The Mayo Leadership Impact Index Adapted for Matrix Leadership Structures: Initial Validity Evidence

Abstract

Jamile A Ashmore,1,2 Anthony C Waddimba,2– 4 Megan E Douglas,5 Stacey V Coombes,6 Tait D Shanafelt,7 J Michael DiMaio2,4,8 1Office of Professionalism and Well-Being, Baylor Scott & White-The Heart Hospital, Plano, TX, USA; 2College of Medicine, Texas A&M University, Dallas, TX, USA; 3Division of Surgical Research, Department of Surgery, Baylor University Medical Center, Dallas, TX, USA; 4Research Development & Analytics Core, Baylor Scott and White Research Institute, Dallas, TX, USA; 5Trauma Research Consortium, Baylor Scott and White Research Institute, Dallas, TX, USA; 6OrganizationRx, Los Angeles, CA, USA; 7Department of Medicine, Stanford University School of Medicine, Stanford, CA, USA; 8Division of Cardiothoracic Surgery, Baylor Scott & White-The Heart Hospital, Plano, TX, USACorrespondence: Jamile A Ashmore, Office of Professionalism and Well-being, Baylor Scott & White-The Heart Hospital, 1100 Allied Drive, Plano, TX, 75093, Tel [phone], Email [email protected]: Physician burnout has reached crisis levels. Supportive leadership is one of the strongest drivers of physician well-being, and monitoring supervisor support is key to developing well-being focused leadership skills. Existing measures of leader support were designed within “direct report” supervision structures limiting their applicability to matrixed leadership reporting structures where direct reports are not the predominant norm. Antecedently, no measure of leadership support is validated specifically for implementation in matrixed leadership structures.Objective: Adapt and validate the Mayo Leadership Impact Index (MLII) for settings with matrixed leadership structures.Design: A psychometric validation study utilizing classical test theory and item response theory.Setting: A tripartite hospital system in the southwestern US.Participants: Physician-respondents to a 2023 cross-sectional survey.Main Outcomes and Measures: After pilot testing, the adapted MLII was examined using a unidimensional graded response model and confirmatory factor analyses. Convergent validity was investigated via correlations with professional fulfillment, perceived autonomy support, self-valuation, and peer connectedness/respect. Divergent validity was tested via correlations with burnout.Results: Of the three candidate revisions of the MLII, the 9-item adaptation was selected for its superior validity/reliability indices. Standardized Cronbach’s and Ordinal alpha coefficients were 0.958 and 0.973, respectively. CFA loadings exceeded 0.70 (p < 0.001), and coefficients of variation (R2) exceeded 0.60 for all items. GRM slope parameters indicated “high” to “very high” item discrimination. Items 2, 5, and 8 were the most informative. Positive correlations of the adapted MLII with professional fulfillment, perceived autonomy support, and peer connectedness/respect were observed, supporting convergent validity. Negative correlation with overall burnout supports divergent validity.Conclusions and Relevance: The findings provide evidence of the adapted MLII’s validity, reliability, and appropriateness for implementation within matrixed leadership settings. Prior to this study, no leadership support measure had been validated for use among the growing number of healthcare systems with matrixed leadership reporting structures.Plain Language Summary: Question: What is the validity and reliability of a well-being centered leadership measure adapted for use in healthcare systems with matrixed, multiform reporting structures?Findings: Classical test theory and item response theory analyses of cross-sectional survey data from 158 physician-respondents supported the adapted measure’s construct validity. All reliability coefficients were strong. Leadership ratings positively correlated with professional fulfillment, autonomy support, self-valuation, and peer connectedness/respect, and negatively correlated with burnout.Meaning: Findings support the adapted measure’s validity and reliability. This study is the first to demonstrate a valid empirical measure of well-being centered leadership behaviors in settings with multiform, matrixed leadership structures.Keywords: physician, psychometrics, well-being, burnout, leader support

Keywords

• physician
• psychometrics
• well-being
• burnout
• leader support