Future of Data Analytics in Healthcare: Part I – Basic Framework

As I talk to my peers in healthcare about data analytics, I find this is a very misunderstood area.  Some people state all health care will eventually revolve around data analytics. Others believe it is not needed to practice good medicine. The answer is probably somewhere in between. Therefore, I wanted to start the discussion on data analytics with some basic concepts. This could seem boring to the data analytics expert, but may help open minds to the concept and how they could be used in a more simplistic manner.

In healthcare, we continuously talk about improving the quality of care, improving patient safety or lowering costs. The problem is this: “How do we actually go about doing this?” The questions are easy to ask, but the answers are sometimes hard to obtain. I want to spend some time talking about the basic framework of data analytics – to help even the non-expert understand the concepts.

The answer to how we do this depends upon different settings, such as acute care hospitals, ambulatory clinics, physician offices, home health or long term care. Even though the answers may be different, the process of determining “what to do” can be quite similar! So, let’s start our journey.

In order to change something, such as patient flow, patient safety, infections or cost, you first have to measure where you are at this point in time. In basic terms, to measure something you need data of some type (like numbers on a ruler to measure length; or number of pounds/grams to measure weight). Then, if you want to change that data, say your body weight, you have to measure how much you weigh over time. Thus, you are collecting a single data element many times, over a period of calendar time. At this point, you have just created a database that can be analyzed for changes in body weight. This is data analytics!

As we migrate away from paper towards an electronic health record (EHR), we will have the ability to collect many different data elements at different times. The collection of these data elements can be part of the normal use of the EHR in the everyday practice of medicine. Thus, you are automatically collecting detailed data elements longitudinally over time.

Before you start building the database, you need to know what problems you’re trying to solve. For each problem or goal, you need to ask key questions such as:

What type of data output or data view do I need to measure the change?

• What data elements should I collect?
• How often do I need to collect these data elements?
• What is the structure of the data elements?  Are they “fixed” or are any of these “free text?”
• Can I collect this data element in the normal course of patient care using the EHR?

My next blog will examine in more detail these questions and some of the typical mistakes made in developing a data analytics database.