Ask the expert: How should our medical staff use rate indicator targets when conducting physician peer review?

Medical staffs can express rate data in a number of ways:

• As an absolute value, such as 23% or 5.3 days
• As a relative value, such as the 50th percentile
• As a statistically significant outlier
   

Of course, the target needs to be expressed in the same measurement format as the indicator. It is not uncommon for the physician profiles to accidentally show the indicator data expressed as percentages but the targets as percentiles.

Although absolute values are easier to use, relative values work better to drive performance improvement because they allow performance expectations to change as the group norm changes without having to change the target every year. Some organizations express the target as an absolute value for ease of understanding, but they select that value using a statistically determined formula. For example, the excellent target for surgical complications could be an absolute, such as 1.3%, but that number would be selected each year by a rule that determines the value for physicians at the 75th percentile of a national database.

When external data aren’t available for rate measures and targets need to be selected based on internal data, using historical data and setting targets based on the mean and the distribution of that data is the best approach. Using percentiles is less useful because it guarantees that in a department of 10 physicians, two will be below the 25th percentile, even if their performance is very close to that of others on the distribution.

This week’s question is from Measuring Physician Competency: How to Collect, Assess, and Provide Performance Data by Robert Marder, MD, CMSL; Mark Smith, MD, MBA, CMSL; Marla Smith, MHSA; and Vicki Searcy, CPMSM.