COMPAS: a computeried patient advice system to direct ventilatory care

We have undertaken a randomized, controlled clinical trial to test a new form of respiratory support which includes Extracorporeal Carbon Dioxide Removal (ECC02R), against traditional continuous positive pressure ventilation for patients with Adult Respiratory Distress Syndrome (ARDS). Since the experimental procedures were complex, detailed treatment protocols were developed by the clinical and research staff to describe monitored variables, frequency of data acquisition, decision logic (treatment protocols), and appropriate respiratory care for all modes of respiratory support in the clinical trial. The treatment protocols had the potential to make the clinical approach to patients more uniform by equalizing the frequency of monitoring and the intensity of care for clinical trial patients. The treatment protocols were converted into computer logic and implemented as a data driven COMputerized Patient Advice System (COMPAS) using a blackboard control architecture within the HELP clinical information system. COMPAS automatically prompted the clinical staff with ventilator or extracorporeal circuit adjustment advice when new observational respiratory therapists, nurses, or members of the radiology, pharmacy, or the blood gas laboratories. Following the availability of new data, COMPAS's therapy suggestions were available at the patient's bedside computer terminal in 2.2 minutes (avg.) (n = 3988) (82.3% available in under 2 minutes). After 624 hours of patient care on the first 5 patients in the controlled clinical trial, the clinical staff carried out 84.4% of the computer generated therapy suggestions (320 of 379). In situations which clearly suggested a therapy reduction, (Fi02 ? 90 mmHg) patients treated with COMPAS therapy suggestions had their Fi02 reduced more promptly (46.7 ± 51 minutes (mean ± SD) n = 20 for COMPAS patients vs 102.6 ± 91 minutes n = 2433 for patients treated in the traditional manner, p < 0.05). The high compliance with COMPAS ventilatory therapy suggestions (88.4%) showed that the respiratory management of complex ARDS patients can be accomplished using COMPAS's computerized medial logic. Both the frequency of monitoring and the intensity of patient care within the clinical trial were more uniform after COMPAS was implemented, thus enhancing the standardization of respiratory care crucial to the interpretation of the results of our clinical trial.