Increasing workloads in radiology: Does it matter?

Many factors contribute to missed diagnoses. Some of these are unique to our specialty, like poor image quality, deficient interpretative skills, suboptimal ergonomics, incomplete review of all images, and delayed or absent communication of findings to direct caregivers, among many others. Another factor that likely has an incredibly negative impact on interpretation quality and accuracy, but essentially goes ignored, is our workload.

Imagine your job this week is to move one brick a day from one spot to another spot 10 feet away. Next, imagine that next week your job will be to move 5 bricks a day from one spot to another spot 10 feet away. Now imagine moving 50 bricks the following week, then 200 the week after that, and so on. Eventually, you are going to drop bricks, get stress fractures, sprain your ankle, crush your toe, sustain heat stroke, or suffer some other painful consequences. As long as you continue to handle the growing workload, more work will be required. It is a fact of life: Good work is rewarded with more work.

I don’t know about you, but I don’t recall ever facing a decreasing workload in my career. There are myriad reasons why radiologist workloads continue to grow: Trying to maintain one’s income; ever-growing numbers of patients, given an aging, generally sicker population; greater paranoia about malpractice claims; and more image-intensive studies, are just a few among many other factors. Most of us have no way to distribute the workload evenly over a given time period. Working in an office with a fixed number of studies per day perhaps is more controllable (notwithstanding the usual add-ons, of course) and less “psycho” than working in a large emergency department where there is no control over study volume, except possibly those limitations imposed by the rate that studies can be performed. Also, in the ED the radiologist is compelled to provide results within rigid time limits. I do not for a moment mean to imply that most other physicians do not have similar workload stresses.

Why would anyone consider it plausible that interpreting more cases per unit of time will not negatively impact diagnostic accuracy? It’s like any other job: Unrealistic time pressure leads to shoddy performance, regardless of one’s best intentions. If you were to give a heart transplant team just 20 minutes to perform the surgery, you would get some hostile pushback. I certainly wouldn’t sell their patient life insurance before the operation.

I exaggerate to make an important point: Has any study correlated radiologist workload with interpretive accuracy in a scientifically valid manner? Does accuracy track inversely with workload and, if so, to what extent? How do you measure workload? How do you best measure interpretive accuracy?

Most likely, measures of workload are made on the basis of RVUs. These do reflect, in some fashion, the degree of intensity or effort required to perform a given study or procedure. But this parameter misses so much additional important information not reflected in RVUs. In emergency radiology, for example, I spend huge amounts of time going over cases with clinicians, answering the telephone, trying to find clinicians who know the patient and are willing to take responsibility for calls, figuring out which patient is STAT and which is SUPER-STAT for MRI, settling fights between technologists and residents (usually not radiology residents), and being constantly interrupted by a zillion other matters. OK, I am exaggerating—just half a zillion. Radiologists in all practices have their own excessive burdens, which they carry while just trying to do their main job of interpreting studies.

I submit that we radiologists are walking a thin line that is getting thinner all the time. Far better comprehension of how to quantify workload and its impact on diagnostic performance is absolutely necessary. How do you know when you have reached your personal maximum work limit—when you faint at the workstation, beat your CT technologist with a stick, or beg her to take a really long dinner break? When you find yourself praying that the incoming radiologist doesn’t notice 30 unread CT scans, or you start dictating in tongues, or describing “really pretty adrenal glands”?

What we do not want as our sign of being on overload is missing important findings that adversely affect our patients. One can make processes efficient only up to a point; pushing beyond that point invites error. There are no absolute criteria for establishing safe levels of demand on radiologists’ productivity; these levels vary among individuals for any number of reasons. Implementing second reads is one way to catch that threshold, but they consume more time within a group. Second reads also do not address the added physical and psychological stress that can bring harm on already hard-pressed radiologists.

In the past, radiology was considered a plush career choice. But our role in patient care has grown tremendously, and our workload, both in numbers of studies and duration of real-time coverage, continues to increase rapidly. Just banging out as many studies as possible may seem like an attractive option when it directly influences your income, but the quality of the product should never take a back seat to that goal.

A valid methodology must be developed to measure workload against interpretation accuracy and serve as a fundamental pillar of quality measurement.