Appropriate Use of Blood Cultures in the Emergency Department Through Machine Learning

Official title

Appropriate Use of Blood Cultures in the Emergency Department Through Machine Learning: a Randomized Controlled Trial

Overview

The goal of this clinical trial is to study whether the use of our blood culture prediction tool is non-inferior to current practice and if it can improve certain outcomes in all adult patients presenting to the emergency department with a clinical indication for a blood culture analysis (according to the treating physician). The primary endpoint is 30-day mortality. Key secondary outcomes are: * hospital admission rates * in-hospital mortality * hospital length-of-stay. In the intervention group, the physician will follow the advice of our blood culture prediction tool. In the comparison group all patients will undergo a blood culture analysis.

Detailed description

Rationale: The overuse of blood cultures in emergency departments leads to low yields and high numbers of contaminated cultures, which is associated with increased diagnostics, antibiotic usage, prolonged hospitalisation, and mortality. Ideally, blood cultures would only be performed in patients with a high risk for a positive culture. The investigators have developed a machine learning model to predict the outcome of blood cultures in the ED. Retrospective and prospective validation of the tool in various settings show that it can be used to reduce the number of blood culture analyses by at least 30% and help avoid the hidden costs of contaminated cultures.

Objective: This study aims to investigate whether the use of our blood culture prediction tool is non-inferior to current practice and if it can improve certain outcomes.

Study design: A randomized controlled non-inferiority trial. Study population: All adult patients presenting to the emergency department with a clinical indication for a blood culture analysis (according to the treating physician).

Intervention: In the control group, all patients will undergo a blood culture analysis. In the intervention group, the physician will follow the advice of our blood culture prediction tool. If the chance of a positive blood culture is \< 5%, the blood culture analysis will be cancelled and the sample destroyed. If the change of a positive blood culture is \> 5%, the blood culture analysis will be performed as usual.

Main study parameters/endpoints: The primary endpoint is 30-day mortality, for which the investigators aim to show non-inferiority. Key secondary outcomes, for which the investigators also aim to show non-inferiority, are hospital admission rates, in-hospital mortality, and hospital length-of-stay.

Interventions

• Device Blood culture prediction tool
  Machine learning based predicition tool

Primary outcome measures

• 30-day mortality [Time frame: 30 days]

Secondary outcome measures (3)

• hospital admission rates [Time frame: 1 day]
• in-hospital mortality [Time frame: 90 days]
• hospital length-of-stay [Time frame: 90 days]

Eligibility criteria

Inclusion criteria

• Age >= 18 years
• Have a clinical indication for a blood culture analysis (according to the treating physician)
• Have sufficient data recorded (laboratory results and vital sign measurements) for a prediction to be made (at least 20% of the needed parameters)

Exclusion criteria

• Central Venous Line (CVL) or Peripherally Inserted Central Catheter (PICC) in situ
• Neutrophil count < 0.5 \* 109/L
• Candidemia or S. aureus bacteraemia in the past 3 months.
• Most likely diagnosis of endocarditis/spondylodiscitis/infected prosthetic material
• Pregnant or breastfeeding patients
• Not capable of giving informed consent

Criteria are shown verbatim from the registry (in English). Final eligibility is always assessed by the study center.

Healthy volunteers: Yes

Study design

Allocation

Randomized

Model

Parallel assignment

Masking

Open label

Primary purpose

Treatment

Study locations

Netherlands · 1 center

• Amsterdam UMC - location AMC — Amsterdam

Publications

• Boerman AW, Schinkel M, Meijerink L, van den Ende ES, Pladet LC, Scholtemeijer MG, Zeeuw J, van der Zaag AY, Minderhoud TC, Elbers PWG, Wiersinga WJ, de Jonge R, Kramer MH, Nanayakkara PWB. Using machine learning to predict blood culture outcomes in the emergency department: a single-centre, retrospective, observational study. BMJ Open. 2022 Jan 4;12(1):e053332. doi: 10.1136/bmjopen-2021-053332. PMID 34983764
• Schinkel M, Boerman AW, Bennis FC, Minderhoud TC, Lie M, Peters-Sengers H, Holleman F, Schade RP, de Jonge R, Wiersinga WJ, Nanayakkara PWB. Diagnostic stewardship for blood cultures in the emergency department: A multicenter validation and prospective evaluation of a machine learning prediction tool. EBioMedicine. 2022 Aug;82:104176. doi: 10.1016/j.ebiom.2022.104176. Epub 2022 Jul 16. PMID 35853298
• van der Zaag AY, Bhagirath SC, Boerman AW, Schinkel M, Paranjape K, Azijli K, Ridderikhof ML, Lie M, Lissenberg-Witte B, Schade R, Wiersinga J, de Jonge R, Nanayakkara PWB. Appropriate use of blood cultures in the emergency department through machine learning (ABC): study protocol for a randomised controlled non-inferiority trial. BMJ Open. 2024 May 31;14(5):e084053. doi: 10.1136/bmjopen-2024-0840 PMID 38821574

Identifiers

NCT: NCT06163781 · NL81971.000.22

Primary sources (government registries)

• ClinicalTrials.gov (US NIH)

View this study on ClinicalTrials.gov ↗