<<  <  >  >>

Page 37

In any testing by any organization, private or public, there

is always the chance of certifying an unsafe device.  Like-

wise, there is always the chance that any single copy of any

manufactured device will cease to function properly.  Per-

fection cannot be the appropriate standard.  The relevant

question is whether we should expect more or less failure,

specific or categorical, under a free market regime than

under a centralized regulatory regime.

Because of the incentives faced by private institu-

tions, a market certification process should result in no

more frequent categorical failures than in the FDA regulato-

ry system, and market access for safe devices should be

faster.  This statement is not to minimize or marginalize

the suffering that results when a device fails or is later

found to be unsafe.  However, keeping useful devices off the

market, banning them, or delaying their market delivery also

causes deaths and prolongs suffering.  Though unsafe devices

will be certified and though samples of safe devices will

fail, as happens under FDA regulation, only market certifi-

cation minimizes the chance of both types of errors.

Supply of Effective Devices

Until now, third-party-certification organizations

within the United States have been primarily concerned with

safety and not with effectiveness or performance.  UL and

other companies conduct performance tests only for products

to protect life and safety, such as smoke alarms and fire

extinguishers.  In contrast, the FDA evaluates the perfor-

mance of every device submitted for approval.  Can third-

party certifiers like UL accommodate performance testing on

a much larger basis?

The certification industry would certainly change if

the public demanded a general effectiveness mandate in

addition to a safety mandate, but the certifying organiza-

tions could adapt.  Adding effectiveness testing will not

change the market incentives for third-party organizations.

The organizations would strive to be involved in research

and development from the earliest stages and to produce

flexible, adaptable certification systems.  The resulting

systems would move to include standards that incorporate

effectiveness requirements.  The standards, based on defined

expectations for performance, would be of the appropriate

quality, reflecting the consensus of consumers, manufactur-

ers, and standards authorities.  Importantly, those compa-

nies would still be dealing directly with their final cli-