Cummings and colleagues compiled data on nurses, patients, and institutions from the Alberta Nurse Survey (a Canadian arm of the larger International Hospital Outcomes Study), from discharge abstracts in the Hospital Inpatient Database, and from regional health authorities to determine 30-day mortality in patients admitted for acute myocardial infarction, congestive heart failure, chronic obstructive pulmonary disease, pneumonia, or stroke. Data were collected for 21,570 patients and 5,228 nurses at 90 hospitals. The nursing leadership style at each hospital was determined based on nurses' responses to 13 questions on the Revised Nursing Work Index. The 13 questions assessed emotional intelligence leadership competencies, and answer choices were based on a 4-point Likert scale (from strongly disagree to strongly agree). Five nursing leadership styles were identified: high resonant, moderately resonant, mixed, moderately dissonant, and high dissonant. High resonant was defined as nurses agreeing or strongly agreeing that all 13 competencies were present at their hospital.

The overall 30-day mortality rate for the 21,570 patients was 7.8%. Rates differed across leadership style, with the lowest mortality rates associated with high dissonant (4.3%) and high resonant (5.2%) leadership styles. The highest mortality rates were associated with mixed (8.1%) and moderately dissonant (8.8%) leadership styles. After controlling for patient demographics, comorbidities, and hospital and nursing factors, leadership style was found to contribute 5% of 72% of the total variance in mortality rates across hospitals, and the lowest 30-day mortality was seen in hospitals with high-resonant leadership (26% lower odds), followed by hospitals with high-dissonant leadership (14% lower odds).