Most patient experience strategy is built on a quiet assumption that almost nobody says out loud: trust is a universal thing that all patients experience in roughly the same way. Design the system well, the thinking goes, and trust follows. Design it poorly and trust erodes. The shape of the patient on the other end of that design barely enters the conversation.

That assumption breaks the moment you look at the data.

Pew Research found in 2024 that 55% of Black Americans say medical researchers today conduct experiments on Black people without their knowledge or consent. 51% believe the health care system was designed to hold Black people back. And the geography matters. 59% of Black adults in urban areas believe medical experimentation on Black patients still happens today, compared to 57% in rural areas and 51% in suburbs. That is the same belief expressed at meaningfully different rates inside the same country.

Massachusetts Health Quality Partners ran a statewide trust survey across 156,000 patients in 2024 and found Black, Asian and Hispanic Patients Report Lower Levels of Trust in Their Healthcare Providers Compared to White Patients. Same state. Same care delivery system. Trust scores vary by race across the entire population.

Trust is not a single number. There are five different conversations a system is having simultaneously with different communities, and the system is usually not listening to them all simultaneously.

The word accessibility shows up in nearly every patient experience strategy deck I have ever seen. It is treated as an operational variable to tune. Hours of operation. Channel coverage. Speed-to-answer. Abandonment rate.

Those metrics measure something. They do not necessarily measure accessibility. Accessibility is a relationship between what the system offers and what the population can use, and the population is not one population.

Consider what accessibility means in five different places:

For a 78-year-old Spanish-speaking patient in Hialeah: Accessibility is whether the system meets her in Spanish at 9pm on a Tuesday. Whether the IVR recognizes her language before it routes her through three layers of decision trees. Whether the path to a clinical answer takes thirty seconds or twenty-two minutes. Most systems will not solve this by hiring their way to round-the-clock Spanish fluency. The math does not work. The question is whether the operational design accepts that and builds around it.

For a Black mother in West Baltimore: Accessibility is whether the system has earned the right to be called at all. The Pew data on Black Americans believing medical experimentation still happens today is not abstract. It is the answer to the question of whether she will pick up the phone. A health system can have the best call center in the country and still be inaccessible to a community that has been given good reason not to trust it.

For a 65-year-old farmer two hours from the nearest hospital in eastern Kentucky: Accessibility is whether the rural hospital is still open. The Commonwealth Fund tracked 424 rural hospitals that stopped offering chemotherapy services between 2014 and 2023. The Chartis 2025 Rural Health State of the State report identified roughly 432 rural hospitals at risk of closure. For him, the patient experience strategy of the closest system is a moot question if that system's nearest facility just shut its doors.

For a 22-year-old immigrant patient in Queens: Accessibility is whether the digital tools the system invested in actually meet her where she is. Patient portals built for desktop browsers in English have nothing to say to a patient who interacts with her entire life through a phone in her preferred language, often passed back and forth with family who help her interpret what she is reading.

For a 70-year-old grandmother in Atlanta with early-stage dementia: Accessibility is whether the system remembers her between visits. Whether the cardiology appointment knows what the neurology appointment said. Whether the patient navigator knows she lives alone now. Whether the discharge team realized she cannot drive home. The phone working at 9pm matters less than the system holding the thread of who she is across the appointments she has already completed.

Five patients. Five completely different definitions of the same operational variable. A system that scores well on its overall accessibility metric is not necessarily accessible to any one of them.

Now extend this to the other dimensions of patient experience. Resolution means something different for the patient whose first call has been ignored for two weeks than for the patient who just needs a refill. Proactivity looks different for the Medicaid mom navigating well-child visits than for the executive on a concierge plan. Continuity is everything for the dementia patient and largely irrelevant for the 35-year-old getting an annual physical.

The patient experience strategy that designs for the average patient is designing for nobody, because the average patient does not exist. The composite is a planning fiction. The population is real, and it is differentiated.

Both of my paternal grandparents were diagnosed with dementia-related diseases.

I watched two of the most foundational people in my life disappear in slow motion. Not all at once. Piece by piece. A name forgotten. A face confused. A conversation that looped back on itself until the thread was gone.

I'd spent years inside healthcare systems. I knew how the machinery worked. I knew what good care looked like. I knew what questions to ask.

And none of that prepared me for the moment I understood the system wasn't built for people who were slowly losing themselves. The scheduling friction. The lack of coordination between specialists. The way every appointment started from scratch, as if the whole history had vanished, just like the person we were trying to protect.

Nobody was cruel. Nobody was careless. But nobody saw the whole person. Every provider saw their piece. Nobody held the thread.

Dementia disproportionately impacts Black and brown communities. The Alzheimer's Association reports that older Black Americans are about twice as likely as older White Americans to have Alzheimer's or other dementias, and older Hispanic Americans are about one and a half times as likely. The gap in cognitive care is real, documented, and largely unaddressed. My grandparents weren't statistics. They were the people who raised the people who raised me.

When I talk about what trust means for different communities, I am not abstracting. I am thinking about a system that did not see them, that started over every time, that asked their daughter to repeat the family history twelve times in eighteen months and never built a record that could carry across visits. The system was not malicious. It was just designed for somebody else.

Most systems are.

All of this lives at scale.

Deloitte's Health Equity Institute estimated in 2022 that inequities in the U.S. health system account for approximately $320 billion in annual healthcare spending, projected to surpass $1 trillion by 2040 if the underlying conditions remain unchanged. A 2023 Tulane University study published in JAMA pushed the 2018 figure even higher, finding racial and ethnic health inequities cost the U.S. economy $451 billion that year, with 69% of that burden borne by the Black population due to premature mortality. Different methodologies, same direction.

That money is not abstract. It is sitting in places executive teams already track. It is sitting in the readmission rate. Patients with limited English proficiency face higher 7-day readmission odds than English-proficient patients, even after adjusting for patient and hospital characteristics, according to

research published in PubMed in 2025 that examined more than 424,000 hospitalized adults. It is sitting in hospital choice. Northwestern's Feinberg School of Medicine found that LEP patients regularly bypass closer, higher-quality hospitals to reach facilities with multilingual staff. The closer hospital is losing revenue it does not know it is losing, because the patient walked past it.

It is sitting in the rural service area that lost its chemotherapy program three years ago and watched its oncology referrals migrate to a competitor 90 miles away. It is sitting in the contact center metrics nobody has stratified by zip code, language, or age. It is sitting in the digital tools that work great for patients who can use them and don't work at all for patients who can't.

Equity is not a values initiative. It is a growth strategy that is being misfiled as a compliance function, and the misfiling is what makes it the first thing cut when the budget tightens.

The natural objection to all of this is staffing. Hiring round-the-clock Spanish-speaking nurses, bilingual care coordinators, dementia-specialized navigators, and cultural liaisons across every community a system serves is not a budget conversation. It is an impossibility conversation.

The shortage is not theoretical. The U.S. healthcare industry experienced a shortage of 84,930 physicians and 250,710 registered nurses in 2025. More than 65% of hospitals and health systems report running below full capacity at some point because of staffing shortages. Roughly 138,000 nurses have left the workforce since 2022 due to stress, retirement, and burnout. No equity strategy that depends on hiring our way out of this is going to survive contact with the operating budget.

Which is where the technology question enters. Not as a replacement for human care. As a way to extend it to communities the current staffing model cannot reach.

Three places where the technology is already mature enough to matter:

Real-time AI-mediated language access. Voice-to-voice translation has moved past the demo stage. Tools like Lexi are already in production at organizations like Lowell Community Health Center, which serves 35,000 non-English-speaking patients across high-demand languages. The 9pm Tuesday phone call that took twenty-two minutes to reach an interpreter is solvable with technology that already exists in 2026. Not perfectly. Not for every language. But solvable for the most common gaps in most U.S. service areas.

Agentic AI for after-hours and overflow. The same year, the major cloud providers all launched healthcare-specific agentic AI platforms. Amazon Connect Health debuted with conversational patient identity verification, EHR integration, and natural-language appointment management available 24/7. Claude for Healthcare and OpenAI for Healthcare launched the same month. These are not chatbots reading from scripts. They are connected to the EHR and capable of carrying a context-aware conversation in the patient's preferred language at 9pm on a Tuesday, escalating to a human when the clinical complexity demands it.

Continuity tooling that holds the thread. The grandparent walking from cardiology to neurology to primary care should not have her history disappear at every handoff. That is a tooling problem, not a clinical one. Ambient documentation, shared care plans, and AI-summarized chart context are pieces that exist today and that Deloitte estimates can give nurses roughly 20% more time for direct patient care when the administrative weight is taken off. That 20% is what gets reinvested into the patients whose journeys are the most complex. Including the ones the average-patient design was never built to hold.

There is a critical caveat. AI as a substitute for trained clinicians and qualified interpreters is not the same as AI as a force multiplier underneath them. A 2026 viewpoint in JMIR Medical Informatics warned that routine use of AI as a replacement for qualified medical interpreters

"would normalize a lower standard of care for people with Non-English Language Preference and reinforce existing health disparities." That caution is correct. Throwing a translation API at a system that was never built for language access is not equity work. It is automating the gap.

The discipline is the same one that has always separated technology that builds trust from technology that just moves transactions faster. The question to ask before any AI rollout aimed at community access is whether it serves the signal or just the speed. Does this tool give the 78-year-old Spanish-speaking caller a clinical answer at the moment she needs one? Does it route the dementia patient's history forward to the next provider? Does it close the gap, or just dress it up in a faster interface?

A system that gets this right does not need to hire its way out of the staffing crisis. It needs to design its way through it. The technology to extend reach is already here. The strategy to use it well is the part that has to be built.

Before the next equity strategy review, before the next DEI dashboard, before the next community benefit report, three questions are worth putting on the table. They are operational, not aspirational.

One. Do you actually know the demographic makeup of the population inside your service area? Not the population that already chooses you. The full population. The patients you serve and the ones who could be choosing you and aren't. Census data, payer-mix data, and zip-code-stratified utilization data all live somewhere in the building. Most of the time, they live in three different buildings.

Two. Do the operational details of how you deliver care match the communication preferences, language needs, and cultural expectations of that full population? Channel options that include patients without smartphones. Digital tools that work in the languages the population actually speaks. Discharge instructions a family member can read. Language access at the moment of clinical need, in whatever form the operation can sustainably staff and supplement.

Three. Do the patients in your service area have the tools they need to engage with you? This is the question most equity strategies skip. Trust is bidirectional. If a patient cannot reach a system in their preferred language at the hour they need it, the failure isn't theirs. The strategic question is whether the operational design is built for a relationship that goes both ways, or for a one-way transmission to the patients who already know how to navigate the system.

If any of these questions takes longer than fifteen minutes to answer in your operation, that itself is the diagnostic.

Equity work has been positioned for a decade as a values initiative, a compliance function, a community benefit report obligation. That positioning is why it keeps getting cut when the budget tightens. A values initiative is a cost. A growth strategy is an investment.

The reframe is operational, not rhetorical. The $320 billion is sitting in the readmission rate. It is sitting in the no-shows. It is sitting in the patients who are calling competitors because the competitor answered in Spanish at 9pm on a Tuesday. It is sitting in the contact center metrics that nobody has stratified by demographic. It is sitting in the digital tools that work great for patients who can use them and don't work at all for patients who can't.

Every line of that revenue can be recovered. Not all at once. Not without operational design choices. But it can be recovered. And the systems that figure out how to recover it first will be the systems that hold their margin while the rest of the industry watches consolidation eat them.

The technology to serve the community already exists. The strategy is the part that has to be built.

If you want to put a number on what these gaps are worth in your own operation, the Trust ROI Calculator runs the math on revenue, retention, and contact center metrics you already have. It returns dollar figures on the gap between what your operation is delivering and what your community is reaching for.

→ Run the Trust ROI Calculator

If you want to understand more about the full Trust Algorithm diagnostic and what it would look like applied to your operation, book a strategy call. We will walk through the five signals, the community trust fit lens, and where the conversation would start for a system like yours.

→ Book a strategy call

The Calculator gives you the headline number. The diagnostic is where the operational map gets drawn.

Until next Wednesday.

Let’s get to work,

Ebony