Health Policy

Predictive Analytics:The Change Agent in Healthcare Industry

Views: 998

Paddy Padmanabhan, SVP “ Healthcare Analytics, Symphony Analytics

There is a shift in perception among healthcare utives today when it comes to analytics. EMR systems have been implemented successfully. But the use of EMR data to gain critical insights remains relatively immature across the healthcare sector.

Hospital utives say that EMR vendors don’t do a good job of enabling analytics, so instead, they have turned to independent vendors and solution providers to fill the huge and emerging demand. They also find it difficult to get data in and out of the closed systems – as data sits inside as well as outside of these systems, there’s a great need to integrate and synthesize it on a real-time basis. But there’s light at the end of the tunnel: more and more Chief Medical Information Officers (CMIOs) and CIOs today are looking past their traditional EMR systems and putting together effective analytics strategies using a combination of third-party solutions and in-house expertise. The vendor landscape is also maturing with promising solutions and offerings. We hear many stories of strategic partnerships between healthcare organizations and solution providers that are delivering value.

But what is driving healthcare analytics?
Clearly, the transformation of healthcare from a transactional to an outcomes-based model – is the force at a macro level. But once we dig deeper, a range of specific issues comes up. For instance, payers are looking at opportunities arising from the individual market, at the same time they are worried about the underwriting risks when they sign up individual members they know little or nothing about. Providers are focused on population health management, but more specifically readmissions, not only just for the financial impact of the penalties, but also the reputational damage that ensues when a hospital is penalized by Medicare for failing to keep readmissions in check. Thus enters the science of predictive modelling.

Faced with the realization that analysing historical data to understand what already happened is no longer adequate, the more progressive healthcare companies are now developing and wielding predictive models as a strategic tool for improved outcomes, reduced utilizations and margin expansions. For example, using real-time integrated data feeds from clinical and financial systems to predict readmissions and then push notifications at point of care to clinicians for effective interventions. By combining structured data analysis with natural language processing (NLP) capabilities, one could predict a range of scenarios to help keep costs and utilization down. Other applications also include reducing ER admissions by tracking dispositions in healthy populations, predicting average length of stay (ALOS), hospital acquired conditions (HAC) and so on.

In general, population health andcare utilization are the biggest analytical issues today in managing healthcare costs. Traditionally, healthcare analytics is used for measuring and analysing operational and tactical matters, focusing on metrics like revenue per bed, PMPM costs and the like. New requirements, such as population health analysis however, will determine the survival of health systems and ACOs alike, where the whole focus is to predict what is likely to happen and thus intervene before it happens. This applies especially to the chronic cases that now sit in the top 15% of the population risk pool.
Armed with predictive modelling and scoring tools, CMIOs can now have conversations with CMOs and physicians about outcome quality and variability in care, based on best practices and how to stratify populations and manage risk pools. This is a whole new world of opportunities forhealthcare thats opening up, for progressive organizations that are looking to thrive in the post-reform era.
On the other hand, C-level utives in mid-sized hospital systems, faced with stagnant to declining in-patient admissions rates, are now focusing on costs to stay afloat. In this uncertain environment, tactical and operational analytics focusing on historical data is no longer a reliable guide for strategic decision-making. Real-time predictive analytics based on integration of data across siloes is almost necessary for survival. Organizations need to be fanatically wed to data analysis to compete in this landscape.

Clearly, it all boils down to making smart investments in analytics infrastructure and computing the returns on this investment. And though there are no clear benchmarks today, but by focusing on specific problems with identifiable cost implications, such as readmissions, returns can be computed using a combination of cost avoidance and reduced utilization levels. What’s encouraging is that CEOs and CFOs are more willing to listen to these business cases.

A handful of healthcare organizations have made aggressive investments in setting up a dedicated analytics infrastructure and building teams of decision scientists and technologists who can take data and turn it into insights. In many cases, they are partnering with external firms to deliver this insight across stakeholder groups on an on-going basis to drive change in behaviour to influence costs.Many payer organizations, especially among the market leaders, are aggressively prowling for analytics companies and snapping them up to give themselves a step up against the competition. For mid-size organizations with limited options, the alternative to partner with a vendor that understands healthcare and brings advanced analytics capabilities that can also build it out over time- is a viable option. Either way, to have a strategy around predictive analytics is no longer “optional” but “required” moving forward.

Most Popular

To Top