Get started Bring yourself up to speed with our introductory content.

Workflow of data in healthcare: From requests to discovery, analysis

Providers must focus their analytics to tackle one question at a time, because of the massive amount of healthcare data our systems generate.

Healthcare analytics comprises the system of tools, techniques and people required to consistently and reliably generate accurate, validated and trustworthy business and clinical insight.

Data, then, is the raw material of analytics.

Therefore, high-quality data -- and the structures and processes in place to achieve it -- is essential for obtaining valid insight into business and clinical operations.

Healthcare providers have always collected and relied on data. Before the advent of affordable home computers and the near-ubiquity of health information technology, however, data in healthcare was mostly stored in paper charts and was difficult to access.

More and better accessible data than ever before

The process of accessing data has changed dramatically at most healthcare organizations. Data is now being generated and used almost continually as healthcare professionals such as clinicians, administrators and analysts rely on health IT systems for providing patient care, conducting research and quality improvement, or managing facilities and practices.

Patients, too, add to available data with their mobile health apps, documenting and managing their health and health-related lifestyle choices. This flood of data in healthcare (and many other industries) is what prompted the term "big data."

Data stored in source systems, data marts, enterprise data warehouses or other storage formats is very rarely useful on its own. Just like any raw material, data must be processed in order to become useful. This processing is how data starts to become the information and insight that is needed to understand the operations of a healthcare organization.

The hidden meaning in a request for data

Interestingly, when a quality improvement specialist, executive or clinician asks for data, the request is rarely for just data. Most people making such requests are not actually asking for a dump from the database, unless they need the raw data for their own analysis. Requests for data typically result from a need to better understand a problem, identify quality and performance issues or evaluate patient outcomes and the effects of quality improvement activities.

To make the most of data requests and help clinicians and other leaders in your provider organization effectively devise quality improvements, it's important to frame the request with a few questions.

What can you do with data?

How can we begin to turn untapped data into meaningful insight that enables better administrative and clinical decision making? The first step in understanding an organization and its processes is understanding the data itself to obtain knowledge of the data's context and how it relates to the business. What follows is requisite information that analysts and decision makers alike need to know about data before it is analyzed.

What the data represents

To what processes, workflows, outcomes, patient-generated data or other measures do the data correspond? To provide meaningful insights that add value to decision makers, analytics must use data that accurately reflects the status of patients and/or the performance and quality associated with clinical and business workflows. It's often said in healthcare settings that "you can't manage what you can't measure." You need to ensure that the data being worked with is an accurate reflection of what you are measuring. For example, if data is storing emergency department length of stay, exactly how are the start and end points of emergency visits defined?

How the data is stored

In what kind of system is the data kept, such as an enterprise data warehouse or other format? How is the data physically stored on the database, meaning is the data type stored by integer, character or date? How might that storage format constrict what can be done with the data? All of these questions should be answered. At the database level, the data type that is assigned to a field controls what kind of information can be stored in that fieldsuch as number, words, character strings or a selection of menu choices. This helps to ensure the integrity stored of data so that when the data is read back from the database, the software knows how to interpret it.

The data type

Regardless of how data might be physically stored in a database, it is important to know what value the data represents in real life. This knowledge allows for meaningful analyses of the data. If the type of analysis performed on data is not appropriate for the data type or what the data represents, the results will likely be nonsensical.

Other healthcare analytics news and tips

Interoperability and analytics hot conference topics

Predictive analytics of data in healthcare the goal of one group

How Lean and Six Sigma made the leap to healthcare

Given the type of data and storage, discretion must be applied when deciding the kind of database manipulations and mathematical operations to be performed so the results are worthwhile. For example, just because data is stored as a numeric data type on a database does not mean that all numeric operations can be applied. Ordinal triage acuity data (such as one, two, three, four and five) may look numeric, but it would be an error to average such scores because the values in this case are used only to establish a ranking order.

The business or clinical problem

Critical to any useful analytics is an understanding of what clinical or business problems decision makers need to solve. With the availability of large volumes of data, and relatively inexpensive computing power that can perform deep data analysis, there is a temptation to take the "shotgun approach" and unleash all available tests and analyses on a data set. I don't mean to discourage this; such data explorations can reveal insight, uncover unknown relationships in data and certainly satisfy intellectual curiosity.

The end result of analysis, however, must be information that drives decision making and enables clinicians, administrators and quality improvement stakeholders to take appropriate action to achieve the goals and objectives of the organization. As my favorite high-school math teacher always implored, "Make sure you answer the question."

Healthcare organizations are generating and using unprecedented volumes and varieties of data. Yet despite advances in data collection, management, analysis and insight-generation, basic principles about data analysis still (and will always) apply: know what data you have, know what it means, and know what you can do with it and be sure to answer the original question.

About the author:
Trevor Strome, M.S., PMP, leads the development of informatics and analytics tools that enable evidence-informed decision making by clinicians and healthcare leaders. His experience spans public, private and startup-phase organizations. A popular speaker, author and blogger, Strome is the founder of, and his book, Healthcare Analytics for Quality and Performance Improvement, was recently published by John Wiley & Sons Inc.

Dig Deeper on Clinical data analytics software and systems

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.