In “Addicted to Food?” Oliver Grimm states that a body mass index (BMI) of a specific value makes a person obese. I disagree: BMI is a simplistic formula based on height and weight that is often inaccurate.

Every time I go in for a physical, the nurse starts to lecture me that I'm overweight based on the calculation of a BMI of 25. Then I interrupt the nurse, and she sheepishly admits that my BMI does not mean I am fat.

I'm under 15 percent body fat and expect to be under 10 percent soon. My mass comes from muscle. I engage in an intense workout at least five times a week.

It worries me that no one ever writes down my body fat percentage. Anyone looking at my medical record would think I'm at risk for obesity. What happens when people start to make medical decisions based on my BMI record?

My experience points to the danger of using BMI; what's true about me is exactly the opposite of how BMI is typically interpreted. And such interpretations are given more credence than they deserve when scientists make statements like “A BMI above 25 indicates obesity” without qualifying them.

Joe Thoennes
San Francisco

THE EDITORS REPLY: It is true that there are more useful ways than BMI to measure health and obesity in individuals. But when used to study large groups of people, BMI provides an accurate snapshot of a population's health—the small number of outliers for whom BMI is misleading do not make a statistical difference. So Grimm's report that “about one third of American adults are overweight, and nearly another third are obese,” based on a survey of BMI, is a fair portrayal of the big picture.


I believe “A Personal Obsession,” by Isabel Wondrak and Jens Hoffmann, was incomplete. The authors state, “Fortunately, celebrity stalkers rarely used violence against their targets.” Unfortunately, some stalkers use violence against themselves. The article speaks of the mental status of the victim but not of that of the stalker. The article also doesn't offer any hope for either party.

One horrific case was David Letterman's stalker, Margaret Mary Ray. She was gripped by the psychotic fantasy that she was romantically involved with Letterman. According to the New York Times, her bizarre behavior, including breaking into his home and stealing his car, was often treated as fodder for comedians. But at the root of Ray's obsession was a very serious mental illness.

Ray was diagnosed with schizophrenia. The illness frequently can be managed with drugs, but getting patients to stay on their prescriptions can be difficult. Ray was in and out of jail. When she took her medication, she would get better and be released; then she would stop taking her pills and be reincarcerated.

The last time Ray stood trial, she was released despite the judge's concerns that no existing law could guarantee psychiatric help for her. Within months, Ray was dead, having thrown herself in front of an oncoming train.

Ray and individuals like her now have the assistance of the Mental Health Court Program, a strict probation and case management program that prevents the mentally ill from being wrongly housed in prison, while also protecting society from their criminal behavior. There is hope.

For more information, please visit www.consensusproject.org.

Rae Packard
Yucca Valley, Calif.


Although Robert Epstein is almost certainly correct when he suggests in “The Myth of the Teen Brain” that “teen turmoil is the result of the artificial extension of childhood” past puberty, he errs when he claims that teen turmoil is a “creation of modern culture, pure and simple.”

It is not so simple. In ancient Roman society, paternal legal rights and inheritance patterns prolonged a dependent preadult state in a significant percentage of young males. (History records much less information about females.) All sorts of adolescent antisocial behaviors were on display in Rome, including drinking, gambling, gluttony, illicit heterosexual and homosexual activities, dabbling in the occult and armed violence. Contrary to the assertions of historian Marc Kleijwegt of the University of Wisconsin–Madison, whom Epstein cites, many Roman authors comment on the turbulent (or lazy and vicious, if they're more unkind) nature of youth. Indeed, there was a Latin phrase describing this stage: lubrica aetas, the “slippery age.”

Blaming modernity for adolescent ills is too easy. It would appear that any society, ancient or current, that exhibits significant economic complexity and social stratification will bring out tendencies toward disorder in the young.

Mark E. Vesley
St. Paul, Minn.


In “Autism: An Epidemic?” [Facts and Fictions in Mental Health], Hal Arkowitz and Scott O. Lilienfeld claim that the rate of autism is not increasing; rather our sophisticated diagnostic procedures have simply detected more autism cases. I would like to offer readers another perspective, as a clinical psychologist.

Some would like to believe that there are not more cases, just more kids encompassed within the “pervasive developmental disorder” (PDD) spectrum, which includes a range of symptom severity. The logic used is that we have changed the diagnostic criteria, thus including a broader population. But DSM-III—the manual used by doctors to diagnose mental health disorders in the 1980s—contains the diagnosis of PDD as well as that of autism. Many of the criteria currently considered for a diagnosis of “autistic” (or a condition along the autism spectrum) are encompassed in the PDD criteria of 1980. Call it what you will: the symptoms were described back then.

In addition, a team of 13 prominent physicians compiled the section of DSM-III that dealt with disorders of infancy and childhood. Their conclusions: infantile autism is “very rare (2–4 cases per 10,000),” and childhood PDD is “an extremely rare disorder.” So what are we to deduce? It appears that we have two options: (a) the physicians, experts in the field of child psychiatry, were poor diagnosticians who failed to recognize thousands of impaired children (if, as Lilienfeld and Arkowitz claim, the rate was one in 166 back then), or (b) the rate has risen.

Randall Strandquist
Spokane, Wash.

ARKOWITZ AND LILIENFELD REPLY: In contrast to DSM-III of 1980, the later editions, namely, DSM-III-R and DSM-IV, include a new category of “PDD not otherwise specified” (PDD NOS), which encompasses many subsyndromal (mild) cases. Research suggests that PDD NOS and other milder variants now account for about three fourths of all autism diagnoses.

The diagnosis of autism has become considerably less stringent from DSM-III to DSM-IV. DSM-III required that all six criteria be met, whereas DSM-IV requires that only any eight of 16 criteria be met. Moreover, as University of Wisconsin–Madison psychologist Morton Ann Gernsbacher and her colleagues noted, DSM-III criteria for autism required “a pervasive lack of responsiveness to other people” in contrast to DSM-IV criteria, which require only “qualitative impairment in social interaction.” Strandquist's conjectures do not explain why research on a sample of more than 10,000 British children showed no increase in autism prevalence between 1992 and 1998, when researchers ensured that the diagnostic criteria remained constant.

Moreover, the apparent rise in autism rates derives from administrative (for example, school-reported) rather than population-based estimates, only the latter of which allow accurate measures of prevalence. Although we should remain open to the possibility of a genuine rise in autism diagnoses pending new data, the evidence for this rise remains uncertain at best.