The recently launched crusade to have every child tested for autism before the age of two has as its reason an opportunity for "early intervention" to treat the condition.

Dr. Scott Myers, a pediatrician, has been quoted by Reuters news service as saying that autistic children who get earlier treatment "do better in the long run."

That may be true if the children are genuinely autistic. But the dangers of false diagnoses of toddlers and preschoolers have been pointed out by Professor Stephen Camarata of Vanderbilt University, who has tested and treated children with autism for more than 20 years and has encountered many cases of inaccurate diagnoses.

A prudent trade-off, as distinguished from a crusade, would weigh the dangers of false diagnoses against the benefits of "early intervention."

There is already considerable evidence of false diagnoses of preschool children as autistic, and the treatments inflicted on them can be abusive, with incalculable negative effects on their development.

What about the positive effects of "early intervention"?

According to Professor Camarata, those children "with true autism" are "very difficult to treat and may never say 'mommy' or learn to take care of themselves without Herculean efforts by their parents and teachers."

The limitations of what can be achieved with even early intervention mean that there can be real heartbreak, whether a toddler or preschooler is either falsely or correctly diagnosed as being autistic.

Much has been made of statistics showing a sharp increase in diagnoses of autism in recent years.

What has gotten much less attention is the changing definition of autism, which raises the question whether there has been an actual change in the real world or simply a change in the way words are used when collecting statistics.

People today are often spoken of as being "on the autistic spectrum," rather than as having autism.

While there are some conditions which are much like autism, there are other conditions, such as having a very high IQ or simply being late in talking, which often include characteristics listed on checklists for autism. These are open invitations to false diagnoses.

We would see the dangers immediately if people who wear glasses were included on "the blindness spectrum" or people with harmless moles were included on "the cancer spectrum."

Blindness, cancer and autism are all too serious -- indeed, catastrophic -- to use loose definitions that fudge the difference between accurate and inaccurate diagnoses.

Loose definitions of autism produce bigger and more newsworthy statistics, which in turn can attract more children into existing programs and attract more money from the government, foundations and other sources to support those programs.

Many parents have told me that they have been urged to let their children be labeled autistic, or on the autistic spectrum, in order to get money for speech therapy or other conditions from grants that are available to deal with autism.

Professor Camarata points out that the "less precise 'autism spectrum'" label "has had the unintended consequence of diluting resources, research and services to those children and families who most need the support" -- that is, families whose children suffer from genuine autism.

Loose definitions also promote the illusion of "cures" for autism, since most late-talking children who were never autistic in the first place "will be miraculously 'cured' because most late talkers who are otherwise unimpaired learn to talk with little or no treatment," according to Professor Camarata.

Parents whose children are late in talking or have other troubling problems would do well to seek diagnoses from the most highly qualified professionals they can find -- but not rely on the facile checklists being promoted in the current crusade for universal diagnosis of infants and toddlers for autism, without facing the question whether or not there are enough people qualified to make such diagnoses.