Five-year evolution of drug prescribing in a university adult intensive care unit

Introduction: Drug prescription is difficult in ICUs as prescribers are many, drugs expensive and decisions complex. In our ICU, specialist clinicians (SC) are entitled to prescribe a list of specific drugs, negotiated with intensive care physicians (ICP). The objective of this investigation was to assess the 5-year evolution of quantity and costs of drug prescription in our adult ICU and identify the relative costs generated by ICP or SC. Methods: Quantities and costs of drugs delivered on a quarterly basis to the adult ICU of our hospital between 2004 and 2008 were extracted from the pharmacy database by ATC code, an international five-level classification system. Within each ATC first level, drugs with either high level of consumption, high costs or large variations in quantities and costs were singled out and split by type of prescriber, ICP or SC. Cost figures used were drug purchase prices by the hospital pharmacy. Results: Over the 5-year period, both quantities and costs of drugs increased, following a nonsteady, nonparallel pattern. Four ATC codes accounted for 80% of both quantities and costs, with ATC code B (blood and haematopoietic organs) amounting to 63% in quantities and 41% in costs, followed by ATC code J (systemic anti-infective, 20% of the costs), ATC code N (nervous system, 11% of the costs) and ATC code C (cardiovascular system, 8% of the costs). Prescription by SC amounted to 1% in drug quantities, but 19% in drug costs. The rate of increase in quantities and costs was seven times larger for ICP than for SC (Figure 1 overleaf ). Some peak values in costs and quantities were related to a very limited number of patients. Conclusions: A 5-year increase in quantities and costs of drug prescription in an ICU is a matter of concern. Rather unexpectedly, total costs and cost increases were generated mainly by ICP. A careful follow-up is necessary to try influencing this evolution through an institutional policy co-opted by all professional categories involved in the process.

Alveolar echinococcosis of the liver: Diffusion-weighted MRI findings and potential role in lesion characterisation

PURPOSE: To report the diffusion-weighted MRI findings in alveolar echinococcosis (AE) of the liver and evaluate the potential role of apparent diffusion coefficients (ADCs) in the characterisation of lesions. MATERIALS AND METHODS: We retrospectively included 22 patients with 63 AE liver lesions (≥1cm), examined with 3-T liver MRI, including a free-breathing diffusion-weighted single-shot echo-planar imaging sequence (b-values=50, 300 and 600s/mm(2)). Two radiologists jointly assessed the following lesion features: size, location, presence of cystic and/or solid components (according to Kodama's classification system), relative contrast enhancement, and calcifications (on CT). The ADCtotal, ADCmin and ADCmax were measured in each lesion and the surrounding liver parenchyma. RESULTS: Three type 1, 19 type 2, 17 type 3, three type 4 and 21 type 5 lesions were identified. The mean (±SD) ADCtotal, ADCmin and ADCmax for all lesions were 1.73±0.50, 0.76±0.38 and 2.63±0.76×10(-3)mm(2)/s, respectively. The mean ADCtotal for type 1, type 2, type 3, type 4 and type 5 lesions were 1.97±1.01, 1.76±0.53, 1.73±0.41, 1.15±0.42 and 1.76±0.44×10(-3)mm(2)/s, respectively. No significant differences were found between the five lesion types, except for type 4 (p=0.0363). There was a significant correlation between the presence of a solid component and low ADCmin (r=0.39, p=0.0016), whereas an inverse correlation was found between the relative contrast enhancement and ADCtotal (r=-0.34, p=0.0072). CONCLUSION: The ADCs of AE lesions are relatively low compared to other cystic liver lesions, which may help in the differential diagnosis. Although ADCs are of little use to distinguish between the five lesion types, their low value reflects the underlying solid component.

Automatic Mallampati Classification Using Active Appearance Models

Difficult tracheal intubation assessment is an important research topic in anesthesia as failed intubations are important causes of mortality in anesthetic practice. The modified Mallampati score is widely used, alone or in conjunction with other criteria, to predict the difficulty of intubation. This work presents an automatic method to assess the modified Mallampati score from an image of a patient with the mouth wide open. For this purpose we propose an active appearance models (AAM) based method and use linear support vector machines (SVM) to select a subset of relevant features obtained using the AAM. This feature selection step proves to be essential as it improves drastically the performance of classification, which is obtained using SVM with RBF kernel and majority voting. We test our method on images of 100 patients undergoing elective surgery and achieve 97.9% accuracy in the leave-one-out crossvalidation test and provide a key element to an automatic difficult intubation assessment system.