'Is some strange magic at work that conjures up empty beds so suddenly? Sadly not - it's just system dynamics'

Every weekday a strange ritual happens in hospital meeting rooms across the country. A group of people gather together for a collective panic about bed availability.

The story is the same wherever you go: accident and emergency is starting to back up and soon the four-hour target will be breached. Unless something is done, cardiology might cancel patients - with target breaches. Bed managers are sent to find empty beds. Two hours later, they return with stories of unreported empty beds, or beds that will be free by teatime. 'It's not enough!' cries the group leader. 'Find some more.' Two hours later, people are calm. Somehow, free beds have appeared and everyone leaves exhausted, knowing this daily ritual will be repeated.

Is some strange magic at work that conjures up empty beds so suddenly? Sadly not - it's just system dynamics. In an article by Kate Silvester, we discovered queues were normally caused by variation in demand and in capacity, rather than a shortage of capacity (see 'Kate Silvester on waiting lists'). Bed shortages are a classic example of this phenomenon. However, there are two mechanisms at play.

First, each department holds a local queue to ensure staff are not kept waiting. This prevents seamless patient progress from one stage to the next. Second, the whole system succumbs to demand amplification, meaning any demand or capacity variation is intensified by over-reaction. So as the queue is passed from A&E to x-ray and back again, both departments are 'swamped' by the higher peaks and lower troughs around average original demand at the front door of A&E. The situation is made worse when there are large time lags in the system and information is uncertain. The system can be destabilised by these effects.

Demand conundrum

Healthcare demand, especially emergency demand, is relatively stable compared with other service sectors, such as selling airline seats. However, the biggest determinant of demand variation in hospitals is artificially introduced. When we look at elective demand, the system, in theory, has time to smooth the number of patients as they enter the hospital from the waiting list. But instead, patients are batched to optimise use of surgical sessions and this magnifies the variation in demand. So Mondays are really busy for electives, and the predictably worse Monday 'no beds' crisis in A&E is utterly self-inflicted.

Now the system dynamics start to get interesting. As batching happens at every step along the clinical pathway, admissions and discharges become massively unsynchronised on daily and weekly periodic cycles. On a daily basis, arrivals happen from early in the morning to late at night. Discharges are still compressed to a relatively narrow time span in the late afternoon and early evening. This is why the daily beds crisis often melts away by 6pm. On a weekly basis, Fridays are almost always the busiest day for discharges, especially Fridays before holidays. Just think about the state of beds on Christmas Eve.

People often blame social services for not dealing with hospital discharges. Is this ever likely when the events we manage convert a relatively smooth demand into chaos?

Paul Walley is associate professor in operations management at Warwick Business School.