Does our attention decrease when we do a monotonous but attention-demanding task for a long time at a time? Not necessarily, a study shows. In experiments with 360 radiologists who were asked to examine numerous mammograms in a row for signs of breast cancer, they found that their accuracy did not decrease over time. On the contrary: the number of false-positive results actually decreased. According to the researchers, the results suggest that experts should be able to freely organize their working and break times for such tasks, instead of giving them fixed guidelines as was previously the case.
Millions of women have mammography screenings every year to check for early signs of breast cancer. In around one in a thousand women, the radiologists who evaluate the images miss pathological tissue in the breast. At the same time, almost one in ten women receives a false-positive result and has to undergo further examinations for clarification. To improve accuracy, many facilities require their radiology staff not to review images for too long at a time. Previous laboratory studies have suggested that human attention declines rapidly after about 30 to 45 minutes of focused but monotonous work.
Research in a real environment
“It has been assumed for over 70 years that as search tasks take longer, people make more errors in recognition and take longer,” writes a team led by Sian Taylor-Phillips from the University of Warwick in Great Britain. “However, the studies on this were almost all carried out in the laboratory and the transferability to real scenarios is questionable.” Nevertheless, the results have been reflected in guidelines for the division of work for numerous professional groups, including airport staff in baggage inspection, military forces in satellite monitoring, and radiologists X-ray or MRI scans found.
To test whether the assumption of declining attention actually applies in a real-world context, Taylor-Phillips and her team evaluated the results of 360 radiologists who examined more than a million mammograms for signs of breast cancer. The researchers took into account how long the medical experts had been evaluating images, how long they needed each image, when they took breaks and how accurate their assessments were. The participants examined up to 200 mammography images in a row for up to two hours.
Fewer false-positive findings
The result: “We observed no signs of decreasing attention. Working for more than an hour without a break resulted in an improvement in overall accuracy rather than a deterioration,” the team reports. “This was due to a reduction in false positives, i.e. false alarms in which a woman without cancer is incorrectly ordered for further testing, causing her anxiety and consuming significant resources.” However, the rate of missed signs of cancer increased not. After a break, the number of false-positive findings increased again, with shorter breaks having a weaker effect than longer ones.
The results indicate that the radiologists’ attention does not wane, but that over time they set a higher threshold at which point they classify an abnormality in the breast tissue as so suspicious that they summon the patient in question for further examinations. Taylor-Phillips and her team also found another benefit of longer periods of uninterrupted work: “In addition to the experts becoming more accurate, the more mammograms they had examined since their last break, the faster they made their decisions,” they report .
More accurate and faster
From the point of view of Taylor-Phillips and her colleagues, the results speak in favor of giving radiology specialists the opportunity to determine the duration of the uninterrupted work phases during the diagnosis themselves. Avoiding mandatory regular breaks could not only make work more effective but also more precise and save many patients unnecessary follow-up examinations.
According to the researchers, it is unclear whether the results can also be transferred to other specialist areas, for example security staff at the airport. In order to clarify this question, they recommend further large-scale practical studies in the relevant areas. Additionally, further research could help elucidate the underlying psychological mechanisms in decision-making.
Source: Sian Taylor-Phillips (University of Warwick, UK) et al., Proceedings of the National Academy of Sciences, doi: 10.1073/pnas.2309576121