About - Part 1 a — I-Sample

Part 1a
Analysis of routine hospital data

Quantitative analysis of routine hospital data from three NHS trusts has been undertaken. A rapid scoping review of predictors of reliable blood culture sampling has been conducted. The aim was to identify factors associated with reliable, high quality, blood sampling for patients prescribed broad-spectrum antibiotics within 24 hours of admission.

Methodology :

What data did we use?: Data were extracted from hospital electronic health records and electronic microbiology records among the three hospitals.

Participants: Adults (> 18 years) receiving intravenous broad-spectrum antibiotics within 24 hours of admission.

Analysis: Data were analysed using logistic regression.

Progress & Key Findings

Datasets from all three hospital sites were analysed.

Initial analysis of data from two datasets found that:

Clinical presentation influenced blood culture collection:

It found that the strongest predictor of whether a blood culture was taken was how unwell the patient appeared at admission.

Lower-severity patients were more likely to miss blood cultures:

Patients with lower NEWS (National Early Warning Score) values or who did not trigger sepsis alerts were less likely to have blood cultures collected.

Our findings suggest that patients who appeared less seriously unwell were more likely to have the recommended blood cultures missed, even when they were treated with broad-spectrum intravenous antibiotics.

Part 1a Analysis of routine hospital data

Part 1a
Analysis of routine hospital data