Riu Marta, Chiarello Pietro, Terradas Roser, Sala Maria, Castells Xavier, Knobel Hernando, Cots Francesc
IMIM, Instituto Hospital del Mar de Investigaciones Médicas, Barcelona, España; Programa de doctorat en Salut Pública, Departament de Pediatria, Obstetrícia i Ginecologia, Medicina Preventiva i Salut Pública, Universitat Autònoma de Barcelona (UAB), Barcelona, España.
IMIM, Instituto Hospital del Mar de Investigaciones Médicas, Barcelona, España; Programa de doctorat en Salut Pública, Departament de Pediatria, Obstetrícia i Ginecologia, Medicina Preventiva i Salut Pública, Universitat Autònoma de Barcelona (UAB), Barcelona, España.
Enferm Infecc Microbiol Clin. 2016 Dec;34(10):620-625. doi: 10.1016/j.eimc.2015.09.007. Epub 2015 Nov 10.
The excess cost associated with nosocomial bacteraemia (NB) is used as a measurement of the impact of these infections. However, some authors have suggested that traditional methods overestimate the incremental cost due to the presence of various types of bias. The aim of this study was to compare three assessment methods of NB incremental cost to correct biases in previous analyses.
Patients who experienced an episode of NB between 2005 and 2007 were compared with patients grouped within the same All Patient Refined-Diagnosis-Related Group (APR-DRG) without NB. The causative organisms were grouped according to the Gram stain, and whether bacteraemia was caused by a single or multiple microorganisms, or by a fungus. Three assessment methods are compared: stratification by disease; econometric multivariate adjustment using a generalised linear model (GLM); and propensity score matching (PSM) was performed to control for biases in the econometric model.
The analysis included 640 admissions with NB and 28,459 without NB. The observed mean cost was €24,515 for admissions with NB and €4,851.6 for controls (without NB). Mean incremental cost was estimated at €14,735 in stratified analysis. Gram positive microorganism had the lowest mean incremental cost, €10,051. In the GLM, mean incremental cost was estimated as €20,922, and adjusting with PSM, the mean incremental cost was €11,916. The three estimates showed important differences between groups of microorganisms.
Using enhanced methodologies improves the adjustment in this type of study and increases the value of the results.
与医院获得性菌血症（NB）相关的额外费用被用作衡量这些感染影响的指标。然而，一些作者认为，由于存在各种类型的偏差，传统方法高估了增量成本。本研究的目的是比较NB增量成本的三种评估方法，以纠正先前分析中的偏差。
将2005年至2007年间发生过一次NB的患者与同一全患者精细诊断相关组（APR-DRG）中未发生NB的患者进行比较。根据革兰氏染色对致病微生物进行分组，以及菌血症是由单一还是多种微生物或真菌引起。比较了三种评估方法：按疾病分层；使用广义线性模型（GLM）进行计量经济学多变量调整；并进行倾向得分匹配（PSM）以控制计量经济学模型中的偏差。
分析包括640例NB住院患者和28459例无NB患者。观察到的NB住院患者平均费用为24515欧元，对照组（无NB）为4851.6欧元。分层分析中平均增量成本估计为14735欧元。革兰氏阳性微生物的平均增量成本最低，为10051欧元。在GLM中，平均增量成本估计为20922欧元，通过PSM调整后，平均增量成本为11916欧元。这三个估计值显示了微生物组之间的重要差异。
使用改进的方法可以改善此类研究中的调整，并提高结果的价值。
