🏥 NurseOptima: Algorithmic Staffing & Compliance Optimizer

NurseOptima is an automated healthcare management solution designed to solve complex staffing challenges using Operations Research (OR). [cite_start]By integrating real-world Payroll-Based Journal (PBJ) data with mathematical optimization, the system generates staffing plans that ensure 100% regulatory compliance while minimizing unnecessary labor hours. [cite: 268, 270, 271]

📊 Project Overview

Healthcare facilities often struggle to balance high-quality patient care with rigid federal staffing requirements. [cite_start]This project provides a sophisticated analytical solution: [cite: 282, 283, 285]

• [cite_start]Automated Compliance: Mathematically ensures all facilities meet the CMS minimum Hours Per Resident Day (HPRD) for RNs, LPNs, and CNAs. [cite: 306, 420]
• [cite_start]Operational Efficiency: Minimizes total labor hours to reduce overutilization and operational costs. [cite: 306, 415]
• [cite_start]Data-Driven Insights: Replaces manual, heuristic scheduling with provably optimal staffing levels derived from real-world datasets. [cite: 289, 549]

---

🧪 Methodology & Mathematical Formulation

1. Data Cleaning & Aggregation

[cite_start]The framework processes raw CMS Payroll-Based Journal data (Nurse_fulldata.csv) to prepare it for the optimization engine. [cite: 346, 385]

• [cite_start]Standardization: Daily records for over 14,000 facilities are cleaned of missing values and zero-census days. [cite: 350, 401, 402]
• [cite_start]Aggregation: Data is grouped by facility ID (PROVNUM) to calculate total resident days and existing staffing hours for the reporting period. [cite: 72, 404]

2. Linear Programming Formulation

[cite_start]The core optimization model minimizes the total hours ($Z$) for each facility $f$ while satisfying federal "care floor" mandates. [cite: 292, 416]

Decision Variables

Let:

• [cite_start]$x_{RN}$ = Total optimized Registered Nurse hours [cite: 412, 413]
• [cite_start]$x_{LPN}$ = Total optimized Licensed Practical Nurse hours [cite: 412, 413]
• [cite_start]$x_{CNA}$ = Total optimized Certified Nursing Assistant hours [cite: 412, 413]

Objective Function

Minimize the Total Staffing Hours ($Z$): [cite_start]$$Minimize: Z = x_{RN} + x_{LPN} + x_{CNA}$$ [cite: 292, 416]

Constraints (CMS Regulatory Minimums)

[cite_start]Each facility must meet the mandatory HPRD thresholds multiplied by the total resident days ($D$): [cite: 420, 424]

• [cite_start]RN Requirement: $x_{RN} \geq 0.75 \times D$ [cite: 298, 421]
• [cite_start]LPN Requirement: $x_{LPN} \geq 0.55 \times D$ [cite: 298, 422]
• [cite_start]CNA Requirement: $x_{CNA} \geq 2.25 \times D$ [cite: 298, 423]
• [cite_start]Non-Negativity: $x_{RN}, x_{LPN}, x_{CNA} \geq 0$ [cite: 412]

---

🛠️ Tech Stack & Implementation

• Language: Python 3.x
• [cite_start]Optimization: PuLP library (Linear Programming Solver) [cite: 20, 389]
• [cite_start]Data Manipulation: Pandas and NumPy [cite: 20, 387, 388]
• [cite_start]Visualization: Matplotlib (Bar charts) and Pillow (Interactive Image Viewer) [cite: 20, 390, 436]

---

📈 Key Results & Evaluation

• [cite_start]Compliance Mastery: The model successfully increased the compliance rate from an initial 4.78% to a perfect 100%. [cite: 170, 171, 516]
• [cite_start]Understaffing Identification: The analysis revealed a massive deficit in current staffing, requiring a net increase of ~26.2 million hours to meet legal safety standards. [cite: 518]
• [cite_start]Resource Optimization: While adding hours where needed, the model also identified thousands of "Overstaffed" instances, allowing for potential reallocation of resources. [cite: 537, 546]

---

📂 Project Structure

Nurse-Staffing-Optimization
├── OR_1.IPYNB          # Jupyter Notebook: Data Cleaning & PuLP Solver Logic
├── OR_Final_Report.pdf  # Final academic report (Methodology & Results)
├── OR_FINAL_REVIEW.pptx # Stakeholder presentation and visualization
├── OR.docx              # Detailed technical documentation and literature review
├── final_results.csv    # Final optimized data used for dashboard/reporting
└── README.md            # Project documentation (This file)