This week we interviewed Joachim Werr the CEO of Health Navigator, a company that helps predict and prevent avoidable urgent and emergency care by combining data and artificial intelligence with proactive health coaching.
We asked Joachim how Health Navigator started:
I was an emergency room physician, trained in Stockholm, and I would regularly see the same patients in A&E. It’s estimated that 1% of the population account for over 30% of all A&E attendances and over 50% of all non-elective bed days.
If you look from a hospital perspective and at the top 50 patients that are visiting A&E, they normally have over 2,500 A&E visits in a single year. The thing that drove me clinically was I would see the same patients coming back and we didn’t have a mechanism to intervene.
The approach we took in 2010, what we did, we developed a statistical algorithm to track patients and assess their risk of coming back to A&E. We just did that as a part of mathematical, statistical work and we took that algorithm and put in on the server of one of the large hospitals in Stockholm, and started to identify patients who are high intensity users.
That’s where it all started and we turned our efforts into what became to be Europe’s largest trial into AI, predictive tele-coaching, and that trial was published in 2016 in the European Journal of Medicine. It was a 12,000 patient fully randomised trial.
In 2015, through academic contacts, we introduced the model here in the UK. It’s basically an algorithm customised to a trust and it identifies patients at high risk. Those patients at high risk you approach with a team of nurses, now employed by Health Navigator. The essence is you find the patients, approach them with a skilled nursing team and make an aggressive (in a positive way) care coordination and coaching programme for these patients to reduce the need for unplanned care. In Staffordshire we have seen a 59% reduction. 2 of our sites have demonstrated that these patients who receive the intervention actually live longer.
We came to the UK and we contacted the Nuffield Trust and we started with them a national clinical trial to evaluate the health navigator intervention. The trial has gone on for 4 years and has recruited 1,700 patients. Interim results are very strong, patients have reported better quality of life and pressure on A&E goes down.
What does the next 12 months look like for Health Navigator?
The next 12 months we are running and completing the clinical trial. We of course want to launch pro-active health coaching as a clinical service to trusts, STPs and CCGs. We now have the data from the trial which is very strong and we want to launch broadly as a clinical service, benefiting patients and the system.
The potential is huge, the NHS estimates there is somewhere up to £6 billion in preventable care in the system.
What challenges are there to overcome?
The IT structure and the way data is organised in a trust is slightly different from trust to trust. The digital readiness varies as well. So one challenge is there is no one structure or one API or interface to plug into, you have to basically somewhat customise your solution for every trust.
Another challenge is the model is based on the data that goes into the model. If there are delays in coding or being behind on records, then it effects the model, you don’t have the data you need. The model is there but there’s no in data.
Smaller companies often underestimate there is no uniform standard or interface in the NHS and there is detective work to be done with every trust first to understand how we can and they can adopt the model.
It’s only when the team of nurses and the tech work as one unit the model works. It’s important to get a model that acts on insight, and front line delivery of nurses with tech, then you can really capture how to improve. That discussion is much harder if you are an isolated tech provider which many SMEs are.
Also the NHS needs to develop strategies for transforming savings in activities to cashable savings. So if A&E attendances go down, how can you take those activity savings and turn into monetary savings. If we take out say 500 bed days, that frees up resources, how do we capitalise on those savings. All funding is done reactive. This is very similar in Stockholm, there’s not a big difference either in the funding challenge there.
