What data-driven thinking is doing to help LifeSpan save lives

While most data professionals are tasked with driving growth for their organisation, for Rachel Green, the opposite is true.

As the director of LifeSpan at mental health research organisation, the Black Dog Institute, Green is using data to tackle a problem that has been stubbornly trending upwards in recent years, and breached 3000 individuals for the first time in 2015: Suicide.

“Each one of those numbers is a person with loved ones, an extended family and a social network and colleagues that are impacted by that loss,” Green says. “We know that for every suicide there are a significant number more who attempt suicide, and we calculate about 71,000 suicide attempts in a year.

“This is a really wicked problem, and the impact it has on the community can last for decades.”

Now Green is leading a data-driven effort that aims to reduce that rate by 20 per cent annually. With expertise from the Centre for Research Excellence in Suicide Prevention and analytics technology provider, SAS, LifeSpan has analysed the effectiveness of suicide prevention studies around the world to create a nine-point prevention strategy.

“These were studies that showed either a reduction in deaths, a reduction in attempts, or an increase in demonstrable protective factors,” Green says. “By combining those strategies and coming up with a framework to put them in place in a meaningful way, we predict a reduction of 20 per cent in suicides and 31 per cent in attempts.”

Bringing disparate data sources together

LifeSpan is now being progressively rolled out through four trial sites in NSW. Selecting the sites and implementing LifeSpan requires the analysis of data from many sources, including coronial data, NSW’s Police’s Computerised Operational Policing Systems (COPS), ambulance data and hospital data.

“We, along with the Australian National University are coding that data and mapping it geospatially over a couple of decades,” Green says. “What this means we can do is start to pinpoint areas where clusters might be occurring.

“This is powerful information, which means for example we can look at a physical location where a number of suicides are occurring and think about evidence based opportunities to improve infrastructure to reduce the likelihood of suicide. This is called means restriction, and it is one of the most effective strategies in suicide prevention.”

Green says bringing together disparate data sources in this way overcomes the current poor understanding of evidence.

“Even health districts and local government agencies may currently have limited access to good information about what is going on in their region,” she says. “We look at the data, map it geospatially, and produce a report about what is going on in their region, combine this with information about distribution of evidence-based suicide prevention services, and then attach to that the evidence about what they could do.”

Another component of the program is the gathering of information via surveys distributed to workers who might have high levels of contact with people who are thinking about suicide, in industries such as aged care, social welfare and mental health services.

“What we are interested in doing is capturing this kind of data across employers and service providers, to starting to find out where the risk is and where the expertise is, and whether they are in the same place,” Green says. “This enables local suicide prevention governance groups to do better prevention planning, to put safety measures in place and develop a regional plan which is really targeted to their local needs.”

LifeSpan is also working to enrol people to an evidence-based training course called QPR that gives them the skills to know if someone is at risk and be able to do something about it.

“Just like CPR, it gives you the basics and how to do it,” Green says.

Wider applications of data-driven thinking

The more data that LifeSpan can gather, the more it can do to bring suicide rates down. Green says other projects have already shown the benefits of a data-driven approach, including one study at Florida State University by psychology researcher, Jessica Ribeiro, that used machine learning to predict suicides of people within the health system.

“They had an 80 to 90 per cent success rate of predicting who was likely to attempt or die by suicide, and the accuracy of the prediction got higher and higher the closer it got to the event,” Green says.

Green says SAS has been instrumental in pulling the program together, through support and assistance including helping LifeSpan set up its data environment, and providing access to data specialists.

“What we are trying to do is look at multiple data sets at the same time, and that is different to looking at linked data, which is where someone has done a particular project lining up different records,” she says. “Often that is not possible. So being able to use the software to analyse and ask the same questions of different datasets makes a whole new level of analysis possible.”

The LifeSpan program is now being rolled out in a step-wedge pattern across four sites at four-monthly intervals, commencing first in Newcastle. The program will run initially for two years in each site, with sites tasked with making the strategies sustainable long term.

“The stepped timing means that from two years’ time onwards we should start to see changes in process outcomes, and also decreases in rates that should roughly follow that step pattern.”

Green is buoyed by the success of a similar multi-factor approach in Nuremberg, which delivered a 32 per cent reduction in suicidal acts.

“No strategy alone is going to be sufficiently powerful to achieve a significant reduction in suicide,” she adds. “But by putting a systems approach in place, multiple actors in a community have different roles and are better skilled at those roles, and therefore more likely to pick up on someone who is at risk and intervene early.”

“It is possible to prevent suicide. We have the evidence, we just need to share and implement it. And data is the key to that.”

Follow CMO on Twitter: @CMOAustralia, take part in the CMO conversation on LinkedIn: CMO ANZ, join us on Facebook: https://www.facebook.com/CMOAustralia, or check us out on Google+:google.com/+CmoAu