Felitti had started the weight-loss clinic in 1980 because many Kaiser Permanente members suffered from health problems, such as diabetes and heart attacks, which were exacerbated by their weight. When patient records did not reveal a pattern, Felitti decided to ask people who had dropped out if he could interview them, and 286 agreed.

Many were morbidly obese. For example, a woman who is 5 feet, 5 inches and weighs 140 pounds is within the normal range of BMI, or body mass index. If she weighs 180 pounds – and she is not a body-builder or athlete whose extra weight is muscle – she has a BMI of 30 and is considered obese. If she weighs 245 pounds, she has a BMI of over 40 and is morbidly obese. “Morbid” in this case means unhealthy or diseased.

During the interviews, two clues surfaced: The first was that none of the participants was born fat – all had normal or below normal birth weights. The second was that people who were severely overweight didn’t gain 10 or 20 pounds a year over several years.

“I had assumed that people who were 400, 500, 600 pounds would be getting heavier and heavier year after year. In 2,000 people, I did not see that once,” says Felitti. When they gained weight, they did so abruptly and then stabilized. If they lost weight, they often regained all of it or more rapidly, within weeks or months.

The turning point in Felitti’s inquiry came by accident. The physician was running through a series of questions with yet another clinic dropout: How much did you weigh when you were born? In first grade? In high school? How old were you when you became sexually active? How much did you weigh when you married.

“I misspoke,” he recalls. “Instead of asking, ‘How old were you when you were first sexually active,’ I asked, ‘How much did you weigh when you were first sexually active?’ The patient, a woman, answered, ‘Forty pounds.'”

He thought he’d misheard. He asked again. She gave the same answer, began sobbing, and added: “It was with my father.”

The moment remains vivid, says Felitti: “I remembered thinking, ‘This is only the second incest case I’ve heard about in 23 years of practice.’ I didn’t know what to do with the information. About 10 days later, I ran into the same thing. It was very disturbing. It seemed that every other person was providing information about childhood sexual abuse. I thought, ‘This can’t be true. Someone would have told me in medical school.'”

Worried that he was somehow introducing observer bias, he asked five of his colleagues to interview the next 100 patients. “They turned up the same things,” he says.

Another piece of the puzzle dropped into place during an interview with another clinic dropout, a woman who had been raped when she was 23. In the year after the attack, she gained 105 pounds. “As she was thanking me for asking the question,” says Felitti, “she looks down at the carpet, and mutters, ‘Overweight is overlooked, and that’s the way I need to be.'”

Felitti began to realize that obese people didn’t see their fat as a problem. For many, it was a solution.

In 1990, Felitti presented his findings at a meeting of the North American Association for the Study of Obesity in Atlanta. Its members told him he was naïve to believe his patients, and they loudly denigrated the study. However, a researcher from the Centers for Disease Control and Prevention (CDC) was intrigued. “He told me that people could always find fault with a study of 100 people,” says Felitti, “but not if there were thousands. I said that wouldn’t be a problem.”

Felitti was introduced to Dr. Robert Anda, a medical epidemiologist at the CDC. Anda had been studying how depression and feelings of hopelessness affected coronary heart disease. Medical researchers and public health experts had known since the 1960s that behavior and disease are strongly linked. Drinking too much alcohol for too long can cause liver disease, for example. Excessive eating that leads to carrying a lot of extra weight is linked with diabetes, which can lead to blindness and amputation of feet and legs. But what physicians and public health experts didn’t understand is why people start and continue to smoke, drink or eat excessively, even when they know it’s bad for them.

Kaiser Permanente in San Diego was a perfect place to do a larger study. In any four-year period, 81 percent of the adult members of Kaiser visit the HMO’s Health Appraisal Center for a comprehensive medical exam. They fill out a detailed biomedical, psychological and social evaluation. Since the exam was a health assessment and not related to illness, the evaluation represented the experiences of the majority of Kaiser members, not just people who were sick. Anda added another series of questions to the evaluation.

From Felitti’s experience with obese patients, Anda selected eight categories of experiences that adversely affect a child. Three categories were actions directed specifically at the child: physical abuse, sexual abuse and verbal abuse. Five others addressed the dysfunctional family environment in which a child witnessed trauma or experienced loss or severe neglect: growing up in a household where someone was in prison; where the mother was physically abused; with a family member who was an alcoholic or used drugs; where at least one biological parent was lost to the child; and where someone was chronically depressed, mentally ill or suicidal.

Each category – not incident – of adverse childhood experience (ACE) counted as one point. That means a person who’s been raped three times by her grandfather has an ACE score of 1. Similarly, a person who grows up with two alcoholics also has an ACE score of 1. Someone who was sexually abused by their stepfather and grew up in a household where the mother was physically abused has an ACE score of 2.

The results of the first survey flabbergasted the physicians.

“Some of the increases (in health problems) are enormous and are of a size that you rarely ever see in health studies or epidemiological studies,” says Anda.

Those with an ACE score of 6 had a 4,600 percent increase in the likelihood of becoming an IV drug user. Most startling of all: 64 percent of the study participants had experienced one or more categories of adverse childhood experiences.

The study is a wake-up call for the medical and public health communities that previously thought that high levels of child trauma, including sexual and physical abuse, were seen only in disadvantaged populations. The 17,000 people who comprise the ACE study are typically American middle-class – 80 percent white, 10 percent Asian, 10 percent Latino. Seventy-four percent attended college; 46 percent graduated from college. Their average age: 57.

“The study is disquieting in its description of the frequency of abuse against children and how often families appear to be dysfunctional,” wrote epidemiologist Dr. William Foege, former director of the CDC, a senior fellow with the Bill and Melinda Gates Foundation and a legend in the field of public health, in an editorial in the American Journal of Preventive Medicine. “It is not what we want to believe about our culture, our neighborhoods, or ourselves. And yet as troubling as the data seem to be, we need to confront the problems described and find an appropriate public health response.”

“It (the study) changed my thinking dramatically,” says Anda, who is still a co-investigator on the ACE study. “In my opinion, because of the many things that ACE leads to, the problem of adverse childhood experience is the most important public health problem I’ve seen.”

Felitti and Anda, with a dozen other researchers who contributed to the work, published their first findings in 1998 and 1999 in two papers in the American Journal of Preventive Medicine and the Journal of the American Medical Association. Since then, they have published more than 30 other papers in other prestigious medical journals, including Pediatrics, Child Abuse and Neglect, and the International Journal of Obesity.

The latest research, published in May in the American Journal of Preventive Medicine, reveals how common child sexual abuse is. Among the 17,337 Kaiser Permanent clients, 25 percent of women and 16 percent of men reported being sexually abused as children. The research corroborates other studies. So, going back to Felitti’s experience with the dropouts in his weight-control clinic, do the results of the ACE study mean that all obese people were sexually abused as children? Not at all, say the researchers.

“We don’t want to conclude that every obese child or adult has experienced child sex abuse,” says David Williamson, a CDC researcher who used ACE study data to look at the link between child trauma and obesity. Just as there are many different reasons for getting cancer, the ACE study is revealing different reasons that people become obese.

The point is, although no one knows exactly how many of the nation’s obese and morbidly obese adults were abused as children, it is a substantial number. Williamson is comfortable with saying that there’s a link between child abuse – physical, sexual and verbal – and obesity in at least 8 percent of the obese population. If there are 70 million obese and morbidly obese Americans, as the CDC says, that means that more than five million obese and morbidly obese people are likely to have suffered physical, sexual and/or verbal abuse during their childhoods.

Williamson’s study did not include the ACE study’s five other categories of family dysfunction. And, so, for a good part of the rest of the obese population, Felitti believes it is likely that some type of trauma marks the starting point of the path to obesity. The question requires more research, he says, but “what we’re convinced of at this point is that we’ve made a major landing on the right beach.”

On this beach, says Felitti, fixing the obesity problem with diets or advice about eating won’t have any effect. “Nutrition is a nice subject and has nothing to do with obesity,” he explains. “Teaching people about nutrition is essentially predicated on the assumption that people get fat because they don’t know any better.”

Regarding the larger issue – the effect of behavior on chronic disease – one thing is very clear to Felitti and Anda. If physicians don’t address how childhood trauma affects people’s health, it’s not likely that patients will change their behavior. They won’t lose the weight, stop the smoking or dig out of the depression that contribute to their diseases.

“The study makes it clear that time does not heal some of the adverse experiences we found so common in the childhoods of a large population of middle-aged, middle-class Americans,” writes Felitti. “One does not ‘just get over’ some things, not even 50 years later.”

Jane Ellen Stevens is a science and technology journalist, and also teaches at the University of California Berkeley, Graduate School of Journalism. She is writing a book about child sex abuse.