PROBIty: How to Reduce Unnecessary Blood Cultures: Construction and Validation of a Predictive Score for Blood Culture Positivity in Intensive Care

Sponsor

University Hospital, Tours (Other)

Overall Status

Recruiting

CT.gov ID

NCT05686577

Collaborator

(none)

2,400

Enrollment

Locations

18.3

Anticipated Duration (Months)

240

Patients Per Site

13.1

Patients Per Site Per Month

Study Details

Study Description

Brief Summary

Prospective observational cohort consisting of all adult patients admitted to participating critical care units (ICU and CCU) during the study period, with blood cultures collected as part of their care, and who did not express any objection to participating.

For each patient, data will be collected prospectively for each blood culture set collected.

Condition or Disease	Intervention/Treatment	Phase
Infection, Bacterial Infection, Fungal Bacteremia	Other: no intervention

Study Design

Study Type:

Observational

Anticipated Enrollment :

2400 participants

Observational Model:

Cohort

Time Perspective:

Prospective

Official Title:

Comment Diminuer Les hémocultures Inutiles : Construction et Validation d'un Score prédictif de positivité Des hémocultures en réanimation.

Actual Study Start Date :

Nov 22, 2022

Anticipated Primary Completion Date :

May 1, 2024

Anticipated Study Completion Date :

Jun 1, 2024

Arms and Interventions

Arm	Intervention/Treatment
learning database group The results of blood cultures and the clinical elements of the patients in the first sample (1600 patients) will constitute the learning database. A supervised learning will be performed on these data in order to select a set of clinical elements allowing to define a predictive score of positive blood cultures	Other: no intervention no intervention
validation database group The score will be validated on the remaining sample (800 patients), independent of the first sample. Both samples will be constituted by randomization.	Other: no intervention no intervention

Arm

Intervention/Treatment

learning database group

The results of blood cultures and the clinical elements of the patients in the first sample (1600 patients) will constitute the learning database. A supervised learning will be performed on these data in order to select a set of clinical elements allowing to define a predictive score of positive blood cultures

Other: no intervention

no intervention

validation database group

The score will be validated on the remaining sample (800 patients), independent of the first sample. Both samples will be constituted by randomization.

Other: no intervention

no intervention

Outcome Measures

Primary Outcome Measures

Area under the ROC curve of the predictive performance of the score for predicting bacteremia or fungemia. [day 1- day 28]

Secondary Outcome Measures

Area under the ROC curve of the diagnostic performance of body temperature to predict bacteremia or fungemia; [day 1- day 28]
Area under the ROC curve of the diagnostic performance of the Shapiro score for predicting bacteremia or fungemia; [day 1- day 28]
Proportion of bacteremia or fungemia identified by blood cultures in patients on antibiotic therapy and broad-spectrum antibiotic therapy, and factors associated with these bacteremias or fungemias; [day 1- day 28]
Comparison of the diagnostic performance of the score in predicting bacteremia or fungemia observed during the first 48 hours of hospitalization to later bacteremias; [48 hours]
Descriptive study of the contribution of blood cultures, whatever their results, to the prescription of anti-infectives. [day 1- day 28]
Efficiency of the new predictive score for blood culture positivity in Intensive Care, as compared with body temperature and Shapiro score, measured by two Incremental Cost-Effectiveness Ratios (ICER): ICER (incremental cost per additional well-detecte [less than 1 month]
A tree decision model will compare three strategies for predicting positive blood cultures: the new predictive score, body temperature and Shapiro score. Within-trial data will be used to define the model parameters. The model will consider the costs related to the implementation of the strategies, blood cultures, and potential delays in patient management. This will result in the estimation of the two ICERs.

Eligibility Criteria

Criteria

Ages Eligible for Study:

18 Years and Older

Sexes Eligible for Study:

All

Accepts Healthy Volunteers:

Inclusion Criteria:

Male or female, age 18 years or older
Patient admitted to an ICU or ICU
patient with a blood culture sample as part of care

Exclusion Criteria:

Patient already included in this study during a previous hospitalisation
Opposition expressed for participation in the study

Contacts and Locations

Locations

	Site	City	Country	Postal Code
1	CHU Brest	Brest	France
2	CHU Dijon	Dijon	France
3	CHD La Roche sur Yon	La Roche-sur-Yon	France
4	CH Le Mans	Le Mans	France
5	CHU Nantes	Nantes	France
6	CHR Orléans	Orléans	France
7	CHU Poitiers	Poitiers	France
8	CHU Rennes	Rennes	France
9	CHIU Strasbourg	Strasbourg	France
10	CHU de Tours	Tours	France	37044