EP-4742259-A1 - METHOD AND DEVICE FOR PREDICTING THE APPROPRIATENESS OF ANTIBIOTIC USE IN PATIENTS WITH INFECTIOUS DISEASES

Abstract

The present invention relates to a method for predicting the appropriateness of antibiotic use in a patient with an infectious disease and a device thereof, comprising receiving a bio-signal data, blood test data, and clinical characteristic data for a subject; and predicting appropriateness of antibiotic therapy for the subject based on the received bio-signal data, blood test data, and clinical characteristic data using a prediction model configured to predict the appropriateness of antibiotic use with the bio-signal data, blood test data, and clinical characteristic data as input.

Inventors

• LEE, YONG SEOP
• HEO, SEOK JAE
• KU, NAM SU
• PARK, SE YOON

Assignees

• INDUSTRY-ACADEMIC COOPERATION FOUNDATION, YONSEI UNIVERSITY
• Industry-University Cooperation Foundation Hanyang University

Dates

Publication Date

20260513

Application Date

20251106

Claims (15)

1. A method for providing information on the appropriateness of antibiotic use implemented by a processor, comprising receiving bio-signal data, blood test data, and clinical characteristic data for a subject; and in a first prediction step, predicting appropriateness of antibiotic therapy for the subject based on the received bio-signal data, blood test data, and clinical characteristic data using a prediction model configured to predict the appropriateness of antibiotic use with the bio-signal data, blood test data, and clinical characteristic data as input, wherein the bio-signal data, blood test data, and clinical characteristic data for the subject are data obtained at a first time point, wherein the first time point is within 5 days after antibiotic treatment, wherein the appropriateness of antibiotic therapy for the subject in the first prediction step is defined as the appropriateness for the first time point.
2. The method of claim 1, wherein the subject is a patient with bacteremia.
3. The method of claim 1, wherein the bio-signal data and the blood test data are time-series data collected from the subject for 5 days after administration of antibiotics.
4. The method of claim 1, wherein the bio-signal data includes at least one of a systolic blood pressure (SBP), a diastolic blood pressure (DBP), a body temperature, a heart rate, and a respiratory rate, and the bio-signal data includes a pair of baseline values before taking an antibiotic and a value obtained at a time point within 5 days after taking the antibiotic.
5. The method of claim 1, wherein the blood test data includes at least one of aPTT (Activated Partial Thromboplastin Time) level, a CRP (C-reactive Protein) level, an ESR (Erythrocyte Sedimentation Rate) level, a Lactate level, a Procalcitonin level, a PT/INR (Prothrombin Time/International Normalized Ratio) level, a Delta neutrophil level, Hematocrit level, Hemoglobin level, MCH (Mean Corpuscular Hemoglobin) level, MCHC (Mean Corpuscular Hemoglobin Concentration) level, MCV (Mean Corpuscular Volume) level, MPV (Mean Platelet Volume) level, PDW (platelet distribution width) level, PDW_fL level, PLT (platelet count) level, RBC (red blood cell count) level, RDW (red blood cell distribution width) level, TMA (thrombotic microangiopathy) score, WBC (white blood cell count) level, albumin level, ALP (alkaline phosphatase) level, ALT (alanine aminotransferase) level, AST (aspartate aminotransferase) level, bilirubin level, BUN (blood urea nitrogen) level, calcium level, cholesterol level, creatinine level, glucose level, Phosphate levels, total protein levels, and uric acid levels, and the blood test data includes a pair of baseline values before taking an antibiotic and a value obtained at a time point within 5 days after taking the antibiotic.
6. The method of claim 1, wherein the clinical characteristic data includes at least one of the following: age, gender (Sex), Charlson Comorbidity Index score (CCI_score), PittScore, whether the subject suffered from diabetes mellitus, chronic liver disease, chronic kidney disease (CKD), congestive heart failure (CHF), cardiovascular disease (CVD), chronic obstructive pulmonary disease (COPD), connective tissue disease (CTD), cancer, or shock, or whether the subject is on ventilator support (Vent) status, continuous renal replacement therapy (CRRT) treatment status, inotropic drugs (Inotropics) treatment status, tigecycline treatment status, 1 st cephalosporins treatment status, 2 nd cephalosporins treatment status, 3 rd cephalosporins treatment status, 4 th cephalosporins treatment status, clindamycin treatment status, linezolid treatment status, Nafcillin treatment status, TMP-SMX treatment status, aminoglycoside treatment status, beta-lactam inhibitor treatment status, carbapenem treatment status, fluoroquinolones treatment status, glycopeptide treatment status, monobactam treatment status, penicillin treatment status, Polymyxin treatment status, Tetracycline treatment status, Piperacillin/tazobactam treatment status, Amoxicillin/clavulanate treatment status, or Ampicillin/sulbactam treatment status.
7. The method of claim 1, further comprising, re-receiving a bio-signal data, blood test data, and clinical characteristic data obtained at a second time point; and in a second prediction step, predicting the appropriateness of antibiotic therapy for the subject based on the re-received bio-signal data, blood test data, and clinical characteristic data, using the prediction model, the second time point is another time point between the first time point and 5 days after antibiotic treatment, the appropriateness of antibiotic therapy for the subject in the second prediction step is defined as appropriateness for the second time point, and the re-received bio-signal data and blood test data includes a pair of baseline values before administration of antibiotics and a value obtained at the second time point.
8. The method of claim 8, wherein the prediction model comprises a first prediction model configured to predict the appropriateness of antibiotic use using the bio-signal data, blood test data, and clinical characteristic data obtained at the first time point as input, and a second prediction model configured to predict the appropriateness of antibiotic use using the bio-signal data, blood test data, and clinical characteristic data obtained at the second time point as input.
9. The method of claim 1, wherein the prediction model is model trained by receiving training data consisting of bio-signal data, blood test data, and clinical characteristic data obtained from a blood-infected subject to which an appropriate antibiotic is administered and a blood-infected subject to which an inappropriate antibiotic is administered, and predicting the appropriateness of antibiotic use based on the training data.
10. A device for providing information on the appropriateness of antibiotic use, comprising, a receiving unit configured to receive a bio-signal data, blood test data, and clinical characteristic data from a subject, and a processor connected to the receiving unit, the processor is configured to predict the appropriateness of antibiotic therapy for the subject based on the received bio-signal data, blood test data, and clinical characteristic data, using a prediction model for appropriateness of antibiotic use configured to predict appropriateness of antibiotic use by taking the bio-signal data, blood test data, and clinical characteristic data as inputs, wherein the bio-signal data, blood test data, and clinical characteristic data is a data obtained at a first time point, the first time point being within 5 days of antibiotic treatment, and the appropriateness of antibiotic therapy for the subject is defined as appropriateness for the first time point.
11. The device of claim 10, wherein the subject is a patient with bacteremia.
12. The device of claim 10, wherein the bio-signal data and the blood test data are time-series data collected from the subject for 5 days after administration of antibiotics.
13. The device of claim 10, wherein the bio-signal data includes at least one of a systolic blood pressure (SBP), a diastolic blood pressure (DBP), a body temperature, a heart rate, and a respiratory rate, and the bio-signal data includes a pair of baseline values before taking an antibiotic and a value obtained at a time point within 5 days after taking the antibiotic.
14. The device of claim 10, wherein the receiving unit is further configured to re-receive bio-signal data, blood test data, and clinical characteristic data obtained at a second time point from the subject, and the processor is further configured to predict the appropriateness of antibiotic therapy for the subject based on the re-received bio-signal data, blood test data, and clinical characteristic data, the second time point is another time point between the first time point and 5 days after antibiotic treatment, and the appropriateness of antibiotic therapy for the subject is defined as appropriateness for the second time point.
15. The device of claim 14, wherein the prediction model for appropriateness of antibiotic use comprises a first prediction model configured to predict the appropriateness of antibiotic use using the bio-signal data, blood test data, and clinical characteristic data obtained at the first time point as input, and a second prediction model configured to predict the appropriateness of antibiotic use using the bio-signal data, blood test data, and clinical characteristic data obtained at the second time point as input.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS This application claims the priority of Korean Patent Application No. 10- 2024-0156550 filed on November 6, 2024 and 10-2025-0088038 filed on July 1, 2025, in the Korean Intellectual Property Office, the disclosure of which is incorporated herein by reference. BACKGROUND Field The present invention relates to a method for predicting the appropriateness of antibiotic use in a patient with an infectious disease and a device thereof. Description of the Related Art Bacteremia can refer to the clinical condition in which the bacteria enter the blood and circulate throughout the body, and bloodstream infection is a synonym. This bacteremia can occur when bacteria proliferate at a rate that exceeds the removal capacity of the reticuloendothelial system(RES). At this time, bacteremia can be classified into transient bacteremia, intermittent bacteremia, and persistent bacteremia according to clinical manifestations. More specifically, transient bacteremia is bacteremia that occurs in the early stages of systemic or local infection and may be accompanied by meningitis, pneumonia, purulent arthritis, osteomyelitis, gonococcal or meningococcal infection. In addition, intermittent bacteremia can occur when there is an abscess in areas such as the abdominal cavity, pelvic cavity, perirenal region, liver, and prostate, and persistent bacteremia can occur when there is bacterial endocarditis and an infection of intravascular catheter, and can be accompanied by typhoid and brucellosis. On the other hand, there may be the use of antibiotics as a treatment of bacteremia, and empirical use of appropriate antibiotics in the early stages can prevent the progression of bacteremia to sepsis and septic shock. A description of the background is written to facilitate understanding of the invention. The matters described in this section should not be construed as admitted prior art. SUMMARY On the other hand, in order to treat of bacteremia it is most effective to clearly identify the causative bacteria that caused the infection and select antibiotics suitable for the bacteria as the types of bacteria that cause human infection are also very diverse, but the blood culture test used to identify bacteria in samples collected from blood or infected tissues takes at least 2 to 7 days, so empirical antibiotics are treated in consideration of the patient's infection diagnosis and clinical symptoms before the exact causative bacteria are identified. Furthermore, when a bacterial infection occurs in the body, an inflammatory reaction due to an excessive immune reaction occurs, and the symptoms rapidly deteriorate in a short time as the bacteria spread throughout the body through the bloodstream, so it is very important to treat appropriate antibiotics in the early stages of infection. However, since pathogenic organism have not been precisely identified, there are many cases where two to three antibiotics are used simultaneously in the empirical antibiotic treatment mentioned above. Excessive antibiotic combination can harm patients, and the use of inappropriate antibiotics has the problem of promoting antibiotic resistance that causes bacteria to become resistant to antibiotics. The inventors of the present invention have noted that hemodynamic indicators and various biomarker levels show different patterns depending on the appropriateness of empirical antibiotics as a way to overcome the aforementioned limitations. Accordingly, the inventors used bio-signal data, blood test data, and clinical characteristic data on a blood-infected subject to which an appropriate antibiotic is administered and a blood-infected subject to which an inappropriate antibiotic is administered as training data to build a model that can predict the appropriateness of antibiotic use even before test results identifying bacteria present in a sample of a bacteremia patient in the early stages of the infection are available. As a result, the inventors have developed a system for providing information on the appropriateness of antibiotic use for treating a patient with bacteremia in the early stage of a new infection based on a prediction model. Accordingly, the inventors expected that the appropriateness of antibiotic therapy could be predicted through the treatment response of antibiotic therapy before the results of the bacterial identification test of patients with bacteremia were reported by the prediction model, so that medical staff could effectively treat bacteremia in the early stages of infection. Accordingly, an object of the present invention is to provide a method for providing information on the appropriateness of antibiotic use and a device using the same, which are configured to predict the appropriateness of antibiotic therapy through a treatment response of antibiotic therapy for the subject based on bio-signal data, blood test data, and clinical characteristic data obtained from the subject using a prediction mo