In some of our recent posts, we’ve been taking a look at how the ACG System’s suite of tools can be used to understand emergency department (ED) visits, which helps users optimize health care utilization and reduce potential costs.
If you read our earlier posts, you know that the ACG System can reveal specific trends in ED visits for a certain population, specifically, patients who visited for non-emergent care or primary care (PCP) treatable conditions. By drilling down into this data, ACG System users can understand root causes of ED use, segment patients into actionable groups and develop an effective strategy to reduce potentially-avoidable visits.
This week, we’ll discuss a method for tracking those ED reduction strategies using control charts.
Control charts are a user-friendly tool that differentiate true change in a metric from random variation that occurs naturally. Control charts help identify meaningful change early and are an engaging visualization for different types of stakeholders. Supplementing the control chart with markers of key intervention dates can help leaders understand the relationship of intervention timing to outcomes. These advantages make control charts the ideal tool to monitor changes in ED utilization. In fact, the ACG System’s granular ED visit export file can help develop effective control charts for internal monitoring purposes.
All this makes sense in theory, but let’s look at a real-world example of how control charts can be used to monitor ED usage in a population. In the below example, the user became concerned about an increasing trend in avoidable ED visits starting in Month 16. The trend was identified via overall increase in ED visits/1000, and once the analytic team drilled down into the trend, it revealed growth in avoidable ED visits as an impactable cost-driver. A suite of interventions to reduce avoidable ED visits was implemented in Month 19.
The analytic team used historic ED visit data, organized by the ACG System’s ED Classification algorithm’s category and month, to generate the control chart with historic mean and measures of variation. The horizontal blue line represents historic average monthly rate of avoidable ED visits. The two red lines represent upper and lower control limits.
Interpreting this graph, the peaks in Month 17 and 19 represent significant variation above historic means, supporting the organization’s interpretation that avoidable ED visits were increasing. Once the intervention was implemented in Month 19, utilization returned to post-intervention means in months 22 and 23. However, had the data points continued near the upper control limit, the organization would have an early-stage indicator that the intervention was not achieving the desired outcome.
Looking out to month 25, utilization of avoidable ED visits crosses the lower control limit, indicating significant variation from the historic mean – in this case, for the better.
The above example demonstrates how the ACG System’s unique tools and granular visit-level data can help an organization use control charts to monitor ED visits in near-time, creating a strong business-level understanding of intervention impact. Ultimately, these control chart tools give users clear, specific data to indicate whether or not a specific intervention is achieving the desired goal. The result? Users have the information and tools they need to make changes that optimize health care resources and reduce costs.
*The ACG team would like to thank Shannon Murphy, MA for concept and development of this ED monitoring application. More details on using Statistical Process Control and Control Charts to monitor health interventions can be found here.