use treatment of heart disease
• To give you a sense of how significant the cost of
delay is
Some Basic Facts
• Heart Disease is leading cause of death for males
over 45
– Mortality was improving before 1970
• Then mortality decline accelerated
– Especially for males
Death Rate from Heart Disease. US. 1960-2012
Per 100,000. 2012 population weights
Some Basic Facts
• Heart Disease is leading cause of death for males
over 45
– Mortality was improving before 1970
• Then mortality decline accelerated
– Especially for males
• Current male mortality rates much lower than 1970
– Or what they would be with previous trend
Death Rate from Heart Disease. US. 1960-2012
Per 100,000. 2012 population weights
The Role of Drugs and Regulation
• 300,000 fewer deaths per year (compared to trend)
in this age group
• New drugs make a modest contribution to this
improvement
– Around 25% of the overall improvement (Lichtenberg,
2009)
– Which  ~ 75,000 fewer deaths per year due to new
drugs
• All these new drugs made it through FDA process
• BUT
Not So Fast
• Recall: proof-of-efficacy adds 1.5 years to
regulatory process directly
– And a similar extra time indirectly
• For every new drug - including those that save lots
of lives
• In this case: each extra year in testing = 75,000
extra deaths
– You would need 15 decades of safety benefits (at a generous
estimate of 5000/decade) to pay for costs of only 1 extra yr of testing
A Big Benefit for Lots of People Becomes a Big
Cost
• If you make people wait even a little bit
• And the benefit from the extra time is not nearly as
large as the cost
• Both are true of current regulation
How Did This Story Play in Washington?
• Most economic analyses since early 1970s show
benefits << costs
• BUT official reaction evolves slowly
– Denial: 1970s & 1980s
– Recognition & Some Response
• Orphan Drug Act (1983)
– Special treatment for drugs with small market
• Prescription Drug User Fee Act (1992)
– Drug companies pay a ‘user fee’ to speed up NDA process
– Both have ‘worked’
– No serious questioning of overall process
Why Does New Drug Regulation Survive Politically?
• If economists are right regulation kills more – many
more – than it saves
• But the regulatory system is politically safe
• Why? 3 protective forces
1. Asymmetry
2. Interest Group Politics
3. Progress
1. Asymmetry: Who Will Speak for the Victims?
• The victims of the regulation are anonymous
– They would not have died if drug was available sooner
– We – and their relatives & friends – never know exactly
who they are
• But the victims of an unsafe drug are visible
– Their relatives, friends - & the world - hear their story
• Avoiding harm dominates politically
2.Interest Groups: Who Would Lobby for
Change?
• Industry? Well organized, but ambivalent
– Each developer would like to avoid costs/delays
– BUT once you get approval…
• Costs/delays are barrier to competition
– AND ‘Big Pharma’ can sell regulatory expertise
• Most biotech startups sell to large player by stage iii
• Consumers? Not organized & ‘rationally ignorant’
– One exception: AIDS activism in 1990s
– ‘Consumer lobbies’ – push for even more regulation
3. Medical Progress
• New drugs do get to market – eventually
• People live longer
– Life expectancy at birth increases steadily
– 71 years in 1970  78 years now
Years of Life Expectancy. US. 1960-2011
78
76
74
72
70
1960
1970
1980
1990
year
2000
2010
3. Medical Progress
• New drugs do get to market – eventually
• People live longer
– Life expectancy at birth increases steadily
– 71 years in 1970  78 years now
• Where’s the crisis?
– Economist: ‘but for the regulation lives would be saved’
– Non-economists do not live in a counterfactual world
In Conclusion
• Economic analysis suggests FDA is neither safe nor
effective
– Many lives could be saved by reduced testing
• Eliminating pre-market approval
• Eliminating efficacy requirement
• But fundamental change is politically inconceivable
• Do not be discouraged
– 15 minutes is a long time