Operations Investment Analysis

Client

A major Brazilian health insurance provider undergoing a broader digital transformation initiative to improve operational efficiency and decision-making across administrative departments.

Engagement Overview

We were engaged to develop a stochastic simulation model and provide technical training for the client’s internal team as part of a process improvement and operational modernization strategy. The focus of the project was a department responsible for evaluating and validating healthcare payment requests, including invoice verification and medical procedure compliance analysis.

The client wanted to better understand operational bottlenecks, identify improvement opportunities, and evaluate the potential impact of automation and process changes before implementing them in production environments.

Objective: Identifying Bottlenecks and Operational Improvement Opportunities

The department handled a high volume of healthcare payment requests, requiring analysts to validate invoices, medical procedures, and associated costs against medical authorizations and internal business rules.

The project aimed to answer key operational questions, including:

• Which stages of the process created the largest delays?
• Where would investments and automation generate the greatest impact?
• Which operational constraints were not actual bottlenecks?
• How could staffing and processing capacity be optimized?

The client also wanted to build internal capability around simulation modeling, enabling their team to continue using and evolving the model after project completion.

Stochastic Simulation Model Development

• Simulation Platform: FlexSim
• Data Source: Historical operational process data from the healthcare request evaluation department.

Methodology

We developed a stochastic discrete-event simulation model representing the full operational workflow of healthcare payment request processing.

The model incorporated:

• Historical processing times and arrival distributions
• Queue dynamics and resource allocation
• Validation and approval logic
• Statistical distributions derived from real operational data
• Scenario testing for automation and policy changes

To ensure reliability and realism:

• Input data was statistically validated before modeling
• Output significance testing was performed to validate simulation results
• Multiple simulation runs were executed to reduce random variation bias
• Operational assumptions were validated directly with the client’s subject matter experts

Key Challenges

• Process Variability: Request complexity and processing times varied significantly depending on medical procedure type, requiring careful statistical treatment and segmentation.
• Data Quality & Validation: Historical operational data required preprocessing and validation to ensure consistency before being incorporated into the model.
• Organizational Decision Support: The project needed to provide not only technical simulation outputs, but also actionable operational insights understandable by both technical and non-technical stakeholders.
• Internal Knowledge Transfer: In parallel with the model development, we trained the client’s team on both the simulation platform and the modeling methodology to support long-term internal adoption.

Outcomes & Impact

The simulation model enabled the client to:

• Identify operational bottlenecks and low-hanging improvement opportunities
• Test operational changes in a risk-free virtual environment
• Demonstrate that relatively simple automations and system parameter adjustments could increase processing speed by approximately 10%
• Evaluate the impact of operational improvements on staffing and headcount requirements
• Avoid investments in areas that were shown not to be true operational bottlenecks

The project helped the client prioritize initiatives with measurable operational impact while reducing uncertainty around process optimization decisions.

Technologies Used

FlexSim, statistical analysis techniques, historical operational datasets, stochastic process modeling methodologies. Python and R for statistical analysis and data treatment.

Collaboration Style

We worked closely with operational stakeholders and technical teams throughout the engagement, combining process discovery sessions, iterative model validation, and hands-on training workshops. The project emphasized transparency, statistical rigor, and practical decision support, ensuring the client could confidently use the model as part of its ongoing digital transformation efforts.