Reading Time: 5 minutes

By Ed Simcox, Healthcare Practice Leader at Logicalis Healthcare Solutions

Analytics” refers to the process of identifying and gathering of all relevant data needed to solve a problem or answer a question and then applying a methodology to that data to make informed decisions. There’s also something called “big data analytics,” and that refers more specifically to identifying and leveraging disparate data sets that may be so complex and large, that they are difficult (or impossible) to manage with a traditional set of software and hardware.

In addition to analytics and big data analytics, there’s also a third approach, which is the simplest, and sometimes all that’s needed. That’s what we call data analysis.

“Data analysis” is the process of identifying, exporting and/or merging small, targeted subsets of data to glean insight and value. And the data analyzed typically resides only in one or two software systems. There are well-known electronic medical record system providers that say they have analytics modules for their customers right now. But for the most part, these modules actually only offer a form of data analysis, where they are doing simple, repetitive reporting on their own native data.

How Healthcare Organizations Use Analytics

Healthcare providers and insurance companies want to use analytics to help them make important clinical, business and financial decisions. One common use is to help healthcare organizations complete their journey from an encounter-based, fee-for-service revenue model toward what is known as a “value-based care model” — where healthcare systems are compensated for providing value and quality rather than volume of care. This shifts the financial mindset and revenue model of the organization from quantity to quality and requires heavy use of data.

If a healthcare delivery organization cannot deliver data-driven insights necessary to get paid in an optimal fashion in a value-based care model, the operating margins will suffer. Revenue is critical for all healthcare organizations, including non-profits.

No margin, no mission.” – Sister Irene Kraus, Founder and Chief Executive, Daughters of Charity National Healthcare System.

Analytics Seems Difficult  

Across all industry verticals, 80-90% of organizations are trying analytics, but only 6% have managed to do it successfully—according to a 2014 Accenture study.1

Analytics is frequently thought of as an IT or informatics function, and therefore it commonly lacks enterprise-level support, ownership and governance.  IT and informatics organizations have become pretty good at setting up canned reports and running those reports over and over to get good trending information. That’s not data analytics, that’s “data analysis.”

But there are certainly pioneers doing great things with analytics in healthcare. Most of them seem to be home-grown, and when they are successful, they may end up being commercialized, like Health Care Dataworks out of Ohio State University and Explorys, an innovation spinoff of Cleveland Clinic.

A “Typical” Analytics Solution

There really isn’t a typical analytics solution. A recent study showed that over 350 software organizations claim to have tools that are essential to complete successful analytics implementations.

So what do a lot of healthcare organizations do? They buy tools without first developing a strategy. Sometimes this is the result of the hospital administration asking the business or IT folks to try to rapidly solve important problems with data without an allowance of time to properly plan. Unfortunately, many organizations purchase a complex set of these expensive analytics tools—hoping to accelerate the process—but never quite understanding how to use them, much less use them together in a cohesive manner to solve complex problems.

However, there are mature healthcare systems that have good organizational buy-in, enterprise-level governance, the right data toolset, and very smart people.

Data Warehouses and How They Relate to Analytics

Analytics typically relies on setting up a data warehouse. A “data warehouse” is a large repository made up of copies of business and/or clinical data from multiple sources. It may contain billing information, patient satisfaction scores, demographics, labs, telemetry, and any other information from which an organization wants to derive value.

Setting up a data warehouse involves identifying target data sets and copying the data into the data warehouse during a period of planned downtime for the systems that contain the data. The data must then be normalized and rationalized. It also needs to be frequently refreshed as new data enters the source databases.

Then, of course, the warehouse must be backed up and/or replicated to protect it. The compute, storage, and networking resources—even for moderately-sized data warehouses—can cost well into seven figures. Setting one up can be expensive, disruptive, and time-consuming.

A New Approach to Analytics

Conventional wisdom is that one must conquer data warehouses or big data repositories to do complex analytics. And until very recently, that has been the case.

Now there is simpler, quicker technique—Data Insights. Its simplicity provides the quickest path to unlocking the immense value hidden inside of healthcare data. This new approach involves a rapid, low-risk way to evaluate and utilize all information types, including big data.

To help visualize this, think of how Google indexes the Internet. They go out and directly “crawl” all kinds of data sources in multiple languages. As Google crawls this data, they index it. Google doesn’t require a data warehouse to do what they do. Their solution is able to directly access any data set it can see and create a small index of it. Much like the index in the back of a book, the Google index includes information about words and their locations. Because of this, Google’s index is much smaller and more efficient than the source data.

In many cases, the Data Insights approach can now leapfrog setting up a data warehouse because we can also directly crawl target data and create indexes of that data. Then, every so often, we can update the index with only the small incremental changes to that source data. We only have to index the entire data set once at the very beginning. This is a very efficient process that requires very little overhead from the source data systems.

Once we have all the data indexed, we can very quickly use it to solve problems. And by using this technique, we can often avoid the need to use data warehouses. However, existing data warehouses are still useful as sources of data using this technique.

And by using this Data Insights technique, organizations can shift resources, such as time, money, and really smart people away from caring for data and focus them on analyzing and utilizing the data in order to make important decisions. By removing complexity, we gain a tremendous amount of efficiency.

Where Analytics Is Heading Next for Healthcare

There is a movement towards near-time and predictive analytics. In the healthcare industry, new technologies are becoming available that enable providers to gather data in near-time and then use predictive analytics techniques to help discover what will likely happen next.

For instance, by monitoring patients with chronic conditions using wearable devices, healthcare organizations can proactively intervene and remotely triage sick patients. This shifts the care from reactive—where patients show up at the ER—to proactive. This also improves care, saves lives, and lowers healthcare costs. This is made possible through near-time analytics.

Data is stale as soon as it is loaded into a data warehouse, and this prevents near-time or real-time data analytics. Proactive care thus requires direct access to data sources (as delivered by Data Insights) for analysis instead of relying on data warehouses as intermediaries. This liberates the data and empowers healthcare subject-matter experts to better use data to significantly improve health outcomes.

The Future is Now

At Logicalis, we’re excited about what we are doing in the healthcare Data Insights arena. We know people who have spent a long time working with data are typically skeptical of our approach. And that’s completely understandable. We offer free proof-of-concepts for our healthcare customers to prove out this new technique. If you are interested in learning now, please contact us.

Citation

  1. Accenture (2014). Industrial Internet Insights Report for 2015.