# Stroke & Bleeding Risk MCP MCP > The CHA₂DS₂-VASc & HAS-BLED Calculator determines stroke risk (CHA₂DS₂-VASc) and bleeding risk (HAS-BLED) for patients with atrial fibrillation. This MCP integrates established clinical guidelines, giving clinicians a single view of both thromboembolic and hemorrhagic risks. It moves beyond simple scoring by generating definitive anticoagulant recommendations based on combined patient data. ## Overview - **Category:** risk-assessment - **Price:** Free - **Tags:** AFib, CHA2DS2VASc, HASBLED, stroke-risk, guidelines ## Description Managing Atrial Fibrillation requires balancing two opposing forces: preventing stroke versus avoiding excessive bleeding. Simply calculating scores separately leaves the clinician guessing how to weigh these risks against each other. This MCP takes a patient's full profile—their age, comorbidities, blood pressure control, and medication use—and runs it through established protocols simultaneously. It doesn't just provide two numbers; it synthesizes those factors into actionable guidance. Your agent calculates the base stroke risk using one tool, assesses bleeding danger with another, and then combines both findings to generate a final recommendation for annual stroke risk percentage and whether anticoagulation is needed. Because this process involves multiple calculations drawing from diverse clinical data points, Vinkius's secure architecture ensures that all credentials pass through a zero-trust proxy; your keys are used only in transit and never saved on disk. ## Tools ### calculate_hasbled_score It calculates the HAS-BLED score, which measures potential bleeding risk during anticoagulant therapy. ### get_risk_recommendations It generates a final clinical recommendation by analyzing both stroke and bleeding scores to guide treatment decisions. ### query_cha2ds2_vasc_score It calculates the CHA₂DS₂-VASc score based on patient details, determining the risk of a stroke in atrial fibrillation. ## Prompt Examples **Prompt:** ``` Run the full risk assessment. Patient is a 78-year-old male, has HTN and DM. No history of stroke/TIA, but does have documented vascular disease and ICC factor. His BP is 140/90; he takes diuretics and has no GI bleeding history. Recommend anticoagulation. ``` **Response:** ``` First, we calculate the base risk using **`query_cha2ds2_vasc_score`** (inputs: Male, 78, False, True, True, False, True, True). Next, we assess bleeding risk with **`calculate_hasbled_score`** (inputs: 140, True, False, YES). Finally, **`get_risk_recommendations`** synthesizes these scores, providing the annual stroke risk and the definitive anticoagulation recommendation. ``` **Prompt:** ``` I need to know the bleeding risk for a low-risk patient who is only 65 years old, female, and not taking anticoagulation. ``` **Response:** ``` We will first run **`query_cha2ds2_vasc_score`** (inputs: Female, 65, False, False, False, False, False, False) to establish the baseline score. Then, we use **`calculate_hasbled_score`** (inputs: Current BP, False, False, NO). The results will guide us on whether further tools are necessary. ``` **Prompt:** ``` What is the overall recommendation for a high-risk patient? Inputs: Male, 85, True (Stroke/TIA), True (HTN), True (DM), False (CHF), True (VascD), True (ICC). BP: 160/95. Diuretic use: Yes. GI Bleeding: No. Anticoagulation Status: YES. ``` **Response:** ``` The full assessment requires three steps: 1) **`query_cha2ds2_vasc_score`** for the high thromboembolic risk. 2) **`calculate_hasbled_score`** to check bleeding risk (inputs: 160, True, False, YES). 3) **`get_risk_recommendations`** uses both results to give a definitive O/A recommendation and annual stroke percentage. ``` ## Capabilities ### Calculate stroke risk It calculates the CHA₂DS₂-VASc score using patient demographics like age, sex, and comorbidities. ### Assess bleeding danger It evaluates bleeding risk by running the HAS-BLED scale against factors like blood pressure control and diuretics. ### Generate final recommendation It synthesizes both calculated scores into a definitive annual stroke risk percentage and anticoagulant guidance (O or A). ### Analyze comorbidities It incorporates multiple chronic conditions, such as hypertension and diabetes, directly into the base risk calculation. ## Use Cases ### Initial patient intake for AFib A new patient arrives with suspected Atrial Fibrillation. The agent first runs `query_cha2ds2_vasc_score` to establish the baseline stroke risk. Next, it uses `calculate_hasbled_score` to check their current blood pressure and diuretic use. Finally, `get_risk_recommendations` spits out a clear recommendation for immediate next steps. ### Adjusting medication after hospitalization A patient was hospitalized and had minor GI bleeding. The clinician needs to know if they can safely maintain anticoagulation. The agent runs both `query_cha2ds2_vasc_score` (to confirm persistent stroke risk) and `calculate_hasbled_score`, allowing the final recommendation tool to weigh safety over absolute prevention. ### Comparing two patients with similar symptoms You have Patient A, who is young but has multiple comorbidities, and Patient B, who is older but healthy. The agent can run both profiles through `query_cha2ds2_vasc_score` and `calculate_hasbled_score` side-by-side to quickly pinpoint which patient's risk profile is more concerning. ### Pre-procedure assessment Before a planned cardiac procedure, the nurse needs an immediate risk summary. The agent gathers all current data points and uses `get_risk_recommendations` to give a rapid consensus on whether anticoagulation should be continued or temporarily paused. ## Benefits - Avoid siloed thinking. Instead of just seeing a high CHA₂DS₂-VASc score, the agent uses `query_cha2ds2_vasc_score` and then immediately balances it against bleeding risk via `calculate_hasbled_score` to give you the full picture. - Stop guessing about guidelines. The final step, using `get_risk_recommendations`, delivers a definitive annual stroke risk percentage and clear 'O' or 'A' guidance, making your protocol recommendations faster. - It handles complex data dependencies automatically. You feed the agent one set of inputs (age, HTN status, etc.), and it manages running three distinct calculations behind the scenes. - The system doesn't just list scores; it synthesizes them. This is crucial because a high thromboembolic risk can be mitigated by accepting a higher bleeding risk—the MCP tells you how to weigh that trade-off. - Confidence in decisions increases. Because the process mimics established clinical guidelines, your agents provide recommendations that are auditable and defensible. ## How It Works The bottom line is that you stop manually cross-referencing charts; you get one integrated answer based on all established guidelines. 1. You feed the agent a patient's full profile: age, sex, comorbidities (HTN, DM), blood pressure readings, and medication history. 2. The MCP runs the data through the first two specialized calculations to get both the stroke risk score and the bleeding risk score. 3. Finally, it combines these two scores using the third tool to generate a singular, definitive recommendation for anticoagulation. ## Frequently Asked Questions **How does the CHA₂DS₂-VASc & HAS-BLED Calculator use `query_cha2ds2_vasc_score`?** The MCP uses this tool to calculate the base risk by assessing factors like age, sex, and comorbidities (HTN, DM). This establishes the patient's initial likelihood of having a stroke. **What specific data does `calculate_hasbled_score` require?** This tool primarily requires information about blood pressure control and whether the patient is on diuretic medications to accurately assess bleeding risk during treatment. **Can I use this MCP if the patient has multiple conditions?** Yes. The system is designed for integrated care, allowing you to input all comorbidities (like HTN and DM) into the initial data set, ensuring the full risk profile is considered. **What does `get_risk_recommendations` provide?** It synthesizes the two calculated scores, giving you a definitive annual stroke risk percentage (based on ESC 2020 criteria) and whether anticoagulation is recommended ('O' or 'A'). **When running both `query_cha2ds2_vasc_score` and `calculate_hasbled_score`, how does Vinkius secure patient data?** Your credentials pass through a zero-trust proxy, meaning your keys are used in transit but never stored on disk. Every single tool call generates a cryptographically signed audit trail for full accountability. **If an error occurs during `query_cha2ds2_vasc_score`, how does the system prevent bad data from reaching `get_risk_recommendations`?** The MCP is designed to handle failures gracefully. If any tool fails, it reports exactly which function caused the issue and prevents incomplete or invalid scores from being used in the final recommendation step. **What AI clients can access this MCP for calculating scores using `calculate_hasbled_score`?** You connect once through any MCP-compatible client, like Claude or Cursor. That single connection gives your agent access to every tool in the Vinkius catalog, not just this risk assessment. **How does the platform optimize calls when I run all three tools, including `query_cha2ds2_vasc_score`?** The MCP includes native token optimization built into every call. This cuts down token consumption by up to 60% compared to running these calculations without optimization.