Accuracy of Smartwatches in Measuring Oxygen Levels in Patients With Pulmonary Hypertension: A Pilot Study

The goal of this pilot study is assess the accuracy of SpO2 measurements in smartwatches when compared to arterial oxygen levels and pulse oximeter readings at rest, following exercise and at simulated high altitude with the use of a hypoxic challenge test.

Patient recruitment:

Patients will be recruited prospectively from the Scottish Pulmonary Vascular Unit at the Golden Jubilee National Hospital, Glasgow, Scotland.

Patients attending outpatient clinic will be screened for eligibility. Additionally, a study poster will be displayed in the clinic waiting area. If eligible, patients will be asked if they would like to learn about the study. After discussing the study overview in person, the patient will be offered the patient information sheet to take away with them and advised that if they are interested in taking part or have any further questions about the study, they can either reply using the opt in form provided or by phoning the study team. Those patients who opt in to the trial will be contacted to provide them with a date to attend for a study visit.

On the day of the study visit, prior to commencing any study activities, participants will see a different doctor from the research team who will clarify their full understanding of the study and will give them the opportunity to complete the written study consent form.

Procedure:

During a study visit, participants will have an arterial line inserted into their radial artery by an experienced intensive care unit doctor to allow easy access to arterial blood sampling. Each paired sample will include: smartwatch oxygen saturations, arterial oxygen saturations, finger probe pulse oximeter measurements taken at the same time. Paired samples will be taken at baseline (while the patient is in a seated, rested state), post-exercise in the recovery period after a six-minute walk test and in simulated high altitude during a hypoxic challenge.

The six minute walk test will comprise shuttles of 25 metres continuously for 6 minutes. 10 paired samples will be taken during the recovery period after a six-minute walk test. The hypoxic challenge will be undertaken by asking the patient to breathe through a 40% Venturi mask driven by 100% nitrogen at a flow rate of 10L/min which will deliver an inhaled oxygen level of 15%. If SpO2 falls below 85% for over 30 seconds the test will be stopped and if required, inhaled oxygen will be delivered to correct hypoxia. The participant will breathe the air/nitrogen mixture for 20 minutes unless the procedure is aborted due to hypoxia prior to that. 10 paired samples will be taken during the hypoxic challenge test. Following completion the arterial line will be removed and pressure will be applied until any bleeding stops. Participants will be monitored for 30 minutes prior to them leaving.

Data collection/management:

Each study participant will be assigned a study number which will be used on all study documentation to avoid using patient identifiable information. All data will be stored on an NHS password encrypted computer. Data will be collected on pseudonymised data capture sheets with a study ID number and all research measurements will be transferred to an Excel spreadsheet for analysis.

Sample size:

The study is a cross sectional study of validity requiring one study visit per participant. Regarding sample size, this is a pilot study to gain initial estimates of the mean bias and limits of agreement of smartwatch measurements compared to the gold standard arterial oxygen levels. The planned sample size of 20 is not based on a sample size calculation, and is considered sufficient to obtain initial estimates, and to allow planning of larger studies if required. The investigators will aim to recruit 20 patients and will recruit up to a maximum of 25 in total if required in the case of incomplete data sets.

Statistical analysis:

The bias of smartwatches and standard pulse oximeter oxygen saturation measurements relative to the gold standard, arterial blood gas, will be compared using Bland-Altman analysis.

Plan for missing data:

In the event that an arterial sample cannot be analysed, the smartwatch and fingerprobe pulse oximeter samples will still be noted and compared. The investigators will aim to obtain 10 complete paired samples per patient after exercise and also 10 full sets during the hypoxic challenge. The investigators will aim to recruit 20 patients but will recruit up to 25 patients as required to achieve complete data sets. In the event that a smartwatch reading is not obtained during a paired sample, this will be documented as 'no measurement obtained'. This will still represent a complete paired sample as inability to obtain a measurement is useful to note as an outcome. In the event that the fingerprobe pulse oximeter is unable to obtain a reading the data will be documented as 'no measurement obtained' but will not require any increased requirement for recruitment as long as smartwatch and arterial gas measurements have been acquired as a paired sample.