Benchmarking

Create your local goal, by using published benchmarks to estimate percentiles of performance

Use published benchmarks to calculate estimates of mean and standard deviation which leads to estimates of the 75th and 80th percentiles.

Calculator
	Lower range (optional)	Median (national or regional rate)	Upper range (best reported)
1. Example using rates of blood pressure control
Example: BP < 140/90	Not needed	65 (from HEDIS)	90 (from Kaiser)
2. For your topic, find: Median/mean values for your clinic measure or topic using HEDIS and other sites HEDIS (NCQA) MIPS (CMS) Minnesota Community Measurement Best reported rates from PubMed. Show details of PubMed search filters Best reported rates from PubMed. Hide details of PubMed search filters Suggested filter based on (Wilczynski 2010; Hempel, 2011). This example searches for benchmarks for hypertension from Kaiser Permanente. (Health Services Administration[major] OR quality) AND (improv* OR enhance* OR increase) AND (intervention OR initiative* OR strategy* OR program* OR Process* OR strateg* OR implement*) AND (hypertension[title] OR hypertension[major]) AND Kaiser[title/abstract] Search this example at PubMed If insufficient results, consider Removing the title/abstract constraint for the institution of interest, or Removing the institutional term completely
3. Enter your values (replace example values for blood pressure with yours)
Range Confidence interval Standard deviation Interquartile range
Lowest rate reported		Median or mean	Highest reported rate

4. Results: 57 65 73 81

Longer list of estimates:

Mean: SD:
Suggested benchmarks to aim for:
25th percentile (z = -0.67):
50th percentile (from the median value you entered [z = 0]):
75th percentile (z = 0.67):
80th percentile (z = 0.84):
90th percentile (z = 1.30):

z scores (the number of standard deviations an observation lies above or below the mean) obtained from https://measuringu.com/zcalcp/ and https://www.mymathtables.com/statistic/z-score-percentile-normal-distribution.html (table of values).

Details from: mean to median, confidence interval, interquartile range, and range. See downloadable spreadsheet for calculations by Wan, 2014 PMID 25524443

Determine the estimated percentile of your local performance

Determine your percentile by comparing your rate to the published published median benchmark value and the estimated standard deviation of published rates (from above)

Alternatively, get percentile from measuringu.com

Enter Z score at here (use One-sided option)

Your observation	Population 50th percentile	Population standard deviation (from above)

Z: Percentile: Details from: stackoverflow.