Transcript Implementation of CAMC - Collaborative Creativity Group

Health Impact Fund
Aidan Hollis
University of Calgary, Canada
January 28 2008
Prizes work well, but…
Prizes solve the deadweight loss
(“access”) problem, and can be directed to
solve important problems, but - as Davis and Davis 2004 note:



“Due to problems of asymmetric information, it
may be difficult to set the appropriate size of
the reward, and/or to pick the most qualified
contestant. Rewards may also be arbitrary…”
And difficult to set a prize for everything!
The problem of prize size
Ex ante, we don’t know the social value of
an innovation.
 Ex ante, we don’t know how much it will
cost to develop an innovation.
 We don’t know, for a given amount of cost,
what the probability of success would have
been.

So setting a prize amount in
advance is guesswork!
but there may be a way around
this problem…
A prize for health impact
Suppose that the thing you actually value
is the health impact of a drug
 And suppose that you can measure this,
perhaps imperfectly, in terms of QALYs
(Quality-adjusted life-years) after the
product has been introduced.
 Then, there is a prize mechanism that will
offer suitable prizes without guessing.

A mechanism
Fix a total reward amount to be paid
annually, funded by governments (e.g. 5bn
euros).
 Firms can get a share of the reward if they
set their product price equal to production
cost.
 Each product obtains a share of the total
reward equal to its share of health impact
over a period of say 10 years.

Optional or mandatory


If mandatory, this is similar to the proposal
embodied in S. 2210 of Senator Sanders.
An optional version of this gives firms the right
either (a) to set the price of the product at a predetermined low level and to receive payments
from the prize fund, or (b) to exploit their usual
patent rights.


This is the Health Impact Fund to be proposed by
“Incentives for Global Health”
I explore the properties of the optional system
here.
Optionality means
Much less money in funding to start with
 Doesn’t overturn the existing system
 Doesn’t achieve as much as the Sanders
proposal.
 But, automatically sets the prize per
product in a reasonable range.

Setting the prize amount

The amount per firm is “market”-driven.


The payment per firm is proportional to health
impact.


Firms compete to obtain a share of the total payout,
based on their own private information about costs
and probability of success.
Eg: A product with 2% of the total health impact
receives 2% of the fund’s payout.
The payment per QALY is comparable to the
reward per QALY under the patent system, since
firms can choose either the patent system or
payment from the fund.
Efficient payment amount
Note that the relative payment per product
is proportional to health impact, which is
exactly as desired.
 And (because it is optional) the absolute
value is set by the possibility of
substitution between the patent and prize
systems.


If the payment per QALY in the HIF is too
high, more firms enter and dilute the fund.
Room for growth?

If the system appeared to be working well,
increased funding could allow the system
to expand.


The characteristics of such a system would
increasingly be similar to the Sanders
proposal as the total amount of the prize fund
increased
If the system appeared to be working
poorly, it could be cancelled, with notice.
What products would be rewarded
under this system?

Firms with products which had relatively high
health impact but small profitability from
exploitation of a patent monopoly would choose
to be rewarded under this system.


This clearly includes drugs for diseases prevalent
mainly in low- and medium-income countries.
Products with relatively small global health
impact but high profitability from exploitation of a
patent monopoly would choose to exploit their
patent.
Summary

An optional reward mechanism granting
cash payments for patented drug
innovations with prices set at production
cost solves the problem of choosing a
prize: firms make the choice themselves
based on private information about costs,
probability of success, and expected
health impact.
Issues for discussion
Require firms to offer open license for
generic production or price below some
limit?
 Fixed fund or fixed payment per QALY?
 Problem of measuring health impact
 Problem of setting the right baseline
technology

License or price control?

Two options for qualification to receive payments
from the Health Impact Fund



Offer open license for all relevant patents
Or agree to maximum price
The latter



requires the price to be set somewhat arbitrarily
May not achieve lowest cost of production
Is attractive to firms ideologically committed to IP
Why fixed fund instead of fixed
payment per QALY
Benefit of fixed payment per QALY is
reduction of risk for firms
 But entails more risk for funders
 Fixed fund uses market-type mechanism
to set price per QALY



payment in line with what is available in
patent system because of substitutability
Better monitoring; less conflict between
fund and innovators.
Measuring health impact
HIF would use estimates of effectiveness
obtained from clinical trials when drugs
were first introduced.
 Once the drug was in use, epidemiological
studies of health effects could be applied.
 These are likely to lead to much more
investment in comparative effectiveness
studies, and more debate and lobbying.
 Need a limited appeals process.

Baseline technology
Need to set a baseline for determining
improvement.
 May be technology as of 2 years before
introduction of a given product.
 If further back, more likely to have multiple
firms with similar technologies, therefore
competing promotional expenditures.
 If further back, less risk.

How much funding?
Commitment to fund 2-5bn euros per year,
increasing at 3% a year, starting in 2012.
 This commitment to be spread between
developed countries and to a much
smaller extent, medium income countries.

Who gets the money?

Innovators holding valid patents on
registered drugs and vaccines, who agree
to charge a price fixed roughly at
production cost.
How much do they get?


The registrant of each drug receives a share of
the prize fund equal to the share of health
improvements attributable to that drug.
For example, suppose that



The measured effects of a particular drug is 10,000
life years saved.
The total effect of all drugs in the system is 1,000,000
life years saved.
Then the drug would obtain 1% of the fund for that
year.
Who would determine health
effects of a drug?
An appointed expert Health Impact
Assessment Committee.
 It would be instructed to use best available
methods to estimate health impact.
 It would be required to publish its
decisions with reasons.

How would this help?
Prices for drugs in the system would be
low.
 Innovators would be rewarded for valuable
innovation. The larger the health impact,
the greater would be the reward.

How many drugs would this system
pay for?
If developing a new drug costs 500m
euros, 2bn euros a year would add
enough funding to support 4 new drugs a
year on an on-going basis.
 It is likely that the cost of developing drugs
for “neglected” diseases is relatively low.

Intellectual Property
The proposed system requires no
changes.
 Patentees could continue in exactly the
same way as before, but with a restriction
on their pricing.

Summary
Patents
 Fairly efficient
incentives
 High prices
Sanders prize fund
 Very efficient incentives
 Low prices
 Complete overhaul of system
AMCs
 Effective
 But limited to
specific wellunderstood
products
Health Impact Fund
 Very efficient incentives only for
drugs in the system
 Low prices for drugs in system
 Automatically calibrated rewards
 Incremental implementation
HIF vs. AMCs

Main difference from AMCs.
Payments based on QALYs per unit,
estimated afterwards, instead of before.
 No pre-specified technical requirements.


Thus this mechanism can deal with any
new medicine, not just those pre-specified.