Posted by: Jenny Laurello
BI, Data analytics, Quality
Guest post by: Trevor Strome MSc, PMP, Informatics Lead, Winnipeg Regional Health Authority; Assistant Professor, Department of Emergency Medicine, Faculty of Medicine, University of Manitoba
Analytics are not a magic bullet
A sales representative from a major Business Intelligence (BI) software vendor recently tried to convince me that with their product, we would “just have to make the information available to everyone via dashboards and reports, and change will happen by itself.” I am a firm believer in promoting transparency in health care by ensuring people have access to whatever data they need to make decisions. I definitely know from experience, however, that analytics alone can’t (and won’t ever) cause meaningful change in health care.
Too much data to take meaningful action
Health care analytics tools (into which I group business intelligence applications, dashboards, simulation, predictive algorithms and intelligent reporting) now provide amazing insight into the operations of a health care organization (HCO). This is especially true now that electronic medical records (EMRs) are becoming more commonplace and are providing a wealth of data for analysis.
With literally hundreds of data elements being generated on some modern EMR systems, it is important to differentiate which of this data is important to analyze and report on for purposes of improving health care, and which should be set aside until needed at a different time. It is likely that only some of the data available is relevant to the current quality and improvement performance goals of a HCO, and even less are directly actionable.
Yet, with so much data becoming available, the temptation is to create numerous indicators and to build a myriad of dashboards to display them all. In a recent post on my blog, I note that the proliferation of dashboards (and other information tools) risks are causing an increase in information overload, which can be counterproductive to the goals of improving the health care system.
Choose metrics that are relevant and actionable
To prevent information overload (especially in health care situations), all information available via dashboards, reports or other analytics tools must be both relevant and actionable. In other words, the information must be useful for understanding the most pressing quality and performance issues facing a HCO, and should also identify what needs to be done to mitigate those quality and process issues.
The question is, though, how do health care executives, unit managers, QI professionals and analytics developers know what information, of all that is available, is important and necessary for making the right decisions?
When choosing indicators, one factor is alignment with organizational strategy. Such alignment is important so that the goals and objectives of a HCO are communicated (and being adhered to) throughout the organization. Most decision-making that impacts process changes and quality, however, occurs at a more “tactical” level.
When working on or close to the front-lines, integration with a methodology for quality and process improvement is essential to ensure indicators and other information is relevant and actionable. In a previous SearchHealthIT article, I stated that, “insight is best gained when analytic and BI tools (such as dashboards, scorecards other information applications) are applied in a strategic combination with quality improvement initiatives via approaches such as Lean and Six Sigma.”
Focusing on the critical few
Lean and Six Sigma are two of the more common quality improvement approaches in health care, and are often used in concert with each other. Lean is a systematic approach to improving quality and value to the patient by eliminating the waste that is present in almost every health care process. Six Sigma goes a step further, and is a rigorous data-driven approach to reduce process variation and eliminate defects in the delivery of health care.
A quality improvement methodology or framework helps provide the necessary context for analytics to help drive appropriate decision making. Practitioners of Lean and Six Sigma strive to identify the most pressing quality issues based on inputs such as patient safety and quality needs (which are the most important factors), the overall goals of the HCO, national standards and legislative requirements. The methodologies then follow structured approaches for addressing and solving the identified quality issues The results of this process naturally helps to filter all the possible metrics and indicators (based on hundreds of available data points) down to the critical few indicators required for true quality and performance improvement success.
An example of what works, and what doesn’t
As an example, a HCO may set a goal of an average Emergency Department (ED) Length of Stay (LOS) of four hours. This metric, while certainly relevant to the ED and HCO as a whole, fails the actionable test. If the ED is not meeting the target (with an average LOS of six hours, for example), LOS as a metric provides little information about how to address the problem, as a plethora of factors actually contribute to LOS.
Process improvement projects using Lean and Six Sigma will break down ED visits into their component parts to identify the root causes of long LOS. Depending on what the root causes are, metrics of interest to the quality teams might include average time to bed availability, consultant wait times and other measures. Indicators identified by process and quality improvement teams — following rigorous methodologies — will help guide process improvement activities and, therefore, should be the focus of dashboards and other analytic tools. Close monitoring of these metrics will provide much more clarity when determining if implemented changes are effective, or if further changes are necessary. Once a change has been implemented and successfully sustained, teams can then move onto other quality issues (likely requiring their own metrics for analysis).
More isn’t always better
In times when health care is feeling the pressure, it is important to provide management and quality improvement teams with the key pieces of information they need to focus on the most important problems, and to make appropriate, timely decisions. In the same manner that a pilot will focus on about six key instruments throughout most of a flight (with supplemental information being provided by other instruments), the critical indicators derived from approaches such as Lean and Six Sigma can guide decision making on the part of the HCO and result in real health care improvement.
About the author:
Trevor Strome MSc, PMP, is the Informatics Lead for the Winnipeg Regional Health Authority, and is Assistant Professor at the Department of Emergency Medicine, Faculty of Medicine, University of Manitoba. You can visit Trevor’s blog at http://healthcareanalytics.info or contact him via email.