Algorithm 2: Case Study 2


Let’s be honest. When it comes to discharge, everyone wants to blame everyone else,” says Tracey Nixon, R.N., nursing director of capacity management. Doctors haven’t written discharge orders, physical therapy hasn’t come, pharmacy is backed up, nurses are busy with other patients . . . the list of finger-pointing goes on. This lack of coordination is not lost on patients waiting for hours to go home. From a system perspective, it’s inefficient. From a payer’s perspective, it’s expensive. “Patient throughput is a choke on hospitals,” says Nixon, also known as the Czar of Patient Flow. 

But what if discharge is not actually the problem? What if the problem is much less intuitive and much further upstream: admissions? No one knows their names or medical condition, but we know – anecdotally and statistically – that new patients will be admitted every day. Yet curiously when they arrive, we are caught completely off guard and frankly inconvenienced. How could something so predictable be so surprising? 

The problem, Nixon explains, is that no one has ever created a workflow that accounts for the time it takes time-strapped nurses to get a patient settled. “This is just one example of how we don’t use historical data in health care to create workflows that are realistic,” says Nixon. Instead we’ve relied on feelings and perceptions. We’ve also misaligned incentives. “By holding nurses accountable to discharge patients by a certain time, we take away the incentive to admit the patient waiting for hours in the ED or the PACU.” 

Guided by demand-capacity management and queuing theory, Nixon set out to pilot changes on two units. If the theories are highbrow, the solutions are low-tech, mostly involving better communication and planning on the unit level and the system level. Being proactive about admissions instead of reactive about discharge had a ripple effect. In just the first two months of launching the program in two units, discharges before 5 p.m. increased 9 percent. Even more impressive, the units were able to predict admissions and discharges with 87 percent accuracy, a 50 percent improvement. 

Thinking about patient throughput was a huge culture change, Nixon says. Now everyone is accountable. “If I ask a physician to write discharge orders and that patient is still sitting there five hours later, I’ve lost all credibility. It’s a deal breaker,” says Nixon. “Everyone owns a piece of the discharge pie. We’re one system. We’re all partners.”