You have ${pages_left} free articles remaining.
Subscribe now to continue getting the industry’s most current and thoughtful review of hospital performance.
July 2019
Spotlight: Measuring Variability in Hospital Operations and Key Performance Indicators
©2019 Kaufman, Hall & Associates, LLC
Monthly fluctuations and variability of operational and financial key performance indicators (KPIs) may be driven by uncertain or seasonal volume swings. Hospitals experiencing more extreme variation in monthly volumes may be more susceptible to correlated swings in volume-normalized revenue and expense metrics that ultimately impact the bottom line.
In last month’s Spotlight, we examined the impact of more rigid cost structures on the nation’s largest hospitals and health systems, as compared with the national hospital cohort. In the January National Hospital Flash Report, we introduced a more appropriate hospital budgeting methodology that focuses first on attaining accuracy in budgeting by building on actual performance in the prior year. This month, we blend key points of these articles to identify and assess the significance of variability in hospital performance.
Overview
Understanding Variation in Data
Conclusion
Overview
Conclusion
Analytical sophistication and process control, beginning with understanding variation, has long been a workhorse of process-dominated industries. Current analytical techniques employed by financial professionals in hospitals can be bolstered and enriched by gradually increasing analytical sophistication. Our example of analyzing standard deviation of Average Length of Stay may be applied to any KPI to enrich the understanding of operational anomalies that arise in a hospital.
Understanding Variation in Data
We collected a sample of several hundred hospitals, stratified across geographic regions and bed sizes, from proprietary Kaufman Hall data. For each of these hospitals, we computed the KPIs from the Volume and Expense sections of the National Hospital Flash Report. The KPIs were computed on a monthly basis and compared to rolling 12-month aggregate figures. For example, for each of the previous 12 months, we computed the absolute and percent differences in each hospital’s monthly Average Length of Stay and the rolling 12-month aggregation. Figure 1 shows the distribution of the percent difference between monthly and 12-month aggregate Average Length of Stay for the sample data.
Figure 1: Percent Difference in Average Length of Stay
Source: Kaufman, Hall & Associates, LLC
As shown by the distinctive bell-curve shape, the KPI data in Figure 1 appear normally distributed. The data also show that the highest concentration of variation occurs between positive and negative 5 percent, and the frequency continues to dwindle as we move toward the tails. In our sample, tapered tails are truncated to exclude extreme data points from our analysis. Further examination of these extreme data points shows that they are less representative of fluctuations in each hospital’s monthly volumes, and more reflective of anomalies related to external factors that do not materially affect the distribution.
From these data, we can compute the variability, or spread, of the KPIs over the past 12 months. Continuing with the Average Length of Stay metric, we can further explore how the metric varies from month to month. By using the standard deviation of the distribution shown in Figure 1, we can identify the number of days the monthly Average Length of Stay varies in the sample. Figure 2 shows the distribution of each hospital’s Average Length of Stay standard deviation, measured in same days.
Figure 2: Average Length of Stay Standard Deviation Over the Past 12 Months
Source: Kaufman, Hall & Associates, LLC
The data in Figure 2 are distinctly dissimilar. While the data in Figure 1 provide a sense of how often percent variation is observed on a monthly basis, the data in Figure 2 show how many hospitals maintain small gaps in Average Length of Stay compared to the number of hospitals that experience large swings in monthly Average Length of Stay, and hence volumes.
The standard distribution of data is always nonnegative. This means that while the standard deviation of a KPI does not give us any information on the direction of monthly movements (e.g., favorable or unfavorable), it does suggest how much the KPI changes absolutely over, in this case, a given time period. When data are normally distributed, as in Figure 1, knowing the mean and standard deviations is the basis for the statistical rule of thumb stating that 68, 95, and 99 percent of all observations will fall in the range of the average plus/minus one, two, or three standard deviations, respectively.
Department managers and financial leaders who know this simple rule can quickly discern activity or KPIs that seem out of the ordinary based on each hospital’s unique, internal operations. In addition to monthly percent changes, using standard deviation to characterize fluctuations in KPIs is key to understanding if your organization has sufficient control to maintain volume and expense ratio metrics when volumes fluctuate. Quantifying and monitoring the variability in monthly metrics dependent on volume becomes more relevant to healthcare providers as volumes decline and operating costs rise. This approach assists in data-driven decisions frequently associated with Average Length of Stay, such as organizational throughput, choice of care management models, cohorting of patients, and rationalizing care in multi-facility systems to establish Centers of Excellence.
scroll_up.svg
ic_arrow_up.svg
Hide Page Sections
ic_arrow_drop_down_black_24px.svg
Show Page Sections
kha_logo.svg
National Hospital Flash Report
menu_icon.svg
mail.svg
Share article
Understanding Variation in Data
Conclusion
Overview
Sign up now for access to the latest news and reports
SUBMIT
Become a subscriber and receive our monthly reports. Choose the option that best fits your needs.
SUBMIT
Monthly subscription: $100/month
Annual subscription: $600/year
Quarterly subscription: $200/quarter