BACTERIUM: Study for a Machine-learning-based Model to Predict Bloodstream Infections

Fondazione Policlinico Universitario Agostino Gemelli IRCCS5,000 enrolled

Overview

An increase of healthcare-associated infections caused by multidrug- resistant organisms (MRDO) is currently observed. One of the main causes of the emergence of a MDRO infection is an overuse of antibiotics. Therefore, saving useless antibiotic treatment is currently a priority from a public health point of view. The evaluation of the risk of having a bloodstream infection will allow both activating faster treatment decisions (when the risk is significantly high) or to save useless resources in terms of diagnostic tests and treatments, also limiting the potential for side effects (when the risk is significantly low).

Study Type

OBSERVATIONAL

Enrollment

5,000

Conditions

Infection, Bloodstream

Interventions

antibiotic treatmentsDRUG

The outcome of the study will be to implement Prediction Models to support clinicians for the assessment of risk of having a bloodstream infection.

Eligibility

Medical Language ↔ Plain English

Inclusion Criteria: 1. adult patients hospitalized at the Gemelli Polyclinic Foundation 2. having at least one performed blood culture 3. starting an antibiotic therapy Exclusion Criteria: 1. blood cultures with contaminants 2. \<18 years old 3. died after less than 48 hours from blood cultures positivity

Outcomes

Primary Outcomes

Rate of appropriate antibiotic therapy in patients with bloodstream infections

1\) number of patients with bloodstream infections with appropriate antibiotic therapy before and after the application of the predictive model "Bacterium"

Time frame: 24 months

Blood cultures

number of blood cultures done before and after the application of the predictive model "Bacterium"