On a Monday morning in what turned out to be my last stretch of full-time primary care, I was handling medication refill requests for my clinic, seven clinicians including myself. About 200 requests, a typical Monday. It’s a lot, but the kind of thing you can build a rhythm around if everything is set up well.
Part of that workflow involves our medical assistants forwarding a block of structured text to aid in refill decision making. Ours was bad. Not just unhelpful, but working against us. A 60 line wall of text with outdated dropdown options, information that wasn’t highlighted when it was relevant so it was unscannable, and a comment field buried so far below the scroll, and so rarely used, that nobody checked it anymore. The information I’d need to speed up a decision was rarely, if ever, included, and it took about 20 seconds to pull manually. Across 200 refills, that’s over an extra hour of work, every week, that broke up clinical flow because of a data entry problem.
So I rebuilt it. Two lines that would populate when that block of text was inserted, with just the things that matter. Everything else our EHR already showed automatically, or it was irrelevant. I brought it to the clinical operations manager and the medical assistant lead. They understood, but they couldn’t implement it. They were worried that the old format was required documentation somewhere, but didn’t know who to ask to make changes, or even confirm their suspicion. They didn’t want their teams to get in trouble for not following a policy they couldn’t even identify. Nothing more to say.
Real problem. Viable solution. No path forward because nobody felt they had permission. I went back to my desk and stared into space for five minutes, wondering what I was doing with my life. Then I went back to the refills.
But after that, I stopped absorbing those 20 extra seconds over and over again. Those details that told me whether someone needed follow-up before automated flags kicked in. On paper that was a dozen medical assistants’ jobs, but in practice it was optional, nobody owned enforcement, and it went undone most of the time. So I had been doing it myself, quietly. Teaching moments throughout had not changed that, because the problem was never the people. The workflow was built so that skipping the step cost nothing for everybody else.
After that Monday, I stopped doing their half of it. I just refilled, based on what was in front of me, and moved on.
I told myself the harm was minimal. It was not. Some of those patients went longer without follow-up than they should have. Continuity and close follow-up are what make primary care worth anything. They are how a clinician builds a real understanding of a patient and a relationship the patient can rely on, instead of running a refill and referral farm. I had been personally funding that work because the system had stopped. That is the part that took me a while to see clearly. I had not lowered the standard that day. I’d finally stopped exceeding the one the clinic defended for everyone else by design.
It wasn’t that a spring broke on that Monday. It had been unwinding for years across different organizations that shared one pattern. I could see what was broken, design a fix, and occasionally be right. But fixing it required authority I didn’t have, incentives I couldn’t change, and adoption I couldn’t enforce. So I absorbed the pressure, the way you do when you care more about the work than about what the work is costing you.
Primary care now optimizes scheduling for availability and utilization over continuity. These metrics are far easier to track. As long as your patient satisfaction rating is high enough, and you’re meeting your productivity goals, you’re considered good at what you do. It doesn’t matter if you’re maintaining continuity, documenting clearly, or practicing good medicine. Each time I raised this concern, in one form or another, I was told that this is the reality and maybe primary care just wasn’t for me.
That response is significantly cheaper, in effort and in dollars, than redesigning the scheduling template.
I’ve done medicine in homeless encampments out of a van with a laptop and a backpack. In tents in disaster zones with a medication chest and a few cots. Both felt right in a way the clinic never did. It took a long time to figure out why, because the clinical work was the same. The dysfunction was too. It’s not like street medicine runs on a clean system.
The difference is acknowledgment. In the clinic, the dysfunction is officially invisible. In the van and the tent, the dysfunction is the explicit task. The work was figuring out what resources exist and what process needs to exist because nothing does yet. I was the whole clinic, so there was no organizational layer between me and the problem. Nobody was pretending the constraints weren’t there. Nobody expected me to absorb them quietly and not have it show up in how I worked.
In a clinic, you’re expected to keep delivering the same quality of care regardless of whether the scheduling system undermines continuity, whether prior authorization adds two hours of unpaid work to your week, whether the walls of text your MAs are required to use are slowing you down. The pressure is real. The company line is that it doesn’t exist or doesn’t matter.
The Monday of the refills wasn’t the moment the spring broke. It was the moment I noticed it had already broken. I’d stopped holding myself to my own standard. I was now doing what everyone else had learned to do. That’s what absorbed pressure actually produces: not burnout, but a slow erosion of the standard I was trying to hold.
Once I understood that, I knew I was done. Not because the problem was unsolvable. Acknowledgment would mean a clinic that builds continuity into its scheduling structure, sane panel sizes, patients routed to their own PCPs, and protected time for the work that does not show up on a productivity report. They chose the alternative: asking clinicians to supply all of it for free.