Bias
This Simple Cognitive Bias Can Have Serious Consequences
Beware of the "implicit left-digit bias."
Posted June 17, 2022 Reviewed by Tyler Woods
Key points
- When interpreting numbers, we implicitly over-value the left most digit.
- Left-digit bias can lead to flawed decision-making in economics and in medicine.
- It is important to recognize when a left-digit difference might be biasing a decision.
Recently, I became acutely aware and impacted by left-digit bias. At first, it seemed benign—a few letters and a freebie from the AARP—but then the discriminatory communications began to roll in: "Schedule your screening colonoscopy!" "Check your lipid panel!" "Time for the shingles vaccine!"
Suddenly, and in the course of a single day, a healthy, middle-aged male had transformed into an at-risk senior. But what had actually changed? A left digit had rolled forward. If you’ve never heard of the risk of implicit left-digit bias, I am here to tell you that it is real, and you could be at risk.
As it turns out, southpaws need take no offense: left-digit bias has nothing at all to do with handedness. Rather, this bias is one that we are all familiar with, whether we realize it or not. It is the tendency to categorize continuous variables based on the left-most digit. For example, our minds trick us into thinking items costing $79.99 are a better deal than those costing $80.00 because when it comes to buying something, a lead “7” is obviously cheaper than “8”. This bias can also be seen steering prices in the used car market, which happens to be booming these days. Prices drop significantly more when the odometer clicks from 79,999 miles to 80,000 than when it advances from 79,899 to 79,900.
Some very sharp analysts crunched numbers from over 22 million used-car transactions (Sydnor, 2011). They found that cars with odometer values between 80,000 and 80,100 miles sold for approximately $210 less than cars with odometer values between 79,900 and 79,999 miles. The drop in price was far less (only 10 bucks) when the odometer moved up from a similarly sized increase (from 79,800-79,899 to 79,900-79,999), but without changing the critical left digit. Left digits drive prices; we intuitively know that. What we may not recognize is that they also influence medical decisions.
Consider this study, recently published in the New England Journal of Medicine that evaluated the rate of coronary artery bypass grafting (CABG) in patients hospitalized with acute myocardial infraction (Olenski, 2020). Knowing CABG is a high-risk procedure, cardiologists and surgeons might think it less suitable for “older” patients. But what birthday counts for being “older”?
The study compared CABG rates between those who had just celebrated their 80th birthday (<2 weeks) with those whose 80th birthday was just around the corner (<2 weeks). The two groups were virtually the same age and were identical in terms of baseline characteristics. You probably can guess what the investigators found. The “older” patients (perceived as being in their 80s) were less likely to receive a CABG than their “younger” counterparts (perceived as being in their 70s), even though the age gap (< 4 weeks) was negligible: 5.3 percent (265 of 5,036 patients) vs. 7.0 percent (308 of 4,426 patients): P<0.001.
Were there similar differences in CABG rates bordering other birthdays, like the 77th, 78th, 79th, 81st, 82nd, or 83rd? None at all. That left digit makes a considerable difference.
And the left doesn’t get it right evidence doesn’t stop there. In a large cohort (105,387) of potential organ donors between 2010 and 2019, the probability of a donor organ being placed dropped precipitously, more than 5 percent from donor age of 69 to 70 (Jacobsen, 2022).
While left-digit bias has not been well-studied in emergency department practice, one can imagine how patients before and after a decade-altering birthday might be impacted. Might a 90-year-old be more likely to receive a palliative care referral than an 89-year-old? Or have a less aggressive diagnostic work-up? Or insufficient analgesia? Or have a surgeon recommend conservative versus aggressive management?
There seems to be a popular trend in the wellness community, and those who peddle and profit from it, to sell assessments of “biologic” age. In fact, it is quite easy to find online calculators that will figure this out for you. While such forecasts seem to be back-of-the-envelope calculations based on some broad epidemiological evidence (smoking and obesity decrease life expectancy), they do have face validity. We know intuitively that, health-wise, some people are “old” 60-year-olds and some are “young” 70-year-olds.
Maybe our hospital record demographics should have two ages entered—both chronological and biologic. This may help us temper any implicit left-digit bias.
This recent victim of insidious leftness took solace in the fact that the good people at “easycalculation.com” considered his “biologic” age on the right side (that is, younger side) of 50. Looks like there will be an opportunity to plan another big birthday party.
References
Sydnor, Justin R., Devin G. Pope, and Nicola Lacetera. "Heuristic Thinking and Limited Attention in the Car Market." 2011 Meeting Papers. No. 105. Society for Economic Dynamics, 2011. https://ideas.repec.org/p/nbr/nberwo/17030.html
Olenski, Andrew R., et al. "Behavioral heuristics in coronary-artery bypass graft surgery." New England Journal of Medicine 382.8 (2020): 778-779. https://www.nejm.org/doi/full/10.1056/NEJMc1911289
Jacobson, Clare E., et al. "Left digit bias in selection and acceptance of deceased donor organs." The American Journal of Surgery (2022). https://www.sciencedirect.com/science/article/abs/pii/S0002961022002173