How One Plastics Unit Used Roster Simulation to Cut 15% in Costs

A Small Team, Complex Constraints

The Plastics Unit at the oldest tertiary hospital in Western Australia operates across public and private care pathways, delivering reconstructive services with a junior doctor team of 9–11 staff. This appears to be a small in headcount, but it is highly complex to roster.

The complexity comes from role-specific competencies, relief staff for leave, public holiday coverage, enterprise agreement compliance, and variable clinic demand. All have to be balanced manually, using spreadsheets. Rostering responsibility sat with administrative staff, while clinical requirements were defined separately. This meant that the two were rarely well aligned.

The deeper problem wasn't the hours spent building rosters. It was that the department had no reliable way to evaluate whether a proposed change to its staffing model would actually work, before committing to it.

Unplanned Overtime Was Absorbing the Workload

The junior doctors did a lot of unplanned overtime after their shifts finished at 4pm. The department believed a dedicated late shift would be better as their would be a clean person for the junior doctors to handover to every day.

Without a way to model the change, the department couldn't answer the questions that mattered: Would it reduce overtime? Would it actually align staffing with demand? What would it cost, and what would it save? What would this change actually look like? What other changes would be needed to make this work?

These were critical questions to answer to create the business case for change, and their manual spreadsheet process offered no way to find out.

RosterLab Built a Model That Actually Works

In February 2026, the department implemented RosterLab, beginning with two foundational steps.

First, establishing a consistent rule framework, translating enterprise agreements, leave policies, and compliance requirements into a structured, repeatable system. Second, modelling the department in detail: shift patterns, service demand, role-specific constraints, and operational edge cases.

The resulting model incorporated more than 28 constraints. What it produced wasn't just a roster, it was a working simulation of how the department actually operates.

Make the Case for Change with Evidence

With the model in place, the department ran a direct comparison between its existing staffing structure and the proposed late-shift model, incorporating updated shift patterns, clinic schedules, fatigue constraints, and fairness rules.

The simulation gave leadership something they hadn't had before: evidence.

The analysis showed the new model would reduce rostered labour time, improve alignment between staffing and demand, and generate approximately 15% in cost savings. The department could move forward with confidence rather than assumptions.

The new rostering system was deployed over six weeks for junior doctor rosters, with iterative refinement at each stage. Roster outputs were validated against real operational requirements, and ownership is progressively handed from RosterLab’s implementation team to internal staff.

By working closely together, the staff at the Plastics Unit will be empowered to run their own regular roster simulations independently, and run a more effective, sustainable service.

Spreadsheets Gone, Overtime Down

The unit now has capabilities it didn't have before:

• Manual spreadsheet rostering replaced entirely
• Ability to simulate and validate workforce changes before implementation
• Improved alignment between staffing allocation and service demand
• Multi-skill rostering based on competencies
• Dynamic public holiday staffing adjustments
• Integrated leave coverage through relief staff
• Consistent compliance with enterprise agreements

Rostering (Scheduling) Done Right Is a Strategic Planning Tool, Not an Admin Task

The lesson from this case isn't just about rostering efficiency. It's about decision-making.

When workforce planning relies on spreadsheets and institutional memory, leaders are making significant structural decisions with incomplete information. A model-driven approach changes that by turning rostering from a reactive administrative task into a tool for proactive workforce design.

If your department is considering a change to its staffing model, and wants to know whether it will work before you implement it, that's exactly what RosterLab was built to answer.

"If rostering still feels like a grind, it doesn't have to be. We've seen what's possible when the right rules and tools are in place, and I'd be happy to show you how it could work for your team." Daniel, Co-Founder, RosterLab

Book a demo → rosterlab.com/book-a-demo