The study plans to monitor around 300 people from different hospitals with COPD for a period of 3 months after they are discharged from the hospital using a smartphone app and a Fitbit device. This device can passively track certain health metrics; this way the research team can research whether it is possible to identify the early warning signs of a decline in health by using these ongoing measurements of vital signs and symptoms. This could allow doctors to intervene early and potentially prevent further deterioration in health decline and hospital admission altogether. The study seeks to investigate how similar these physiological measurements are when collected in the real world rather than just in the hospital setting, and what influence environmental factors have on a patient's health and experience of their condition.
Study Type
OBSERVATIONAL
Enrollment
300
Observational
Stoke Mandeville Hospital
Aylesbury, United Kingdom
RECRUITINGRoyal Sussex County Hospital
Brighton, United Kingdom
RECRUITINGChelsea and Westminster Hospital NHS Foundation Trust
London, United Kingdom
RECRUITINGNottingham University Hospitals
Nottingham, United Kingdom
RECRUITINGTo assess the volume and quality of the data collected in terms of:
Total number of subjects: number of subjects who completed the on boarding stage of the study
Time frame: 3 months
To assess the volume and quality of the data collected in terms of:
Dropout rate: proportion of subjects who dropped out (withdrew from study or stopped using the app/connected device prior to the off-boarding process or had to exit the trial due to deterioration)
Time frame: 3 months
To assess the volume and quality of the data collected in terms of:
Median time to dropout, where time to dropout is the number of days between subject's enrolment date and drop-out date (either date of withdrawal/exit from the study or date when subject stopped using the app and connected device entirely)
Time frame: 3 months
To assess the volume and quality of the data collected in terms of:
Number of participants who provided passive measurements for minimum of 50% of the study period
Time frame: 3 months
To assess the volume and quality of the data collected in terms of:
Number of participants who provided minimum of 33% of measures requiring active input from user
Time frame: 3 months
To assess the volume and quality of the data collected in terms of:
Data completeness: proportion of missing and total number of data points
Time frame: 3 months
To assess the volume and quality of the data collected in terms of:
Data consistency across sources: where similar information is recorded in multiple modules (e.g., breathlessness scale and symptom tracker), proportion of records in which consistent answers were provided will be reported
Time frame: 3 months
To assess the volume and quality of the data collected in terms of:
Proportion of data within admissible value range, where admissible value range will be determined based on literature or clinical guidance
Time frame: 3 months
To assess the volume and quality of the data collected in terms of:
Similarity of collected data distribution to expected data distribution, where expected data distribution will be determined based on literature and similarity of the two distributions evaluated by a suitable statistical technique (e.g., Kolmogorov-Smirnov test)
Time frame: 3 months
To ascertain whether marked physiological events can be detected using smartphone and connected device sensors in a remote setting.
Using clinical endpoints such as exacerbation events and readmission to predict exacerbation episodes
Time frame: 3 months
To assess the relationship between patient-generated data gathered from smartphone and connected devices and conventional clinical measures at point of readmission.
The prediction of physiological measures at readmission (e.g., pulse rate, respiratory rate, pH, FBC, CRP, and CXR appearance) can be addressed as a regression task and evaluated with metrics such as root mean squared error (RMSE).
Time frame: 3 months
To assess the relationship between patient-generated data gathered from smartphone and connected devices and patient reported functional status.
The prediction of reported outcome measures (CAT; EQ-5D; SGRQ-C) can also be addressed as a regression task, evaluated with RMSE, as detailed above.
Time frame: 3 months
To assess the change in passively generated data at the time of further community intervention (HCP review and/or prescription for corticosteroids or antibiotics).
Acquired physiological and environmental data before and after community intervention will be compared using appropriate statistical tests to identify whether effects of these interventions were detectable in the acquired physiological data. We will also attempt to use machine learning models for the classification tasks of predicting corticosteroids, antibiotics, or HCP review outcome) using the physiological and environmental data in the time window prior to the specified community intervention outcome.
Time frame: 3 months
To evaluate the usability and acceptability of patient-generated data gathered from smartphone and connected devices in a remote setting in patients with COPD.
Summary of the outcomes measured in the HCP mHealth app usability questionnaire (MAUQ) and other app analytics will be generated using standard summary statistics measures (mean/median, standard deviation, confidence intervals). This data will be assessed in relation to app and usage analytics such as compliance rate, drop-out rate, and device wear time.
This platform is for informational purposes only and does not constitute medical advice. Always consult a qualified healthcare professional.
Time frame: 3 months