Editor’s note: Organizations that are under more pressure than ever to quickly develop Big Data insights and demonstrate business value enabled by their technology investments experience appreciate substantive and rapid wins with Predixion’s easy-to-use, self-service predictive analytics solutions. Launched in 2009, the Orange County, Calif.-based analytics provider reduces complexities and optimizes a collaborative end-user experience; its analytics solutions use Microsoft Excel-based modeling tools and a deployment framework that automates predictions from source data to line-of-business applications, dashboards or other portals.
Predixion’s solutions are especially hot these days in helping health plans, providers and Accountable Care Organizations manage the complex risks of healthcare reform’s mandate to switch from a pay-per-service model to a preventative and wellness model that improves patient care while reducing costs. In this article, Predixion’s CEO Simon Arkell discusses how the company’s solutions enable data-driven decision making.
Part of your positioning in the market today is that your solutions are designed for self-service. Was that your vision when you entered the analytics space?
Simon Arkell: We specifically started the company to exploit predictive analytics, which was at that time a hole in the business intelligence industry. We wanted to build a platform that allows companies of all sizes (not just the biggest ones with big budgets and expensive technical people on staff) to build predictive models to solve really important, expensive problems.
We wanted it to be a collaborative, end-to-end solution that is easy for non-technical people such as truck drivers, clinicians and bank tellers to use. Our tool is an add-in to Excel, and people can download our software and create really powerful, accurate predictive models using that workbench within Excel.
Our co-founder, Jamie McLennon, had been the head of predictive analytics software development at Microsoft for many years, and our engineering team in Redmond is the former data mining or predictive analytics team from Microsoft.
In 2010 we released the world’s first general predictive analytics cloud service. It was received with great fanfare and put us on the map.
Solving really important, complex, expensive problems is what the U.S. healthcare industry is facing today because of the shift to preventative care models in healthcare reform. How extensive was the tweaking you had to undertake in your general analytics platform to address the healthcare industry’s needs?
Simon Arkell: Yes, in healthcare there are many very expensive problems. In fact, just preventable hospital readmissions is a $31 billion problem. And the industry’s problems are very high value in that solving them can save lives.
We started focusing two years ago on the healthcare industry and worked with payers such as Aetna, Carolinas Healthcare System and Kaiser Permanente to understand how to modify our platform for healthcare.
This week you announced Predixion’s partnership with the Indiana Health Information Exchange (IHIE), which is an information exchange provider with a network of 90 hospitals and physicians. Several of them are Accountable Care Organizations (ACOs), where risk-sharing is a key component to success. How will your predictive analytics solution help minimize their risks? Can you give me an example?
Simon Arkell: Sure. One of the biggest challenges for a lot of ACOs is that they inherit a patient population that they had nothing to do with earlier. Imagine inheriting 30,000 new patients because of the new health insurance exchanges being launched thanks to healthcare reform. An ACO is compensated on the success of improving the health and wellness of its patient populations, so they need to get their arms around the serious financial implications and business risks they face from a patient population.
They need to know how healthy the patients are and which patients are high risk and need some intervention in their care path in order to have a positive outcome. Also, are they going to providers outside the network because the network doesn’t offer certain services, or is it because the patients are not aware of the network’s services? These are just a few of the things they must quickly discover.
They need data mining and predictive analytics to understand their risks and know how to take action. ACOs also need information about how they’re performing compared to their peers and the government’s KPIs.
Using Predixion’s analytics solutions, how fast can healthcare organizations discover the information they need?
Simon Arkell: It varies, of course. But we recently did a proof of concept with an ACO that has a brand new patient population and was trying to understand where their risks are. They already knew that within the coronary artery disease (CAD) segment they were performing very poorly compared to their peers. They knew that was a risk and recognized that there would be a financial impact in a couple of years.
Within a series of predictive models and a POC that took less than a week, we were able to show them that eight out of 800 of the doctors in their network were not prescribing aspirin for their CAD patients. So they had actionable information to help eliminate the financial risk from those providers.
That’s definitely fast — time to value and risk mitigation in just one week.
Simon Arkell: That ACO client told us that without analytics they likely wouldn’t have found that anomaly in even two or three years since it was only one percent of the doctors.
In another case our client told us that our prime competitor said it would take them 24-30 months to deploy the predictive analytics solution. But we deployed it in 90 days.
How are you able to deploy so much faster than your competitors?
Simon Arkell: First, we take a laser focus on our product development, ensuring simplicity. Second, agility has been part of the root of our success. We always build our technology innovations as a cloud approach. Even though a lot of our healthcare deployments actually end up being on-premises deployments, the fact is they were built to be a very scalable rapid architecture. The on-premises deployments of Predixion are a localized version of our cloud product.
Third, when it comes to deployability, it’s one thing to have data and have a robust predictive model and a nice report; but it’s a completely different thing to get it in the hands of clinicians. Some of the things we’ve done in this area definitely keep us heads above the rest of the market.
We connect to clinical data warehouses in real time and output the data to a really nice business intelligence type of interface on a tablet computer that a nurse, case manager or other clinician can use in real time with a patient. It’s a really nice graphically rich interface with drill-down capabilities on individual patients.
Having an accurate predictive model doesn’t mean anything unless you make it deployable. The way we deploy our solutions really resonates with our clients.
Have you found that as the manufacturer of these solutions you really are doing a lot more than developing software? Are you a go-between between the health plans and the hospitals/physicians, helping them collaborate?
Simon Arkell: I think that’s fair. The data is something that payers and providers are starting to share much more often, and I think that the analytics is a unifying factor. The industry is moving towards data-driven decision-making. In the past, the health plans really struggled because so much patient data was hands-off and it was difficult to question a clinician’s decision making.
I think that as the clinicians start to make more data-driven decisions and everyone starts to understand more and more about what went into the decision-making process, especially in preventative and proactive investments in the care paths, it will be a positive thing for the industry and the patients. Collaboration between health plans and providers is definitely happening now; there are problems that can’t be solved by just one or the other.
What’s on the horizon for your company? Anything beyond healthcare? How are the rapidly evolving trends in the analytics market impacting your go-to-market strategies?
Initially we’re looking at taking some of the solutions that we’ve already built and offering them to IHIE’s customer base. The first one will be predictive hospital readmissions, which is a big challenge in the industry.
Long term, the vision is much more exciting. We will combine our resources with IHIE’s internal data science resources to go through their data and innovate together. A couple of the larger hospitals in their network and some academic institutions also will join that effort so we can collaborate on innovations that really push the limits of what predictive analytics can do in healthcare.
Also, late last year we realized that some of the solutions that we were innovating have a place into other software providers’ products. There are a lot of established healthcare technology companies and software providers (EMR, EHR, billing systems, etc.). They have a unique position in the marketplace because they have the eyes of clinicians, they have established incumbent investments or vendor positions within these providers or payers and they have a fantastic amount of data (just like IHIE does).
So we’ve had some very interesting conversations recently, and we’re looking at opportunities to take our technology and some of the solutions that we’ve built and embed it into other vendors’ software. This is one of our key go-to-market strategies.
In addition, a couple of weeks ago we closed a $20 million SeriesC round of investment led by Accenture and participated in by GE. And we have entered into a global strategic relationship with Accenture. They are building industry solutions across a number of different vertical industries based on our Predixion platform and will distribute those to their customers globally.
It’s very exciting for us to have a global partner like Accenture that is already trained on implementing the Predixion platform and will be applying these solutions to very big problems facing their clients in various industries around the world.
Simon Arkell, co-founder and CEO of Predixion Software, is an entrepreneur with extensive global technology management experience. Before Predixion, he was a partner with Triton Pacific Capital Partners. Prior to Triton, he was president of a UK-based consulting firm developing Oracle solutions. Earlier he was a principal with Gramercy Venture Advisors. In 1998 he co-founded content-management software company Versifi Technologies (acquired by Reef SA). At Versifi, he led acquisition of AdaptiveInfo, a predictive analytics vendor.
Kathleen Goolsby is managing editor of SandHill.com.