After grounding me, my primary care physician (“PCP”) explained that my next trip anywhere would be to a hospital emergency room. I called my wife, Kit, to let her know what was happening and that I would be waiting near our front door where she should pick me up and drive to the ER. As she arrived, I received another call from my PCP, telling me that the sugar level in my blood was highly elevated, but he wasn’t sure how that related to my likely internal bleeding.

My wife knew the answer to that one. “Pancreas,” she said when I relayed the information to her in the car. It would be another two weeks before any doctor made the connection.

As I relate my story, I won’t name any individuals involved in my care. That’s because every medical person whom I encountered has been caring and conscientious. They are truly trying to help patients.

But the constraints within which those individuals operate — and the manner in which those constraints inhibit the delivery of patient care — are not unique. Rather, they are pervasive throughout America’s medical system. I know this, in part, because I’ve been researching that system. In many of the same ways that an undue emphasis on short-term metrics has become a central contributor to the legal profession’s most unfortunate trends (as I describe at length in The Lawyer Bubble – A Profession in Crisis), similar short-termism infects the medical profession, too.

Stated simply, behavior follows incentive structures. If you sense problems with what happens as I make my way through the very top tier of our medical system at one of the finest facilities in the world, remember that those problems are endemic. That’s a good reason to consider them carefully.

My first hospital admission ever happened on January 28, 2015. Everyone assumed that I had a lower GI bleed. For that, you get a colonoscopy, locate the bleed, and cauterize it. In my case, they didn’t find any bleed. So they did an endoscopy, too. Again, no sign of a source of the bleeding.

All of that was okay. Often, tiny blood vessels in the bowel seal themselves. One other noteworthy item would become important later: during the preparation process for the colonoscopy — and there’s definitely no reason to go into details about that — my hemoglobin level (a proxy for the volume of blood in the body) dropped dramatically. I required an infusion of two units of blood.

Three nights after walking into the ER, I went home. Things returned to normal for a few days as I watched the Super Bowl in my living room. On the following Tuesday and Thursday, I taught my undergraduate course. But by the end of my Thursday class, I was noticeably fatigued, again.

By Saturday morning, I decided on another trip to the ER to check my blood levels. I contacted my PCP; he agreed. I also contacted the GI physician who had done my colonoscopy. His “Fellow” — a kind of junior specialist in her late 20s or early 30s — returned my call and said she’d meet me in the ER. Blood tests revealed that the GI bleed persisted.

Meanwhile, my PCP remained active in his efforts to monitor my situation. But the days of his involvement in my care were numbered. To be sure, specialists with expertise in diagnosing and treating problems like mine are vital. It’s appropriate for PCP’s to defer to them. But once you’re in the hospital, the PCP has little or no influence over what happens to you.

The resulting lack of continuity in medical care would become a central theme of my hospital experiences. Anyone who does not actively manage whatever doctors in a hospital want to do — and fails to make sure that the physicians at the top of the food chain are intimately involved in diagnostic and treatment decisions — risks poor outcomes.

The attending ER physician said my low blood levels required hospital admission. Later, the GI Fellow saw me in the ER. After asking questions about symptoms, she voiced her conclusion:

“We’ll give you another colonoscopy.”

Confirmation bias had clouded her judgment. She had locked into a particular version of what was happening to me. I watched her recharacterize my description of symptoms — which suggested an upper GI bleed — to fit her minds’s model of a lower GI bleed.

Aware that the colonoscopy preparation process had caused my blood levels to crash during my previous hospital stay, I was pretty sure that another such procedure would not solve my problems. But now was not the time or place to start that discussion. The GI Fellow would be unwilling to hear whatever I had to say on the subject anyway.

On a gurney, I returned to the same inpatient floor that I had left a week earlier. Except for a third-year medical student, there would be no continuity in my care there.