PRECISE-HBR Score Calculator

A SMART on FHIR clinical decision support tool for assessing bleeding risk after percutaneous coronary intervention (PCI).

Clinical Indication

Intended Use: Post-percutaneous coronary intervention (PCI) bleeding risk assessment to guide dual antiplatelet therapy (DAPT) duration decisions.

This tool is designed for use in patients who have undergone or are planned to undergo percutaneous coronary intervention (PCI) with stent implantation. The PRECISE-HBR score helps clinicians:

  • Quantify the individualized 1-year risk of major bleeding (BARC type 3 or 5) after PCI
  • Identify patients at high bleeding risk (HBR) who may benefit from shortened DAPT duration (1-3 months instead of standard 6-12 months)
  • Balance bleeding risk against ischemic/thrombotic risk to optimize antiplatelet therapy
  • Support shared decision-making between clinicians and patients regarding post-PCI antithrombotic strategy

Target Population: Adult patients (age ≥18) undergoing PCI with coronary stent implantation, including both elective and acute coronary syndrome (ACS) presentations.

Introduction

The PRECISE-HBR Score Calculator is a web-based clinical decision support tool designed for cardiologists and interventional medicine clinicians. It implements the PRECISE-HBR (PREdicting bleeding Complications In patients undergoing Stent implantation and subsEquent dual antiplatelet therapy — High Bleeding Risk) scoring model.

This application connects to FHIR-compliant electronic health record (EHR) systems using the SMART on FHIR standard to automatically retrieve patient data and calculate the risk score. Clinicians can also manually adjust input values to perform "what-if" analysis.

Scoring Formula

The PRECISE-HBR total score is computed as the sum of a base score, four continuous variable components, and three binary factors:

Total Score = Base Score + Age Score + Hemoglobin Score + eGFR Score + WBC Score + Bleeding Score + Anticoagulation Score + ARC-HBR Score
Base Score
Base Score = 2 (constant, applied to all patients)
Continuous Variables

Each continuous variable is first truncated (clamped) to a clinically valid range, then a linear formula is applied:

Parameter Truncation Range Formula Direction
Age 30 – 80 years (Effective Age − 30) × 0.25 Older → higher score
Hemoglobin 5 – 15 g/dL (15 − Effective Hb) × 2.5 Lower Hb → higher score
eGFR 5 – 100 mL/min/1.73m² (100 − Effective eGFR) × 0.055 Lower eGFR → higher score
WBC 3 – 15 ×10³/µL (Effective WBC − 3.0) × 0.8 Higher WBC → higher score

"Effective" value means the raw value clamped to the truncation range. For example, if a patient's age is 85, the effective age used in calculation is 80.

Calculation Examples
Parameter Raw Value Effective Value Calculation Score
Age 65 years 65 (65 − 30) × 0.25 8.75
Hemoglobin 12 g/dL 12 (15 − 12) × 2.5 7.5
eGFR 45 mL/min/1.73m² 45 (100 − 45) × 0.055 3.025
WBC 10 ×10³/µL 10 (10 − 3.0) × 0.8 5.6
Binary Factors
Factor Points if Present Description
Previous Bleeding +7 History of spontaneous bleeding (GI bleeding, intracranial hemorrhage, or bleeding requiring transfusion)
Long-term Oral Anticoagulation +5 Chronic use of oral anticoagulants (Warfarin, DOACs such as Apixaban, Rivaroxaban, Dabigatran, Edoxaban)
ARC-HBR Factors (≥1) +3 Presence of one or more ARC-HBR criteria (see ARC-HBR Factors section below)
Full Example: A 65-year-old patient with Hb 12 g/dL, eGFR 45, WBC 10, prior bleeding, on OAC, no ARC-HBR factors:
Score = 2 + 8.75 + 7.5 + 3.025 + 5.6 + 7 + 5 + 0 = 38.875 → rounded to 39
Risk Level: Very High Bleeding Risk (≥27)

Risk Classification

The total PRECISE-HBR score is classified into three risk categories:

Non-High Bleeding Risk (Score ≤ 22)

Estimated 1-year BARC 3/5 bleeding risk: < 4%
Standard DAPT duration (6-12 months) is generally appropriate.

High Bleeding Risk (Score 23 – 26)

Estimated 1-year BARC 3/5 bleeding risk: 4% – 6%
Consider shortened DAPT (1-3 months) followed by P2Y12 inhibitor monotherapy.

Very High Bleeding Risk (Score ≥ 27)

Estimated 1-year BARC 3/5 bleeding risk: ≥ 6%
Strongly consider abbreviated DAPT (≤1 month) or alternative antithrombotic strategies.

BARC = Bleeding Academic Research Consortium. Type 3 = clinical bleeding requiring intervention; Type 5 = fatal bleeding.

ARC-HBR Factors

The PRECISE-HBR model incorporates ARC-HBR (Academic Research Consortium — High Bleeding Risk) criteria as a composite binary factor (+3 points if ≥1 factor present). The following conditions are assessed:

ARC-HBR Factor Criterion Classification
Thrombocytopenia Platelet count < 100 ×10&sup9;/L Major
Chronic Bleeding Diathesis Hemophilia, von Willebrand disease, or other coagulation factor deficiency Major
Liver Cirrhosis Liver cirrhosis with portal hypertension Major
Active Malignancy Active cancer diagnosed or treated within past 12 months (excluding non-melanoma skin cancer) Major
Chronic NSAID/Corticosteroid Use Chronic use of NSAIDs or corticosteroids (e.g. ≥4 days/week) Minor
Recent Major Surgery or Trauma Major surgery or trauma within 30 days before PCI Minor

Note: Condition and medication detection is automated via FHIR data (ICD-10, SNOMED CT, medication codes) but should be secondary confirmed by clinicians.

Core Features

Automated FHIR Data Retrieval

The application automatically retrieves the following data from the patient's EHR via FHIR R4 API:

  • Demographics: Age, gender, name
  • Lab Results: Hemoglobin (LOINC: 718-7), WBC count (LOINC: 6690-2), eGFR (LOINC: 33914-3) or Creatinine (LOINC: 2160-0) for calculation
  • Conditions: ICD-10/SNOMED coded diagnoses for bleeding history, liver cirrhosis, malignancy, bleeding diathesis
  • Medications: Oral anticoagulants, NSAIDs, corticosteroids detected by medication code and keyword matching
  • Procedures: Recent surgical procedures within 30 days
Interactive What-If Analysis

All score components are editable. Clinicians can manually adjust lab values or toggle condition flags to observe the real-time impact on the total score and risk classification.

Bleeding vs. Ischemic Risk Trade-off Analysis

For patients identified with high bleeding risk (score ≥ 23), the application provides an interactive trade-off analysis tool based on the ARC-HBR model. This visualizes the balance between estimated 1-year bleeding risk and ischemic/thrombotic risk across different DAPT durations.

How to Use the Application

Standalone Launch (for testing or direct use):
  1. Navigate to the application's home page.
  2. You will be presented with a "Standalone Launch" screen.
  3. Enter the FHIR Server URL (the iss parameter) for the EHR system you wish to connect to.
  4. Click "Launch".
  5. You will be redirected to the EHR's login and authorization page.
  6. After successfully authorizing the application, you will be redirected to the main risk calculation page, pre-filled with the selected patient's data.
EHR Launch (integrated use):

When the application is registered within an EHR system (e.g., Epic, Cerner), clinicians can launch it directly from a patient's chart. The launch process is seamless, and the application will open directly to the risk calculation page for that patient.

References

  1. Gragnano F, van Klaveren D, Heg D, et al. PRECISE-HBR Score for High Bleeding Risk Assessment in Patients Undergoing Percutaneous Coronary Intervention. Circulation. 2025;151:693-706. doi: 10.1161/CIRCULATIONAHA.124.072009
  2. Urban P, Mehran R, Colleran R, et al. Defining High Bleeding Risk in Patients Undergoing Percutaneous Coronary Intervention: A Consensus Document from the Academic Research Consortium for High Bleeding Risk (ARC-HBR). Circulation. 2019;140:240-261.
  3. Valgimigli M, Bueno H, Byrne RA, et al. 2017 ESC Focused Update on Dual Antiplatelet Therapy in Coronary Artery Disease. Eur Heart J. 2018;39:213-260.

Technical Stack

  • Backend: Python 3.11 with Flask web framework, Gunicorn WSGI server
  • FHIR Integration: fhirclient library for Python, FHIR R4 / DSTU2 compatible, SMART on FHIR v2
  • Authentication: SMART on FHIR OAuth2 with PKCE, OpenID Connect id_token validation
  • Security: Flask-Talisman (CSP/HSTS), Flask-Limiter (rate limiting), CSRF protection, SSRF prevention, BOLA protection
  • Frontend: HTML5, Bootstrap 5, JavaScript with Chart.js for visualizations
  • Deployment: Google App Engine (Standard Environment), Docker support
  • Regulatory: IEC 62304 / ISO 14971 compliant development process for TFDA SaMD certification