Development of the emergency physician job satisfaction measurement instrument.

The objective of this study was to develop a valid and reliable instrument to measure the job satisfaction of physicians practicing emergency medicine. A prospective survey involving four separate stages (an item evaluation and reduction stage, a factor analysis stage, a construct validity stage, and a reliability stage) was distributed in Canada to full-time emergency physicians. Three separate survey instruments were administered (an initial draft instrument with 228 items, a pilot instrument with 142 items, and the final instrument with 79 items). Construct validity of the final instrument was tested by evaluating the correlation between physician scores on the instrument, and scores on two instruments measuring the same construct, and three measuring different but related constructs. A draft instrument with 228 items and six hypothetical domains was tested on 61 physicians. Evaluation for frequency endorsement, redundancy, and homogeneity reduced the item pool to 157. The remaining 157 items were used as a pilot instrument and tested on 223 physicians. Factor analysis eliminated 66 items from the pilot instrument, creating a final instrument with 79 items, 11 factors, and six domains. Cronbach's coefficient alpha for the final instrument domains is 0.81, and all domain-total correlations are greater than 0.4. All correlations between the final instrument and the construct validity instruments were statistically significant (P < .001), but not so high that they appeared to be measuring the same thing. Correlations between instruments measuring the same construct were higher than those measuring related but different constructs. Correlations between the final instrument and the CES-D scale, emotional exhaustion, and depersonalization subscales of the Maslach Burnout Inventory were negative. A test-retest reliability study on 42 physicians showed Pearson's correlation coefficients for individual domains were all greater than 0.7 and greater than 0.8 for the final instrument. This study has produced a valid and reliable instrument for measuring emergency physician job satisfaction, which is both internally consistent and stable.