Employing a results-based algorithm to reduce laboratory utilization in ACTH stimulation testing

BACKGROUND: The High Dose Adrenocorticotropic Hormone (ACTH) Stimulation Test is the gold standard to diagnose adrenal insufficiency. Normal adrenal function is defined as a peak cortisol response to pharmacologic stimulation with cosyntropin of ≥18 μg/dL. Our practice was to obtain cortisol levels at 0, 30 and 60 min after cosyntropin administration. Once a value of ≥18 μg/dL has been obtained, adrenal insufficiency is ruled out and there is little diagnostic utility in subsequent stimulated levels. METHODS: We aimed to decrease laboratory utilization by developing a results-based algorithm in the electronic medical record (EMR). Cortisol levels were analyzed on the 0 and 60 min samples; then an EMR discern rule automatically generated an order to analyze the 30-min sample if the 60-min cortisol level was subnormal. RESULTS: Exclusion of adrenal insufficiency was excluded using one stimulated cortisol level in 8% prior to algorithm development. After several plan-do-study-act cycles, 99% of normal tests were performed using only one stimulated cortisol level. CONCLUSIONS: This laboratory-based algorithm resulted in reduced laboratory utilization, and aligned our practice to recommendations of the Pediatric Endocrine Society. Similar algorithms could be created for other dynamic tests to reduce unnecessary laboratory utilization.