Blog/From the team
Healthcare Still Thinks in Visits. The Future Thinks in Patient States
Why the patient state is the true atomic unit of care
The visit as the atomic unit of care
For decades, healthcare has organized itself around a simple assumption: the visit is the fundamental unit of care.
Patients schedule appointments. Clinicians evaluate them during those visits. Documentation, billing, and decision-making are all structured around that encounter.
This assumption runs so deep that nearly every modern healthcare technology platform, including the newest wave of AI tools, still revolves around the same idea. AI may summarize the visit, generate notes from the visit, or suggest diagnoses during the visit.
But the visit itself remains the center of the system.
That assumption is the problem.
If the visit is the atomic unit of care, then the healthcare system is forced to treat medicine as a series of disconnected moments. Information must be gathered quickly. Decisions must be made before the appointment ends. Documentation must reconstruct what just happened.
But human health does not unfold in fifteen-minute increments.
Health is continuous.
And if we want healthcare systems and AI to make better decisions, we have to start by modeling the patient continuously as well.
The real atomic unit of care is not the visit.
It is the patient state.
The Patient State: A Higher-Resolution Model of Health
Instead of treating each visit as a self-contained event, a better approach is to represent the patient's health as a continuously evolving state.
The patient state is the system's current understanding of a person's health at a given moment.
It includes:
- Symptoms and symptom history
- Medical conditions
- Medications and treatment responses
- Laboratory results
- Vital signs and biometric signals
- Prior clinical decisions
- Contextual factors such as lifestyle and environment
Importantly, the patient state is not static.
It evolves over time as new information arrives. Symptoms change. Treatments produce effects. New signals appear. Context becomes clearer.
The patient state is constantly updated as the system learns more.
When clinicians make decisions using this evolving model, they are not relying on a snapshot from a single visit. They are reasoning from a continuously refined understanding of the patient.
This leads to something medicine has historically struggled to achieve: high-resolution clinical context.
Why Resolution Matters in Medicine
Every medical decision depends on how clearly we understand the patient.
When information is sparse, decisions become guesses.
When information is detailed and contextualized, decisions become more reliable.
Traditional visit-based care often produces low-resolution information. Clinicians must compress the patient's story into a short interaction and make decisions with incomplete context.
Important details may be forgotten. Symptoms may be oversimplified. Prior records may not be fully reviewed.
The result is a snapshot of the patient rather than a model of the patient.
When care is built around the patient state instead of the visit, the information landscape changes.
Symptoms accumulate over time rather than being recalled from memory.
Medication responses can be observed rather than inferred.
Patterns become visible across weeks or months.
Signals can be interpreted in context.
The resolution of the patient's health state increases dramatically.
And when resolution increases, clinical reasoning improves.
From Episodic Care to Longitudinal Care
The visit-based system also creates another limitation. It treats healthcare as a series of isolated encounters.
In practice, most conditions are not episodic.
Hypertension evolves gradually.
Diabetes requires ongoing adjustment.
Digestive disorders fluctuate over time.
Autoimmune diseases wax and wane.
Mental health conditions change with life circumstances.
These conditions require longitudinal understanding, not isolated visits.
When the patient state is continuously updated, clinicians can see how health evolves over time rather than reconstructing the story from occasional appointments.
Patterns emerge that would otherwise remain hidden.
A medication that seemed effective during a single visit may show declining benefit over months. Subtle symptom patterns may reveal early disease signals. Changes in laboratory values may indicate trends before they become dangerous.
By modeling the patient continuously, healthcare can align itself with the true nature of chronic disease.
AI That Understands Patients, Not Just Conversations
This shift also changes the role of artificial intelligence in healthcare.
Most AI systems today are designed to assist with documentation or conversation analysis. They process the language of a visit.
But language is only a small part of the clinical picture.
A more powerful approach is to use AI to help maintain and interpret the patient state itself.
When the system continuously integrates symptoms, records, signals, and clinical decisions, AI can reason across a much richer dataset.
Instead of asking:
"What diagnosis matches the symptoms described in this visit?"
The system can ask:
"How has this patient's health evolved over time, and what patterns are emerging?"
This allows AI to support clinicians with deeper insights, not just faster documentation.
A Foundation for Proactive Medicine
Modeling the patient state also enables something the traditional healthcare system struggles to provide: proactive care.
Visit-based medicine is inherently reactive. Patients seek care when symptoms become severe enough to prompt an appointment.
But many diseases begin with subtle signals that appear long before symptoms become obvious.
When patient information accumulates continuously, these signals can be detected earlier.
Blood pressure trends can reveal emerging hypertension before it becomes severe. Medication side effects can be recognized before they lead to complications. Changes in symptoms can prompt earlier intervention.
Care shifts from reacting to illness toward preventing its progression.
Why This Matters Now
For years, healthcare systems have tried to improve care by optimizing the visit.
But the visit itself is the constraint.
If the atomic unit of care remains the appointment, even the most advanced technologies will struggle to overcome the limitations of episodic medicine.
A better approach is to start with a different unit of care altogether.
The patient.
By modeling the patient state continuously, rather than compressing information into isolated visits, healthcare systems can finally operate with the resolution needed to manage complex, longitudinal health conditions.
The Next Step
This article is the second in a series exploring the foundations of Actually Health.
In the first piece, we explored why existing solutions and approaches aren't focused on the right problem.
In this article, we introduced the idea that the patient state, not the visit, is the atomic unit of care.
In future articles, we will examine how asynchronous care improves quality, how problem graphs and decision bundles allow clinical reasoning to scale, and how to create a true learning health system.
Key takeaways
Healthcare is still organized around the visit. Appointments define how care is delivered, how decisions are made, and how data is captured. Even modern AI tools follow this model by summarizing visits, generating notes, and assisting during encounters.
But the visit is the wrong unit of care.
It forces medicine into disconnected snapshots. Clinicians make decisions with limited context, reconstructing a patient's story in real time rather than working from a complete picture.
Health does not work this way.
It evolves continuously. Symptoms change. Treatments produce effects. Patterns emerge over time.
The correct unit of care is the patient state.
A continuously updated model of the patient that integrates symptoms, conditions, treatments, labs, signals, and context as they evolve.
This shift increases the resolution of clinical understanding:
- Information accumulates instead of being recalled
- Treatment response is observed, not inferred
- Patterns emerge across time, not just within visits
And that changes what becomes possible:
- Better clinical reasoning
- True longitudinal care for chronic disease
- AI that understands trajectories, not just conversations
- Earlier detection and more proactive intervention
Most healthcare systems and most AI are still optimizing the visit.
But optimizing the wrong unit will never produce the right outcome.
The future of care starts by modeling the patient, not scheduling the appointment.