Paul Alper points us to this op-ed, “Don’t Visit Your Doctor in the Afternoon,” by Jeffrey Linder:
According to the study, published in JAMA Network Open, doctors ordered fewer breast and colon cancer screenings for patients later in the day, compared to first thing in the morning. All the patients were due for screening, but ordering rates were highest for patients with appointments around 8 a.m. By the end of the afternoon, the rates were 10 percent to 15 percent lower. The probable reasons? Running late and decision fatigue.
The research article, “Association of Primary Care Clinic Appointment Time With Clinician Ordering and Patient Completion of Breast and Colorectal Cancer Screening,” by Esther Hsiang, Shivan Mehta, Dylan Small, et al., says:
Results Among the 19 254 patients eligible for breast cancer screening, the mean (SD) age was 60.2 (6.9) years; 19 254 (100%) were female, 11 682 (60.7%) were white, and 5495 (28.5%) were black. Screening test order rates were highest at 8 am at 63.7%, decreased throughout the morning to 48.7% at 11 am, increased to 56.2% at noon, and then decreased to 47.8% at 5 pm (adjusted odds ratio [OR] for overall trend, 0.94; 95% CI, 0.93-0.96; P < .001). Trends in screening test completion rates were similar beginning at 33.2% at 8 am and decreasing to 17.8% at 5 pm (adjusted OR, 0.95; 95% CI, 0.94-0.97; P < .001). Among the 33 468 patients eligible for colorectal cancer screening, the mean (SD) age was 59.6 (7.4) years; 18 672 (55.8%) were female, 22 157 (66.2%) were white, and 7296 (21.8%) were black. Screening test order rates were 36.5% at 8 am, decreased to 31.3% by 11 am, increased at noon to 34.4%, and then decreased to 23.4% at 5 pm (adjusted OR, 0.94; 95% CI, 0.93-0.95; P < .001). Trends in screening test completion rates were similar beginning at 28.0% at 8 am and decreasing to 17.8% at 5 pm (adjusted OR, 0.97; 95% CI, 0.96-0.98; P < .001). Conclusions and Relevance Clinician ordering of cancer screening tests significantly decreased as the clinic day progressed. Patient completion of cancer screening tests within 1 year of the visit was also lower as the primary care appointment time was later in the day. Future interventions targeting improvements in cancer screening should consider how time of day may influence these behaviors.
All those p-values are just horrible—I suppose JAMA made them do it—but the findings are clear enough in these graphs:
Alper writes:
I am sure that your bloggers will draw differing morals from this data. Those of us who view (over) screening suspiciously will opt for later appointments.
Continuing on this theme, perhaps the topic is not “decision fatigue” in the afternoon but an itchy trigger finger in the morning. Why be so sure that the rate of screening in the morning is the appropriate baseline?
Two guesses:
– Some doctors might be taking more difficult cases in the morning. It makes sense to schedule this way and I wold be surprised if it is not already a standard practice.
– “Subject selection bias”. Something along the lines, “If it is only a routine check, i do it in the afternoon, but if I experience severe symptoms, I do it asap”.
I am reminded of those Israeli judges who reviewed the cases of pro-se defendants right before they went on lunch-break, leading to papers claiming they were harsh because they were hungry.
Standard medical practice is that people who get sick overnight or who have been getting sick or been sick tend to come into the office in the morning, And there’s another burst after lunch. So one obvious question is the extent to which this is true of cancer screenings. Or rather cases that lead to cancer screenings. Or even the ‘bias’ that doctors might think a case early is likely to be worse because other early cases tend to be sicker. I’d like to see how many cases are needed to cut the gap to 0. But I’d want to understand something of the typical case that gets referred and why. It’s fairly easy to think about that for colon cancer, less easy for me to understand breast screening because I would expect that’s near mandatory by age and history. It would help to compare how many screenings were positive because then you could say, for example, that it looks more like worse cases come in early. If the rate is the same, then I’d wonder if the screenings weren’t part of a larger diagnosis process – or even if it isn’t ‘defensive medicine’ protecting against the odd chance and thus the malpractice suit. But otherwise, like you say, how do you know?
I guess my background matters. I grew up in a medical family. If you call a doctor, they’ll ask what’s wrong and say come in now if you’re really sick and come in after lunch if you’re sick but can wait. That allows the day to flow without the office being bogged down by all your sick patients coming in at once, a sort of basic triage around which the usual stuff flows. But I have no idea how that might affect breast screenings.
“All those p-values are just horrible—I suppose JAMA made them do it—but the findings are clear enough in these graphs:”
Well they showed the graphs too (and CIs) so no big deal. It is good to have some actual number because “clear enough” isn’t well-defined, or what is “clear enough” to some might not be to others.
Justin
Justin:
See Zad’s comment below. The problem with the p-values isn’t so much with the p-values—they’re a waste of time but, hey, some people might like them!—but rather that they come with a dichotomization of results into “statistically significant” or “not statistically significant” (or a trichotomization, “highly statistically significant,” “marginally significant,” “not significant”) which is equivalent to adding a huge layer of noise to already-noisy data.
JAMA also makes authors ignore any actual differences if P isn’t magically below 5%. Or maybe it just attracts all the folks who prefer to ignore the differences.
https://jamanetwork.com/journals/jamasurgery/fullarticle/2752842
“A total of 60 009 patients (mean [SD] age, 62.8 [13.9] years; 32 139 [53.6%] male) were included in the study. The overall SSI incidence for clean wounds was 0.87% before policy implementation and 0.83% after policy implementation, which was not found to be significant (odds ratio [OR], 0.96; 95% CI, 0.80-1.14; P = .61). After accounting for possible confounding variables, a multivariable analysis demonstrated no significant reduction in SSIs (OR, 0.85; 95% CI, 0.71-1.01; P = .07). During the postintervention study period (26 months), a total of 2 010 040 jackets were purchased, which amounted to a cost of $1 709 898.46.”
Nice circular reasoning. Why do doctors order fewer screenings? Because of decision fatigue. How do you know that you observed decision fatigue? Because doctors order fewer screenings.
The New York Times piece by Jeffrey A. Linder shows an additional example: Why do doctors prescribe fewer unnecessary antibiotic prescriptions for respiratory infections first thing in the morning, while unnecessary prescriptions gradually increase over the day? Because of decision fatigue in the afternoon. How do you know that you observed decision fatigue? Because preescriptions increase over the day.
So, any increase or decrease of a behaviour during the day can be “explained” by decision fatigue; and decision fatigue can be demonstrated by any change that takes place?
Such circular explanations might sometimes impede a more thorough thinking about what is going on.
“Why be so sure that the rate of screening in the morning is the appropriate baseline?”
Badabing! Lots of times I start out very carefully while working on something but after a while I realize I was being more careful than necessary and I speed up – usually without negative consequences.
Also my experience with physicians is that they have a strong inclination to order unnecessary treatment. My health care cost makes the same argument.