Burn-out: artificial intelligence could spot the first signs more easily | Engineering Techniques

Burnout affects many employees in France. But its identification is complex, because the symptoms can be superimposed on other diseases or syndromes. Thanks to AI and natural language processing, Swiss researchers have achieved a high level of detection.

According to the Health Watch Institute, 480,000 people in France are in psychological distress at work and burnout concerns 7%or 30,000. Considered by the WHO as a syndrome linked to chronic stress at work and not an illness, this deterioration in mental health is not without repercussions.

It has a strong impact on physical health. People in psychological distress are twice as likely to report a somatic problem. Thus, 40% of those questioned report sleep disorders (+5% in one year), according to the 7th edition of the Human Footprint barometerspecialist in the prevention of psychosocial risks, carried out from April 30 to May 10, 2021 with 2,007 people.

Biased results

One of the key elements in the prevention of burnout is the speed with which it is detected and therefore treated. Various warning signs, both emotional (irritability, increased emotivity, etc.) and physical (pains and various disorders, malaise, lack of sleep), but also difficulty concentrating and loss of memory must be identified as quickly as possible to alert the person and his surroundings.

Currently, these warning signals are detected through psychological tests in which patients complete a multiple-choice questionnaire. One of the best known is the “Maslach Burnout Inventory”, consisting of 22 questions divided into three categories: burnout, depersonalization/loss of empathy and personal performance evaluation.

Example of a typical question and answer: “I feel exhausted at the end of my working day: never / sometimes / every day”. Some people do not dare to tick the answers “never” and “every day” or are tempted to lie to influence the results.

The use of natural language processing (NLP) methods has the potential to mitigate the limitations of this detection. Mascha Kurpicz-Briki, professor of data engineering at the Bernese University of Applied Sciences in Bienne, Switzerland, and her team of scientists supported by the Swiss National Science Foundation (SNSF) have developed a method based on automatic text analysis.

93% of burnout cases were detected

This team analyzed texts from Reddit anonymously. It is a social news aggregator, that is to say a mixture between a discussion platform and a link distributor, which is divided into forums, the “subreddits”.

Although there is a subreddit dedicated to burnout, the number of entries was considered by the Swiss to be too low to provide a large enough data set. The team therefore added texts from various thematic forums. And the result was impressive: their method correctly identified 93% of burnout cases.

To achieve this result, the team first classified the collected text extracts. Those in threads about burnout were manually categorized, to exclude those where burnout referred to something else. Texts from other threads, not related to mental health, have been labeled as not related to burnout.

Based on these examples, she trained several models. Each used different configurations to determine whether a text (never seen by the model) contained indications of burnout or not. These models were then pooled as part of the diagnostic method, which therefore proved to be very effective.

However, the Swiss team specifies that the collaboration of AI experts and medical experts remains essential to verify the conclusions of this study on real cases of burnout and on a representative sample of the population.

We want to say thanks to the author of this article for this remarkable content

Burn-out: artificial intelligence could spot the first signs more easily | Engineering Techniques


Visit our social media profiles along with other related pageshttps://www.ai-magazine.com/related-pages/