The Use of Technology in Suicide Prediction and Prevention

Written By: Michael F. Armey, PhD, Research Psychologist, Butler Hospital on February 2, 2022


For more than a decade, the suicide rate in the United States has been on the rise. As of 2019, the last year for which CDC data are available, suicide remained the second to the fifth most common cause of death for individuals between the ages of 10 to 54. While there are some encouraging data to suggest that death by suicide has been less frequent during the COVID-19 pandemic, suicide still accounts for roughly 20% of all injury-related deaths, which is more than twice the rate of homicide.

There are many reasons why suicide is a complex problem to solve. Suicide is an overdetermined behavior, which simply means that there are many different ways that an individual can develop a risk for suicide. For example, some people might experience traumatic life experiences, others might have difficulty managing or regulating their emotions, and still, others might struggle with substance use or dependence. Unfortunately, this is far from a complete list, and most people at risk for suicide have a combination of factors that contribute to their unique experience of suicide.

Further complicating the detection and treatment of suicide is the manner in which most treatment for suicide is delivered. Most people receiving treatment see a therapist every week or two, sometimes in combination with medication. While this type of treatment can be helpful for a great many people, some of the people at greatest risk for suicide need help at times when a therapist might not be available. Likewise, people at risk for suicide sometimes miss or deny the subtle warning signs that they may be headed for a suicide-related crisis in the future. In short, there are gaps in both our understanding of suicide risk as well as our ability to deliver treatments when and where they are needed the most.

One promising solution to these problems lies in the use of technology, specifically smartphones, smartwatches, and activity monitors, to predict and respond to suicide risk. Using technology, we are able to address many of the limitations of traditional detection and treatment of suicide risk.

Eye-tracking technology, which uses a camera to observe how a person’s eye looks at images on a computer screen, can provide important information such as the location and amount of time an individual view part of an image. Working with patients who have been recently hospitalized for suicide risk, our team has found that these eye movements are associated with risk for suicide attempts up to six months later. This technology is particularly useful as it requires no feedback or response from the person observing images, meaning that it can potentially detect risks that individuals are either unaware of or unwilling to disclose to others.

Another promising technology is called ecological momentary assessment, or EMA. Through EMA, research participants or patients install an application on their smartphone that asks a series of questions throughout the day. These questions, assessing emotion, thoughts, and behaviors, provide unique and individualized insights about a person’s life experiences. In our research, we have asked patients who are being discharged from a psychiatric hospital for suicide risk to participate in EMA for three weeks. This research has so far yielded two important findings. First, patients tend to experience increased negative emotions in the weeks following hospital discharge that strongly predict increases in thinking about suicide. These increases correspond to prior research finding that risk for rehospitalization peaks roughly two weeks following hospital discharge, providing one explanation for these rehospitalizations. Second, positive emotions like happiness seem to be protective against suicidal thinking – that is, individuals who reported higher levels of positive emotions reported less thinking about suicide and were, in turn, less likely to be re-hospitalized.

Granted, these findings might seem obvious; however, this technology is innovative and potentially impactful as it allows us to better understand not just if, but when, individuals experience increased risk for suicide, potentially permitting the delivery of treatments outside of the therapist’s office when they are needed the most. These treatments, known as ecological momentary interventions, are a new and growing area of research, with great potential to improve our treatment of suicide risk.

One final technology worth mentioning is called digital phenotyping. A phenotype is defined as a set of behaviors that theoretically relate to an individual’s biological makeup or genetics. Through digital phenotyping, or the use of digital data to infer biological processes, data from a smartphone, smartwatch, or activity tracker sensors (i.e., motion/activity, heart rate, location, light levels, sleep, audio data, and social media and/or text messaging data) is used to construct a behavioral profile by identifying which digital data are most strongly predictive of suicidal behavior. As with eye-tracking data, the benefit of digital phenotyping is that data are collected passively, without the knowledge of the person being profiled.

Despite the benefits, it is clear that the use of technology to predict and track human behavior such as suicide also has enormous potential for abuse. Social media companies such as Facebook have been criticized by consumer groups and scientists for using this sort of data without permission from users. As scientists and clinicians alike embrace these promising technologies, it is important that potential patients and research participants understand the goals of the research and consent to the collection of these data.

However, despite these risks, the use of technology in the prediction of suicide risk is an important and growing area of clinical research with great potential to improve treatments and save lives.

Explore Our Services

Sign up for latest updates in mental health and wellness