<<  <  >  >>

Page 29

reason to assume that product performance standards will be

any different.  The FDA will look for one standard to apply

in every case for particular devices.  Products will be

evaluated and research conducted to meet exactly that stan-

dard and no other.  Few incentives will exist for companies

to develop devices that outperform the standard.  Whether

the device far exceeds the standard or barely passes the

test, it will still get the same FDA approval.  Meeting a

government-mandated standard passes along a fixed amount of

information about quality in every case.  The manufacturers

cannot readily internalize the benefits from outperforming

the standard when the FDA holds a monopoly on recognition.

In any case, use of performance standards and self-certifi-

cation (in some specific cases decided by the secretary of

HHS), by itself, will not change the FDA's behavior.  The

FDA will still possess the legal power to require submission

of data and information for its review.

Product performance standards may introduce another

bias into research.  Given the cost of developing new devic-

es, manufacturers may concentrate their efforts on producing

devices that clearly comply with certain product standards

and avoid the cost-increasing uncertainty of innovative

device development that may involve classification delays or

writing a new standard.

Still a further issue arises.  How is the FDA, or a set

of its employees, to choose the best or most appropriate

standard from among the collection of good standards?  In

the market for other sorts of goods, consumers, by their

choices of what to buy, determine the standards for safety,

effectiveness, and quality.  Many different marketplace

standards exist simultaneously, and the market provides a

wide range of goods of varying quality.  In a monopoly, the

monopolist sets the standards; currently, the FDA has a

legally protected monopoly.  The question is, how the FDA,

or who at the FDA, will be able to make the decision that is

best for all people in all circumstances regarding the most

appropriate standard.  In a specific example, will a stan-

dard that requires 70 percent effectiveness for 90 percent

of all patients always be better than a standard that is 90

percent effective for 70 percent of all patients?

The incentives for the FDA will not have changed, and

the FDA will still overinvest in minimizing Type I errors.

The FDA's primary use of other organizations' standards will

not change the incentives.  The FDA will adopt those stan-

dards that do not force the agency to be more concerned with

Type II errors.