Significance of on-call errors by nuclear medicine residents: Results of a 3-year analysis

Objectives (1) To determine the frequency of discrepancies between preliminary results given by nuclear medicine residents on-call for an academic 700-bed university hospital, and the official scan interpretation by the attending physician for emergency nuclear medicine procedures. (2) To determine how often a discrepancy resulted in a change in patient management.

Methods Over a 3-year interval, we prospectively recorded the total number of emergency nuclear medicine examinations during non-working hours (night, weekends) performed whenever a nuclear medicine resident was on call.

Results A total of 254 emergency nuclear medicine scans were interpreted by on-call nuclear medicine residents on both inpatients and emergency room patients. Of these, there were 31 scans (12.2%) whose preliminary reading by the resident differed from the official final reading. Most of the disparate interpretations involved V/Q scans in which the discordance ranged from the "normal" to "very low probability" to low probability" scans. However, the changes in interpretation never resulted in a significant impact on patient management when reported to the referring physician.

Conclusions (1) At our institution, there was concurrent agreement in roughly 88% (223/254) of on-call scan interpretations between the resdient on-call and the attending physician. (2) Of the interpretation discrepancies, none resulted in a change in patient management.