Demographics and Covid-19 outcomes
A total of 19 patients were approached of whom 8 declined recruitment and 3 were found to be ineligible on full eligibility review. Reasons for declining participation included the perceived complexity of the system, not wanting additional attachments (patient had nephrostomy/colostomy), and wanting to focus on recovery from Covid-19/cancer treatment.
Eight patients (6 female, 2 male) with a median age of 62 years (range 40–73) gave informed consent and became study participants (Fig. 3). Two patients completed 21 days on study. Two patients were admitted to hospital due to a drop in the oxygen saturations; in both cases the devices were used to highlight falling oxygen saturations to the medical team, prompting admission. Four patients stopped wearing the sensors between 2–15 days after commencement for various reasons which are explored more in the semi-structured interview section of the results. No patients died during the study period.
Participant demographics, oncology diagnosis, comorbidities and study outcomes are summarised in Table 1.
Table 1
Patient characteristics and study outcomes
Patient No | Gender | Cancer type | Age | Days on study | Comments |
001 | Male | Lung | 68 | D8 – hospital admission | Hospital Admission – triggered by low oxygen saturations |
002 | Female | Ovarian | 57 | D2 | Unable to work the technology and increasingly anxious. Patient and CI decision |
003 | Female | Colorectal | 47 | D8 | Patient decision – felt overwhelmed with diagnosis, C19 and study |
004 | Male | Lung | 68 | Completed 21 days | Completed 21 days on study |
005 | Female | Lung | 73 | D5 – hospital admission | Hospital Admission – triggered by low oxygen saturations |
006 | Female | Breast | 40 | D15 | Patient decision – sensors no longer tolerable, G2 skin reaction |
007 | Female | Ovarian | 67 | Completed 21 days | Completed 21 days on study |
008 | Female | Breast | 47 | Day 14 | Patient decision - experiencing side effects from cancer treatment requiring hospital visits |
Qualitative interview results
Seven out of eight participants underwent semi-structured interviews. Three main themes were identified from the data: 1) Thoughts on the device itself; 2) Implementation into clinical practice; and 3) Patient impact. Each of these themes included subthemes which are presented in Fig. 5.
Thoughts on the device itself
Overall impressions of the device were mixed. Some patients liked the idea and did not have any concerns about it, others felt the device was similar to other wearable health technology already commercially available and was therefore not needed. Participants were generally not concerned about the appearance of the devices although one participant felt the sensors could have been smaller.
Patients discussed a number of issues related to usability. Some participants found the Gateway™ tablet easy to use once they had familiarised themselves with it. Others found the device complex and often had problems when required to change the sensors. Patients found the complexity of the system and the difficulties they experienced frustrating and often required support from the hospital team or family members to help them. Some patients felt their lack of familiarity with technology made it difficult for them to use.
‘…you know new devices with me don't…they don't mix, it takes a bit to cotton on [understand].’ 05
Generally, patients found the sensors comfortable to wear and didn’t have any concerns about wearing them for the duration of the study. A small number of patients described minor skin irritation due to wearing the sensors. Many of the patients had problems with adhesion of the sensors and required additional tape to keep the sensors in place. This was a particular problem for some during the night and caused the device to lose connectivity.
‘No, I just think it was the night-time, it just kept on falling off. And obviously you’re in bed, you’re not going to get up and start messing about, you just stick it back on the best as you can, and you’ve lost the connection by then’ 04
Implementation into clinical practice
There were a number of factors which patients felt influenced the amount of time they stayed on trial. Finding solutions to these problems would encourage routine engagement in the future. Some patients were admitted to hospital or felt unwell and stopped using the device at that point and one patient withdrew early due to skin sensitivity. Other patients reported feeling overwhelmed with everything going on and the complexity of the device added to this feeling.
‘I should have said no right from the off because I’d only just been diagnosed. It was my first chemo, I had COVID, I had appendicitis. I shouldn’t have really agreed to it.’ 03
Some of the specific benefits of use during cancer treatment mentioned by participants included increased monitoring, identification of problems, increased communication with clinical team, highlighting problems the patient may have dismissed, and decreasing anxiety.
‘It would make you worry less, because you’d know someone would be, like, looking after you when they weren’t really here.’ 06
‘just if there’s somebody else monitoring it and, like you say, picked it up, something that you might find I’d dismissed and they might say, no, that’s a bit more serious.’ 03
Participants felt that the demographics of the patient group and the complexity of the device would impact engagement.
‘if it's the old…then I would think you're on a loser’ 01
‘and I find that if people can't figure it out in five or ten minutes, then they lose interest and they don't proceed’ 01
Patients made a number of suggestions as to how the training could be improved. Some patients felt that they needed more time during the training session to allow them to fully understand how to use the device. Others felt more of a live ‘hands on’ demo would have been beneficial or the provision of very simple step by step instructions.
‘You know, not like the proper instructions that you get given, just like a simple version, like, press this button, this is what you do.’ 06
Patient impact
The majority of patients did not feel wearing the device impacted their daily activities. Some patients completely removed the device for showering whereas others left the sensors attached to their body, neither of which caused a problem. Some patients found the sensors disrupted their sleep, either because they were uncomfortable to wear or the light disturbed them.
‘A little bit awkward at night because, you know, you toss and turn in bed. You know, you, sort of, leant on it or you could feel it’ Patient 4
‘At night…because they light up, I had to cover them.’ Patient 6
Patients discussed how they had benefited from taking part in the trial. For one patient, information recorded by the sensor was given to medical staff when the patient felt unwell. For another patient, wearing the device had alerted them to low oxygen levels allowing them to seek medical advice.
‘So I think it was through that device, that's how we found out my oxygen levels were low.’ Patient 5
Some patients felt being monitored gave them an extra sense of security and reassurance and meant they didn’t need to worry about their health. Others mentioned they were more aware of things because it was being monitored but overall it didn’t make them feel anxious.
‘Oh yeah, yeah, it did, it put our minds at rest, you know, because chemo, because the third day I had a rough day. I only have one rough day normally, and, you know, I think it would put our minds at rest.’ Patient 4
‘It made you a bit more, like, conscious of, like, what…do you know what I mean, if you’ve been for a walk or something, you could be like, oh, look what my heart rate was when I was walking.’ Patient 6
Physiological data analysis
Physiological data were analysed retrospectively and are presented as a whole elsewhere (19). By way of illustration, data from patient 1 who was readmitted to hospital on day 8, is presented in Fig. 6. The early warning score (EWS) was calculated with temperature excluded as temperature for all patients fluctuated between 30 and 35 degrees within a fairly short space of time. This was not felt to be physiological and therefore likely due to sensor issues with adhesion. A longer-term trend in increasing respiratory and heart rate can be seen from Day 6, two days prior to the decrease in oxygen saturations on day 8 that triggered the admission.
In addition, the raw data were used as part of an exploratory analysis of a novel algorithm to identify early signs of deterioration. This work showed that trends in a patient’s vital signs over time, collected via wearable sensors, can be potentially informative if incorporated in early warning systems. To analyse the trajectory of a patient’s condition, the heart rate and respiration rate data were segmented into three-hour long sections and compared using the Dynamic Time Warping (DTW) distance measure. The most dissimilar epochs exhibited the greatest pairwise distance to all other epochs, and were characterised as outliers. Of the 1% greatest outliers, 2 of 3 epochs corresponded to patients who were readmitted for treatment, showing the potential for this method to effectively identify abnormal physiological signals (20).