In my career, I have spent many hours writing doctors’ rosters – every month, reviewing individual requests and leave entitlements, ensuring equity, and balancing preferences. In most emergency departments there will be several people undertaking this for senior and junior medical rotas and nursing rotas, taking up many hours of clinicians’ time each week. And then there are unexpected absences and vacant posts. No wonder it is not a popular role. In my early career, I was an early adopter of spreadsheets and produced a rolling rota for the whole six months of a junior ED attachment. They then had to swap internally. It made it easier for me – but did they like it, did it lead to more “sick leave”?There is a real skill to writing a good rota with so many conflicting priorities:
- The individual has shift preferences but also specific areas of interest, training needs, and home needs. Individuals also have their own pace of work.
- The team needs to work coherently to be safe, so all starting together and having regular shift mates is an advantage for safety and efficiency but a disadvantage for close matching of capacity and demand.
- The department needs to ensure a full range of skills is available each shift.
- The organization needs to match the staffing levels to the workload demand to optimize patient safety and flow.
But is this too complicated for one person to juggle in to a fair and efficient roster? It is complicated but not complex (complicated problems can be hard to solve but once solved can be addressed with a set of rules in the future; complex problems have many interrelated variables and so may need a unique solution on each occasion). Most manual rosters do not account for the variation of attendances by day of week, month of year, or predictable events. However, it is not just about numbers but also case mix (e.g., the proportion of very sick patients increases overnight). Most rosters I have seen are based on historical data, despite the knowledge that there has been a 4–6% increase in attendances for the last few years – potential the roster may need to be different every day of the year!
It has been solved in other sectors. Network Rail has adopted a digital solution for over 25,000 of its signaling and maintenance workers with very different rostering cycles, by optimizing which employee works when, and on what part of the rail network, factoring in compliance requirements, job priority, and employee preference. It uses a demand-driven model to match resources/capacity to demand over one-, two-, four-, and 13-week rostering periods. It uses competency and skills from the HR system when creating a roster and checks the expiry date of the skill for each person and warns the roster manager. It has many rostering rules that incorporated the European Directive working time rules and added the HSE fatigue index algorithm. Finally, it creates a timesheet incorporated overtime and sends data to payroll for 25,000 staff. The similarities with healthcare rosters are obvious.
While in Canada a couple of years ago, I came across a similar solution that had been developed by an ED team. MetricAid has been building rosters based on capacity and demand as well as individual preferences for Canadian hospitals for nine years. MetricAid has developed a scheduling platform that collects patient flow data from hospitals, and self-scheduling data from clinicians, merging this data to build rota that are balanced for all considerations (flow, patient safety, cost, and staff satisfaction).
MetricAid was motivated in the Canadian market to lower wait times in the ED, which it has done for more than a dozen large hospitals there. MetricAid’s capacity and demand algorithm had this sorted out by 2013. Since then, MetricAid has worked on its self-scheduling component, which allows clinicians to plan out their rota one shift at a time, to ensure that they a) will be happy enough with the published rota that there are reductions in shift trades and sick time, and b) work exactly the shift MetricAid wants them to work, based on the needs of the department.
MetricAid has been deploying at the Royal Derby Teaching Hospitals Trust since 2017, and was able to hit Lord Carter’s recommendation of implementation for October 2018. The rota for October 2018 saw a 50% reduction in alterations to shifts (either filling vacant shifts, filling sick calls, or facilitating trades), and a 25% reduction in HR overspend (£51,000).
As well as the predicted advantages of giving key clinicians more time for caring for patients, it has also resulted in a workforce that feel more in control of their working life. In the words of Richard Branson, “Clients do not come first. Employees come first. If you take care of your employees, they will take care of the clients.”
So, in 2025, when you work an emergency department this may be the story.
This week, I have worked my usual shifts which fit around my personal preferences and my home commitments. I had to swap one shift but that was easily done on my app, which found someone with a matching skill set who liked working Wednesday evenings. The shifts were fully staffed as the system automatically checks and finds someone able to do an extra shift. The system even predicted the increased workload because of the weather and added an extra nurse practitioner. As Clinical Director, I spent fifteen minutes validating next week’s roster and saw the alerts for staff whose mandatory training was about to expire and therefore would be removed off the roster. Most importantly, the team all feel in control of the work-life balance and so are much happier and therefore give more compassionate care.
Digital solutions should be about making life easier for staff and more efficient for the department as well as the direct benefits for patients. All these improve outcomes.
Matthew Cooke is the Chief Clinical Officer, Capgemini, UK. You can visit his expert profile here or write to him at email@example.com
Les Blackwell is the CEO of MetricAid, Canada. You can write to him at firstname.lastname@example.org