The Cliff Problem

There is a moment in the lifecycle of every new clinician entering home health that the organization rarely sees clearly — because it happens alone, in a car, somewhere between the last day of orientation and the first solo patient visit.

The preceptor is gone. The cohort has dispersed. There is a real home, a real kitchen table, a real person who needs something. And for the first time, the new clinician is the only one there.

I've spent a long time thinking about what happens at that moment. Not because I have a clean answer to it, but because I think it might be one of the most consequential and underexamined transitions in healthcare operations — and I keep seeing a version of it surface in places that seem entirely unrelated.

We call it the cliff. And I'm not sure most organizations know they've built one.

Orientation ends well. In the organizations I've been close to, real care goes into the design of the onboarding experience — the clinical preparation, the cultural introduction, the procedural training. By completion, most new hires are genuinely ready. Technically.

What hits them at the cliff isn't clinical complexity. In my experience, that's almost never where things come apart. What lands all at once — without warning, without scaffolding, without anyone having specifically named it as the thing that would be hard — is the documentation and administrative load.

In a supervised environment, someone else catches what falls through. In the real world, nothing catches anything. The visit note, the authorization, the coding decision, the schedule coordination — all of it lands simultaneously, on a person who is also trying to be present with a patient in a home they've never been in before. That is a different kind of complexity than clinical training prepares anyone for.

Here's what I noticed, over time, about the ones who were heading toward the door.

They went quiet.

Not dramatically. Not in a way that announced itself. The person who used to send a quick message after a complicated visit stopped sending them. The person who used to flag something in a team huddle stopped flagging things. It wasn't disengagement as event — it was disengagement as drift. Behavioral withdrawal before explicit decision.

The problem with silence as a signal is that it requires relationship infrastructure to detect. You can't notice a change in someone's communication pattern if you've never established a baseline. And you can't establish a baseline if every interaction is transactional — productivity, schedule, metrics. Silence only means something when you know what the person sounded like before.

I've thought about this a lot: what would it actually take to build an organization that could hear that kind of silence? Not react to it after the resignation letter — hear it while there's still time to respond. The answer isn't technology. It's not a more sophisticated dashboard. It's deliberate, non-transactional contact at specific moments in the first 90 days. Conversations where someone is listening for texture, not output.

That's expensive in attention. I'm not sure it's expensive in anything else.

There was one signal I came to trust more than any other as an indicator of long-term retention.

When someone referred a colleague into the organization.

Not because of any formal incentive — because they wanted to. Because they called someone they knew and said, without being asked, come work here.

I think about what that act actually requires of a person. You're putting your own credibility on the line. You're making a promise, implicitly, about what someone's experience will be. Nobody does that if they have one foot out the door. Nobody does that if they're tolerating the job rather than belonging to it.

The referral isn't a loyalty metric. It's a signal of arrival — the moment when someone crosses from employee to member of something. And I've never seen anyone who had crossed that threshold leave shortly after.

I built recruiting infrastructure around that observation. Whether it's transferable elsewhere, I genuinely don't know. But I find myself wondering: what's the equivalent signal in a Medicare Advantage member population? What does arrival look like when the relationship is clinical rather than professional?

I've been paying attention to the MA industry's current conversation about new member economics. The financial framing is clear and consistent across earnings calls and analyst reports: new members are dilutive, returning members are profitable, the gap between year-one and year-three cohort performance is the core operating challenge.

What I don't hear very often is the question I keep wanting to ask, which is: what does the new member's first 90 days actually feel like from the inside?

A new member on January 1st has just navigated Annual Enrollment Period, selected a plan, and now needs to figure out — often alone — how to use it. New provider relationships, new authorization processes, new care management contacts, new benefit structures. It is, structurally, a transition moment. Not unlike orientation completion.

I'm not suggesting the situations are identical. They're not. But I find myself wondering whether the same dynamics might be operating — whether behavioral withdrawal before explicit decision might look similar, whether the distance between enrolled and genuinely activated might be wider than the data suggests, whether the plans that are compounding profitability figured something out about that transition window that others haven't.

I don't know. I'm genuinely asking.

The thing I've come to believe about retention systems — not just clinician retention, but any system designed to keep human beings engaged and present — is that the ones that work are not the ones with the most sophisticated components.

They're the ones that are built for reality rather than for the plan.

The most elegant onboarding process in the world will encounter a new hire who doesn't move through it predictably. A member who has competing priorities. A provider who doesn't engage with the platform the way the implementation assumed. A manager who is carrying too many direct reports to execute the touchpoint cadence at the designed frequency.

A brittle system breaks at those moments. A resilient one bends and recalibrates — not because someone made a judgment call, but because the design anticipated deviation and built in ways to respond to it. Escalation paths that terminate in action, not notation. Feedback loops that distinguish between a touchpoint that was completed and one that actually changed something.

I've spent a long time thinking about what makes a system non-brittle. I don't have a formula. I have a strong suspicion that it has less to do with the sophistication of the system's components and more to do with whether the people accountable for it are measured on outcomes rather than activities. Whether the organization treats early-stage engagement as a strategic priority rather than a service function.

Whether someone is listening for the silence.

I'm not sure the MA industry's member retention problem and the home health industry's clinician retention problem are as different as they appear from the outside. The settings are different. The regulatory environments are different. The vocabulary is different.

But the human dynamic at the center of both — a person navigating a new and complex relationship, hitting an unexpected wall, going quiet before they go — that part feels very similar to me.