To evaluate the operating characteristics of a previously published decision rule (Lowe's criteria) for the ordering of the stat serum electrolyte panel (SEP) and to identify features from the history and physical examination that are predictive for clinically significant electrolyte abnormalities (CSEA) in older emergency patients.
Design: A cross-sectional study using a physician questionnaire, record review, and telephone follow-up.
Setting: An urban, university hospital emergency department with approximately 70,000 patient visits a year.
Type of participants: A convenience sample of patients 55 years of age or older who presented for care for any reason were identified by a research associate stationed in the ED.
Interventions: Physicians managing 1,766 patients were interviewed to determine the presence or absence of 20 dichotomous clinical variables, ten of which constitute Lowe's criteria set. All patients were followed by chart review and~or telephone interview
Measurements and main results: The determination of clinical significance was made on record review using pre-established guidelines. The SEP was ordered on 800 of the patients on the index visit. The yield of CSEA in this group was β’6%. Lowe's criteria operated with a sensitivity of 0.95 (95% confidence intervals, 0.89 -0.98) and a specificity of 0.I0 (95% confidence intervals, 0.07 -0.12) for predicting CSEAs in this population. Logistic regression analysis showed that impaired ability to communicate, acute seizures, vomiting, and prior abnormal electrolytes were independent predictors for CSEA.
Conclusion: Lowe's criteria showed useful sensitivity but poor specificity in this populationl The criteria may be used to encourage selective test ordering by physicians when a low pretest probability for CSEA exists and as par t of an algorithm for emergency nurses who order tests to expedite patient care.