Return to 2001 Abstacts List
*203. Measuring a Safety Culture: Preliminary Results
P Ebright, Indiana University School of Nursing; ML Render, VA Midwest Patient Safety Center of Inquiry; U Meyer, University of Cincinnati College of Nursing; ES Patterson, Ohio State University Institute of Ergonomics; A Eisenlohr, VA Midwest Patient Safety Center of Inquiry; C Deets, University of Cincinnati College of Nursing
Objectives: There is general agreement that creating a culture of safety within our healthcare institutions is critical, but few if any validated tools measure important elements of a culture of safety. Investment of resources in this important facet of safety will be limited as long as the effect of strategies cannot be measured. Characteristics of a safety culture include a just system, ability to learn from error, a non-punitive and safe reporting environment, and organizational commitment. (Kizer,1998; Reason 1990). We developed a "Safety Climate Questionnaire " to measure changes in safety culture following planned interventions.
Methods: Over 3 iterations, a 37-item Safety Climate Questionnaire to measure the specific aspect of "safety climate" in health care organizations that guides the thinking and behaviors of health care providers was developed and tested. Items represented subscales derived from safety culture theory. Respondents indicated relative agreement with each item on a 5-point Likert- type response scale. The instrument format, item structure, content validity and reliability (internal consistency) were evaluated using data from a total of 150 respondents (about 50/iteration) drawn from physicians, nurses, and pharmacists. Factor analysis was performed using the conservative alpha method for subscale determination.
Results: Analysis of the first administration resulted in deletion of items with unnecessary duplication, rewording of items to improve singular focus, and creation of new items. The reformulated tool contains 37 items. In the second administration internal consistency for the 37 items was .88 (Cronbach alpha). Preliminary factor analysis identified four separate characteristics of a safety culture: organizational commitment, openness about mistakes, communication about errors, and mechanisms for learning from error.
Conclusions: The Safety Climate Questionnaire appears unidimensional with a good internal consistency reliability. Further study is needed to determine construct validity, test-retest reliability, and the validity of subscales.
Impact: Utilization of a tool to assess dimensions of a safety culture will allow empiric evaluation of organizational interventions and support investment.