The last post was about catching every patient who reaches out. This one is about the harder thing that comes right after: deciding who you actually take on. Not everyone who contacts a practice is a good fit for it, and for a solo practice in particular, saying yes to the wrong patient is a mistake that costs everyone, the patient most of all. This post is about eligibility, screening, and the underrated skill of saying no well.

It took me a while to understand that "no" is a feature, not a failure. Early on, every inquiry felt like something to convert, and turning someone away felt like leaving care, or money, on the table. What I learned is that a wrong yes is far more expensive than a clean no, and that the practices that run well are not the ones that take everyone. They're the ones that are good at matching the patient to the right place, which sometimes is you and often isn't.

Why "no" costs a solo practice more

A large practice can absorb a bad fit. There's slack, there are other clinicians, there's a front office to route around the problem. A solo practice has none of that slack. My time is the entire constraint, and a wrong-fit patient does not consume an average amount of it. They consume a wildly disproportionate amount, in messages, in visits that don't go anywhere useful, in worry. Every hour that goes to a patient I can't actually help is an hour taken from a patient I can, and there is no bench to pick up the difference.

So screening isn't gatekeeping for its own sake, and it isn't about being choosy. It's the mechanism that protects the thing the whole practice runs on, which is one person's finite attention, and it protects the patients who are a good fit by making sure their doctor isn't buried under the ones who aren't.

There are two different "no"s

The word "no" hides two very different decisions, and conflating them is where practices get screening wrong.

The first is a clinical no. This person is not a safe or appropriate candidate for what the practice offers, because of a contraindication, or because they need a level of care beyond what a telehealth practice can responsibly provide, or because what they need is simply not what this practice does. This no is non-negotiable and the single most important thing screening exists to catch, because getting it wrong isn't an inconvenience, it's a safety problem. The right answer here is not "we're full," it's "you need something different, and here's where to look."

The second is a fit no. This person could be served safely, but they're not the right match, most often because of where they live and where I'm licensed, or because their expectations and what the practice actually does don't line up. This no is softer and its handling is different: the goal is to redirect kindly, not to screen for safety.

Keeping these two separate matters because they need different rigor. The clinical no has to be caught every single time, with no exceptions on a busy day. The fit no can be handled with more grace and more flexibility. Building both into the same undifferentiated "eligibility" bucket is how the important one gets missed.

Screen before the visit, not during

The core principle is that the clear no's should be caught structurally, up front, before anyone's time is spent, so the actual visit slots go to people the practice can help.

That's what the eligibility intake is for. It asks the questions that determine fit, and it asks the ones that would rule someone out early enough to give a fast, honest answer rather than discovering the mismatch halfway through a first visit. A structured intake does this consistently in a way a human never quite can: it asks every applicant the same screening questions in the same order, so no one slips through because it was a busy afternoon and the question didn't get asked.

There's a real design tension here that I still think about. Front-loading the screening questions, before I've even asked for a name, is efficient for catching a no fast, but it can feel cold, like being interrogated before being greeted. Leading with the friendly, identifying questions feels more human but means collecting a full profile from people who were never going to be eligible. I don't think there's a perfect answer. You decide which failure you'd rather have, and you make the screening itself feel as much like care and as little like a bouncer as you can.

Saying no kindly

A no should almost never be a dead end. The difference between screening that feels like care and screening that feels like rejection is entirely in what happens after the no.

For the person who might be a fit later, but not now, there's an automated sequence that keeps the door open: useful information, a clear path back when their situation changes, a relationship that doesn't end just because the timing was wrong. For the person who isn't a fit at all, the right thing is a genuine redirect, pointing them toward the kind of care or the kind of provider that actually matches what they need, so they leave with a next step instead of a closed door.

This is the part most screening gets wrong. It treats "not eligible" as the end of the interaction, and the person walks away feeling turned away rather than helped. Done well, even the people you say no to should feel like the practice did right by them, because it did: it told them the truth quickly and pointed them somewhere useful, which is worth more than a yes that was never going to work.

The false economy of the wrong yes

It's worth being concrete about what a wrong yes actually costs, because the temptation to say yes is strongest exactly when you shouldn't.

A wrong-fit patient rarely just quietly doesn't work out. They generate more messages, more visits that circle without resolving, more after-hours worry, and more risk, and they do it while occupying a slot a good-fit patient needed. One wrong yes can distort an entire week. And it usually ends the way the mismatch predicted at the start, except now with more time invested, more frustration on both sides, and a patient who is further from the care they actually needed than when they first reached out. Taking everyone feels generous and is actually the opposite: it's a slower, more painful way of failing the people you couldn't help, while shortchanging the ones you could.

Screening is a safety function

I want to be clear that the most important reason to screen well is not to protect my time. It's patient safety.

The eligibility gate is where you catch the person for whom the treatment is not appropriate, or who needs a higher level of care than a telehealth practice can give, or who is in a situation that calls for a different kind of help entirely. Missing that is the kind of mistake that matters, and it's exactly the kind that gets more likely when a tired human is doing the screening at the end of a long day. Software doesn't get tired. A structured screen applies the same criteria to everyone, every time, and it flags the cases that need a human's careful judgment instead of quietly letting them through. That consistency is a safety feature, and it's the part of screening I would never automate away and never cut corners on.

Why this matters for a solo practice

Saying no well is one of the highest-leverage things a solo practice does, and it's almost never talked about, because it feels like the opposite of growth. It isn't. It's what makes the yeses count. Every clean no protects the finite attention the good-fit patients depend on, catches the safety cases that must be caught, and, done kindly, still leaves the person better off than a false yes would have.

The software's role is to make the no consistent, early, and humane: consistent so the safety cases are never missed, early so no one's time is wasted, and humane so the answer, even when it's no, feels like care. A solo practice can't be everything to everyone. The skill is being honest about that in a way that serves the people you turn away almost as well as the people you keep.

Next time

Once you've said yes to the right patients, you run into the next problem: getting them to actually show up. The next post is about scheduling and the no-show, the quiet tax on every practice, and the mix of automation and expectation-setting that turns a booked slot into a kept one.