Accessibility links

Doctors Hope to Predict Patients’ Suicide Risk

  • Carol Pearson

Suicide is hard to predict, even for close friends and family of victims.

When Dorothy Paugh's son, Peter, took his life at the age of 25, her family was devastated.

"We had no idea. None of us," she said. "We had gotten together a week or so before for his older brother's birthday. There were really no signs. He never said a word to anybody."

Suicide is shrouded in stigma and secrecy, even though it impacts countless individuals, families and communities. Someone commits suicide every 40 seconds across the globe, according to the World Health Organization. Data from the Centers for Disease Control show that 40,000 people kill themselves every year in the U.S.

Researchers at Massachusetts General Hospital developed a study aimed at pinpointing those most at risk.

According to Thomas H. McCoy Jr., director of research at the hospital's Center for Quantitative Health who led the study, researchers drew on typical factors, "but we also added a newer risk predictor. And that was the words that the doctors had written about the patient at the time of discharge."

Researchers examined the hospital discharge statements for nearly half a million patients over a period of several years. In the study, the researchers noted that a patient has an elevated risk for suicide immediately after being discharged, so doctors’ notes might provide some life-saving information.

"The primary objective of this study was to try to build a model to understand a little bit more about risk factors for suicide,” said Dr. Roy Perlis, a psychiatrist who worked on the study. “In particular, what we wanted to understand was, could we get additional information out of what doctors write in the chart when patients leave the hospital?"

The researchers analyzed the notes using a computer that was sensitive to the doctors’ language. They found that certain words, such as "unfortunate" or "poorly," were linked to patients with a higher risk for suicide. The computer provided information that researchers might miss if only patients' data — diagnostic codes and medicines, ages, genders and so forth — were examined.

The doctors acknowledge they are just at the beginning of their research. Next, they hope to examine discharge notes from other hospitals in the United State and worldwide, to see if other doctors’ language offers similar clues as to suicide risk.

The study was published in JAMA Psychiatry, a publication of the Journal of the American Medical Association.