Innovation
Text: Patricia Michaud

Burnout and Big Data

Burnout is on the rise but remains difficult to diagnose. A research team from Bern is devising a method to identify the disease using artificial intelligence.

In Switzerland, nearly three in ten employees are in a state of what is called critical stress. That means these employees experience more work-related stress than they have the resources to cope with, according to the latest edition of the Job Stress Index*. Initial observations indicate that the share of the working population experiencing stress – which was one in four in 2014 – likely increased during the pandemic. The study estimates that workplace stress costs 7.6 billion Swiss francs a year. On top of that, the amount of time off work has exploded in recent years. In 2020, insurers Swica and PR Rück compiled statistics showing that workplace absences due to burnout or depression have risen 50% since 2012.

Is burnout the disease of the 21st century? That is at least what is suggested by the title of the book published by psychologists Anny Wahlen and Nadia Droz**. This epidemic continues to spread in our age of companies go digital, multitasking becomes the norm, and the pace of work accelerates. Unfortunately, detection of burnout has not advanced as fast. This is mainly due to
the difficulty in diagnosing it, at least based on symptoms alone, Nadia Droz says.

“People experiencing burnout have symptoms that can be emotional, cognitive, behavioural, social or psychosomatic. I haven’t yet seen two people with exactly the same symptoms” says the Lausanne-based practitioner, who leads burnout prevention courses at CHUV.

The most common symptoms of the disease include headaches, sleep disorders, reduced self-confidence or gastrointestinal problems.

Automatic language processing

Experts say that detecting burnout – defined as exhaustion due to chronic workplace stress rather than mental health issues – is a key challenge of modern society. Against that backdrop, findings from a study conducted at the Bern University of Applied Sciences and Arts and published in April 2022 in the journal Frontiers in Big Data attracted attention. Mascha Kurpicz-Briki’s research team has developed an innovative approach based on automatic language processing.

“These days, burnout is often diagnosed using psychology questionnaires with a Likert scale format, such as ‘I feel exhausted at the end of my working day: never/sometimes/every day’,” Mascha Kurpicz-Briki notes. The problem is that respondents may be tempted to skew results by not ticking the most extreme answers. “Of course, experts also use open-ended questions and more comprehensive questionnaires, analysis is time-consuming,” the project leader adds. “Our staff wondered whether artificial intelligence could help to avoid these shortcomings, or at least be applied towards detecting burnout.”

The team began using automatic language processing to sift through texts on the English-language online forum Reddit, which offers discussions by topic. “We analysed more than 13,000 anonymous excerpts, with some from discussions on burnout and some on other topics,” Mascha Kurpicz-Briki says. “As a result, the team developed a method to assess whether the language contained in texts involves a situation of burnout. And it worked. “Burnout was correctly identified in 93% of cases.”

No recognition as a disease

As promising as it is, this approach still needs to be consolidated. “The next step is to test our method in a clinical setting.” In collaboration with experts on burnout, that will involve drawing up a list of open-ended questions that will be used to survey a representative sample of the population. The method will then be tested on the answers, ideally in the national languages. However, Mascha Kurpicz-Briki warns that, “even if our findings are validated, it doesn’t mean we’ll be able to put out a tool on the market overnight that HR managers can use to detect burnout in companies with three clicks or a quick self-test for employees.” The researcher designs software for psychologists and doctors to assist them in their approach and supplement existing methods.
“In this sense, we should refer to ‘augmented intelligence’ rather than artificial intelligence, as it guides humans, but ultimately humans are responsible for the diagnosis.”

Nadia Droz confirms that “currently, the most effective method of detecting burnout is anamnesis, i.e. the medical history of the person concerned.” Nevertheless, any form of exploration – like the study from the University of Applied Sciences and Arts in Bern – that can advance diagnosis and classification “is welcome”, the psychologist says. Burnout is not considered a disease by the Swiss authorities. In June 2019, the Swiss National Council rejected a parliamentary initiative calling for coverage of burnout by the employer’s accident insurance instead of the employee’s health insurance. Meanwhile, although the WHO has not gone that far, in 2019 the organisation officially redefined burnout as a syndrome resulting from chronic workplace stress and no longer merely as a state of exhaustion.

Changing corporate culture

Unlike in other countries such as Italy, workplace burnout is not recognised as an occupational condition in Switzerland. However, officially classifying it as such has many advantages. For example, it would facilitate the reintegration process for employees with burnout, and they would feel less guilt (as responsibility for the disease would be placed on the employer). Also, by improving overall visibility of the condition (through reporting requirements), it would consequently be easier to identify companies where burnout is most common.
In addition, this recognition from accident insurance could encourage organisations to introduce prevention and early detection measures.

And that could make all the difference. The later it is treated, the longer burnout lasts. “On average, burnout cases mean three to six months off work,” Nadia Droz says. But prevention also has its price. “For it to be effective, companies would have to completely transform their corporate culture, truly listen to their employees and debrief regularly.” For example, managers should be encouraged to take action as soon as repeated complaints about workload or working hours are brought to their attention. “The problem is that this goes against the still very prevalent ‘Kleenex’ philosophy in business. Companies are happy to hire super-motivated people then later dispose of them without giving it much thought when employees become burnt out due to stress.”

Compared to the previous generation, however, society now considers burnout to be a serious issue, especially in companies, and it is taken more seriously. /

Identify burnout to prevent it

Burnout symptoms are similar to those experienced with depression or anxiety. Experts have also found that victims of burnout tend to underestimate early warning signs or try to conceal them for fear of being perceived as less resilient.

Warning signs

→ Feeling concerned about work during holidays

→ Being less productive despite higher involvement

→ Neglecting relationships with friends and family to give more attention to work

→ Difficulties concentrating

→ Memory problems

→ Difficulty falling asleep

→ Increased use of stimulants

Preventing burnout

→ Make sure primary needs are met (food, sleep, physical activity)

→ Set aside time for rest and recovery

→ Make sure that demands imposed are justified

→ Set priorities

→ Delegate tasks

→ Seek support from managers or colleagues



Share:

 
 
 

* The study was published by Promotion Santé Suisse in 2020.

** Burnout, la maladie du XXIe siècle ? (Is burnout the disease of the 21st century?) Anny Wahlen and Nadia Droz, 2018, Éditions Favre.