Can an algorithm stop suicide?


The algorithm is based on an analysis of thousands of previous suicides in the VA database from 2008. The computer mixes and mixes numerous facts from the medical records – age, marital status, diagnoses, prescriptions – and determines the factors that together are strongest are associated with the risk of suicide. The V.A. The model integrates a total of 61 factors, including some non-obvious factors such as arthritis and statin use, and creates a composite score for each person. Those who score at the top of the range – the highest percentage of 0.1 – are classified as high risk.

"The concentration of risk for people under the top 0.1 percent in this regard was about 40 times," said John McCarthy, director of data and surveillance, in Suicide Prevention, VA Office of Mental Health and Suicide Prevention. “That is, they died 40 times more often than the average person from suicide”.

Bridget Matarazzo, the director of clinical services at Rocky Mountain Mental Illness Research Education and the Clinical Center for Veteran Suicide Prevention, said of Reach Vet, "My impression is that it identifies some people who were previously on providers' radar, but others who weren't. "

At the end of 2018, a V.A. Dr. McCarthy-led team presented the first results of the Reach Vet system. During a six-month period that Reach Vet was used, high-risk veterans stopped using V.A. Services. In contrast, a comparison group followed six months before Reach Vet was installed, the use of V.A. Services stayed about the same.

The Reach Vet group also had a lower death rate during this period – although it was an overall rate, including all causes of death. The analysis showed no difference in suicides, at least up to this point. "It's encouraging, but there is a lot more we need to do to see if we get the effects we want," said Dr. McCarthy.

Ronald Kessler, Professor of Healthcare and Policy at Harvard Medical School, said, "Currently, this and other models predict who is at greatest risk. What they don't tell you is who is most likely to benefit from an intervention. If you are don't know that, you don't know where to put your resources. "

However, for doctors using the system, it has already led to a rethinking of risk assessment. "You end up with a lot of older men who are really struggling with medical problems," said Dr. Goodman. "You are quietly unhappy, in pain, often alone, with financial problems, and you don't see them because they don't come in."