Auckland Tertiary Hospital Improves Fairness for On-Call Roster

TL;DR

• Auckland Tertiary Hospital’s radiology department transformed its on-call rostering process using RosterLab, cutting scheduling time from 15-20 hours per year to less than 30 minutes based on a 10 week roster cycle.
• AI-generated rosters helped remove perceptions of bias, significantly improving fairness and transparency.
• More leave and preferences were met through optimisation, contributing to safer staffing patterns and higher team well-being.
• Year-long planning and automated fairness balancing ensured equitable distribution of on-calls across the entire year.
• Improved visibility and transparency in workload distribution strengthened trust, reduced admin burden, and supported better clinical safety.

Background

Auckland Tertiary Hospital’s radiology department manages on-call coverage for 10 consultants providing acute care for shifts between 5:30 PM and 8:00 AM. Junior consultant Dr. Rukshan was responsible for building these rosters manually - a task that became increasingly unsustainable and time-consuming to manage, compounded by the persistent burden of addressing fairness concerns.

Strict constraints also added to the complexity:

• No back-to-back on-calls
• Equitable distribution of fellow support
• Balanced on-call load over multiple periods

“Before RosterLab, I would have to email staff back and forth, sending out multiple versions, with responses like: No, this doesn't work, have you missed this?... and then I would need to change the roster and start all over again… So it would take about three or four hours, and require multiple email exchanges.” - Dr. Rukshan, Consultant, Auckland Tertiary Hospital - Radiology.

The Challenge

Auckland Tertiary Hospital’s radiology roster required 4 hours of rostering administration every 10 weeks, plus additional time for revisions. While the roster itself was relatively small, the hidden cost to clinical safety and team well-being was substantial and impossible to ignore.

Creating each roster was a time-intensive process. Even small, nuanced changes often meant large portions of the roster had to be rebuilt from scratch. Communication occurred largely over email, which made it difficult to keep track of the latest version and dangerous gaps in coverage weren’t always caught until it was too late.

But the real cost wasn’t just administrative - missed constraints could result in consultants working back-to-back overnight shifts, increasing fatigue-related errors during acute decision-making. When workload becomes too imbalanced during high-demand periods, consultants face burnout that affects both their well-being and their ability to provide optimal patient care.

Fairness was another major concern. The manual nature of the allocation process made it vulnerable to perceived bias, and it was difficult to evaluate fairness across multiple roster periods. Visibility into workload distribution was also limited, especially around holidays. Optimising for fairness was largely manual, leading to situations where “three people get hit with massive amounts of on-calls,” requiring multiple future periods to compensate. As Dr. Rukshan notes, “It’s easy for people to blame the roster maker… and say it’s not really fair.”

Additionally, the system lacked flexibility when it came to leave management. Accommodating all requests manually was challenging, and there was little visibility into alternative scheduling solutions that might meet everyone’s needs.

The RosterLab Solution

“I thought there must be an easier way to do this. There must be some rostering software that can do this.. and then I found RosterLab! I really liked it and could see how it could make my job a lot easier.”

After Dr. Rukshan received approval to use RosterLab, he was able to implement AI rostering to help automate the full on-call scheduling process.

“It was doing a good job, everyone entered their preferences, and I hit the generate button, and there is very little for me to do after that..Without it, you're wasting hours where I could be doing clinical work, but I'm doing this instead.”

The key solutions RosterLab was able to implement include:

• Time savings - what once demanded hours of manual effort is now completed in a fraction of the time. Those reclaimed hours translate directly into greater clinical capacity and reduced operational cost. It also means a higher patient throughput for Dr. Rukshan.
• Automatic fairness distribution for on-call load across consultants - this helped ensure appropriate fellow pairing, prevented back-to-back on-calls, and maintained equitable on-call distribution across periods.
• Self-rostering - consultants enter unavailability directly into the staff mobile app, which eliminates slow, error-prone email workflows.
• Enhanced transparency - clear metrics for on-call distribution, fellow pairing, spacing between calls, easy statistical retrieval, and sharing

Dr. Rukshan also notes the new and improved capabilities of the algorithms used to discover all the best possible shift combinations to solve the problem - “Before RosterLab, it was much harder to give people all the leave that they wanted, because we weren't seeing all the possible solutions to the problem. Allowance of leave definitely impacts safety.”

Ongoing Benefits

Reflecting on the transition, Rukshan shared, “I used to spend hours trying to make everything fit… now I actually enjoy doing it.”

What began as an experiment has quickly become an essential part of Auckland Tertiary Hospital’s radiology department's daily operations.

Now that the team feel comfortable using RosterLab, they have recently started long-term planning for the year ahead - maintaining core fairness practices for the year but re-rostering each period when there are changes required.

The ongoing benefits include:

• Equitable on-call shift across the year: which supports annual fairness balancing, allowing re-rostering each period without losing overall equity.
• Staff confidence and engagement: with objective, transparent on-call statistics - staff confidence and trust has grown in the fairness of the roster.
• Better leave accommodation: AI routinely uncovers combinations and solutions that manual rostering would miss. The result is safer patterns, more approved leave, and less administrative burden.
• Clinical safety improvements: by reducing manual workload and enabling smarter allocation, the system supports safer staffing levels and enhances the quality of patient care.
• Department-wide momentum: after seeing the results, other services, including respiratory, cardiology, gastroenterology, and pediatric radiology, have expressed strong interest in adopting the same approach.

Together, these benefits have reshaped expectations across the department. What used to feel like an unavoidable strain has become a streamlined, trusted, and clinically safer process.

Why This Matters

Auckland Tertiary Hospital outcomes demonstrate that AI rostering isn’t just an efficiency upgrade - it directly improves patient care, staff well-being and other hidden costs. By removing manual bottlenecks, reducing perceptions of bias, and supporting safer staffing patterns, the department unlocked capacity that goes back into clinical work.

This shift also sets a new expectation for what modern workforce management can deliver: transparent processes, equitable workloads, and the flexibility to adapt quickly as demands change.They also continue to pilot new features like automated shift swaps and our new AI-powered features like Otto.

If your rostering still depends on spreadsheets, email chains, or “gut feel”, this is exactly the type of problem RosterLab is built to solve. Book a demo to see how AI rostering would work with your team’s rules, fairness targets, and leave constraints.