Accuracy Matters When Compensation for Hospitalists Is at Stake

Not long ago, I received an email from a hospitalist group leader who was working with her CMO on a new compensation plan. The CMO, wanting to ensure that the proposed compensation per unit of work was appropriate, had taken the MGMA national median annual compensation for internal-medicine hospitalists ($234,437) and divided it by the national median annual work RVUs (4,185) to arrive at a targeted compensation per wRVU of $56.01.

The hospitalist leader, however, had the benefit of referring to her 2012 State of Hospital Medicine report, in which Table 6.30 reported an MGMA median compensation per wRVU for internal-medicine hospitalists of $58.28. That variance of more than two dollars per wRVU could mean an additional $8,000 or so in annual compensation to her and her colleagues, so she was seeking to understand why the report has a different number than the one calculated by her CMO.

The answer is that the CMO got caught in a common error of logic: The CMO assumed that the compensation median and the wRVU median were derived from exactly the same population, failing to consider that the underlying data sets might be different. Here’s what happened: Compensation data were reported for 3,192 internal-medicine hospitalists, but wRVUs were reported for only 2,389 of those hospitalists. So the analysis of compensation per wRVU can be accurately calculated only for those 2,389 hospitalists for whom both compensation and wRVUs were reported. The other 803 hospitalists for whom no wRVUs were reported had to be excluded from the ratio calculation. The CMO’s error was to calculate a ratio of two medians based on different data sets, rather than calculating the individual comp-to-wRVU ratios, then determining the median for that smaller data set.

A similar thing has happened over the years with nocturnist data. In SHM’s 2007-2008 compensation and productivity survey, and again in the 2011 SHM/MGMA State of Hospital Medicine report, the median compensation reported for nocturnists actually was lower than that reported for all adult hospitalists. In my work with hospitalist practices across the country, I’ve only run into one or two where the nocturnists earned less than the daytime doctors, so I was flummoxed by this finding. Turns out, I was making the same mistake of assuming I was looking at “nocturnist” and “all adult hospitalist” compensation for the same hospitalist groups. But the adult medicine groups using nocturnists are actually a small subset of all adult medicine groups, and the nocturnist data likely included at least a few pediatric hospitalist nocturnists. Because the underlying data sets are different, the two medians aren’t directly comparable.

When all is said and done, we don’t really care whether the average nocturnist earns more or less than the average non-nocturnist hospitalist. What we really want to know is, Do the nocturnists in a given group earn more than the non-nocturnists in the same group? That’s why this year SHM asked groups to report the average percent compensation differential between nocturnists and non-nocturnists in their groups. It turns out that groups serving adults only reported a median of 15% higher compensation for nocturnists, a far different result than users of previous surveys inferred.

The bottom line: Make sure you understand how the State of Hospital Medicine survey results are calculated. Many of the formulas used are described in Appendix B of the report, and if you have questions about others, feel free to contact SHM and ask.

Leslie Flores is a partner in Nelson Flores Hospital Medicine Consultants.

Not long ago, I received an email from a hospitalist group leader who was working with her CMO on a new compensation plan. The CMO, wanting to ensure that the proposed compensation per unit of work was appropriate, had taken the MGMA national median annual compensation for internal-medicine hospitalists ($234,437) and divided it by the national median annual work RVUs (4,185) to arrive at a targeted compensation per wRVU of $56.01.

The hospitalist leader, however, had the benefit of referring to her 2012 State of Hospital Medicine report, in which Table 6.30 reported an MGMA median compensation per wRVU for internal-medicine hospitalists of $58.28. That variance of more than two dollars per wRVU could mean an additional $8,000 or so in annual compensation to her and her colleagues, so she was seeking to understand why the report has a different number than the one calculated by her CMO.

The answer is that the CMO got caught in a common error of logic: The CMO assumed that the compensation median and the wRVU median were derived from exactly the same population, failing to consider that the underlying data sets might be different. Here’s what happened: Compensation data were reported for 3,192 internal-medicine hospitalists, but wRVUs were reported for only 2,389 of those hospitalists. So the analysis of compensation per wRVU can be accurately calculated only for those 2,389 hospitalists for whom both compensation and wRVUs were reported. The other 803 hospitalists for whom no wRVUs were reported had to be excluded from the ratio calculation. The CMO’s error was to calculate a ratio of two medians based on different data sets, rather than calculating the individual comp-to-wRVU ratios, then determining the median for that smaller data set.

A similar thing has happened over the years with nocturnist data. In SHM’s 2007-2008 compensation and productivity survey, and again in the 2011 SHM/MGMA State of Hospital Medicine report, the median compensation reported for nocturnists actually was lower than that reported for all adult hospitalists. In my work with hospitalist practices across the country, I’ve only run into one or two where the nocturnists earned less than the daytime doctors, so I was flummoxed by this finding. Turns out, I was making the same mistake of assuming I was looking at “nocturnist” and “all adult hospitalist” compensation for the same hospitalist groups. But the adult medicine groups using nocturnists are actually a small subset of all adult medicine groups, and the nocturnist data likely included at least a few pediatric hospitalist nocturnists. Because the underlying data sets are different, the two medians aren’t directly comparable.

When all is said and done, we don’t really care whether the average nocturnist earns more or less than the average non-nocturnist hospitalist. What we really want to know is, Do the nocturnists in a given group earn more than the non-nocturnists in the same group? That’s why this year SHM asked groups to report the average percent compensation differential between nocturnists and non-nocturnists in their groups. It turns out that groups serving adults only reported a median of 15% higher compensation for nocturnists, a far different result than users of previous surveys inferred.

The bottom line: Make sure you understand how the State of Hospital Medicine survey results are calculated. Many of the formulas used are described in Appendix B of the report, and if you have questions about others, feel free to contact SHM and ask.

Leslie Flores is a partner in Nelson Flores Hospital Medicine Consultants.

Not long ago, I received an email from a hospitalist group leader who was working with her CMO on a new compensation plan. The CMO, wanting to ensure that the proposed compensation per unit of work was appropriate, had taken the MGMA national median annual compensation for internal-medicine hospitalists ($234,437) and divided it by the national median annual work RVUs (4,185) to arrive at a targeted compensation per wRVU of $56.01.

The hospitalist leader, however, had the benefit of referring to her 2012 State of Hospital Medicine report, in which Table 6.30 reported an MGMA median compensation per wRVU for internal-medicine hospitalists of $58.28. That variance of more than two dollars per wRVU could mean an additional $8,000 or so in annual compensation to her and her colleagues, so she was seeking to understand why the report has a different number than the one calculated by her CMO.

The answer is that the CMO got caught in a common error of logic: The CMO assumed that the compensation median and the wRVU median were derived from exactly the same population, failing to consider that the underlying data sets might be different. Here’s what happened: Compensation data were reported for 3,192 internal-medicine hospitalists, but wRVUs were reported for only 2,389 of those hospitalists. So the analysis of compensation per wRVU can be accurately calculated only for those 2,389 hospitalists for whom both compensation and wRVUs were reported. The other 803 hospitalists for whom no wRVUs were reported had to be excluded from the ratio calculation. The CMO’s error was to calculate a ratio of two medians based on different data sets, rather than calculating the individual comp-to-wRVU ratios, then determining the median for that smaller data set.

A similar thing has happened over the years with nocturnist data. In SHM’s 2007-2008 compensation and productivity survey, and again in the 2011 SHM/MGMA State of Hospital Medicine report, the median compensation reported for nocturnists actually was lower than that reported for all adult hospitalists. In my work with hospitalist practices across the country, I’ve only run into one or two where the nocturnists earned less than the daytime doctors, so I was flummoxed by this finding. Turns out, I was making the same mistake of assuming I was looking at “nocturnist” and “all adult hospitalist” compensation for the same hospitalist groups. But the adult medicine groups using nocturnists are actually a small subset of all adult medicine groups, and the nocturnist data likely included at least a few pediatric hospitalist nocturnists. Because the underlying data sets are different, the two medians aren’t directly comparable.

When all is said and done, we don’t really care whether the average nocturnist earns more or less than the average non-nocturnist hospitalist. What we really want to know is, Do the nocturnists in a given group earn more than the non-nocturnists in the same group? That’s why this year SHM asked groups to report the average percent compensation differential between nocturnists and non-nocturnists in their groups. It turns out that groups serving adults only reported a median of 15% higher compensation for nocturnists, a far different result than users of previous surveys inferred.

The bottom line: Make sure you understand how the State of Hospital Medicine survey results are calculated. Many of the formulas used are described in Appendix B of the report, and if you have questions about others, feel free to contact SHM and ask.

Leslie Flores is a partner in Nelson Flores Hospital Medicine Consultants.