Results of Multivariate Linear Regression Models of Burnout and Engagement as Predictors of Patient Experience Scores, Adjusted for Covariates and Clustering by Clinic, (n = 151)*
Variable | Adjusted β | Robust Standard Error | 95% CI | P value |
---|---|---|---|---|
Dependent variable: Proportion of patient ratings in top box for primary care provider communication | ||||
Model 1: High burnout† | 0.006 | 0.01 | −0.02–0.03 | .55 |
Model 2: High engagement‡ | 0.008 | 0.01 | −0.02–0.04 | .54 |
Model 3: Burnout-engagement phenotypes | ||||
High burnout-low engagement | Reference | |||
High burnout-high engagement | 0.04 | 0.02 | 0–0.08 | .05 |
Low burnout-low engagement | 0.001 | 0.02 | −0.04–0.04 | .95 |
Low burnout-high engagement | −0.002 | 0.01 | −0.03–0.03 | .89 |
Dependent variable: Whether in top quartile of patient overall rating of primary care provider | ||||
Model 1: High burnout | 0.0005 | 0.01 | −0.03–0.03 | .97 |
Model 2: High engagement | 0.007 | 0.03 | −0.05–0.07 | .81 |
Model 3: Burnout-engagement phenotypes | ||||
High burnout-low engagement | Reference | |||
High burnout-high engagement | 0.04 | 0.02 | 0.009–0.08 | .02 |
Low burnout-low engagement | 0.01 | 0.03 | −0.04–0.07 | .62 |
Low burnout-high engagement | −0.0004 | 0.03 | −0.06–0.06 | .99 |
Dependent variable: Proportion of patient ratings in top box for recommendation of clinic | ||||
Model 1: High burnout | −0.003 | 0.02 | −0.04–0.03 | .87 |
Model 2: High engagement | 0.008 | 0.02 | −0.03–0.04 | .64 |
Model 3: Burnout-engagement phenotypes | ||||
High burnout-low engagement | Reference | |||
High burnout-high engagement | 0.05 | 0.03 | 0.0002–0.10 | .049 |
Low burnout-low engagement | 0.02 | 0.03 | −0.05–0.09 | .55 |
Low burnout-high engagement | 0.002 | 0.02 | −0.04–0.04 | .90 |
↵* This table illustrates the results of multivariate logistic regression analysis. Three multivariate regression models are represented for each outcome in this table, with results adjusted for covariates and clustering by clinic. Burnout, engagement (likelihood to recommend the clinic), and the phenotypes were tested in separate regression models that included only 1 of these predictors at a time, along with the covariates. Covariates include complexity adjusted panel per clinical full-time equivalent, position (physician vs nurse practitioner), years in the health system, type of health system (county-based vs university-based), and whether working at a residency-teaching practice.
↵† The term “high burnout” is defined as reporting either high exhaustion or high cynicism.
↵‡ The term “high engagement” is defined as a score of 9–10 on the “the likelihood to recommend clinic as place to work.” This item has a possible range of 0–10, which 10 being the most positive rating. This definition is consistent with scoring for the “Net Promoter Score,” which defines a score of 9–10 as being a promoter.
CI, confidence interval.