WEARABLES: Wearable Technology and Machine Learning for Early Detection and Risk Assessment of Unacceptable Toxicities in a Paediatric Oncology Cohort
Trial Snapshot
- Phase
- Not Applicable
- Status
- Recruiting
- Enrollment
- 150
- Locations
- 1
- Primary Endpoint
- Changes in cardiac electrical activity patterns on Electrocardiogram (ECG)
Study Overview
Brief Summary
Data collection study to establish a predictive model of infection observed during childhood cancer therapy using data captured by wearable technology.
Detailed Description
Rationale: Development in treatment for childhood cancers has improved remarkably with the 5-year survival rate now exceeding 80% in developed countries. However, these treatments are not without their adverse effects. The international Childhood Cancer Survivor study revealed that 62.3% of survivors had at least one chronic health condition and 27.5% had a severe or life-threatening condition as a direct result of their cancer treatment (DOI: 10.1056/NEJMsa060185). One of the adverse events experienced by 90% of children treated for cancer is infection. Septic shock, the most severe of infection outcomes, is characterized by life-threatening organ dysfunction, is the most and carriers a mortality rate of 41 to 46% (DOI: 10.1016/j.jped.2023.01.001). Beyond mortality, delayed first antibiotic administration (> 1 hour from fever onset >38 degrees) is associated with intensive care admissions, prolonged hospital stays, and adverse outcomes. Fluctuations in physiology can precede fever onset by 72 hours in patients with infection. This may provide a window for early detection of infection via wearable technology. The WEARABLES study will combine wearable technology with machine learning to develop an infection prediction model to allow earlier detection and reduce the suffering of children with cancer.
Trial Design: This is a non-interventional silent pilot trial to establish a predictive model for infection observed during childhood cancer therapy using data passively captured via wearables.
The study will be conducted in patients (5-18 years) with a new cancer diagnosis, currently receiving treatment at The Royal Children's Hospital, and have access to an iPhone (either themselves as an adolescent and young adult or via their parents/guardian). Once consented the wearable device will be paired to the patients or parent/guardians phone, and the WEARABLES app will be downloaded onto the phone. Once the device has been set up correctly, the wearable device will collect a range of vital signs for the duration of the study (4 weeks), and a weekly survey will be sent to check for symptoms and/or hospital admissions for infection. At the end of the 4 weeks, participants will receive a final survey to evaluate the feasibility of using a wearable device for toxicity detection. No further involvement will be asked of participants for this pilot trial. All data collected will be utilized to develop a machine learning model for sepsis/infection before being prospectively validated in a second trial.
Study Design
- Study Type
- Observational
- Observational Model
- Case Only
- Time Perspective
- Prospective
Eligibility Criteria
- Ages
- 5 Years to 18 Years (Child, Adult)
- Sex
- All
- Accepts Healthy Volunteers
- No
Inclusion Criteria
- •Paediatric, adolescent or young adult diagnosis of cancer AND receiving therapy placing them at risk of infection
- •Receiving cancer treatment at The Royal Children's Hospital
- •Patients aged 5-18 years at time of the eligibility screening
- •If aged < 16 years, parent or guardian able to provide consent
- •iPhone 8 or later (iOS must be up to date/updated at time of enrolment)
- •At least 10MB of iPhone storage for WEARABLES app and data collection.
- •Willing and able to wear a wearable device for a period of 4 weeks (during waking hours).
- •Consent to data being shared to the WEARABLES app (owned by the research team).
Exclusion Criteria
- •<5 years of age.
- •<16 years of age without guardian or parent consent.
- •Aged 16-18 and unable to provide consent.
- •Participant did not consent to wearing Apple Watch for a period of 4 weeks.
Outcomes
Primary Outcomes
Changes in cardiac electrical activity patterns on Electrocardiogram (ECG)
Time Frame: Baseline (Day 1), Day 8, Day 15, Day 22, Day 29
ECG data will be collected once per week over a 4-week period for each participant to identify changes in cardiac electrical activity that may be associated with early infection. These data will be used as input features for a machine learning model aimed at predicting infection risk in children receiving cancer treatment.
Changes in physical activity
Time Frame: Baseline (Day 1) and every 15 minutes until Study completion at Day 29
Exercise data will be collected every 15 minutes over a 4-week period for each participant to identify changes that may correlate with early signs of infection. These data points will be used as input features for a machine learning model aimed at predicting infection risk in children receiving cancer treatment.
Changes in heart rate
Time Frame: Baseline (Day 1) and every 15 minutes until Study completion at Day 29
Heart rate data will be collected every 15 minutes over a 4-week period for each participant to identify changes that may be indicative of infection. These data points will be used as input features for a machine learning model aimed at predicting infection risk in children receiving cancer treatment.
Changes in heart rhythm
Time Frame: Baseline (Day 1) and every 15 minutes until Study completion at Day 29
Detection of irregular heart rhythms that may reflect early signs of infection. These data points will be used as input features for a machine learning model aimed at predicting infection risk in children receiving cancer treatment.
Changes in blood oxygen saturation
Time Frame: Baseline (Day 1) and every 15 minutes until Study completion at Day 29
Blood oxygen saturation data will be collected every 15 minutes over a 4-week period for each participant to identify changes that may be indicative of infection. These data points will be used as input features for a machine learning model aimed at predicting infection risk in children receiving cancer treatment.
Changes in respiratory rate
Time Frame: Baseline (Day 1) and every 15 minutes until Study completion at Day 29
Respiratory rate data will be collected every 15 minutes over a 4-week period for each participant to identify changes that may be indicative of infection. These data points will be used as input features for a machine learning model aimed at predicting infection risk in children receiving cancer treatment.
Infection-Related Hospital Admission
Time Frame: Baseline (Day 1), Day 8, Day 15, Day 22, Day 29
An infection survey will be completed by patients once per week over a 4-week period to identify confirmed episodes of infections requiring admission to hospital. This data will be used to determine when patients are controls (non-infectious) vs cases (infectious), and used as an input feature for a machine learning model.
Secondary Outcomes
- The acceptability of using (and not using) wearable devices in children receiving cancer therapies(Day 29)