Machine learning of cell population data, complete blood count, and differential count parameters for early prediction of bacteremia among adult patients with suspected bacterial infections and blood culture sampling in emergency departments

Recent Advancements in Machine Learning-Based Bloodstream Infection Prediction: A Systematic Review and Meta-analysis of Diagnostic Test Accuracy

“…The selection process, including reasons for exclusion at each stage, is outlined in the PRISMA flow diagram (Figure 1). Notably, the majority (29 out of 30) of the included articles employed retrospective study designs, with only two conducting prospective validations of their models [40, 41]. A single study presented a prospective cohort design [29].…”

“…Most of the reviewed studies were conducted within a single institution; however, three studies utilized datasets encompassing two hospital systems [20, 22, 44], and two studies expanded their analysis to incorporate multi-center data [41, 49]. External validation, which is critical for the generalizability of findings, was performed in four studies [38, 40, 41, 44]. Compliance with the Transparent Reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD) guidelines, which enhance the reliability of predictive modeling, was confirmed in five studies [20, 28, 37, 39, 41, 50].…”

“…Within this context, three studies utilized the publicly available Medical Information Mart for Intensive Care (MIMIC) database [41, 44, 48], while four studies included Complete Blood Count/Differential Count (CBC/DC) data [24, 25, 34, 40]. Furthermore, two of these studies also incorporated Cell Population Data (CPD) [24, 40], highlighting the integration of detailed hematologic parameters. Unstructured data utilization was observed in one study that utilized textual data [23].…”

“…Within inpatient settings, nine studies examined general populations [20][21][22][23][24][25][26][27][28], and others focused on specific patient cohorts including those with central venous catheters (CVC) [30,31], systemic inflammatory response syndrome (SIRS) [29], and various other conditions. In ED settings, seven studies were identified [37][38][39][40][41][42][43], two of which addressed specific patient groups. ICU-focused research numbered six studies [44][45][46][47][48][49], again with two concentrating on specific patient subsets.…”

Advancing Bloodstream Infection Prediction Using Explanable Artificial Intelligence Framework

“…Within the inpatient group, the studies varied, with nine examining general populations [14][15][16][17][18][19][20][21][22], two targeting inpatients with central venous catheters (CVC) [23,24], and others focusing on specific patient categories such as, hemodialysis (HD) patients [26], cancer patients [27], maternity patients [28], patients with low procalcitonin levels (PCT ≤2.0 ng/ml) [29], and HIV patients [30]. Bacteremia was the primary condition under study in 24 articles, including varied focuses such as fungemia [14] and Candidemia [27,43], while three studies aimed at predicting central line-associated bloodstream infections (CLABSIs) [33,34,41] and one on hospital-acquired BSI (HA-BSI) [22]. All articles reported high prediction performance (AUROC > 0.7) except for one article [20] which reported poor performance metrics.…”