Prediction Model for Urinary Tract Infection

Overseen By Mofetoluwa Abraham Oluwasanmi

Age: 18+

Sex: Female

Trial Phase: Academic

Sponsor: Case Western Reserve University

Must be taking: Antibiotics

Disqualifiers: Male, Catheter, ESI 1-2, others

No Placebo Group All trial participants will receive the active study treatment (no placebo)

Compensation: Varies

Number of Visits: 2

Duration: 2 weeks

What You Need to Know Before You Apply

What is the purpose of this trial?

This trial tests a new method to reduce unnecessary antibiotic use for urinary tract infections (UTIs), aiming to combat antibiotic resistance. The focus is on a prediction model (Decision Aid-prediction model) and a callback system to assess whether patients truly need antibiotics after an emergency room visit. Women treated in the ER for UTI symptoms and sent home with antibiotics may qualify for this study. As an unphased trial, it offers participants the chance to contribute to innovative research that could enhance UTI treatment and decrease antibiotic resistance.

Will I have to stop taking my current medications?

The trial information does not specify whether you need to stop taking your current medications. It's best to discuss this with the study team or your doctor.

What prior data suggests that this prediction model is safe for use in reducing unnecessary antibiotic use?

Research has shown that a decision aid with a prediction model can safely assist doctors in making better treatment choices for urinary tract infections (UTIs). This method employs machine learning, a type of artificial intelligence, to predict the presence of a UTI based on routine urine tests and other information, aiding doctors in determining the necessity of antibiotics.

Studies have found that these models enhance the accuracy of UTI diagnoses, reducing the likelihood of unnecessary antibiotic prescriptions. This is crucial, as overusing antibiotics can lead to resistance. Notably, using this decision aid does not involve new medications or risky procedures; it simply uses smart tools to guide treatment decisions.

No reports of harmful effects have emerged from using these prediction models. They are well-received because they provide additional information to healthcare providers, helping to avoid unnecessary antibiotic use, which benefits patient safety and public health. ^{1 2 3 4 5}

Why are researchers excited about this trial?

Researchers are excited about the Decision Aid-prediction model for urinary tract infections (UTIs) because it offers a new way to predict and manage the condition. Unlike traditional treatments that focus on antibiotics to fight infections, this model uses data and algorithms to help doctors accurately predict the likelihood of a UTI in patients who show up at the ER with symptoms. This approach could lead to more precise and personalized care, reducing unnecessary antibiotic use and improving patient outcomes. By providing doctors with a powerful decision-making tool, this model has the potential to enhance how UTIs are diagnosed and treated.

What evidence suggests that this prediction model is effective for reducing unnecessary antibiotic use in UTIs?

Research has shown that machine learning can predict urinary tract infections (UTIs) using lab and urine test results. One study found that a decision tree model correctly identified 87% of actual UTI cases, even with uneven data. Risk score models also help identify patients at high risk. AI tools have reduced incorrect antibiotic prescriptions, which is crucial for combating antibiotic resistance. This trial will evaluate a decision aid-prediction model for patients presenting to the ER with UTI symptoms. These findings suggest that prediction models could reduce unnecessary antibiotic use and improve treatment accuracy for UTIs. ^{2 5 6 7 8}

Who Is on the Research Team?

David Sheyn, MD

Principal Investigator

University Hospitals Cleveland Medical Center

Are You a Good Fit for This Trial?

This trial is for individuals who visit the emergency department with symptoms of a urinary tract infection (UTI), such as burning during urination, frequent urge to pee, or cloudy urine. The study excludes specific details on eligibility criteria.

Inclusion Criteria

I am older than 18 years.

Discharge ICD code consistent with a UTI diagnosis

I was prescribed antibiotics for a UTI when I left the hospital.

Exclusion Criteria

Pregnancy (confirmed with a negative pregnancy test ordered in the ER)

I need a catheter for urination long-term.

Patients who have an Emergency Severity Index (ESI) of 1 and 2

Timeline for a Trial Participant

Screening

Participants are screened for eligibility to participate in the trial

1-2 weeks

1 visit (in-person)

Treatment

Participants receive antibiotics based on a prediction model and are monitored for UTI symptoms

2 weeks

1 visit (in-person), follow-up calls

Follow-up

Participants are monitored for safety and effectiveness after treatment, with a focus on antibiotic resistance

4 weeks

Follow-up calls

What Are the Treatments Tested in This Trial?

Interventions

• Decision Aid-prediction model

Trial Overview The trial tests a prediction model designed to guide antibiotic use in UTI treatment. It aims to reduce unnecessary prescriptions by using this model along with a system that informs patients if they don't need antibiotics after culture results are analyzed.

How Is the Trial Designed?

Group I: Presenting to ER for Urinary Tract Infection (UTI) Experimental Treatment 1 Intervention

Patients presenting to the a UH ER location for UTI symptoms.

Who Is Running the Clinical Trial?

Case Western Reserve University

Lead Sponsor

Trials

314

Recruited

236,000+

Citations

1. pmc.ncbi.nlm.nih.gov pmc.ncbi.nlm.nih.gov/articles/PMC11715496/

Prediction of urinary tract infection using machine learning ... This study showed the potential of machine learning strategies for prediction of UTI according to laboratory and urinalysis results.

2. sciencedirect.com sciencedirect.com/science/article/abs/pii/S0195670125001264

and community-associated urinary tract infections using ... The Decision Tree model demonstrated the highest sensitivity, particularly in handling the highly imbalanced data of HAI, with a sensitivity of 87%.

3. journals.plos.org journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0290215&type=printable

Risk score models for urinary tract infection hospitalization Our findings emphasize the effectiveness of risk score mod- els as practical tools for identifying high-risk patients and provide a quantitative ...

4. nature.com nature.com/articles/s41746-024-01400-5

AI driven decision support reduces antibiotic mismatches ... This study evaluates the impact of “UTI Smart-Set” (UTIS), an AI-driven decision-support tool, on prescribing patterns and mismatches in a large outpatient ...

5. pmc.ncbi.nlm.nih.gov pmc.ncbi.nlm.nih.gov/articles/PMC11057209/

Interpretable machine learning-based decision support for ... In this study, we present four interpretable machine learning-based decision support algorithms for predicting antimicrobial resistance.

6. pmc.ncbi.nlm.nih.gov pmc.ncbi.nlm.nih.gov/articles/PMC12113611/

Machine Learning for UTI Prediction in Real-World Laboratory ... This study aimed to develop and evaluate machine learning (ML) models that predict urine culture outcomes using routine urinalysis and demographic data.

7. sciencedirect.com sciencedirect.com/science/article/pii/S2001037024002502

A step forward in the diagnosis of urinary tract infections The aim of this study was to improve UTI diagnostics in clinical practice by application of machine learning (ML) models for real-time UTI prediction.

8. medrxiv.org medrxiv.org/content/10.1101/2023.08.06.23293723v1.full-text

Risk Score Models for Unplanned Urinary Tract Infection ... In this study, we develop two risk score models to predict unplanned hospitalizations for UTI using claims data from Centers for Medicare and ...