Computerized laboratory alerting system to warn of life-threatening events

A computerized laboratory alerting system (CLAS) was developed using the decision-making capabilities of the HELP computer system at LDS Hospital in Salt Lake City, Utah. The CLAS generated alerts on life-threatening conditions indicated by laboratory test results and other patient data stored in HELP's computerized patient database. An effective alert feedback mechanism was developed which promoted quick review and acknowledgement of alerts by health care personnel. An evaluation of the effect of the CLAS on patient care and outcome was carried out. Results of the evaluation showed that general nursing floor patients with alerts for hypokalemia, falling potassium, hyperkalemia, hypokalemia--patient on digoxin, hyponatremia, falling sodium, hypernatremia, hypoglycemia and hyperglycemia were treated appropriately 68.1% of the time before CLAS implementation ('pre' study period) and 83.8% of the time after CLAS implementation ('post' study period, p = 0.017). The average number of hours which patients spent in the life-threatening conditions decreased from 30.4 hours in the 'pre' study period to 15.7 hours in the 'post' study period (p = 0.0125). The average patient length of stay also decreased from 350.6 hours in the 'pre' study period to 211.9 hours in the 'post' study period (p = 0.025). Results of the evaluation did not show significant differences in the number of times patients were handled appropriately, or in patient length of stay for general nursing floor patients generating alerts for metabolic acidosis. There was, however, a significant decrease in the average number of hours spent in the life-threatening condition (44.3 hours in the 'pre' study period vs. 26.5 hours in the 'post' study period, p = 0.033) for patients with metabolic acidosis alerts on the general nursing floors. Nurses were the major users of the CLAS (53.0%), followed by physicians (28.5%), ward clerks (11.5%), and other users (7.0%). The overall assessment of the value of the CLAS by users was positive. The CLAS was classified as very useful by 25.6% of users, frequently useful by 24.4% of users, somewhat useful by 34.6% of users, seldom useful by 7.7% of users, and not useful by 7.7% of users.