Did you know that punctual patients can save a hospital tens of thousands of dollars? One Australian hospital found that it could save over A$66,000 (US$49,967) a year when patients turned up for their appointments on time.

“We make assumptions when we design any government service, government policy. We assume people will respond in a certain way,” says Dr Alex King, Director of Behavioural Insights (BI), Department of Premier and Cabinet, in the Australian state of New South Wales. In reality, our behaviour is quite complex and sensitive to the situation we’re in, he says.

King’s unit brings in a growing body of knowledge about how people behave, which includes fields like psychology, behavioural economics, sociology, anthropology and neuroscience. “To my mind, it’s really about bringing in a more realistic model of how people behave,” he explains.

The unit examines how changes to government services, communications and policies can ‘nudge’ people in the right direction, like paying their taxes on time or cutting energy use. GovInsider spoke to King to find out what has worked, and how.

How SMSes can save $66,000

Alex King

Let’s go back to the Australian hospital. How did they do it? St Vincent’s Hospital estimated that it loses A$125 (US$94) for each missed appointment, despite sending SMS reminders to patients. However, the content of the messages matters a great deal.

The hospital tried seven different text messages, and found that one cut missed appointments by 19%, King says. This message simply said: “If you attend, the hospital will not lose the $125 we lose when a patient does not show up.” It told patients that “the cost of you not turning up is actually real money that is incurred to the hospital”, he explains.

In another project, the BI team worked with the Cancer Institute of New South Wales to get women to attend cervical cancer screening appointments. It found that tweaking the reminder letter could lead to up to 7,500 more women attending their appointments within three months of receiving it. “For the very small cost of just changing the template, it is a huge gain,” King says.

Helping injured teachers

One of the more complex trials the team has run involves getting injured teachers back to work. With the right kind of communications, they were able to get teachers to return to work 27% faster.

The result came from a combination of different techniques, but there are two of most interest, King says. One is to focus on positive messages of recovery and wellness, rather than reminding patients of their injury. In the past, they received a letter five days after their injury saying “You have a serious injury”. The new letters encouraged people to take steps like taking medication or attending physiotherapy.

The second technique was to get the teachers to plan for their own recovery. “If you want someone to do a thing, it’s much better if they have thought through how they are going to do that thing,” King explains. Injured teachers were urged to make their own plan and commit to steps towards recovery.

How do they do it?

The team follows a three-step process in all of their projects. First is to understand the way people currently behave. They use a combination of data, interviews, workshops and surveys with frontline workers and customers. They also look at academic literature on the topic and speak with experts in the field.

Next is the “build” phase, where they try to design a change in the system or create a new aspect of the system which will encourage people to behave in the desired way. This change can range from something as simple as rewriting a text message, to more complex tasks like “reforming an entire service” or “retraining a lot of staff”, King says.

The unit works closely with the partner agency to design this change. It must be something they can implement and scale. “There’s often many things that we can do, and we’ll try to pick those that really only are viable for them,” he adds.

The third is to test this change on a portion of users to see if it works. “By running a trial, we can show that having tangible benefits for them and they can understand the costs and benefits of doing that,” says King.

Ideally, they would use a method called randomised controlled trial (RCT). Let’s say a clinic wants to test a new appointment reminder. Patients are randomly divided into two groups, one group gets the new reminder and the other gets the original. The clinic collects data on the results from each group and then compares them to see if the new reminder made any difference.

However, the team is open to using other piloting and trialling techniques to build an evidence base of what works. Statistically, RCTs give the “best possible evidence”, but sometimes it’s not possible to carry them out, King says. There can be legal or ethical barriers that do not allow officials to run such experiments. There are also practical hurdles like IT systems not capable of handling the tests.

True to their roots

NSW’s BI unit itself is a big experiment. “We were, in the best tradition of our own field, set up originally as a trial,” King says. They were able to demonstrate the value of their work through a number of early wins, allowing the unit to grow.

The unit was set up at the end of 2012 by contracting members of the UK government’s behavioural insights team, which was then the pioneer in the field. It has now grown to about 12 full-time staff with backgrounds in behavioural sciences, government policy, sociology, market research, qualitative research and data analysis.

King has been running the unit for the last one and half years. His original background was in astrophysics, and he’s spent time working in the UK Prime Minister’s Strategy Unit and USA’s NASA. This role, he feels, is taking him back to his roots as a scientist. “It’s all about bringing science into government. A slightly different form of science than I started with, but nonetheless, still science,” he says.

His team is currently set up as a centralised unit, but there’s no telling that it will remain this way. There are discussions around whether it should move towards individual teams embedded within agencies. But “it’s still fairly early days” and “we haven’t really decided on the best course yet”, King concludes.

True to their own field, the nudge team must constantly look for evidence of what works best and adapt to that model.

Main image by by Seb Zurcher – CC0