Participant Characteristics
A total of 124 interviews with 99 participants were conducted across the three sites. Forty interviews were conducted at the 3-month timepoint (UK, n=15; Spain, n=15; The Netherlands, n=10), 42 at the 12-month timepoint (UK, n=16; Spain, n=16; The Netherlands, n=10) and 42 at the 24-month timepoint (UK, n=15; Spain, n=16; The Netherlands, n=11). Seventeen participants took part in an interview at two time points; four participants interviewed across all three. Participant characteristics by timepoint are displayed in Table 3.
Themes
This study aimed to explore the subjective experience of long-term engagement with RMTs over a two-year follow-up period. We present our results under five themes: 1) research-related factors 2) utility of RMTs for self-management, 3) technology-related factors, 4) clinical-related factors, and 5) system amendments and additions.
Research-related factors
When considering initial motivations for engaging with an RMT study, contributing towards novel research findings was the most prevalent reason for taking part. Across all timepoints, research team support was also a key facilitator of sustained engagement with the study.
Altruism and academia
Taking part in the study was an opportunity to use personal experiences of depression to help others, to advance scientific understanding, and to ‘give back’ to the system:
“I’ve suffered with depression the whole of my adult life, I’ve obviously had a lot out of the system. If I can do anything to put back, do you see what I mean—I will.”
P30, 24 months, UK
Taking part for “the future, for the people who come after me” (P8, 3 months, Spain) was a strong theme that arose in all sites when discussing reasons for enrolling in the research study. Altruistic motivations continued across later timepoints, regardless of whether participants felt they had experienced any direct benefits:
“I am actually quite proud to say that I am doing this as part of research. Some people will ask me what it is [the wearable], and I say well it is good if more people get to know about it. And for the long-term benefits, might not be for me but for other people, because it might show.”
P18, 12 months, UK
With regards to the RMT aspect of the study, some mentioned that it “piqued my interest” (P37, 24 months, UK) and “I was very intrigued by a study that kind of has consistent monitoring” (P39, 24 months, UK). However, many participants signed up with limited knowledge of the study procedure, or of the use of RMTs for healthcare monitoring. Thus, lack of prior understanding of RMTs was not a barrier to initial engagement.
Privacy did not present as a barrier to participants upon entering the study or throughout their participation. A key reason for this was that the research was being undertaken in a clinical, academic setting. In the Spanish cohort, one participant viewed the study as parallel to their clinical care:
“…it's not data about, about privacy, things about you, no, it's related to a medical condition, isn't it? A case of depression, that's what it's about. So if they ask you for medical data, well, it's normal.”
P25, 24 months, Spain
Any initial privacy or data security concerns had been largely alleviated by the 3-month point through conversations with the research team. At later timepoints, privacy was not discussed frequently.
Research team support
Support from the research team was a facilitator to continued engagement with the RMTs. This was primarily practical; at 3-months, the research team provided support on how to use the devices and study apps, which was often imperative to successful enrolment into the study:
“I tried it once [the wearable] and wasn't able to... to... put it on the phone. If it hadn't been for [researcher name]'s help I wouldn't have made it.”
P1, 3 months, Spain
The need for practical support remained a key theme at 12-months, this time concerning technological malfunctions. Ability to contact the research team through various methods, and receiving a timely reply, was important. Some felt comfortable with initiating support themselves: “I didn't need that much contact personally, I could get in contact easily, if it were necessary” (P21, 24m, The Netherlands). Others wanted more contact, e.g., more points of researcher-initiated contact, or specific contact from specialists. At-hand support was essential for continued participation:
“I think it is really important to have the practical support ‘cause you don’t want to be offline or not working for long than is necessary. Otherwise it goes against the purpose of the study really”
P18, 12 months, UK
There was a consensus at all timepoints that the research team were approachable, patient and reassuring, helping to alleviate technological concerns.
The research team also provided emotional support. Some participants sought comfort in the knowledge that they were being monitored as part of a study: “I liked it a lot because, jeez knowing, I felt safe, you know? Because knowing that you were there…” (P25, 24 months, Spain). Others had specific examples of receiving mental health support from the research team. One participant in the British cohort received direct signposting, which was noted in both their 12-month and 24-month interview as a significant part of their study experience:
“because of the letter from [researcher] to the GP clinic I was able to get an immediate referral, and the problem is if you're the system it’s great, if you're not in the system it’s difficult to get in. I couldn't have done it on my own.”
P27, 12 months, UK
Benefits of RMTs for self-management
Despite primarily engaging with the study for altruistic reasons, many participants experienced unexpected benefits of using RMTs for symptom monitoring during their time in the study. These comprised symptom awareness and communication, both of which was integrated into depression self-management.
Symptom monitoring and awareness
Across all three timepoints, the most frequently reported benefit was an increase in awareness of symptoms. Monitoring various factors related to depression, e.g., mood, sleep, exercise, increased self-reflection, and the ability to identify patterns. For example, having access to objective sleep data provided clarification and reassurance:
“I loved that [the wearable data], I found that so reassuring to just relax, of course you’ve slept and then you go ok, the next time you’re lying in bed you go I’m not ever gonna sleep again but actually you have, you’ve seen that you do I think that’s brilliant, really reassuring.”
P14, 3 months, UK
Though the app did not provide feedback on symptom scores, many felt that the act of answering questionnaires prompted them to analyse how they had been feeling:
“I'm more aware of it, the questions on the questionnaire, especially those that ask how I'm feeling right now raise my awareness, I feel quite average or, I'm feeling not great, sometimes you ignore these things. And if you can take more time to think about these things... maybe I need to meditate more, I really feel self-conscious...”
P10, 3 months, The Netherlands
For some, answering the questionnaires and viewing the Fitbit data simply provided an understanding of their experience of depression: “I have noticed that my answers have gotten more positive throughout the year” (P22, 24 months, The Netherlands). For others, these data directly motivated behaviour change. At 3 months, discussion focused on the motivational effects of the Fitbit data; participants felt encouraged to complete their daily step count or achieve target physical activity ‘badges’. Towards the later timepoints, these data came to act as prompts for self-care, for example increased exercise or relaxation.
“…wearing a watch and knowing that my activity matters, you know? I mean, like the steps I take have a direct effect on my health, both physical and mental, all my activity makes me more aware of it, more conscious of it and it has also been like a driving force for me to put my batteries in sport or stress management… a habit forever, so I do not want to do without it”
P26, 24 months, Spain
This became especially apparent during the 24-month interviews, where the Fitbit data were used to monitor sleep and mood symptom changes during the COVID-19 pandemic. Disruption to usual routines during this time allowed some to reflect more than ever on the benefit of monitoring exercise:
“I knew in theory, exercising and getting out and so on was good for your mental health, but over Covid, the monitor helped, and the benefit would have been even better. I think I might have been worse during Covid without it.”
P36, 24 months, UK
Communication
At each timepoint, RMT data was also used for communicating personal experiences to others. Participants used their increased understanding of their depression to inform significant others: “For the first time it kind of occurred to me to let me partner know when I could feel it was starting… so if you see my behaviour change or I’m unresponsive this is why” (P39, 24 months, UK).
Having access to the Fitbit data also facilitated joint decision making, both for immediate symptom management and longer-term strategies:
“There are also days that I don’t reach 5000 steps, which will make me think oh I haven't done that many today… my spouse will say that too, go for another walk.”
(P2, 3 months, The Netherlands)
Overall value and utility
There was a consensus throughout that the benefits of participating in the study outweighed the costs, of which there were relatively few. Many had not envisioned any personal benefits when enrolling, as they were aware that they would not receive personalised outcomes, however had been pleasantly surprised by the integration of RMT data into their depression self-management, as early as the 3-month timepoint:
“I think its empowering to know more about myself to understand more so I think once I can see more what the data is from collecting from data when the other apps are working and being able to see what the data is and notice any correlations then I think that will be really valuable.”
P12, 3 months, UK
Technology-related factors
Experience of the technology used in the study (smartphone apps and Fitbit) was the most widely cited theme across all sites. This covered the convenience of integrating the RMTs into daily life, the usability of the technology, technological malfunctions that occurred, and extent to which participants found the technologies intrusive.
Convenience
Using a mobile phone and wearing a watch was already an integral part of many participants’ daily routine. The Fitbit device, “it’s basically wearing a watch” (P7, 3 months, UK), collected data passively without the need to input information, and continual wear, syncing and charging was integrated into routine as early as the 3-month timepoint. Reminder messages across the system were useful in the process of longer-term integration.
One aspect that participants found more difficult to integrate into their routine was the app questionnaires. Timing of the questionnaires was often inconvenient, for example when at work, driving, or in social situations: “Obviously I’m less likely to stop my conversation to be like oh this questionnaire, because that’s a bit rude” (P4, 3 months, UK). Frequency of the ESM questionnaires was also too high from some: “it's impossible to have a routine with that. If you have a full-time job, it's always a bother” (P17, 24 months, The Netherlands). Participants were rarely able to change their routine to accommodate answering the questionnaires, which sometimes caused guilt. One participant in the Spanish cohort reflected on how work affected their ability to respond to app notifications over their two-year participation:
“At the beginning it was a bit difficult because I was working, then as I was on sick leave for two years, the truth is that I've been able to adapt quite well. And in the end, when I went back to work again, it was a bit difficult…”
P1, 24 months, Spain
Usability
For those who received a smartphone upon enrolment, a large technological barrier was the process of ‘re-learning’ a new operating system. This was described by some as ‘more difficult than anticipated’ (P3, 3 months, UK), particularly during the 3-month interviews, due to adapting to a new user interface and decreased connectivity with other devices. At 24-months, some had adjusted to using the new device, whereas others planned to swap back upon study completion:
“No, my only peeve was that I’m an Apple user and having this bloody awful Android phone, the first thing I shall do on April 1st is take my SIM card out of the Motorola thingy.”
P35, 24 months, UK
Technological malfunctions
Participants reported a range of technological malfunctions that impacted their participation in the study. Issues with the study apps were particularly prevalent during the 3-month interviews, owing to ongoing technological challenges during the early phases of the study. These included not receiving notifications, the apps crashing, the apps logging out, and difficulties with re-scanning QR codes. Participants sometimes had limited time or motivation to report issues to the team:
“I tried opening a questionnaire I wouldn’t be able to see it, I wouldn’t be able to do it and there was no way of saying this is happening or why this is happening so maybe I should have contacted you about it but I just kind of ignored it.”
P4, 3 months, UK
Issues with missing data persisted throughout the three timepoints. Participants were aware of times when the active app had been unable to submit completed data, or the passive app had ceased monitoring. Such malfunctions often led to anxiety or guilt that they were not ‘correctly’ participating: “Well, yes, when it didn't work, I became a bit nervous...” (P15, 3 months, Spain).
Participants also reported frequent missing data with the Fitbit, caused either by a syncing error or inaccurate recording. These issues caused some to question the integrity of the study: “It just didn't work and that's not what you expect from a research study” (P18, 24 months, The Netherlands).
A participant in the Spanish cohort reflected on how these technological malfunctions affected not only their ability to participate in the study, but also their experience of being able to use the resulting data:
“there is data that I have missed here, and of course I was analyzing it with me in important situations of how I was, and that I have missed them, for more than a month.”
(P32, 24 months, Spain)
Intrusiveness
Generally, the concept of remote monitoring, or the use of the technologies, was not regarded as intrusive. Rather, passive data collection was noted as a preferable method because “at some point you don't notice it. You don't notice that you're wearing it anymore” (P18, 24 months, The Netherlands).
One area that caused disruption, however, was the wearability of the Fitbit device. Several issues with the Fitbit strap were reported, including skin irritation, increased sweating and allergic reactions. Some had briefly chosen to remove the device while experiencing discomfort, others had purchased straps with alternative materials. At 12-months, many reported that their strap had broken, and by 24-months, some had had to apply for a full device replacement. One participant felt guilty when asking the research team for their device to be repaired:
“I know that the money allocated to research programs or projects is minimal, and of course, when the strap broke or the Fitbit wouldn't charge me and then I felt really bad because I thought "oh my God, now they have to change my Fitbit"”
P26, 24 months, Spain
Waiting for a replacement strap or device meant that participants were unable to continue to use the Fitbit for self-management:
“if I was going to continue and for the others who will be continuing, it will probably begin to happen more and more depending on how much people are actually exercising with them on. It only grows, that’s the problem, in my experience with the other Fitbit, that definitely happens.”
P3, 12 months, UK
Clinical-related factors
Participants were asked to reflect on whether and how they could see RMT data being used in a clinical setting. Discussions included the extent to which participants felt comfortable sharing the data, how they envisioned clinicians using the data, and how feasible this was in the current climate.
Views on data sharing
At the 12 and 24-month timepoints, participants were specifically asked to comment on data sharing with medical professionals. In general, allowing trusted clinicians to view RMT data alongside medical records was acceptable, or even essential: “let's say my whole history, my doctor already has it, if she has it more extensive, then all the better for me.” (P30, 24 months, Spain). There was some discrepancy over whether this data should automatically be available to clinicians or mediated by the patient. Some thought that medical professionals “would be in a better position to evaluate what they needed from it than me to decide that” (P32, 24 months, UK). Others worried about interpretation of the data without context:
“I suppose, [I would like to] understand what it is that is proposed to be shared, and if there’s something there that would not be appropriate at that time, because I don’t know what it is until I see it, then yes, I would like to have a choice… I would want to make sure that my health record reflects actuality rather than something that can be interpreted by people incorrectly”
P31, 24 months, UK
Clinical uses of RMT data
Participants suggested several ways in which they might expect RMT data to be beneficial in clinical care. These included i) allowing the clinician to view the ‘whole picture’ of individual experience, ii) allowing the clinician insight into new symptoms, iii) as a way for patients to report specific areas of concern, and finally iv) as a basis for making decisions about suitable treatment or care. Importantly, treatment decisions should be reached as a joint decision involving the clinician, the patient, and the data:
“I think they could actually look at the data that’s being produced, and that could assist them in helping me to come to another decision. Like, if I was deciding that I would like to move my medication down, but they’ve got the data that says, no you’re not…. but if it backs it up as well, so it can work both ways, so I think it does have those benefits”
P33, 24 months, UK
Sleep data was repeatedly cited as a data stream that would effect change in treatment. Participants from all sites gave examples of conversations had with their mental health clinicians. One participant in the British cohort also discussed their experience of integrating the sleep data into their sleep clinic appointments:
“It’s too expensive for the NHS to keep on doing [sleep tests]… I said, well, actually, I can show you any time in the last six months or so… an indication of when I’m sleeping… It helped them choose what exercises I needed to do and what therapy was required, so, yes, it was extremely helpful.”
P22, 12 months, UK
Presentation of objective sleep data was seen as helpful ‘proof’ of the participant’s recent experiences:
“You can tell your GP that you sleep terribly, but of course your GP can also think that you're just worried, but with the data it's a fact that you can prove, so that's nice, that you have concrete info… whether you worry or complain about it or not doesn't matter, the facts are there.”
P10, 12 months, The Netherlands
Current clinical utility of RMTs
While the potential for RMTs in clinical care was recognised, two key barriers to implementation were envisioned. First, the level of technological acceptance of medical professionals influenced views on the long-term utility of the data. Participants in the Spanish cohort, who were recruited through their clinical care, generally reported acceptance of the study from their clinicians: “even my psychiatrist here and in Barcelona had the same way of thinking and saw that this was very useful for me and encouraged me” (P9, 24 months, Spain).
Others described more negative experiences, often causing them to question the use of the data: “I thought it would be more relevant for my neurologist, but my neurologist wasn’t particularly interested when I told him about what I was doing in the study.” (P17, 12 months, UK).
Second, lack of funding, resources and time was perceived as a major roadblock to using RMT data in appointments. This was particularly apparent in the British cohort, with regards to the National Health Service (NHS). For the data to be monitored and reflected on, new procedures would need to be put in place:
“I would be amazed if there was sufficient funding for that… I don’t believe that the NHS have got the resources to have people monitoring this sort of stuff.”
P22, 12 months, UK
Given the perceived lack of resources to effectively use RMT data in the NHS, some considered how best to come to a compromise:
“I think realistically, if they had that [data] and I went to them with a problem, then I would like them to be able to use it at that point. But I don’t see it as something that they would be—so, for example, if I went to them with something and if somehow, it was a part of my NHS records, if they could access that, that might be helpful to them. But I don’t see them using it other than that really.”
P32, 24 months, UK
System amendments and additions
Participants discussed various changes or additions to the RMT system used in the study to further encourage long-term engagement. These comprised suggestions for questionnaire data collection and feedback.
Data collection
Across all sites and timepoints, the most prevalent suggestions for changes to the study design was the content of aRMT questionnaires. Participants felt that they were frequently being asked to complete the same questions, particularly within the ESM schedule, which often prompted them to provide the same answers, for example with regards to mood changes. This affected motivation:
“At first, I was more excited about it, but as time has passed, sometimes I don't feel much like answering since the same questions get repeated.”
P19, 12 months, Spain
Some also suggested the ability to postpone questionnaires if feeling too low to complete them, and the ability to provide contextual information. As early as the 3-month timepoint, some noted that external factors affecting their mood were not being monitored within the validated mood and self-esteem questionnaires: “I notice that when my home situation isn’t great, I also fill in the questionnaires less positively” (P5, 3 months, The Netherlands). On reflection, some would have liked to have given more information at certain points:
“The answers are very closed, so you can't really answer what you feel. You know? It's very... it's very up in the air”
P1, 24 months, Spain
Data feedback
When asked how they might wish to view their symptom data in future use, the majority felt this was best displayed visually, through in-app graphs. Many also expressed that this would need to be accompanied by a “human explanation for what those things mean” (P3, 12 months, UK). There was a discrepancy between when this data would be best received; some only expected to receive it at the end of the study, some felt that it would be more useful in real-time, whereas others were cautious that receiving data during periods of low mood would be detrimental:
“If I'm well I want to see it, if I'm unwell, no. If I was reporting that I was feeling suicidal I don't think I’d want to revisit it.”
P27, 24 months, UK
Further to this, some participants considered the potential for RMT data to provide feedback on symptom patterns and changes over time, correlations with other factors, and depressive relapse prediction. Specific examples included relationships between exercise and mood, sleep and mood, and mood and concentration: “At some point I had a burn out. I’m very curious as to how my ability to concentrate changed, and if that maybe shows on the THINC-it app” (P3, 24 months, The Netherlands).
It was generally accepted that having access to data of this nature would be useful for both self-management and integration into clinical care. Looking forward at the 24-month time point, one participant at the British site explained their hopes for the future of this field:
“I think trends are really quite important for me in managing what is going on… I think one of the things I am thinking would be good to come out of this is an ability to see patterns over time and then maybe being able to use that as a predictor or, I need to do some intervention here so that I don’t end up there again if that makes sense.”
(P30, 24 months, UK)