Mobile Health Technologies for Continuous Monitoring of Cancer Patients in Palliative Care Aiming to Predict Health Status Deterioration: A Feasibility Study

Abstract

Background: Unplanned readmissions or emergency visits (EVs) after discharge from hospital are frequent in patients in palliative care. Strategies to anticipate and prevent rapid deterioration of health are needed. Objective: Assessing feasibility and predictive ability of remote monitoring using wearables. Design: Prospective observational feasibility study in a single center. Setting/Subjects: Thirty cancer patients with an estimated life expectancy of >8 weeks to <12 months, aged >18 years and being discharged from inpatient to outpatient care were included. Measurements: Patients were provided with a smartphone, including the preinstalled ''Activity Monitoring'' application and a sensor-equipped bracelet. Follow-up was 12 weeks. Both devices recorded several features (e.g., vital signs). Visual analog scale (VAS) for pain and distress was reported once daily and European Organization for Research and Treatment of Cancer Quality of Life Questionnaire-Core 30 (EORTC QLQ-C30) once weekly. Statistical methods were applied to explore relationship between sensor data, self-reports, and EVs or readmissions or death. Results: Between February 2017 and May 2018, 30 patients were included. Twenty-five of 30 participants (83%) completed 12 weeks of follow-up. On average, bracelet was worn on 53% and smartphone on 85% of study days. Completion rate of daily digital questionnaires for subjective ratings was 73%. Eight unplanned hospital readmissions occurred. Ratings of pain, distress, and QLQ-C30 scores were not associated with readmission, whereas resting heart rate, resting heart rate variability, as well as speed of steps differed significantly in patients with and without readmission. Conclusions: Monitoring of palliative cancer patients using wearables is feasible. First results indicate that mobile health features might be promising biomarkers to predict unplanned readmissions.

Harnessing consumer smartphone and wearable sensors for clinical cancer research

“…In a sample of 10 gynecologic cancer patients, Fitbit and smartphone sensor data were actively monitored for anomalies; this led the research team to identify a patient with severe nausea and vomiting whose symptoms were then able to be managed over the phone as well as a patient who took fewer steps and spent less time away from home prior to an emergency department visit 46 . Finally, a recent study collected both smartphone and wearable sensor data as well as patient-reported pain and distress ratings and quality of life measures in 25 palliative cancer patients 47 . Mobile sensor data were significantly correlated with pain and distress ratings.…”

A Review of AI and Data Science Support for Cancer Management

“…PROs were added to this table when used in 3 or more papers. In this review, the most commonly used validated PRO questionnaire for collecting HRQoL data is the EORTC QLQ-C30, used in 28% of the 50 included papers [10][11][12][13][14][15][16][17][18][19][20][21][22]. It is a European widely used core generic questionnaire for cancer patients, and comprises most dimensions of wellbeing as described by Linton et al [6], excluding spiritual wellbeing and personal circumstances.…”

“…Several PRO questionnaires were used to collect data regarding patient reported symptoms. The EORTC QLQ-cancer specific modules were used to collect cancer-specific patientreported symptoms in 7 out of 50 included papers [6,7,13,14,17,20,21]. These modules are specific to tumour site, treatment modality, or a HRQoL dimension, to be administered in addition to the core questionnaire (EORTC QLQ-C30) [39].…”

“…Data from the smartband was being collected 3 to 7 days before surgery, during the patients' stay at the hospital and approximately 2 weeks after their release from the hospital. Finally, Pavic et al [21] described a system where patients were provided with a smartphone with preinstalled ''Activity Monitoring'' app and a commercial sensor-equipped bracelet (Everion). The app recorded motion sensor data (GPS, acceleration), phone call statistics (anonymized), and selected vital signs measured by the bracelet, which was automatically connected to the smartphone via Bluetooth Low Energy.…”

“…As can be seen from the above examples, the majority of systems measure the PA of patients. This information is collected based on step count, step speed, which are collected by accelerometers [11,21,43] and burned calories [11]. Solutions that are more complex measure heart rate, oxygen saturation [43] or weight, blood pressure, oxygen saturation, pulse and temperature [45].…”