Since 1999, MGMA has published benchmarking metrics for “better performing” practices — medical groups selected for having superior productivity, profitability and revenue cycle performance.
Originally the results from these better performing practices were published in a separate report, the Performance and Practices of Successful Medical Groups Report. With the availability of online survey data, the better performer metrics are now included in MGMA DataDive Provider Compensation and MGMA DataDive Cost and Revenue, and the business practices and behaviors these groups employ to achieve success are part of MGMA DataDive Practice Operations.
The better performer criteria were originally developed by a panel of experts and the metrics were held stable for comparison purposes. However, healthcare has changed over the past 19 years, so it is important to confirm the statistical basis for the better performer metrics and to understand how well the metrics are associated with desired financial and productivity outcomes. IBM SPSS statistical software was used to analyze the 2017 Cost and Revenue Survey database to determine how changes in key revenue cycle metrics affect another metric, total bad debt per full-time-equivalent (FTE) physician. The key revenue cycle metrics were:
The analysis used regression, a statistical method that determines the strength of the relationship between a dependent variable (in this case, total bad debt per FTE physician) and a series of other independent variables (the three revenue cycle metrics).
The three scatterplot graphs — representing data from 687 practices of all types and specialties — show a bi-variant view of how each metric is related to the amount of total bad debt per FTE physician. The patterns show a strong relationship between the variables and bad debt. The graphs also display the statistical relationship — the coefficient of determination expressed as “R-squared” or the degree that variation in bad debt is explained by the linear regression line — with a higher value implying the metric on the vertical axis is more strongly related to the amount of bad debt per FTE physician shown on the horizonal axis.
Examining the three graphs, the plot with total A/R per FTE physician (Figure 1) has the greatest association with the amount of total bad debt per FTE physician, with an R-squared of 0.437, essentially saying the 43.7% of variation in total bad debt per FTE physician can be explained by the total A/R per FTE physician with higher amounts of A/R being strongly associated with greater levels of bad debt.
The adjusted collection percentage (Figure 2), defined as the amount of billed charges after contractual adjustments collected from insurers and patients, is almost as predictive with an R-squared of 0.412. Even with some outlying practices, the relationship is obvious: As the adjusted collection percentage decreases, the amount of bad debt increases.
Less predictive but probably just as important as a management tool are total days of gross charges in total A/R (Figure 3), with an R-squared of 0.250. As the scatterplot displays, as the number of total days of gross charges in total A/R increase, bad debt also increases.
While each of these metrics is associated with the amount of bad debt, the three revenue cycle metrics taken together, paired with practice ownership demographics (e.g., independent or part of a health system), creates a model that is extremely predictive of bad debt. Figure 4 displays information from the SPSS multiple regression statistical analysis, showing each variable with the computed statistics of the adjusted R-square, b coefficients and the statistical significance of how a change in each independent variable affects the dependent variable, total bad debt per FTE physician.
The multiple regression shows that the three revenue cycle metrics have a statistically significant relationship to bad debt. With only ownership having a measurable probability that the results could be random, the probability was less than 3%.
The multilinear regression created a mathematical model that predicts the amount of bad debt based on information provided on the four explanatory variables. The adjusted R-square for the model, 0.681, states that 68.1% of the variation in bad debt can be explained by the model.
The b coefficients shown in Figure 4 can be used to predict bad debt, using a simple mathematical formula. Begin with the constant: If the practice is hospital owned, add an additional $3,271.63; add nothing for an independent practice. Multiply the adjusted FFS collection percentage by -2,772.57; multiply the total A/R per FTE physician by 0.15; and multiply days gross FFS charges in A/R by -$268.52 and sum the results. The answer statistically approximates the bad debt of a practice.
While creating a statistical model is interesting, the analysis most importantly confirms that practices that exceed the median for the three key revenue cycle metrics are truly better performers. Knowing that just three revenue cycle metrics in combination can predict almost 70% of the variability in bad debt gives a practice executive better information needed to manage the organization’s business office.
Originally the results from these better performing practices were published in a separate report, the Performance and Practices of Successful Medical Groups Report. With the availability of online survey data, the better performer metrics are now included in MGMA DataDive Provider Compensation and MGMA DataDive Cost and Revenue, and the business practices and behaviors these groups employ to achieve success are part of MGMA DataDive Practice Operations.
The better performer criteria were originally developed by a panel of experts and the metrics were held stable for comparison purposes. However, healthcare has changed over the past 19 years, so it is important to confirm the statistical basis for the better performer metrics and to understand how well the metrics are associated with desired financial and productivity outcomes. IBM SPSS statistical software was used to analyze the 2017 Cost and Revenue Survey database to determine how changes in key revenue cycle metrics affect another metric, total bad debt per full-time-equivalent (FTE) physician. The key revenue cycle metrics were:
- Total accounts receivable (A/R) per FTE physician
- Adjusted fee-for-service (FFS) collection percentage
- Days adjusted FFS charges in A/R
The analysis used regression, a statistical method that determines the strength of the relationship between a dependent variable (in this case, total bad debt per FTE physician) and a series of other independent variables (the three revenue cycle metrics).
The three scatterplot graphs — representing data from 687 practices of all types and specialties — show a bi-variant view of how each metric is related to the amount of total bad debt per FTE physician. The patterns show a strong relationship between the variables and bad debt. The graphs also display the statistical relationship — the coefficient of determination expressed as “R-squared” or the degree that variation in bad debt is explained by the linear regression line — with a higher value implying the metric on the vertical axis is more strongly related to the amount of bad debt per FTE physician shown on the horizonal axis.
Examining the three graphs, the plot with total A/R per FTE physician (Figure 1) has the greatest association with the amount of total bad debt per FTE physician, with an R-squared of 0.437, essentially saying the 43.7% of variation in total bad debt per FTE physician can be explained by the total A/R per FTE physician with higher amounts of A/R being strongly associated with greater levels of bad debt.
The adjusted collection percentage (Figure 2), defined as the amount of billed charges after contractual adjustments collected from insurers and patients, is almost as predictive with an R-squared of 0.412. Even with some outlying practices, the relationship is obvious: As the adjusted collection percentage decreases, the amount of bad debt increases.
Less predictive but probably just as important as a management tool are total days of gross charges in total A/R (Figure 3), with an R-squared of 0.250. As the scatterplot displays, as the number of total days of gross charges in total A/R increase, bad debt also increases.
While each of these metrics is associated with the amount of bad debt, the three revenue cycle metrics taken together, paired with practice ownership demographics (e.g., independent or part of a health system), creates a model that is extremely predictive of bad debt. Figure 4 displays information from the SPSS multiple regression statistical analysis, showing each variable with the computed statistics of the adjusted R-square, b coefficients and the statistical significance of how a change in each independent variable affects the dependent variable, total bad debt per FTE physician.
The multiple regression shows that the three revenue cycle metrics have a statistically significant relationship to bad debt. With only ownership having a measurable probability that the results could be random, the probability was less than 3%.
The multilinear regression created a mathematical model that predicts the amount of bad debt based on information provided on the four explanatory variables. The adjusted R-square for the model, 0.681, states that 68.1% of the variation in bad debt can be explained by the model.
The b coefficients shown in Figure 4 can be used to predict bad debt, using a simple mathematical formula. Begin with the constant: If the practice is hospital owned, add an additional $3,271.63; add nothing for an independent practice. Multiply the adjusted FFS collection percentage by -2,772.57; multiply the total A/R per FTE physician by 0.15; and multiply days gross FFS charges in A/R by -$268.52 and sum the results. The answer statistically approximates the bad debt of a practice.
While creating a statistical model is interesting, the analysis most importantly confirms that practices that exceed the median for the three key revenue cycle metrics are truly better performers. Knowing that just three revenue cycle metrics in combination can predict almost 70% of the variability in bad debt gives a practice executive better information needed to manage the organization’s business office.
Dig deeper
MGMA DataDive is the premier data benchmarking tool in healthcare. Access the industry’s largest benchmarking datasets in topics such as compensation, operations and cost and revenue. For more information on how MGMA DataDive can help your organization, contact sales@mgma.com or call 877.275.6462, ext. 1801.Learn more
MGMA Consulting provides practices with solutions to their operational issues, including physician compensation models and examining revenue cycle to improve financial performance. Learn more: Call 877.275.6462, ext. 1877, or email consulting@mgma.com.
Explore Related Content
More Insight Articles
