Jonathan Landsman, CISA, MBA
Senior Director of Big Data Strategy, MiraMed Global Services, Jackson, MI
Over the past few years we’ve heard the terms “Big Data” and “Analytics.” Data and reporting aren’t new ideas. But in recent years it seems like we’ve gone into data overload. There’s too much data, too many reports and too little time to make sense of it all. Where did all of this data come from? How did it get so big? And what is analytics? Over the past five decades data reporting has evolved from almost no data to too much data. To understand how we got to where we are today, perhaps we have to understand where we came from.
The collection of data has been around since the start of the information technology age. As technology evolved, even more data was captured. Early on, the cost to store data was expensive. Data was stored on punch cards, magnetic tape, cartridges and disk storage. These types of media required that they be kept in clean, cool, low humidity, static-free environments. As a result, data and data storage was a high-cost premium, so little data was captured or retained. Data retention rules were focused on minimum necessary data to be retained at one, three, five or seven year intervals. Anything beyond basic financial reporting or summary records was regularly deleted or lost. However, over time, Moore’s law proved that significant reductions in the costs as well as the doubling the speed and capacity of technology every 18 to 24 months also impacted data collection and storage.
In the 1980s PCs offered five megabytes of capacity that equated to approximately 550 eight-page documents.
- In the mid-1990s PC capacity was one gigabyte (112,000 ct. eight-page documents).
- By the mid-2000s PCs capacity was 500-gigabytes (56 million ct. eight-page documents).
- In 2007, more than 50 years after the first computers, hard disk drives for business/commercial computer use finally reached the size of 1 terabyte (TB) (the storage of 112 billion ct. eight page documents).
- In 2009, only two years later, the first 2 TB hard drive became available.
- Today, a typical home laptop has a 1 TB hard disk.
Over time, the cost of storage dropped and the capability has grown. Data capture and retention has shifted from the minimum necessary, to how much more data could be kept. Today, almost no data is deleted.
But what’s the benefit of all of this data? How does one extract value? The answer is through business intelligence (or BI). BI is four levels of capabilities that range in complexity and increasing business value. Analytics is a part of the fourth level in this range:
- Level 1 is basic reporting. Its focus is on retrospective. It tells you what happened this past quarter, how does that compare to last quarter or same quarter last year?
- Level 2 is analysis. It attempts to explain why something happened.
- Level 3 is monitoring. Through the use of dashboards and score cards you can actively monitor what is happening now.
- Level 4 is predictive analysis, analytics and modeling. Through the use of analytics you can anticipate or plan what might happen.
It’s been said that “past performance is an indicator of future performance.” Although situations change, predictive analytics can be used to understand why things in the past may have occurred. We can create mathematical algorithms that use this historical data to describe why these situations occur. Using these algorithms to follow current data on a real-time or near-real-time basis gives us the opportunity with definable, repeatable processes to improve our operations when these specific circumstances recur, resulting in improved outcomes—actionable insight in action.
Applying variables to these mathematical algorithms gives us the opportunity model to anticipate changes and trends, which, in turn, allows us to shift gears in our operations as situations evolve.It’s been said that “past performance is an indicator of future performance.” Although situations change, predictive analytics can be used to understand why things in the past may have occurred. We can create mathematical algorithms that use this historical data to describe why these situations occur. Using these algorithms to follow current data on a real-time or near-real-time basis gives us the opportunity with definable, repeatable processes to improve our operations when these specific circumstances recur, resulting in improved outcomes—actionable insight in action.
So how is predicative analytics and actionable insight being used today, specifically in healthcare? At the recent HIMSS 15 conference, as elsewhere, analytics has been the hot topic. Although data plays a role in most IT-based solutions, for analytics solutions it’s the primary cornerstone. There were a number of educational seminars focused on analytics. Two such seminars provide perspective into analytics which can be used in both clinical and administrative settings.
- Creating predictive analytics to improve clinical outcomes is a question that most providers are focused on. The Carolinas HealthCare System (CHS) hoped to solve some of today’s toughest healthcare challenges by leveraging predictive analytics. Their goal was to minimize readmissions.
They used real-time data and clinical analytics to predict which patients are at the highest risk for readmission. Their nurses provide stepped-up interventions with these high-risk patients before they leave the hospital to minimize the possibilities of readmission.
- Revenue cycle management payment models continue to evolve. These changes have and will continue to have a significant impact on all providers. One aspect of the revenue cycles is that the patient share or patient responsibility continues to grow. Models, methods and processes to collect the patient share and/or self-pay must continue to evolve as it becomes a greater percent of the payment ratio mix. Northwestern Memorial Hospital (NMH) discussed their use of advanced predictive analytics and improved automated workflow that optimized their revenue cycle performance in comparison with conventional manual approaches.
NMH used analytics to understand the prior payment habits of repeat patients to optimize payments with patients that regularly use NMH facilities. Their analytics allowed them to categorize some recurring patients into a number of categories. Lessons learned were to: minimize calls and letters to probable paying patients and to focus on patients that required greater attention by reaching out to them sooner.
From these and other examples, it’s clear that analytics offers great promise to address a number of different challenges providers face whether they are practices, hospitals or health systems. Analytic solutions are based on the data and the understanding that past performance is an indicator of future performance. Adoption of analytics is transitioning from the early adopter phase to that of the mainstream. Analytics-based solutions will continue to evolve, as will the challenges that analytics addresses. The value of big data isn’t just about insight, but also to leverage that insight, that is, actionable insight in action.