Trust Is Not One Thing

A while back, I wrote about the diagnosis the healthcare industry keeps failing to make. A cleared technology stalls. The science is sound. The market knows the product exists. And still the adoption curve refuses to bend. The argument was that low adoption is rarely one problem. It is one of three structurally distinct barriers: awareness, usability, or trust. Each looks identical from the outside. Each requires a different intervention. And the dangerous error is treating one as another.

This is the sequel. Because of those three barriers, trust is the one that resists being treated as a single thing.

Start with my own first encounter with it, long before I had the vocabulary. I was twenty-one, holding a diagnosis and a pamphlet, sitting with a question I could not yet name: why does this feel like it was not built for me? The structural answer came later. The research behind that diagnosis had not been required to include women like me. My doubt was not a feeling. It was pattern recognition, and the pattern held.

A refusal I watched from the other side

Two decades later, I watched a different refusal, this time from the other side of the table.

I was working on a digital health platform built to deliver a sleep-apnea care pathway to commercial truck drivers. Roughly one in three of them carries the condition, and untreated sleep apnea behind the wheel is a danger to the driver and to the people sharing the road. The pathway was clinically sound. Telehealth visits, home sleep tests, equipment shipped to the door. The team had engineered the friction out. By the logic of the playbook, drivers needed only to know it existed and find it easy to use.

Both assumptions were wrong.

For a commercial driver, a documented sleep-apnea diagnosis that is not perfectly managed can put his federal medical certification at risk. His paycheck depends on staying behind the wheel. His family depends on the paycheck. Sitting at the kitchen table, looking at the telehealth link on his phone, he was not evaluating the clinical quality of the pathway. He was weighing the chance that a diagnosis on his record would cost him his license against the chance that his untreated condition would cause a problem before his next physical.

The clinical solution worked. The pathway was seamless. And the technology was asking a man to make himself visible inside a system that could punish him for the visibility. That is not irrational. That is a person responding to a structure that has not yet earned his trust.

Two refusals, pointing in opposite directions

Set the two side by side.

The first refusal looks backward. The doubt is about origin: was this built for me, was the evidence assembled with someone like me in the room. It rests on a history the patient did not invent and cannot unsee.

The second refusal looks forward. The fear is about consequence: what does saying yes set in motion, and can the system be trusted with what it learns about me. The driver believed the technology worked. That was not the question.

Same word. Two different refusals. One is a verdict on the past. The other is a forecast of the future. A person can know a technology exists, find it easy to use, believe the evidence, and still decline, for reasons unrelated to the product, and reasons that point in opposite directions.

This is why trust resists a single fix. It is not one belief sitting in one place. It is a set of distinct judgments, held by different people, formed for different reasons. Treat it as one thing and the intervention lands on a deficit that was not binding, while the real one keeps working underneath.

I have told this through two patients, because that is where I first felt it. The directions are not theirs alone. A clinician who has read your data and remains unconvinced it holds for her patients is voicing a backward-facing doubt. An administrator weighing what adoption exposes the institution to is running a forward-facing one. And the procurement committee that writes either concern into the vendor rubric is how a single doubt becomes standing policy. Trust is distributed across the room. The two directions travel with it.

The fix that makes it worse

Here is where capital gets destroyed.

When adoption stalls, the trained reflex is to increase visibility. Run the campaign. Buy the reach. Get the product in front of more people. For an awareness problem, that reflex is correct. For a trust problem, it is the precise wrong move, and it does not fail quietly.

Aim a visibility campaign at the backward-facing refusal, and you broadcast a credibility gap to a wider audience. A community that has already concluded an institution does not understand it reads the campaign as confirmation: money spent on being seen, not on being trustworthy. The spend becomes evidence for the doubt.

Aim the same reflex at the forward-facing refusal, and you move more people toward a decision the structure was not built to protect them through. A driver who did not encounter the program is unaffected. A driver who entered it, met the risk, and closed the browser is worse off than before, because now the program is a confirmed threat. The intervention did not fail. It succeeded at the wrong thing.

This is the part that belongs in the boardroom. The cost of getting the diagnosis wrong is not flat. It is directional.

Get awareness wrong, and you waste capital. You build credibility in a room few people are watching, or you broadcast to a market that already knows. The money is spent, but the damage is contained, and the position is recoverable.

Get usability wrong, and you create damage. A trial driven into a broken experience manufactures a firsthand witness who can tell colleagues the product did not work. Those references are durable. But they are bounded by the number of people who actually touched the product, and the experience can be repaired.

Get trust wrong, and the deficit compounds. A trust judgment does not stay with the person who formed it. It travels. The clinician raises it at the department meeting. The community member raises it at the kitchen table. The concern enters the networks where decisions are actually made, and it arrives carrying the credibility of the person repeating it, not the credibility of the company it is about. NRC Health put a number on the upside of the same mechanism: consumers are close to 300 percent more likely to recommend a healthcare organization when they trust it. The networks that carry a recommendation when trust is present carry the refusal when it is absent. The infrastructure runs in both directions.

That asymmetry is the argument for diagnosing before deploying. Not because trust appears more frequently than the others. Because it is the one that keeps working against you while you treat something else.

What the diagnosis asks of you

The work is not to build more trust in general. Trust in general does not exist. The work is to identify which refusal you are facing, whose it is, and what it is about, before a dollar is committed to closing it.

A backward-facing refusal asks the organization to demonstrate, structurally, that the technology was built for the people being asked to adopt it. A forward-facing refusal asks the organization to redesign the architecture around the diagnosis, so that saying yes does not expose the person to a risk the product did not account for. These are different problems with different remedies. The campaign that feels like progress addresses neither.

The leaders who get this right are rarely the ones outspending the room. They are the ones who slow down long enough to ask which barrier is actually in the way, and who resist the comfort of the familiar lever. Trust is not a message you send. It is a structure you build, and it is built one refusal at a time.