Using smartwatches, smart rings, and AI to predict who may be readmitted after surgery

Predicting Hospital Readmission for Surgical Patients Using Deep Learning Models With Smart Watch and Smart Ring Sensors Data

This project will try to see if data from a smartwatch and a smart ring can help predict which adults having medium-to-large elective surgery might need to return to the hospital after discharge.

Quick facts

• Study type: Observational
• Enrollment: 300 (estimated)
• Ages: 18 Years and up
• Sex: All
• Sponsor: Getúlio Vargas University Hospital (Government)
• Locations: 1 site (Manaus, Amazonas)
• Trial ID: NCT07349901 on ClinicalTrials.gov

What this trial studies

This observational study will continuously collect physiological and behavioral data before and after elective medium or large surgeries using a smartwatch and a smart ring, capturing metrics such as heart rate, ECG traces, oxygen saturation, sleep patterns, blood pressure trends, bioimpedance-based body composition, and stress indicators. Collected signals will be linked to clinical records and 30-day readmission outcomes and used to train and test deep learning models to identify patterns associated with unplanned return visits. The workflow emphasizes noninvasive, wearable monitoring combined with AI to detect early or subtle signs of postoperative complications that routine checks may miss. Study activities occur at Getúlio Vargas University Hospital with device fitting, data collection, and follow-up performed locally.

Who should consider this trial

Good fit: Adults (18 years or older) scheduled for medium or large elective surgery at Getúlio Vargas University Hospital who are conscious, oriented, able to use wearable devices, and who consent to participate are ideal candidates.

Not a fit: Patients with wrist or finger skin conditions that prevent wearable use, known allergies to device materials, implanted cardiac devices, pregnant or lactating individuals, or those unable to use the wearables are unlikely to benefit from this approach.

Why it matters

Potential benefit: If successful, the approach could enable earlier detection of complications so clinicians can intervene sooner and potentially reduce unplanned readmissions and improve recovery.

How similar studies have performed: Artificial intelligence approaches have shown promise in predicting clinical deterioration and readmission in surgical settings, but combining continuous smartwatch and smart ring biosignals with deep learning for readmission prediction is a relatively new and emerging application.

Eligibility criteria

Inclusion Criteria:

* Adults over 18 years of age;
* Hospitalization for medium and/or large elective surgery at HUGV;
* Conscious and oriented patients who have sufficient understanding to answer questionnaires and use wearable devices for the study;
* Have minimal skills in the use of wearable technologies;
* Patients who have agreed to participate voluntarily in the research by signing the Informed Consent Form (ICF).

Non-inclusion Criteria:

* Presence of tattoos or any other skin condition (skin pathologies or skin diseases such as vitiligo, lupus, and atopic dermatitis, among others) that affects the area of the wrist or finger where the wearable sensors are located;
* Presence of any type of sensitivity or allergic reaction, of any degree, to the materials of the wearables (smartwatch and smartring);
* Pregnant and lactating women;
* Participants with implantable cardiac devices, such as pacemakers, cardioverter defibrillators, and resynchronization devices;
* Participants in drugs abuse;

Exclusion Criteria:

* Severe medical conditions and decompensations prior to surgery;
* Patients who die before hospital discharge;
* Patients with an expected postoperative hospital stay of more than 10 days.

Where this trial is running

Manaus, Amazonas

• Getúlio Vargas University Hospital — Manaus, Amazonas, Brazil (Recruiting)

Study contacts

• Principal investigator: Robson Luís Oliveira de Amorim — Getúlio Vargas University Hospital
• Study coordinator: Robson Luís Oliveira de Amorim, PhD

Conditions Patient Readmission; Hospital Readmission; Postoperative Complications; Artificial Intelligence; Machine Learning; Wearable Electronic Devices; Physiological Monitoring; Sleep Disorders