Are automated anesthesia records better?

STUDY
OBJECTIVE: To determine whether data recorded by an information management system is significantly different from that recorded manually.
DESIGN: A comparison was made between 13 handwritten and 13 computer-generated anesthesia records by calculating the frequency with which recorded variables were outside predetermined acceptable ranges. Five physiologic variables [systolic blood pressure (SBP), diastolic blood pressure (DBP), heart rate (HR), end-tidal partial pressure of carbon dioxide (PETCO2), and oxygen saturation by pulse oximeter (SpO2)] were compared during the initial 1 1/2 hours of operation.
SETTING: Surgical suite at a university-affiliated hospital. PATIENTS: Thirteen adult patients scheduled for operations that required general anesthesia for longer than 1 1/2 hours. INTERVENTION: In addition to the traditional handwritten anesthesia records, an information management system (ARKIVE Patient Management System, DIATEK, San Diego, CA) was used to collect data from each case.
MEASUREMENTS AND MAIN RESULTS: No significant differences were found between the methods in the frequency of elevated SBP, elevated DBP, and tachycardia. However, the manual records showed low SBP, DBP, and HR with a significantly lower frequency (2%, 11%, 1%, respectively) than the automated records (6%, 26%, 5%, respectively; p < 0.01). The automated PETCO2 readings were higher than the upper limit (40 mmHg) with a higher frequency (18%) than the manual records (3%; p < 0.01). On the automated records, SpO2 was noted to be 90% or less on two occasions, but significant desaturation was noted only once on the manual charts.
CONCLUSIONS: Observer bias, missed readings, and errors of memory, which affect manual anesthetic records, may cause significant inaccuracy and may be avoided by using automated records generated by information management systems.