One morning, in the chaos of getting two physicians and three small children out the door, our ceramic sugar bowl slid off the kitchen table. Although the bowl was unscathed, a pale green chip came off the lid. Raised by an engineer with a thrifty New England ethos, I knew the damage was reparable. I tucked the missing chip and lid into a drawer for a later procedure, left the lidless bowl on the table, and rushed to carpool and clinic. That evening, my husband opened the kitchen cabinet and blinked in surprise.

The sugar bowl sat front and center. Our cleaners had placed a pale green pear directly on top of the sugar. The pear’s diameter neatly covered all visible sugar; the peak rose proudly above the bowl’s rim. I quickly explained about the morning’s mishap and the chipped lid. Well, queried my husband with an anesthesiologist’s penchant for protocols and lists, had I told the cleaners not to put a pear in the sugar bowl? I tilted my head in blank disbelief. How could I possibly warn someone not to do something it would never have occurred to me they would do?

If you dig into the way patients take medicine, there are a plethora of pears in sugar bowls. A patient on a triphasic combined oral contraceptive pill came in for irregular bleeding. She asserted that she was taking one pill every day. I carefully confirmed that she took the Sunday pill, then the Monday pill, then the Tuesday pill. Correcting me, she explained she took a Sunday pill, and then a Monday pill, but not always from the same row. Color coding aside, she picked any Tuesday pill she wanted from any of the rows. After asking other patients, I discovered her style was not unique.

Sometimes, we give clear instructions without predicting which part patients will prioritize. My selective serotonin reuptake inhibitor prescriptions say, “Take one pill at the same time each day.” My intention in this directive is to stress each day. Patients with anxiety disorders sometimes emphasize at the same time. They dutifully explain that, upon realizing at 9:15 they had missed their 9:00 dose, they did not take the pill that day. Priorities that seem self-evident to the author are not intuitive to the reader.

Too often, frustrated by non-compliance, physicians dismiss these patient mistakes as ridiculous one-offs. How can we possibly predict every error a patient will make? There are endless pears in sugar bowls; we have neither time nor energy to warn against them all. With medication compliance, the stakes are higher. A pear in the sugar bowl is irritatingly quirky, but missed oral contraceptives and selective serotonin uptake inhibitors carry significant risks.

In truth, these errors are not random. Each patient makes decisions based on a perfectly rational schema, but their schema is not transparent or easily predictable. Decades in medicine have narrowed my ability to imagine other decision trees. I have seen so many oral contraceptive pill packages that it would never occur to me to do anything but take the pills in order; I know enough pharmacology to understand that a 15-minute difference is trivial compared to a skipped dose. Unfettered by medical training, patients make decisions using logic systems divergent from ours.

The pear in the sugar bowl was not happenstance. The cleaner prioritized matching the lid’s color and covering the exposed sugar. A solution that seemed arbitrary to us had internal logic, just as patients often have thoughtful explanations for behaviors we label as non-compliant.

Even if we acknowledge that these (pears or pills) are logical choices, they are often still unpredictable. Where we have seen patterns, we can preemptively explain our thought processes. I now draw pictures of oral contraceptive pill packages with directional arrows and specifically tell patients to take their anxiety medicine at the same-ish time.

Yet the very essence of the pear in the sugar bowl is that we cannot reliably anticipate every alternative decision strategy. However, we can come home each evening and metaphorically open the cabinet by asking patients exactly how they take their medicines. We should ask patients why they take (or do not take) their medicines and understand the factors driving their decisions. Then, we can design individual solutions to align our logic and goals.

It took me several months to glue the ceramic chip back on the sugar bowl lid. Until then, I carefully covered the open sugar bowl with aluminum foil every Monday morning in preparation for the cleaners. I came home each Monday evening to a covered sugar bowl and pears resting quietly in the refrigerator.