The psychometric assessment of a new instrument for measurement of the construct of nurse-physician collaboration in making specific patient care decisions. Collaboration and Satisfaction About Care Decisions (CSACD), is reported. Content validity for the tool was supported by literature review, nurse and physician experts, and potential subjects. In a pilot study 58 neonatal intensive care nurses' and resident physicians' responses showed variance; alpha reliability of the collaboration questions was 0.95. Construct validity of the collaboration questions was supported by finding expected correlational patterns and by factor analysis revealing a single factor that explained 75% of the variance in collaboration. Further psychometric testing of the instrument is proposed, particularly in the context of substantive studies that would allow investigation of its association with patient outcomes.