Blog/From the team

    Article 5 · Decision Bundles

    The Real Bottleneck in Healthcare Is Clinical Reasoning

    Decision Bundles and the Architecture of Scalable Care

    Blog
    Simon Mathews, MDCo-founder & CCSO, Actually Health

    Healthcare has spent decades trying to scale access to clinicians. We have built larger hospitals, expanded clinic networks, created telemedicine platforms, and introduced digital tools designed to make clinicians more efficient. The goal has always been the same: help more patients receive care.

    But scaling healthcare has proven far harder than scaling most other industries.

    The reason is simple. Medicine is not just a service. It is a reasoning process.

    Every patient presents a unique combination of symptoms, history, context, and risk. Clinicians interpret that information, identify the most likely problems, and decide what to do next. Those decisions depend on judgment, experience, and careful analysis.

    This reasoning process is the heart of clinical care. It is also the reason healthcare does not scale easily.

    If every decision depends entirely on an individual clinician's judgment in the moment, then scaling care requires simply adding more clinicians. The system grows linearly with labor.

    To build a healthcare system that can scale safely, the reasoning itself must become structured.

    This is where decision bundles come in.

    The hidden problem in modern healthcare

    When people talk about scaling healthcare, the conversation usually focuses on visits.

    • How many patients can a clinician see per day?
    • How quickly can notes be written?
    • How efficiently can visits be scheduled?

    Technology has improved many of these processes. Telemedicine has removed geographic barriers. AI can now generate clinical notes automatically. Scheduling systems have become more sophisticated.

    But none of these changes address the core bottleneck.

    The real bottleneck in healthcare is clinical reasoning.

    During every encounter, clinicians must answer a set of fundamental questions:

    • What problem does this patient have?
    • How certain are we about that problem?
    • What intervention is appropriate?
    • What needs to be monitored?
    • When should the plan change?

    These decisions determine outcomes. They determine safety. They determine cost.

    Yet in most healthcare systems, the reasoning behind those decisions is largely invisible.

    It exists inside individual clinicians' heads and appears only partially in documentation. When two clinicians evaluate the same patient, their reasoning may differ significantly.

    This variability is one of the main reasons healthcare quality is inconsistent.

    To improve quality at scale, the reasoning process itself must become explicit and structured.

    From clinical judgment to structured decisions

    Medicine will always require clinical judgment. No system can replace the expertise of a trained clinician.

    But the structure of clinical decisions can be standardized without removing judgment.

    This is the idea behind a decision bundle.

    A decision bundle is a structured unit of clinical reasoning that represents a specific medical decision. Each bundle includes several elements:

    • the clinical question being addressed
    • the relevant patient signals and context
    • the possible decision paths
    • the criteria that guide those decisions
    • the actions associated with each outcome
    • the follow-up conditions that determine what happens next

    In other words, a decision bundle captures how clinicians reason about a particular problem. Instead of relying entirely on implicit knowledge, the decision logic becomes visible.

    Why medicine is full of decision bundles already

    Clinicians already think in decision bundles, even if the system does not formally recognize them.

    Consider a common example such as hypertension management. A clinician reviewing a patient with elevated blood pressure typically considers several questions:

    • Is the blood pressure consistently elevated?
    • Are there secondary causes?
    • Is medication required?
    • Which medication class is appropriate?
    • How should the patient be monitored?

    Each of these steps represents a structured reasoning process. Certain inputs lead to certain actions. Follow-up depends on how the patient responds.

    The problem is that these reasoning patterns exist informally. They are taught during training and refined through experience, but they are rarely represented explicitly within healthcare systems.

    Because the reasoning is not structured, it cannot easily be scaled or improved systematically.

    Decision bundles make these patterns explicit.

    Interactive

    Peel back one clinical recommendation

    One line in a visit note sits on top of six layers of structured reasoning. Drag the slider to reveal them.

    Patient · 58yo, T2D, BMI 31 · referral for BP · home avg 148/92 / 14d

    What the note shows

    Start RAASi+CCB single-pill combination, max tolerated dose, once daily. Recheck BP at 8 weeks.

    Usually only this line is visible—the rest stays implicit.

    NoteStructure
    1Question2Signals3Options4Criteria5Action6Follow-up

    Structured reasoning

    Reading the demo. The recommendation stays fixed while layers underneath encode the question, signals, ruled-in and ruled-out options (with rationale), cited criteria, a typed action, and follow-up rules—together, one decision bundle. The shape is stable; only the contents change patient to patient.

    Why structured reasoning improves quality

    When clinical reasoning becomes structured, several things happen.

    First, the logic behind decisions becomes transparent. Instead of relying solely on individual interpretation, the system can represent the reasoning that guides care. Clinicians can see how decisions were made and evaluate whether the logic is sound.

    Second, consistency improves. When similar clinical situations are evaluated using similar reasoning frameworks, patients receive more reliable care. Variability caused by incomplete information or rushed decision-making decreases.

    Third, learning becomes possible. If decisions are structured, outcomes can be observed and linked to the reasoning that produced them. Over time, the system can identify which decision paths lead to better outcomes and refine the logic accordingly.

    This creates a feedback loop that improves care continuously.

    Enabling asynchronous care

    Decision bundles also make asynchronous care possible.

    In a traditional visit model, clinicians must gather information, analyze it, and make decisions within a single appointment. The reasoning process happens quickly and often under time pressure.

    When care is asynchronous, information can accumulate over time. Patient symptoms evolve. Records are reviewed. Signals become clearer.

    Decision bundles provide a structured way to evaluate that information as it arrives.

    Each bundle represents a specific clinical question that can be addressed when sufficient context exists. Once the necessary inputs are available, the clinician can evaluate the bundle and determine the appropriate action.

    This approach separates clinical reasoning from the constraints of the visit.

    Decisions are made when the information is ready, not when the appointment ends.

    Scaling clinical expertise

    One of the most important advantages of decision bundles is that they allow clinical expertise to scale.

    In traditional healthcare systems, expertise is tied directly to individual clinicians. The quality of care depends heavily on who happens to see the patient.

    Structured decision bundles allow that expertise to be represented within the system itself.

    Experienced clinicians can help define the logic behind bundles. Best practices can be encoded into decision pathways. Safety checks can be built into the reasoning process.

    As the system learns from outcomes, bundles can evolve and improve.

    The result is a system where clinical reasoning becomes a shared resource rather than an isolated skill. Clinicians still make the decisions, but they do so with a consistent framework that captures the collective knowledge of the system.

    A foundation for scalable medicine

    Healthcare has long struggled with a fundamental challenge. High-quality medicine depends on careful reasoning, but reasoning does not scale easily.

    Decision bundles provide a way forward.

    By structuring clinical decisions into explicit units of reasoning, healthcare systems can improve consistency, enable asynchronous care, and learn from outcomes at scale.

    This does not replace clinicians. It supports them.

    Clinicians still interpret complex situations, apply judgment, and communicate with patients. Decision bundles simply provide a structured foundation for that reasoning.

    The next step

    This article is part of a series exploring the foundations of Actually Health. In this article we introduced decision bundles, the structured units of reasoning that allow clinical care to scale safely.

    Previous articles discussed why finding the right problem to solve in healthcare is important, why the patient state, not the visit, is the true atomic unit of care, how asynchronous care improves quality by removing the time constraints of traditional visits, and how problem graphs extend these ideas by modeling how medical problems evolve across time.

    The next step is understanding how these elements work together.

    If the patient state represents what we know about the patient, and decision bundles represent how we reason about care, the question becomes:

    How do these components interact to produce a continuously learning healthcare system?

    About this series. Actually Health is building an AI-native, asynchronous care platform organized around the patient state, problem graphs, and decision bundles. This is Article 5 in a series exploring the foundations of that operating model.