A method for predicting the utilization of medical technology in neonatal care

Abstract

The utilization of 12 neonatal intensive care technologies was recorded in 193 patients. An index of utilization (P_u) was adopted to classify these technologies into three sub-groups corresponding to low (P_u < 15%), medium (15% ⩽ P_u ⩽ 85%), and high (P_u > 85%) intensity of use. No models were developed for technologies with high intensity of use (intravenous hydration, warming, micro-haematocrit, heart rate monitoring). Multiple linear regression was used to predict the utilization of technologies with medium intensity of use (gasometry, oxygen hood, CPAP, mechanical ventilation), and multiple logistic regression was used with the same purpose for technologies with low intensity of use (exchange transfusion, bicarbonate, adrenaline, echocardiography). Both methods yielded significant models ( $\mathit{P}\mathrm{<}0.05$ ) which can be used to improve planning and management of technology in neonatal intensive care units.