Theravive Home

The more hours worked in a stressful job, the greater the risk of depression. 

A study of doctors published in the New England Journal of Medicine found that medical residents who worked 90 or more hours a week experienced changes in their depression symptom scores that were three times greater than the change experienced by those working 40 to 45 hours a week. 

The researchers found a “dose response” effect was found between hours of work and depressions symptoms. The average symptom increase was 1.8 points on a scale for those working a 40 to 45 hour week, compared with up to 5.2 points for those who worked more than 90 hours. 

“That means the increase in depressive symptoms has a positive correlation with the number of work hours, not only in extremely high work hour range, but across the span from lower work hours to higher work hours,” Yu Fang, M.S.E., lead author of the study and a research specialist at the Michigan Neuroscience Institute told Theravive. 

In undertaking their study, the researchers examined data on over 17 thousand first year medical residents. The data came from the Eisenberg Family Depression Center and the Intern Health study, and spanned 11 years. 

The data follows recently graduated doctors to track their depressive symptoms, sleep and their work hours during their intern year (the first year of residency). 

The researchers utilized a study design known as an emulated clinical trial, which mimics a randomizes clinical trial in circumstances in which conducting a real trial is not possible. 

“From our data we have always observed the strong correlation between long work hour and depression risk. The novel method of emulated target trial, which could simulate a clinical randomized trial among observational data, provides a way to consolidate the statistical evidence. It is important to use this information to find ways to improve training physicians' well-being, which is beneficial to both the physician themselves and the quality of care they deliver to the patients,” Fang said. 

Of the young doctors studied, the average age was 27 and just a little over half were women. One in five were in training for surgical disciplines of medicine and 18% were from an ethnic or racial group that traditionally have been underrepresented in the field of medicine. 

At the beginning of the intern year, less than 1 in 20 of the doctors met criteria for moderate to severe depression.

The most commonly reported work hours by the interns were between 65 to 80 hours a week. 

The researchers found that there was higher percentage of interns who met the criteria for a diagnosis of moderate to severe depression in those who worked longer hours, compared with those who worked fewer hours. 

Rates of depression among physicians are high, and national organizations like the National Academy of Medicine and the Association of American Medical Colleges are attempting to address this. 

The Accreditation Council for Graduate Medical Education sets an 80 hour limit on the number of hours a resident can work, but this can be averaged over a four week period and can be subject to exceptions in some cases. 

The same council limits the length of a single shift a resident can work and the number of continuous days a resident can work. This has had mixed results in reducing rates of physician depression. 

The researchers of the study published in the New England Journal of Medicine argue their research emphasises the need to continue to reduce the number of hours young doctors work.

“Our study suggests that implementing strategies to reduce work hours, for example, by increasing work efficiency, reducing unnecessary work burdens, would be beneficial to training physician's well-being,” Fang said.

But the study authors anticipate it is not only doctors who would see an increased risk of depression with number of hours worked. They say a similar effect could be seen in other professions, particularly those with long hours. 

“Parallel studies should be conducted in these professions to further prove the generalizability of these results,” Fang said.