2000 OPEN FORUM Abstracts
PULSE OXIMETRY PERFORMANCE CAN AFFECT CLINICIAN EFFECTIVENESS
Charles G. Durbin, Jr., MD, FCCM, Stephanie K. Rostow, RRT; University of Virginia Health System, Charlottesville, VA
INTRODUCTION- Recent interest in human error and patient safety has identified "untrustworthy alarms and indicators" as being one of many "latent conditions" leading to human error [1,2]. We investigated the impact of an oximeter using a new signal processing technology [Masimo SET pulse oximetry (MSO) Masimo Corp, Irvine, CA] on the frequency of oximeter malfunction (downtime and/or inaccurate alarm conditions).
METHODS- We prospectively compared MSO to conventional pulse oximetry (CPO), the Ohmeda 3740 (Datex-Ohmeda, Louisville, CO). After obtaining Human Use Committee approval, 68 adult cardiac surgery patients with good preoperative ventricular function, following CABG surgery, were enrolled. On arrival in the ICU, both CPO and MSO were attached to the same hand of each patient. The output from both monitors was continuously recorded to a computer recording system until 4 hours following extubation or for a maximum of 24 hours postoperatively. The digits to which the monitors were attached were randomly chosen by the bedside clinician. Patients were randomly assigned to have the display of only one of the devices available to the bedside caregivers with the other device "blinded". No other clinical management was altered. From the computer records we determined episodes of device failure (reading 0% saturation), inaccuracy as determined by coincidental ABG (difference > 5%), and obvious artifacts were identified and tabulated as downtime, the total time the device was non-functional. The percentage of monitoring time during which each device was non-functional in each patient was then determined. Differences were analyzed using Student's t test. Significance was determined at p<0.05.
RESULTS- Total monitoring time per patient was 867 ± 362 (mean ± SD) minutes. Non-functional time (when used as caregiver device): CPO 3.6 ± 5.6 % (CI, 0.1 - 20.7 %); MSO 0.3 ± 0.4%, (CI, 0 -- 1.2%) p=0.03. Non-functional time (blinded device): CPO 9.4 ± 11.7 % (CI, 0.1 -- 38.7 %); MSO 0.3 ± 0.5%, (CI, 0 -- 1.8%) p=0.01.
DISCUSSION- Increased non-functional monitoring time, which was significantly greater with the CO device, resulted in caregivers needing to respond to the device failure, diverting them from patient care. It is assumed that clinicians had to devote additional time evaluating or questioning the functionality of either device, however this could not be quantified in the current study. Inaccurate, invalid and incorrect pulse oximetry consumes caregivers' time when they are forced to care for the monitor and not the patient. Masimo SET pulse oximetry provides significantly less oximeter non-functional time than CPO and reduces the number of "untrustworthy alarms and indicators". This performance benefit existed regardless of blinding, implying the potential for improved monitor reliability in unattended settings. Recent studies on human error and patient safety point to caregiver cognitive overload and distraction (termed latent conditions) as one cause of patient injury or error. Data from this pilot study suggests larger studies may reveal the extent of the impact of these decreased latent conditions and the benefit of increased clinician confidence on human error and patient safety.
 BMJ 2000; 320:768-70.
 Institute of Medicine. To Err is Human, National Academy Press, 1999.
Some technical support and equipment was provided by the Masimo Corporation, Irvine, CA