Skip to main content
. Author manuscript; available in PMC: 2020 Dec 11.
Published in final edited form as: Commun Stat Simul Comput. 2019 Mar 28;2019:10.1080/03610918.2019.1595649. doi: 10.1080/03610918.2019.1595649

Figure 5:

Figure 5:

Comparison of overall flagging performance based on (gray) empirical null distribution/adjustment vs. (white) nominal p-values (unadjusted) for FE models. Given are sensitivity and specificity by provider volume (small, medium, and large) and for varying baseline readmission rates (BRR, low, medium and high).