Machine learning to predict how long spinal anesthesia lasts for total knee replacement

Comparative Evaluation of Machine Learning Algorithms for Predicting Spinal Anesthesia Termination Time

This project will try machine learning to predict how long spinal anesthesia will last and when patients can begin mobilizing after total knee replacement under spinal anesthesia.

Quick facts

• Study type: Observational
• Enrollment: 140 (estimated)
• Ages: 18 Years and up
• Sex: All
• Sponsor: Kocaeli City Hospital (Government)
• Locations: 1 site (Kocaeli, İzmit)
• Trial ID: NCT07256548 on ClinicalTrials.gov

What this trial studies

This observational study collects demographic, surgical, and anesthetic data from adults undergoing total knee arthroplasty with spinal bupivacaine anesthesia at Kocaeli City Hospital between November 2025 and March 2026. Machine learning models are trained on these clinical variables to estimate the time of spinal anesthesia termination and the patient’s readiness for mobilization. Model performance will be measured using standard accuracy and error metrics and compared for practical usability in perioperative workflows. The goal is to determine whether ML predictions can be integrated into routine anesthesia planning and postoperative care.

Who should consider this trial

Good fit: Adults aged 18 or older with ASA physical status I or II who are scheduled for total knee arthroplasty under spinal anesthesia at Kocaeli City Hospital and can provide informed consent.

Not a fit: Patients converted to general anesthesia, those requiring postoperative ICU care, those with surgical complications preventing planned mobilization, or those unable to complete pain assessments are unlikely to benefit from the predictive tool.

Why it matters

Potential benefit: If successful, the predictions could help tailor intraoperative planning and timing of analgesia and mobilization, improving safety, recovery speed, and patient comfort.

How similar studies have performed: Previous machine learning work has shown promise for predicting perioperative events like hypotension and complications, but using ML specifically to predict spinal anesthesia duration and mobilization timing is relatively novel and less tested.

Eligibility criteria

Inclusion Criteria:

1. Patients scheduled to undergo total knee arthroplasty between November 2025 and March 2026 at the Kocaeli City Hospital Operating Theaters.
2. Patients who have provided written informed consent to participate in the study.
3. Patients whose surgery is planned under spinal anesthesia.
4. Patients for whom complete clinical data can be obtained during the study period.
5. Adults aged 18 years or older, classified as American Society of Anesthesiologist's (ASA) Physical Status I or II.

Exclusion Criteria:

1. Patients who were converted to general anesthesia during surgery or initially operated under general anesthesia.
2. Patients who required postoperative intensive care unit (ICU) admission following anesthesia.
3. Patients who developed surgical complications and for whom postoperative mobilization could not be planned.
4. Patients with cognitive impairment preventing them from completing pain assessment scales in the postoperative period.
5. Patients with neuropathic pain, multiple sclerosis, or other neuromotor disorders will be excluded from the study.

Where this trial is running

Kocaeli, İzmit

• Kocaeli City Hospital — Kocaeli, İzmit, Turkey (Türkiye) (Recruiting)

Study contacts

• Study coordinator: Sıddık Varolgüneş, MD

Conditions Spinal Anesthesia; Machine Learning; Knee Arthroplasty, Total; Spinal Anesthesia Duration; Postoperative Care; Postoperative Acute Pain; spinal anesthesia; machine learning