Suicide Prevention Algorithm

The Veterans Administration (VA) has developed an artificial intelligence (AI) decision-making algorithm to identify those veterans who are at a high risk for suicide. Called ReachVet, the system has prompted doctors to reassess their methods for isolating the factors that lead to suicidal events.

The algorithm was developed from a database of thousands of previous suicides. It integrates 61 factors — including diagnoses, prescriptions, age and marital status — and uses predictive modeling to derive a score for each person. Those in the top 0.1% have been shown to be 4o times more likely to die of suicide. Consequently, these veterans are flagged and contacted to arrange for an appointment to review prevention strategies with a clinician. The algorithm generates a new list each month.

With ReachVet in use since 2018, providers have observed that the algorithm identifies not only patients who were previously considered at risk, but also many who were not. In addition, it appears to encourage high-risk vets to increase their use of VA services.