Feasibility of in-home monitoring for people with glaucoma: the I-TRAC mixed-methods study

Clinician decisions regarding patient suitability for glaucoma home monitoring

Across the scenarios, there were varying rates of agreement regarding the suitability of each patient scenario for home monitoring (Figure 1 and Table 3). Based on the levels of agreement defined (i.e. > 60% reporting a patient as suitable), only one patient, scenario 4, was deemed suitable for home monitoring. In the scenario, 61% (n = 30) of clinicians would refer the patient to use the iCare tonometer and 65% (n = 32) the home visual function app assessment. Scenario 4 reported Ms McEwen, a low-risk patient with normal tension glaucoma (NTG) and mild disease who had not progressed in 5 years.

FIGURE 1.

Clinician agreement on home monitoring using digital technology within patient scenarios.

Clinicians were asked to suggest optimal time frames for the frequency and duration of home monitoring (of both IOP and visual function) through open-ended questions in each scenario. Similar to the variation in responses to clinical scenarios, a wide spectrum of monitoring frequencies and durations for each scenario were suggested (Table 3). Despite this, greatest consensus appeared within low-risk scenarios (2 and 4) where for both IOP and visual function monitoring, participants reported opting for reduced frequency of monitoring (every 2–6 months) but over an increased duration (7–24 months). For high-risk scenarios (1 and 3), frequency and duration for visual function monitoring were consistent and similar to low-risk scenarios (2 and 4). However, in relation to IOP monitoring, participants reported increased frequency (every 1–7 days) but for a lower duration (1–6 months).

Explaining clinician glaucoma home monitoring decision-making

The justifications provided by the clinicians when reporting that Ms McEwen, scenario 4, would be suitable for home monitoring stated this was due to the nature of her stable disease and her low risk of progression. They felt that she could be safely monitored at home without high risk of missed progression, which could allow increased clinic capacity for individuals with advanced disease, ‘[f]reeing up capacity in the hospital eye service, allowing better use of resources and enabling better care of high-risk patients. Low-risk patients may prefer not having to come in the hospital’ (Consultant, 5–10 years of experience).

Among the participants who predicted that this patient would be unsuitable for home monitoring, the justification was that it may be a waste of resources due to the stable nature of her current condition. This was a contradiction with other participants who deemed her suitable due to the exact same reasoning, ‘stable for 5 years, waste of effort’ (Consultant, over 10 years of experience).

Table 4 summarises the justifications given by clinicians when deciding a patient described in the clinical scenarios would be unsuitable for home monitoring. When evaluating the rationale behind clinicians deeming the other scenarios unsuitable, the main concerns contradicted each other. However, many clinicians (n = 32) highlighted that they may be more hesitant to recommend home monitoring to high-risk individuals. They predicted that they would rather assess these patients in clinic to confidently determine their IOP and visual function. This was expressed to be important as it would avoid any discrepancies or impaired clinical judgement due to potentially ‘unreliable’ readings. They also felt that these advanced cases may require additional resources such as imaging, ‘[h]igh risk of further vision loss within lifetime – young, advanced VF defects bilaterally. Would prefer to see in clinic and discuss surgery at each visit’ [regarding scenario 1] (Optometrist, > 10 years of experience).

In addition to measuring clinical agreement within each clinical scenario regarding home monitoring of glaucoma patients, we also investigated agreement across scenarios to determine whether the disagreement in scenarios is related to the digital technology or to which patients should be monitored. There was a lack of consensus relating to which patients should be monitored using the tonometer, with 23 clinicians (47%) reporting that home monitoring of patients would be useful in at least three of the four clinical scenarios. This was also true for visual function, with 22 clinicians (45%) believing it to be useful for three out of the four scenarios.

Clinicians’ acceptability of home monitoring of glaucoma within NHS Care pathways

The fifth clinical scenario in the questionnaire described a clinical care model that combined home monitoring into the current hospital-based system. Regarding acceptability, no agreement was reached on this scenario, with just over half of clinicians (52%, n = 26) reporting that this model was acceptable and around one-third (37%, n = 18) reporting that it was unsuitable; 10% (n = 5) did not respond. Thematic content analysis of free-text responses in relation to this scenario resulted in the following seven themes: Resources, Patient Characteristics, Clinician Confidence in Home Monitoring, Perception of Risk, Wider Benefits, Accessibility, and Disease Suitability of Patient. Within each theme, anticipated advantages and disadvantages in relation to glaucoma home monitoring were identified, as shown in Table 5.

Resources

From a resource perspective, many clinicians believe that implementing home monitoring could increase clinic capacity for review of high-risk patients and believe that this model could be better than the current alternative of no or limited monitoring. There was also a belief that access to timely care via home monitoring could prevent irreversible severe visual loss, promoting patient health and well-being. However, many doubted the cost-effectiveness of glaucoma home monitoring. Some felt that this model could reduce clinic capacity for patients by the time this service was staffed in relation to the human resource needed to train patients and review the data collected.

Patient characteristics

Clinicians had concerns about patients’ cognitive, physical and mental capacity for home monitoring. While concerns were broadly shared about physical and cognitive capacity, opinions were divided in relation to patient anxiety; while some felt home monitoring could alleviate anxiety in relation to glaucoma progression, some felt it had potential to increase this. The majority reported concerns that home monitoring could lead to reduced patient compliance with regular testing and concerns about how this would be monitored. For the few, compliance could increase as regular feedback on disease status was anticipated to be motivational for patients.

Clinician confidence in home monitoring

We found that while there was some optimism about more frequent monitoring, home monitoring was largely anticipated to cause a decrease in clinician confidence in the quality of care delivered to patients. This was related to concerns about the reliability, standardisation, and compatibility of the devices, with and against hospital equipment, and concerns about relying on a narrower range of measures, as opposed to additional measures obtained in clinic (e.g. optical coherence tomography). There was some optimism, however, that more frequent monitoring of IOP and visual function, as would be possible with home monitoring, could detect progression sooner.

Perception of risk

While several clinicians reported more timely monitoring being a positive, there was a strong concern about the risks posed by the home monitoring of glaucoma. Risks raised reflected patient harms (e.g. missing glaucoma progression, resulting in loss of vision) and data and technology risks resulting from events such as failing equipment.

Wider benefits

A couple of wider benefits to home monitoring were raised. For example, the reduced travel arising from monitoring at home offers an environmental advantage. For some, home monitoring was expected to be more convenient for patients. Some also raised the advantage of being able to monitor patients who currently are not able to be monitored in clinic, such as those who are bedbound and/or residing in care homes.

Accessibility

Several disadvantages which could be potential barriers to accessing home monitoring were reported, namely language barriers and technology/internet access. These were expected to result in low uptake of home monitoring among some population groups.

Disease suitability of patient

Similar to the variation in responses to the clinical scenarios (1–4) there was little consensus in responses made in relation to a patient’s medical suitability for home monitoring. Most agreed that those at high risk of progression were unsuitable for home monitoring. Certain disease classifications (OHT and NTG) were frequently considered good candidates. The use of these technologies for glaucoma screening and phasing purposes was often frequently suggested as having potential in addition to or instead of regular monitoring.

Enthusiasm tempered with uncertainty

When asked how they felt about the I-TRAC home monitoring interventions for glaucoma patients, expert glaucoma clinicians were generally enthusiastic about the opportunities this intervention presented, described through terms such as ‘excited’, ‘delighted’, and ‘glad’.

I think it’s a great idea, I think we need to know this. We’ve got real problems with capacity in glaucoma and if we can have a trial that helps us better understand the patient acceptability, clinicians’ views on acceptability and the meaningfulness of the data and so on and so forth then it’s to be welcomed.

P002, Optometrist, Focus Group 1

These feelings were often reported in relation to home monitoring being a potential solution to national capacity problems across HES. However, enthusiasm was tempered with caution – as indicated by a number of conditional statements such as ‘I think it would be a great thing if it worked well and it was reliable’ (P008, Consultant, Focus Group 2) – demonstrating the connection between affective attitude and perceived effectiveness.

Concerns about negative affect for patients and staff

Expert glaucoma clinicians also reported that they felt many patients would be welcoming of this approach, highlighting how this could reduce patient stress and provide comfort, ‘some patients genuinely like writing the numbers down and they like to know what they were before and they get great comfort in knowing there are numbers written down that they have control over’ (P018, Consultant, Interview).

However, among the positive feelings, some clinicians reported feelings of anxiety and frustration as a risk of such interventions, which could be experienced by both staff and patients.

Digital data, I mean it does sound like this panacea and it’s wonderful, but actually when it doesn’t go quite so right it can be extremely difficult and frustrating, and I’m thinking about the age group of our patients, how hard it might be for them.

P004, Consultant, Focus Group 1

Ethical risks of remote and commercial data collection technologies

When asked about ethical issues this intervention may pose, many clinicians discussed aspects of data security. There were concerns about how to protect these systems from external threats, and the need for secure data exchange. Many felt that NHS systems of governance for IT should be adequate to adopt this; however, one clinician raised the need to consider how commercial entities also manage data governance:

[O]ne thing we’ve not necessarily quite touched on is the link with the commercial sector. So OKKO Health are a commercial sector organisation, whoever’s come up with [home visual field intervention] presumably is as well. So who owns the data, who’s responsible for the governance of the data, all of that.

P002, Optometrist, Focus Group 1

Also related to the commercial side were concerns about what would happen to data should a private enterprise withdraw from service, posing the risk that patient data could be lost and no longer governed by NHS data governance. There were concerns that these data could then be misused. There were also some concerns about safety, with calls for regular data auditing being required to establish the safety of this service. One issue raised by a number of clinicians connected to data control is the risk that patients may allow the device to be used by others. The potential risks from this include poor treatment decisions from non-patient data, issues about consent if researchers hold data that are not from the consented patients, and the potential for detecting eye diseases in others (not the patient). However, having user logins to prevent non-patients from using the devices was suggested as an approach that could overcome this, suggesting addressing this concern was not insurmountable.

I wonder… the temptation would be to try it on your family and friends. I mean, I don’t know. But they might say, ‘Have a go. This is what I have. Look at what the hospital’s given me’ which will mess up your data.

P010, Consultant, Interview

Intervention fit with principles of good care

There were a number of references to how this intervention fits with the principles of good care. For some, home monitoring fits well within the self-management framework, where it is considered appropriate to empower patients to manage their own health care. For others, there was concern that home monitoring could lead to reducing glaucoma clinics to little more than data monitoring hubs, leading to a loss of personalised and holistic care.

. . . there’s the reduction potentially of… it could be that glaucoma clinics are seen to be reduced down to data collection on instruments. We know that many of these patients have things they want to talk to us about, we know that they have things wrong with their eyes other than glaucoma and there’s a loss of the holistic approach to patients when things are reduced down to maybe a few questions on a proforma and some data supplied intermittently versus a face-to-face interaction.

P002, Optometrist, Focus Group 1

For others, limitations in the current system are contributing to the overtreatment of glaucoma, whereby clinicians, knowing it could be some time before they review a patient again, tend to provide medical treatment, on a ‘just in case’ basis. Being able to offer home monitoring was viewed as a solution to this and could reduce unnecessary medical treatment.

I think actually virtual clinics you often overtreat and also I guess many of you have these, what we call follow up pending lists, and certainly COVID’s made them go much higher, and you think a patient may have a problem and you say, ‘Okay, we’ll see you again in six months’, but six months maybe 18 months, and, so on that basis I think you sometimes overtreat because you just don’t know when you’re going to see this patient again. So the fact you could have a facility which can help with monitoring patients that you haven’t got capacity to see within the hospital eye service may enable you to treat less patients and therefore get less morbidity from their treatments.

P005, Consultant, Focus Group 2

For one clinician, while supportive of home monitoring, they wondered if this focus upon already diagnosed patients was missing a bigger problem in glaucoma care, specifically, the low level of detection of the disease in its earlier stage.

I think it’s fine this sort of study, but there is a bigger picture that there are lots of people in the UK and particularly in the developing world who don’t get picked up in the community and there has to be solutions to that that could involve these mobile technologies or whatever.

P009, Consultant, Focus group 2

Managing equity in patient selection

Clinicians frequently reported concerns related to equity and equality, particularly in relation to making decisions as to who should obtain home monitoring equipment, and what fallout there may be from those decisions. Frequent references to factors which may make it difficult for some to participate include accessibility, language, education, and technical abilities (discussed in detail in Perceived effectiveness). However, these led to ethical consequences: how to select and prioritise patients for home monitoring (particularly where there will be resource constraints), impact upon those who are not selected for home monitoring and the risk of creating a two-tier system. The latter was in relation to an expectation that patients may have to contribute financially towards equipment, which opens the door to one system for those who can afford and one for those who cannot.

Well actually, prioritisation. If say someone’s neighbour got this and someone else didn’t get it, word goes around in the glaucoma community: ‘Why did she get it? How come I didn’t get it. Am I more less important? Are you more important?’ That’s not going to go well, I suppose. Ethically how do you choose because you have a limited resource, so that limited resource ethic problem. Even if you gave them the option to buy it, then it’s also another ethical dilemma, because the rich are getting better monitoring. But, having said that, it frees up space, frees up one iCare for another person who can’t afford it, so I mean it’s that ethical dilemma of privatisation versus, you know, but… so that same story. But if they want to buy it, I think they should be able to buy it.

P010, Consultant, Interview

However, some argued that such a two-tier system would be inevitable and necessary to accommodate those for whom home monitoring is not suitable regardless of affordability. Clinicians’ accounts report the need for clear guidance as to how to select patients for home monitoring.

I suppose the other ethical issue is, and we’ve kind of touched on it with patients that are able to use these devices and patients that aren’t able to use these devices, if this becomes the gold standard of treatment for patients and then you have a patient that is unable to do it, how can you ethically not let them have treatment, do you then pass(?) that on to go and take the measurements as regularly in-house. That’s obviously years down the line.

P015, Consultant, Focus Group 3

Autonomy to determine the who, when and where of home monitoring

When asked about the advantages and disadvantages of this intervention, several attributes became apparent. The first was that there was a desire among clinicians for autonomy to decide how best to integrate home monitoring into usual care.

So I don’t think we could decide or should decide as to who can use it. I guess if the underpinning evidence is that these are meaningful measurements, then it’s about the training and accreditation and ability to autonomously make decisions by whoever is looking after the case mix of patients where that service is being commissioned, and that could be primary care or secondary care.

P002, Optometrist, Focus Group 1

Clinicians discussed varied purposes and patient scenarios (varied clinical situations and parameters) where these technologies could be helpful, and it became apparent there was no consensus among clinicians as to the ideal clinical scenario or patient for glaucoma home monitoring. Suggested clinical scenarios included phasing, reducing overtreatment, monitoring for progression, risk-stratifying/referral strategies, promoting patient self-management, increasing localised care (reducing hospital visits) and increasing clinic testing capacity for higher-risk patients.

[T]he way I envisaged it is more as a trigger to find the patients that need intervention. So this isn’t going to be the be all [and] end all care of a patient’s pathway, it’ll be a service where you can gather whatever data you can regularly that gives you a trigger to say, ‘Well actually that patient’s doing absolutely fine, don’t need to get involved’, whereas another patient, ‘Oh, they need to come in hospital, we need to have a proper look at them’.

P015, Consultant, Focus Group 3

It could also be valuable across patient groups as some felt the increase in data collected would be relevant and beneficial for all groups.

[A]ny data points you can get will enable you to have a better picture and diagnosis of what’s going on with a patient, but I think the more points you can get, the more data you can get, the better. I think that could be taken to any patient group.

P015, Consultant, Focus Group 3

Clinicians generally agreed with the use of home monitoring technologies for patients considered at low risk of progressing to significant visual loss, but some could see value in considering these technologies for higher-risk patients in specific circumstances.

[C]hecking your pressure having changed treatment in somebody who’s relatively low risk and your next clinical decision probably isn’t going to be surgery. Then again, you’ve got your ocular hypertensive, in fact we discharge ours to community optometrists, but if you haven’t managed that service with your local optometry committee then you could argue very low risk glaucomas could be monitored quite long term with fields and IOP.

P005, Consultant, Focus Group 2

I think the patient… the people who we see in the hospital services are high risk that need to be seen, because I think the addition of… additional pathology, the very nature of the discussions we have to have with the patients that a home monitoring system isn’t ideal. However, for… even with our high-risk patients, I think again it comes down to how accurate the data is. This might help with anxiety for some patients who are always worried about what their pressures are doing and obviously you can’t see patients frequently, but if you get a baseline with this then patients being able to measure their own pressure at home a couple of… two or three times a week, it might ease that.

P004, Consultant, Focus Group 1

There were several statements suggesting that defining an ideal patient was perhaps not the right approach and that there should be flexibility afforded to individual services as to for whom and how home monitoring is used.

So I suppose each local… well it will have to be very individual to each place and how they work isn’t it as to who looks at the data and which group of patients, that kind of thing as well.

P016, Consultant, Focus Group 3

Some mentioned that age may be a deciding factor. Partly this was in relation to ability to use the devices, partly this was in relation to those considered at risk of the greatest QoL losses from losing visual function at younger age, and partly this was in relation to a subgroup of glaucoma patients who may still be working, and therefore for whom attending clinics for testing is perceived as being more burdensome.

I’m thinking of young normal tension glaucoma patients, so I mean 50. They’re the ones I worry about because they’ve got a lot of life to live and so they’ve got higher likelihood of going blind. And the normal tension glaucomas have trans-central scotomas, so they’re very sensitive. As soon as they lose one decibel that’s it, the whole vision is really bad, so those might be the ones.

P010, Consultant, Interview

Some clinicians felt that home monitoring may not be able to replace usual care for complex cases (e.g. multimorbidities) where information gathered from face-to-face observations is really important.

[T]he people who we see in the hospital services are high risk that need to be seen, because I think the addition of… additional pathology, the very nature of the discussions we have to have with the patients that a home monitoring system isn’t ideal.

P004, Consultant, Focus Group 1

Data relevance and integration

Distinct but complementary to concerns about data ‘ethicality’ was the intervention coherence concerns about the reliability and relevance of the data collected, and how these data will be integrated with electronic medical records. Those claiming to be less familiar with the evidence appeared sceptical that these technologies would produce reliable data to make meaningful decisions about patients’ care. There was disagreement about what aspects of visual health need to be measured; some agreed with both IOP and VF, others felt only IOP is required, others VF only and others worried about the lack of data in relation to other visual health domains, such as disc imaging and acuity, often measured in clinic. While it was agreed that data must be meaningful to clinicians (i.e. the measures being assessed are useful to make clinical judgements), there was little agreement regarding exactly what measures are relevant to make home monitoring meaningful, ‘[u]nless you can make the disc images cheaper and the pressures cheaper, then you can’t just rule out fields. I mean, fields itself are known as not very useful, if you catch my drift’ (P010, Consultant, Interview).

Some were so enthused by the evidence for glaucoma home monitoring that they were keen to see what future technology developments could achieve and made suggestions as to what else this home monitoring intervention could look to incorporate and overcome limitations from just assessing IOP and VF (data-related opportunities to improve for the future). For example, several clinicians discussed a contact lens device which can measure IOP or technologies permitting home disc imaging.

I was just wondering if when we’re talking about disc photographs, there are a few devices now that you can attach to smartphones to give you a disc photograph, I’m not sure would it be worth expanding to that you know, and you get your disc photograph.

P019, Consultant, Interview

Some participants reported being worried about how these data would integrate with existing medical records and any associated time required to input the data into existing systems. This links with perceived burden, discussed below. A solution put forward by many was to make the intervention software compatible with electronic patient records so that results can be immediately integrated without additional effort. A further solution proposed by some participants is the future use of AI for helping review the data as it comes in, making it quicker for clinicians to interpret the data collected remotely.

I’m pretty sure with artificial intelligence, we may be able to even define that what has changed has really change[d] or not, and then we can fast-track that… the results can be interpreted with the AI, I think I can see in five, ten years’ time, this is the way we do.

P001, Consultant, Interview

Support for patients

There was also significant importance placed upon the need to ensure the intervention is supportive for patients to address potential concerns. Participants felt many patients would require professional reassurance in understanding this intervention, particularly in terms of educating patients to understand that readings can fluctuate day to day and that it is the overall trend that the clinicians are concerned with rather than daily readings. Clinicians felt this would be important to prevent unnecessary worry.

You know, as long as they know that someone’s going to tell them I want a report of it or something, then all that hard work’s not for nothing, then I think that’s fine… then you could tell them that everyone’s pressure varies with… there’s a variation every day etc, and warn them not to worry about it, and the whole point is that we gather data over time so that we can make a judgement at the end of the day, so don’t worry about that. If you give them a clause, that will be fine….

P010, Consultant, Interview

Limiting access to results or reconsidering how results are presented to patients was one area of patient-related opportunities to improve in the future suggested by several participants. For example, more generic feedback, such as informing the patient that they had completed the tests correctly, was suggested to be a compromise to maintain engagement and prevent worry from results that may not be fully understood by the patient, causing unnecessary worry.

[W]e should just say in the app, ‘You have correctly answered 75% and you are among the top grade who have done the test very nicely’, or ‘You’ve done test reasonably, but it could be improved if you pay attention to these things’, or, ‘Your test needs to be repeated’. That is the only feedback going to the patient, and then we say, ‘Okay, now, you’ve done the test, it will be reviewed by the doctor… medical team, and we will come back to you.’ So that when we inform them that the situation is getting worse or the visual function is getting worse, we give them the solution there and then that, ‘Okay, we reviewed your results, your results shows deterioration, we have looked at your management plan, and we suggest you should be changing this, and after this you will be reviewed again by somebody.’

P001, Consultant, Interview

Another solution proposed is to offer a simpler intervention in the community for those struggling with the technology, for example facilities at shopping centres, etc.

This could even be done by a trained station in a supermarket, and we can just tell that person to go… you can’t do the technology for somebody, maybe there to help you guys, they don’t need to come to the hospital. They could go for their weekly shopping, and they go to a booth where there is one trained person who could have these appointments for these eighty, eight five… and those people who cannot cope with technology. And it is then outside the hospital, so we basically take it to the community in real sense.

P001, Consultant, Interview

In addition to clinician-provided support, participants felt this intervention would likely require utilisation of social support, making use of family and friends. For example, reminding patients to perform their assessments or physical support to use the devices.

[I]n the same way that we’re quite content to have, a lot of our elderly patients have their partner put their eye drops in, well it can be a team effort. Sometimes elderly people cope rather better when, ‘Oh, yes, my husband always remembers to do this’, who locks the door, there’s a certain way, people survive together don’t they and I just wonder whether there’s something in that for some of our patients. It goes beyond this technology really, but yes, you’d probably need to involve the younger generation.

P002, Optometrist, Focus Group 1

Beliefs about cost-effectiveness

Two clinicians reported that they believed the proposed intervention would be good value for money.

Well, it could potentially reduce the cost. If you’re not having to bring patients in, you don’t require transports, you won’t need technicians to gather the data of the patients as they can gather the data themselves. It would be an actual virtual data collection, and fewer staff costs.

P018, Consultant, Interview

Several clinicians specifically stated that this approach is not good value for money, and many were concerned about the cost of the equipment, particularly the home tonometer. Additional costs raised were maintenance of this equipment, having spares in case of damages, and the staffing to train patients and review data.

[W]e could be looking at wasting our resources looking after too much data points from patients who are well and were not really at risk of going blind with their glaucoma and are we diverting our resources be it instrumentation, time looking at the data and collecting the data which could be used more effectively for other patients.

P017, Consultant, Focus Group 3

I think I prefer something which would be a bit even easier than Home iCare tonometer at the moment, and perhaps less expensive as well, because we give it to such a big population and if you can just give it to one person they will take it for a week, so just calculate for one person the population how many Home Care eye tonometer you need, and then if you extrapolate that cost and the overall benefit you are going to get from that, I’m not convinced that it is for everyone.

P001, Consultant, Interview

Some felt this would make it difficult to obtain management or clinical commissioners’ support for such a service.

I think the distribution and the (inaudible) the devices and persuading the health and social care service that we want to spend thousands of pounds on purchasing these things, I think that could potentially be an issue.

P019, Consultant, Interview

Adjustments to address affordability

However, several clinicians perceived that it could be made more affordable, firstly by looking at app services and dropping the iCare tonometer or asking patients to purchase their own equipment. One clinician felt that the costs would reduce as the technology improved and became more mainstream.

Yeah, that’s where your visual field app may be more cost effective in terms of economics and access to the piece of equipment that the patient has to take home, perhaps an app or something that they can download on to a tablet with a licence, it may be much more generalisable than to take home an iCare HOME.

P009, Consultant, Focus Group 2

Clinician burden

Burden upon clinicians was perceived to stem from several aspects, most commonly from concerns about finding the time to manage services and review and action patient data.

[S]o you’ve got all of these other notes and scans and things to look at from virtual clinics and I think this will add to it, even though these patients aren’t clogging up your clinic they are sending you lots of data which you need to be on top of, and if you don’t action it then they’re going to miss out anyway. So for example, if someone’s in trouble you need to be able to action that and find a space in your clinic for them.

P011, Consultant, Interview

There were also concerns about the potential to increase patient caseload and unscheduled care. This was mostly in relation to spurious results and anxious patients making contact concerned about their results. One clinician raised the burden of adapting to change and learning new skills, likening it to when electronic medical records were introduced, and discussing the time and effort it took to learn a new way of working.

But I suspect what will happen, potentially, is that as the numbers go up, we’re able to deal with more patients and you’ll just end up with a bigger workload, more patients under your care, I guess.

P018, Consultant, Interview

Conversely to these many concerns, one clinician felt that there would not be additional burden from this service, but instead a change in burden, away from the footfall of patients towards an increase in administrative burden, ‘It’ll lighten the burden in terms of footfall perhaps, but it increases the burden in terms of additional or admin work that you have to do’ (P011, Consultant, Interview).

Several clinicians discussed burden more in terms of its impact upon services rather than themselves as individuals. This was often qualified with statements such as ‘I’d have to delegate’ as clinicians felt they did not have time to take on these roles. Tasks to be delegated included maintaining and distributing equipment, training patients, and reviewing data.

Then you have to find the people that are suitable and then you have to train someone who’s able to train the patients on how to use it and someone to distribute the devices, to maintain them, to chase them up when they’re not brought back, to clean them, and then someone to look at the information. So probably quite a lot of work actually.

P019, Consultant, Interview

Patient burden

Several references were made in relation to burden upon patients. While many felt this service could reduce patient burden (linked to and discussed in Intervention coherence), there was some concern about the psychological burden patients may experience, particularly in relation to those who may worry about their results (see Perceived effectiveness). There was also a suggestion about the burden of additional tasks this would involve for the patient, such as collecting equipment from the clinic. One clinician wondered whether, by the time patients collect equipment and learn how to use it, they could have just had their tests done in the clinic, ‘So you’re kind of thinking if they have to come and pick up an iCare [tonometer] you might as well do it’ (P019, Consultant, Interview).

Anticipated outcomes

In line with different perceptions of the purpose of home monitoring, several expected outcomes were reported: improved management from increased data collected, detecting progression quicker and preventing sight loss, either directly from the monitoring patients or indirectly as home monitoring allowed clinicians to spend more time with high-risk patients in clinic. Several discussed how this intervention would be more convenient and less stressful and burdensome for patients. However, one clinician felt that some patients would demand in-clinic services regardless.

Balancing benefits and harms

There was mixed certainty across the clinicians’ accounts as to the degree of clinical effectiveness that could be achieved. Some clinicians were confident that patients would benefit from this intervention. However, a number of clinicians focused upon the potential unintended consequence of increased patient anxiety arising from monitoring. The impact of this was unclear, but clinicians were concerned that rather than empowering patients with responsibility and knowledge, it could add psychological burden. Some were concerned that any benefits in terms of clinic capacity could be eliminated by increase in patient contacts from the ‘worried well’ or ‘unreliable data’. Statements about expected outcomes were often tempered with ‘potentially’ reflecting uncertainty.

This might help with anxiety for some patients who are always worried about what their pressures are doing and obviously you can’t see patients frequently, but if you get a baseline with this then patients being able to measure their own pressure at home a couple of… two or three times a week it might ease that. But at the same time sometimes it can make it worse (over speaking) get phone calls unnecessarily coming back to you so it’s going to be a double-edged sword, I think.

P004, Consultant, Focus Group 1

A number of additional factors which could enhance or limit effectiveness were raised, falling under categories of data and technology limitations, patient demographics and health, and service support and culture.

Impact of perceived data and technology limitations upon intervention effectiveness

An overarching theme was doubt as to whether current technologies are reliable enough to home monitor safely and effectively. Specifically, many clinicians were not convinced that they or their patients would find these technologies acceptable, adversely impacting effectiveness through low engagement.

I still don’t know how reliable these technologies are at the moment so I can’t really comment whether we are going to have meaningful data from this at the moment, certainly with the visual field at home I don’t know personally, yeah… So I think yes, certainly there is a role for this in future but I think there is still a way to go before it comes common practice based mainly on the technology.

P017, Consultant, Focus Group 3

We have to be clear about how meaningful it is, they’re measuring visual function as has been suggested in patients who’ve got a range of other conditions, and clinicians as I say can disagree on routine perimetry. So I’m going to take some persuading, I feel, that the visual function measures are meaningful and valid for glaucoma care personally. That doesn’t mean I’m saying that… I just feel that the jury’s out and we need that data to be confident.

P002, Optometrist, Focus Group 1

Patient characteristics impacting intervention effectiveness

Several patient demographic and health characteristics were stated to have potential impact upon how effective these home monitoring interventions could be. Clinicians were concerned about the physical and cognitive abilities required for this intervention, fearing that a substantial proportion of the glaucoma population would either be excluded from this intervention or would find it unacceptable and would not engage in the first place. This would limit its impact.

Yeah, they have to have the manual dexterity for it as well. So I suppose that will exclude a few people from the clinic able to do that, particularly for the elderly patients.

P019, Consultant, Interview

There are groups of people who are frail, vulnerable, who are serial DNA’ers [Do Not Attend] or whatever and they might be tricky to persuade to do these home-based tests….

P002, Optometrist, Focus Group 1

Impact of service levels factors on effectiveness

There were also several references to effectiveness being related to wider service issues, such as support from NHS Trusts to staff such a service appropriately and the culture of the work environment. One clinician discussed how adoption of this intervention would likely be gradual, requiring time for people to learn and adapt.

So certainly there’ll be a learning curve and I’m guessing after the learning curve people might abandon in it or… like in anything that’s implemented. Certainly like electronic patient records, people hated them to start with. I think now we’re in a situation I really can’t bear to look at paper notes, but it did take us a while.

P016, Consultant, Focus Group 3

Low self-confidence and confidence in patients to use technology

There were fewer statements reported in relation to self-efficacy, likely due to the hypothetical nature of the interviews in relation to the use of these interventions. Clinicians were generally not confident about delivering these home monitoring interventions. Confidence concerns were often related to low confidence in the technologies and doubt regarding service support.

. . . have to say our Trust is very… anything would need a lot of planning and a lot of business plans and those kind of things. I don’t know whether it’s just my Trust, but I’ve certainly seen making anything happen in this Trust is really hard.

P016, Consultant, Focus Group 3

I’d have to say without having all the data, the validity data to hand I’d have to say not very at the moment.

P004, Consultant, Focus Group 1

Several clinicians discussed concerns about the self-efficacy of patients to participate in this intervention. Several aspects were perceived to be quite challenging for patients’ confidence, including worries about using the equipment, risks of damaging equipment, and some patients really struggling with other aspects of eye care.

I don’t know how much an iCare HOME costs, but I can imagine that there’d be some patients who’d be frightened of breaking it or damaging it and then having to kind of pay for it.

P018, Consultant, Interview

I’m really quite select(?) who I give it to because it’s amazing actually how many people just don’t like anything to do with their eyes and it’s all they can do to put their eye drop in.

P019, Consultant, Interview

Home monitoring participants

Baseline case report forms collected demographic information (age, gender, ethnicity, employment status), current use of technologies (smartphones, tablets/iPads, laptops), preference for text or e-mail reminders, clinical information about their glaucoma (e.g. glaucoma type and severity, treatment, surgeries), and baseline IOP and VF measurements. Data were collected by a trained member of site staff through participant report (demographics and technology use) and clinical information recorded in medical records. IOP and VF tests were only performed if they had not been conducted in clinic (with results available in medical records) within the last 3 months.

Site staff collected the follow-up case report form data on IOP and VF measurements conducted at follow-up and information on any changes in treatment. Data were also collected by site staff through participant report relating to satisfaction with the home monitoring intervention, satisfaction with training received, and overall preference between home and clinic monitoring. Participants were also asked to complete questions covering health resource use during the home monitoring period, which asked about their actual use of health services during that time, and if they considered contacting/using health services, to measure whether home monitoring led to any change in use of health services. Again, data were collected verbally from the participant by a trained member of site staff. IOP and VF tests were performed by a trained member of site staff at the follow-up visit.

Data were downloaded from the tonometer device via USB cable to dedicated software (iCare LINK) when the device was returned at the follow-up assessment. The data were transferred as an Excel sheet (.csv file) and added to the secure study database. The OKKO app data were stored on the OKKO Health portal and provided to the research team on request.

Site staff were asked to record any contact they had from the participant throughout the home monitoring period. Data collected included: date, method of communication, a brief description of the reason for the contact, and the approximate time the site staff member spent to resolve this issue.

Interviews

Data collection for the interviews with patient participants recruited to the home monitoring study was informed by a topic guide with questions framed around the constructs of the TDF. The TDF is an established behavioural framework that integrates 33 theories of behaviour into 14 domains that inhibit or enable behaviour (Knowledge, Skills, Social/Professional Role and Identity, Beliefs about Capabilities, Beliefs about Consequences, Optimism, Reinforcement, Intentions, Goals, Memory/Attention/Decision-making Processes, Environmental Context and Resources, Social Influences, Emotion, Behavioural Regulation). 50 For the purposes of I-TRAC, the target behaviour of interest was broad in its definition and was considered as use of the digital technologies for monitoring their glaucoma at home. Interviews were conducted by the Research Fellow over the phone.

Site staff focus groups and interviews were informed by a topic guide framed around key questions from the ADePT framework to assess pragmatics of running the study and the feasibility of progressing this study to a full-size trial. 51 Focus groups and interviews were conducted via MS Teams.

All focus groups and interviews were recorded for verbatim transcription by an external transcription company.

Home monitoring

Frequency of patient participant contacts throughout the study period (for assessing resource impact) was collected via staff self-report and presented using frequencies. Adherence to intervention was calculated for each participant and for each device. This was calculated based on the number of weekly measurements they performed over the 12-week monitoring period. The start of the monitoring period was classified as the date the participant was due to perform their first measure plus each week for the subsequent 11 weeks (12 measures in total). A reported measurement was considered adherent to schedule if it was conducted ± 3 days of the scheduled measure. Overall adherence for each device was calculated as a percentage based on the number of measurements adherent to schedule, divided by 12 (total number of possible measures). Those scoring ≥ 80% were considered ‘adherers’ to the device. Combined adherence scores were categorised as ‘adherent’ based on ≥ 80% adherence to both devices individually with all other variations (i.e. adherence to only one device or to neither) classified as non-adherent or combined adherence.

Participant contact records were collated by the Research Fellow and the total number of contacts, mode of contact, reason for contact, and total time spent by site staff resolving any questions, for each participant, summarised.

Interviews

Interviews with patient participants were analysed using the TDF. A coding guide was developed to describe relevant data to be coded under each domain (see Appendix 4). Each transcript was reviewed for data relevant to any of the TDF codes but also concurrently for any utterances relevant to patient experience of glaucoma home monitoring but that did not fit TDF domains. Three members of the research team (KG, TC, CS) reviewed and double coded transcripts (KG n = 2, TC and CS n = 3). Coders met to review coding decisions with any disagreements resolved through discussion with an independent TDF expert for arbitrating unresolved disagreements. The coding guide was updated to reflect these discussions and then applied to the remaining transcripts.

Site staff focus groups and interviews were analysed using the items from the ADePT framework to help directly inform decisions about progression to full trial. A coding framework for the ADePT items was developed to describe the relevant data to be coded under each ADePT item (see Appendix 5). A second member of the research team checked coding of two of the four transcripts and disagreements were resolved through discussion between the two coders (CS and KG). Data under each ADePT item were reviewed to identify themes, subthemes and relationships between these. Analysis was further refined during writing with both coders reviewing and agreeing the final thematic framework.

Any changes from inception to the design of this phase of the research (and linked phases) are reported in Appendix 6.

Sample characteristics

Across the 3 sites, 42 patient participants were recruited. Differences in recruitment approaches were apparent in the screening logs received from each site. In sites 1 and 2, one strategy used was to prescreen clinic lists in advance to identify potentially suitable participants for the clinician to discuss with the patient at their clinic appointment. Another strategy involved the Research Nurses/other suitably qualified persons approaching patients in the clinic waiting area. Due to delays in opening and the requirement for time-efficient recruitment, site 3 purposively selected glaucoma patients who were known to the research team as being research active (e.g. having participated in one or more previous research studies). With regard to participation rates, for site 1, all patients approached by the PI at their clinic appointment agreed to participate in the study. For site 3, all patients approached by Research Optometrists agreed to participate. For site 2, where the Research Assistant approached patients waiting in clinic waiting rooms, the participation rate was 16% (18/116). While most did not provide a reason for declining participation, of those who did, 13 patients declined due to a reported lack of confidence with technologies, 10 due to a lack of interest in technologies, 5 due to the travel requirements to attend additional clinic visits and 4 due to concerns about burden related to multiple comorbidities. One person declined due to disagreement with the concept of home monitoring and another declined due to glaucoma testing being anxiety provoking.

Figure 2 provides an overview of participant recruitment and retention.

FIGURE 2.

In-home Tracking of glaucoma: Reliability, Acceptability, and Cost participant flow chart.

The patient participant-level characteristics shown in Table 8 indicate a mean age of 67 years in our cohort; the majority were white, and there was equal gender representation. The majority were familiar with some form of electronic device.

Baseline clinical characteristics

Median baseline IOP using GAT was 17 mmHg (IQR 12–20) and 16 mmHg (IQR 12–20) in right and left eyes respectively. Among the 42 patient participants, 26 (62%) had glaucoma in the right eye and 30 (71%) had glaucoma in the left eye. Other diagnoses are summarised by eye in Table 9.

Median baseline VF mean deviation (MD) was −1.8 dB and −3.8 dB in right and left eyes. Median MD in the worse eye (by baseline MD) was −4.65 dB (IQR −10.2 to −1.5, range −30.5 to 0.75). Median MD in the better eye was −1.3 dB (IQR −4.2 to −0.1, range −30.5 to 1.2). The patient participant with a MD of −30.5 dB had VF data entered for only one eye. No other patient participants had VF data entered for only one eye. For almost all patient participants the VF test was deemed reliable by the clinicians (for 40 of 41 tests and 41 of 42 tests in right and left eyes, respectively).

Baseline topical IOP-lowering medications are summarised in Table 9. Out of 42 patient participants, 11 (26.2%) were not using regular eye drops for glaucoma; 9 patient participants had previously undergone laser trabeculoplasty in their right eye and 10 in their left eye; 16 patient participants had had previous surgery to their right eye and 20 previous surgery to their left eye. The majority of patient participants had a diagnosis of glaucoma (26 right eyes and 30 left eyes), but patient participants with suspected glaucoma, OHT and primary angle closure or primary angle suspect were also included in the study (Table 9). A total of 34 patient participants had glaucoma in at least 1 eye.

The majority of patients had mild to moderate glaucoma severity. The VF MD values also allow severity of glaucomatous VF loss to be determined. In all, 23 of 42 patient participants had mild to moderate glaucoma in their right eye and 24 mild to moderate glaucoma in the left eye. Ocular comorbidities including cataract and age-related macular degeneration were present in eight right eyes and eight left eyes.

Table 10 shows the clinical data obtained at the 3-month follow-up visit. Median Goldmann IOP was 17.0 mmHg in the right eye (IQR 13.5–19.0) and 17.0 mmHg in the left eye (IQR 13.0–19.5). VF MD was −2.0 dB (IQR −4.8 to −0.6dB) and −3.2 dB (IQR −7.3 to −1.2) in right and left eyes, respectively. Only 2 of 42 patient participants had a change in glaucoma treatment instigated by their clinician while participating in the study.

Three patient participants withdrew from the study during the monitoring period, two from site 2 and one from site 1. One was due to health reasons (unrelated to glaucoma) and two were due to difficulties with the technology, with one withdrawing from the home monitoring intervention only and completing follow-up data collection and the post-intervention interview.

Adherence to intervention

Adherence to the tonometer alone was 67% (n = 28) compared to 60% (n = 25) of participants being adherent to the OKKO app (Table 11). Overall adherence to both devices (i.e. ≥ 80% adherence to both devices) was 55% (n = 23), with 31% (n = 13) considered as non-adherent to both devices and 14% adherent to only one device.

Additional patient participant contact with site

Of the 42 participants, 20 (48%) contacted their respective study site during their home monitoring period, averaging 1.7 contacts per participant who required support, and were spread across sites (Table 12). In total, 503 minutes (average of 15 minutes per contact) were estimated by site staff as being required to answer or resolve issues. The most common reasons for contact were problems with usernames and passwords for the OKKO app (n = 5), and patient participants having difficulties using both devices (n = 5) (Table 13).

When asked about satisfaction with the home monitoring, the most frequent response (40%, n = 17) was that patient participants were ‘very satisfied’, and similarly, a majority (45%, n = 19) reported being ‘very satisfied’ with the training they received on the digital technologies (Table 14). When asked to report their preference for future monitoring, 50% of patient participants reported that they would prefer to be monitored at home.

Decision to participate (goals; intentions; knowledge; social influences; social professional role and identity)

Patient participants reported that entry into the I-TRAC study was often precipitated by a discussion with their glaucoma specialist to participate in a research study that would explore the potential of home monitoring as an alternative to traditional monitoring. Participants did not report basing their decision to participate on the input of others, such as friends and family. Instead, they were often motivated by their own history of glaucoma to join the study or were interested in engaging with a new technology or research project, generally. Participants typically understood the rationale behind the trial (i.e. to test the feasibility of home monitoring as an alternative/supplement to in-clinic monitoring). However, some participants struggled with the concept that the devices themselves, particularly the tonometer, were also being refined through the I-TRAC study before they could be implemented into standard clinical practice.

R – In terms of your confidence, it sounds like you felt quite confident at that first contact about the study that this would be something that you would be able to do.

P – Yes, but I probably didn’t quite understand what you’re trying to achieve. I thought what you were trying to do was sort of find out… not quite sure. I didn’t realise that it was actually sort of testing the equipment. I thought it was more the idea of regular pressure checks, that sort of emphasis. I suppose I got the emphasis wrong.

Pt 12007

Use of the devices (behavioural regulation; environmental context and resources; skills)

When considering participants’ experiences using the devices, several patient participants discussed their impressions of the intervention training. Most patient participants reported that training for the use of the device was generally well received and believed to instil the skills necessary for participants to monitor on their own. However, delivery of training was not consistent across the sample, with some participants stating that the session was ‘rushed’ and did not cover all aspects of monitoring to their satisfaction. At times, this appeared to be attributable to time and space pressures within the I-TRAC sites, which either did not allow sufficient time or a suitable environment to conduct training. Some participants also voiced concerns about the difficulties staff conducting the training were having aligning the tonometer to take a measurement. Lastly, some participants felt that the interval between the training session and active monitoring was too long, leading to them forgetting many of the skills they had acquired.

Well, I just had met up with this one nurse and she was struggling to find a room where we could do it, so … but then found one, and we had a limited amount of time, but it probably wasn’t ideal for her. […] I think it’s somehow sort of a … I think in order to go from a situation where you’re shown something to you’re doing it, you need a sort of continuity, and I don’t think that was there in the session, and because of the holiday break, I wasn’t following it through so there wasn’t a sort of trans- … a smooth transition from trying it to using it.

Pt 12007

Regardless of training, many participants felt that the devices were accessible enough to engage with through step-by-step, trial-and-error perseverance, assisted by the manual and demonstration videos, until they were reasonably confident they were conducting the monitoring as intended. Engaging with the devices through practice also allowed participants to individualise their process, such as performing it in certain rooms or with additional materials (e.g. stacks of books), so that they could reliably achieve readings in less time and/or fewer attempts. Remembering to monitor was not described as difficult, with participants successfully incorporating the process into their established routines, most often in tandem with taking medications or self-care activities.

R – How easy or difficult was it to remember to do your weekly monitoring each week?

P – I found it pretty easy but I’ve got a reasonably decent memory. And I also put it near the eye drops so when I was doing my eye drops, I would see it and think, ‘Is it Tuesday? No, it’s not.’ But it would be there to be seen, it was somewhere it could be seen and not just away in a cupboard out of sight.

Pt 13005

Potential challenges to home monitoring (environmental context and resources; memory, attention, and decision processes; reinforcement; skills)

Challenges to using the devices were predominantly focused on the use of the tonometer. Difficulties opening and mounting probes, as well as positioning the tonometer in front of the eye, were described often and appeared to be a persistent frustration for some, particularly those with dexterity issues. Feedback from the tonometer on accurate positioning and the pressure readings themselves was deemed insufficient or believed to be absent. Issues performing the visual function tests on the iPad were comparatively fewer, as use of a tablet was more familiar to most participants. However, a frequent criticism of the VF testing was that it was ‘boring’ or ‘tedious’ and required efforts to maintain concentration on the task. Overall, however, use of the devices appeared to be a surmountable challenge for most participants.

Then the tonometer. I normally looked at the instructions to make sure I was doing the thing correctly. I found it very difficult to have the piece of equipment correctly on the eye because it was so sensitive to every little movement, but I did eventually manage to get the green circle of light, but then often when I went to press the button to record the measurement, the piece of equipment moved, so you lost the [measurement]. That was one of the little difficulties. Then it went on. I can think of all sorts of things that didn’t happen that shouldn’t have done.

Pt 13001

Perceived advantages and disadvantages of home monitoring (beliefs about consequences; emotion; environmental context and resources; knowledge; social professional role and identity)

Participants could readily volunteer both potential advantages and disadvantages to home monitoring. Speaking first of advantages, there were perceived benefits to the individual as well as to the larger healthcare system. The most frequently mentioned benefit of home monitoring was its promise to relieve pressures on clinics and eye-care providers for providing all aspects of routine monitoring. Participants could appreciate that taking on some aspects of that monitoring could free up appointments for those with more acute health concerns or otherwise allow providers to reallocate their time to improve standards of care. Having control over how often one could monitor was also seen as advantageous, particularly when waiting times exceeded several months in clinic settings. Participants believed that more frequent monitoring could be beneficial to catch any early indicators of advancing disease and seek treatment before their condition worsened. Finally, the logistics of travelling to in-person appointments were described as a practical motivator to engage in home monitoring to reduce the number of clinic visits necessary.

It’s certainly much more convenient to sit here than have to trundle up to the eye clinic. I know that’s necessary because they need to see you as well, but maybe not just as often. I think it’s a good idea and having undertaken it, I would recommend it.

Pt 14002

Conversely, several of the perceived disadvantages were related to deviating from the standard treatment pathway. Firstly, several participants were concerned about the overall accuracy of the home monitoring device and how it compared to monitoring within professional settings. These concerns appeared to be predicated either on the accuracy of the home monitoring devices themselves or in the participant’s belief in their capability to use the equipment as intended. In either case, the potential for inaccurate results drew concerns that an individual’s eye health could be deteriorating without their, or their provider’s, knowledge. Even when concerns over accuracy were not as relevant, delays between recording high pressures and being able to access care were seen as a source of anxiety. Even nominal readings were felt to be of little value if they were not contextualised by the knowledge of the healthcare provider, who could provide reassurance that values were trending normally or that a plan of care needed to be enacted.

I think for me at first, it was just because I’d be scared, thinking it’s really high. Say it was like eight o’clock at night when you measured and for me, it would be something like 37 and I would have to massage my eye. I would then think, ‘Now I’ve got to wait until nine in the morning to contact them if it goes up.’ It just affects you psychologically that way. Sometimes I’d think I’d rather not know. I mean, you could phone a doctor at the hospital, but you know it’s not an acute emergency so you’re actually then very anxious. That was probably the bit that scared me the most at the beginning.

Pt 12011

Recommendations to improve monitoring (behavioural regulation; environmental context and resources; reinforcement; social influences)

Several aspects of the devices and the home monitoring process were identified from this sample. The coordination required to align the tonometer for readings was cited as a barrier for those with dexterity or other hand–eye coordination issues. Participants believed that this could be streamlined in some way to promote easier alignment of the tonometer with the eye that was not as susceptible to fluctuations in hand steadiness. Some participants also mentioned difficulties in accessing their pressure readings from the device, possibly due to a display time-out. It was suggested that readings should be available in an easier-to-read format that preserves the history of recordings, as well.

Potential improvements to training were mentioned by several participants. As mentioned above, some participants felt there was too long an interval between the training session and the first ‘live’ monitoring session. They suggested that training be held as near as possible to this first session to increase the likelihood that skills practised during training will transfer to real-world use. ‘Top-up sessions’ (i.e. condensed training focused on reiterating or improving skills already learned) were seen as a useful addition to the course of home monitoring if participants were finding it difficult to apply their skills from one session. Additionally, some participants suggested that ongoing technical support would have been helpful, but their attempts to solicit that help within this feasibility trial were not addressed. As for any other additional content of the training, some participants were unaware of the demonstration videos and suggested those be emphasised as an available resource in the course of training. There were also mentions that the training manual provided, while helpful for most, should be reviewed for readability, as some participants complained of not being able to utilise it.

R – So, one suggestion that’s come in from a couple of participants is that it might be useful to have a lead-in period, so you get maybe a week or two just to practise at your leisure with the equipment and then commence the study. What would you think of that suggestion?

P – I think that’s a very good idea because when you’re only doing it weekly, from one week to the next, you can forget how you managed to do it correctly. So, if you had a week where you could just use it anytime, say every day, just to get into the swing of it, I think that would be a good idea.

Pt 13009

Finally, some improvements to the process of monitoring were suggested. In order to help build confidence in their ability to conduct monitoring, participants suggested the inclusion of a ‘lead-in’ period. This period would be an opportunity for participants to practise with the device without fear of ‘getting it wrong’ and affecting trial results. Once in active monitoring, as mentioned above, participants would have also liked easier-to-read and easier-to-understand feedback from the devices so they could track the progress of their readings. Further resources on how to interpret results may also prove valuable in this sense. A final comment on the active monitoring period was the anxiety that may be precipitated by too-frequent measurements. To pre-empt this, participants suggested that monitoring should be restricted to only what is necessary, for example once a week.

I think it would be valuable if you could do it monthly or something like that, just once, no more than that if it was stable. Then that way it could be fed into some computer to see the trends or something. But any more, I’m just not sure, just like my blood pressure. Psychologically, you just might get neurotic.

Pt 12011

Site staff knowledge, beliefs and feelings towards home monitoring

All site staff reported positively when asked to reflect upon how they felt when first approached to help deliver the I-TRAC study. All staff were optimistic about a novel approach for glaucoma care which they believed could improve patient care.

Always positive about doing the stuff you know, there’s something new coming it’s always nice to have something new isn’t it. But particularly I think the it’s the fact it’s a device study and you’re interested in how the devices will actually improve patient care. So it’s about excitement of being able to introduce something new that’s going to hopefully make a difference really, yeah.

P037, Research Nurse, Site 2

Several site staff members reported having one or two concerns prior to the site opening. One concern related to patient participants asking about clinical implications of data collected, as this was not something being assessed within I-TRAC. Another found some of the flexibilities in the protocol (e.g. eligibility criteria, what actual tests were done for assessing IOP and VFs in clinic prior to the study) a bit worrying to start, due to past experiences on tightly controlled trials (compared to the I-TRAC feasibility study).

Yeah, there is also the fear, going back to feelings or thoughts, that if you’re doing this as part of a study as well, the patient wants to know, ‘All that effort I put in taking the measurements, what does it mean? Am I okay?’ that creates a challenge as well particularly when you’re just looking at feasibility and you’re actually not that interested in the actual data they’re collecting.

P036, PI, Site 1

Site staff perceived a number of benefits, particularly for patients, arising from home monitoring. These included a sense of control of their glaucoma through self-monitoring, greater convenience from not having to attend clinic as often, and increased chances of detecting significant changes more quickly, leading to better outcomes in terms of vision. Across site staff accounts, there was a strong sense of buy-in and a desire to see the home monitoring interventions work for patients.

Well that’s the thing, I mean it could be useful for people who live far from the hospital where there is a remote… access issue, it might be helpful, people perhaps with some mobility issues about them getting to the hospital but not stopping them from using the home monitoring device. Yeah, I think there are niches where it would be helpful, but I’ve always had in my mind that this would be a device for continuous monitoring for some reason, and that ultimately the aim is to keep the patient out of the hospital.

P035, PI, Site 2

The ease of running the I-TRAC study and the compatibility of its components

The findings around study conduct were themed in relation to ease of running the study and were composed of two subthemes related to study deliverability: (1) easy-to-run and low-burden study; and (2) quality and usefulness of training.

1. Low-burden, easy-to-run study

Overall, all participating site staff felt that I-TRAC was easy to run and its components (i.e. study processes) worked well together. This was qualified by reports that the I-TRAC study was well explained, supported and was perceived to be relatively low burden for participants and sites.

A lot easier than I thought it was going to be. I think as well, I did sit down the day before and go through the paperwork and just have it in my head, ‘This is what I’m going to say, this is how I’m going to do it’, and have it laid out. I think you have to do that to seem confident in front of the patient. Otherwise, if I’m faffing about, they’ll be like, ‘You don’t know what you’re doing’, and I’ll not know what I’m doing. I think the combination of the training plus the written material that was there too, it was fine.

P038, Research Optometrist, Site 3

Site staff reflected on I-TRAC being surprisingly easy to recruit for and felt this was due to patient participants having a keen interest in something that was novel and that had a clear benefit for them in terms of reassurance, control and burden (being able to monitor from home instead of travelling to clinics).

think maybe for this particular study, home monitoring, I don’t know if that’s such a big issue because we found that recruiting was really easy and we recruited just from our face-to-face clinic, and really we recruited the majority of the patients for each phase in a day. It was literally asking consecutive patients and most people said, ‘Yes, I’d be very interested in this’, and that was surprising, that’s certainly not an experience that is very common in studies. I know there were… I mean it was a very wide inclusion criteria, but nevertheless you still expect perhaps a fair few to decline.

P036, PI, Site 1

Site staff reported few challenges in terms of study conduct. Those challenges raised involved logistics (equipment return and appointments), an issue relating to accessing data from the tonometers (i.e. one site forgot to download the data from the tonometers before reissuing the devices to the next set of participants), and research and information governance issues prior to site opening.

I think we may have had some challenges in getting the equipment back and how the patients… how we got patients back into the system, because I think that was potentially an issue in terms of clinicians’ time because there’s also the idea of getting… picking up all the other data like the visual fields and things like that, and these things all… the organisation of those things and organising when the patients were coming back about picking suitable times where clinicians were available and the actual patient was available, the logistics of that with lots of people.

P037, Research Nurse, Site 2

1. Quality and usefulness of training

The training of patient participants to use both pieces of home monitoring equipment was a critical and central component of this intervention. The consensus among site staff was that the time to train patient participants reduced with practice, initially requiring more time than had been allocated (sometimes up to 2 hours) but reducing to the allocated time (30–45 minutes) after they had trained their first few patient participants. Site staff reflected that they felt the quality of their training improved after the first few participants, as they gained in confidence and developed the skills and tips to help patient participants learn to use the devices. They reported that the patient participant-facing training materials (a step-by-step handbook with images, and several YouTube videos) were a good resource for patient participants. However, some site staff reported that they felt not all participants were using the provided support, resulting in a number of contacts from patient participants requiring additional support during the monitoring period. Site staff reported that they found the site training materials and site initiation visit training (protocol and study instruction manuals) to be useful, coherent and generally to cover all that was needed, ‘Yeah, the step-by-step guide was very, very detailed, the fact that there was one for the patient and one for the practitioner as well was really handy’ (P033, Research Optometrist, Site 3).

However, one aspect of site training, the instructional training for using the tonometer, was delivered by a representative from the commercial partner. The consensus across site staff was that this training did not help them to teach patient participants how to use the tonometers. Several mentioned that they felt the representative was there in a sales capacity, not to deliver training, and had no practical experience of using the device so could not answer a number of questions the site staff had.

We need somebody who’s thorough, who knows all about it, who knows from a patient perspective as well. I’ve had some fantastic training in the past with other companies and real trainers but I kind of feel there was a sales element there. So no, that wasn’t very good for explaining.

P039, Research Nurse, Site 1

Many mentioned that the YouTube videos and written instructions were more helpful than the commercial partner training in terms of helping them master how to use the equipment and train patient participants to use the equipment.

Yes, because I mean I actually learned more from watching that YouTube training than I did actually from the sale rep side of things, so, yeah. Yeah, certainly the YouTube thing helped once we’d got to grips with the YouTube address.

P037, Research Nurse, Site 2

The feedback from sites was that equipment training should be developed and delivered by those with experience of using the equipment and training others (preferably patients) to use the equipment, so that they have knowledge and awareness of common problems patients have and how to overcome these.

Eligibility, recruitment and representativeness of sample

All three sites reported that they found it easier to recruit than they had anticipated at the outset of the I-TRAC study. Across the sites, staff generally felt, however, that their samples were not representative of the typical glaucoma population. The three sites reported how they approached recruitment differently.

Site 1 recruited from a general face-to-face clinic and participants were first approached by the clinician/PI and if agreeable, referred to the Research Nurse. The PI reported that the selected participants were really keen to participate. The Research Nurse in one site, however, noted a trend where they felt older participants appeared more determined to master the technology, and the younger participants, particularly those still working, had other priorities. This was echoed by the PI from site 2. Conversely, some staff from other sites believed younger participants were easier to train, perceiving older participants as being more time-intensive to train and support with new technologies.

I must say though, when I was phoning some people, maybe one or two of the younger ones, they said, ‘We’ve not been doing it, just not got round to it’, etc., so that was a wee bit of a disappointment… They were more elderly in the first cohort but they were determined they were going to do it. I don’t know if it’s that generation, I’m not sure.

P039, P039, Research Nurse, Site 1

Site 2 recruited participants through their virtual clinic, which typically deals with straightforward glaucoma cases (i.e. glaucoma patients considered at low risk of progression due to stable disease and few comorbidities which may complicate glaucoma management). Site 2 staff reported that while recruitment was easy given the access to a large pool of suitable participants, the approach to recruitment involved the Research Associate spending time in the waiting area of the clinic with patients not pre-notified about the study by the clinician/PI. The PI discussed how while time-efficient (a high volume of patients to speak with in one half-day session), recruiting from this clinic only, as they did, would not represent all glaucoma patients. This is because those whose condition may be more complex would be seen in a different clinic and, conversely, those patients with straightforward glaucoma and OHT would be discharged to community services.

[W]e have discharged quite a few of our patients to the community, the very, very straightforward ones like ocular hypertensions, so they weren’t… there might’ve been one or two but they’re usually… if they’re ocular hypertensions in the hospital they’re usually there for a reason. So actually quite a lot of them… I’m sorry, that group would be missing.

P035, PI, Site 2

The Research Assistant discussed the difficulties of having around half of participants decline participation, with reasoning related to the person approached having no interest in a technology study.

But yeah, generally I think my group is diverse-ish, but obviously there’s that inherent lack of diversity you get by just sitting down outside a clinic saying, ‘We want you to take some technology home’, and then half of the patients say, ‘Oh, I don’t do tech’, and just don’t want to talk to you ever again.

P031, Research Assistant, Site 2

The Research Assistant also discussed how they had found it challenging to recruit a diverse sample of participants, especially in relation to ethnicity. An issue reported by several members of site staff was that of language, perceiving language issues as being a significant barrier not just to accessing the information leaflets but also learning to use the devices in a second language.

Yeah, I did recruit, I did very happily manage to recruit a lovely I think Caribbean lady who’s part of the study, but other than that I have approached a lot of people and I think sometimes there have been language issues, the fact that it’s all in English, the information sheet’s in English, that’s been an obstacle I think a couple of times, although it hasn’t always been explicitly stated that that is the issue, yeah.

P031, Research Assistant, Site 2

The PI from this site questioned whether it would be more helpful to have a more diverse study team, or at least a more diverse team of recruiters, as he felt a shared ethnicity may go some way towards the study having a more inclusive approach to recruitment of glaucoma patients.

Well I think there is a certain racial thing, I mean if you had a black guy, black doctor asking black people to participate in a study I think that’s much better than some old white fella like me asking them to do it, and I think that just makes it much more accessible.

P035, PI, Site 3

For site 3 the seven participants were specifically targeted as they were known as individuals with the educational and technology familiarity needed to participate and would enhance recruitment efficiency. Site 3 staff reported they felt their participants were younger than typical glaucoma patients. The following excerpt presents this site’s views when asked about how representative their sample was.

Well I don’t… well I expected ours [when considering the I-TRAC sample of patients] to be older….

P032, Research Optometrist, Site 3

Yeah, not in terms of age and probably also not in terms of familiarity with electronic devices.

P033, Research Optometrist, Site 3

But then I suppose it’s good to have a range and different types of glaucoma as well as part of the study, but on the whole they’re not.

P032, Research Optometrist, Site 3

Unintended consequences arising from the intervention

Three members of site staff (one from each site) discussed how they had initially been concerned about the intervention causing patient participant anxiety arising from increased focus on their disease. Despite these concerns, there were no reports from any site staff of any actual or suspicions of induced anxiety within their patient participants. One site reported two patient participants who required treatment, an outcome that was triggered by patient concerns relating to the tonometer readings which resulted in contact with their consultant. In both cases, the site staff member reported that the patient participants did require treatment and that both patient participants were glad they had participated so it could be detected.

There were two participants who actually had to come in as a result of the pressure that they measured. One of them went to the optometrist and then she came in to us, and the other one just came directly to us because… if they hadn’t had the equipment, they wouldn’t have known their pressure had gone up…. Both these people liked the equipment, they liked using it, they thought it was worthwhile.

P039, Research Nurse, Site 1

Adaptations or considerations required for a future trial

Site staff reflected on several aspects of study conduct that would require careful review prior to progressing to a full trial. This included: (1) the additional resources that would be required; (2) aspects of study design to be improved (i.e. determining the ideal follow-up duration); (3) participant selection criteria; and (4) improvements to intervention training. These subthemes are presented below.

1. Increased human and financial resources

In relation to resource use, sites reported that to feasibly run I-TRAC as a full trial with more participants, increased logistical and technical support would be required to manage the devices and assist participants with any queries. There were also concerns about the perceived (significant) costs of acquiring a large number of devices and accessories (e.g. batteries), and how a large trial would operate if equipment had to be rotated between participants in an effort to reduce technology costs.

Does it mean you’re going to follow patients for a year? Does that mean you’re going to follow patients for… if you have to follow them for a year, well essentially you need equipment for every patient in the trial, that’s costly I expect, and you need backup to manage the technical aspects that they may have, repairs, all of those sorts of things. I suppose it’s all manageable, it’s just understanding it before you do it, which I guess is what this is about.

P035, PI, Site 2

While it was not conducted as part of the I-TRAC study, it was recognised that the human resource requirements for reviewing and acting on the data received from the devices in a future trial would have implications for delivery. Site staff were concerned about the substantial time required to process this amount of data. Suggestions to make this achievable included exploring methods for efficient data review (e.g. making use of automated algorithms that would trigger a message to the clinician only if a test is abnormal, or AI technologies) and integrating data into electronic patient records. There were also concerns about the impact on resources required to address readings from the home monitoring devices that suggest the need for further intervention.

I guess at the moment it would be… perhaps it’s not very efficient because it’s all very innovative. But I guess, you know, you can incorporate this into electronic records and there will be, let’s say, the potential of having the information displayed in a different way, that makes the clinicians’ tasks easier. It depends, you know? At the moment it will be time-consuming because we don’t have a good system to process this information. It would need to be well thought to make it efficient.

P034, PI, Site 3

1. Changes to study design

Most suggestions related to study design were improvements to the intervention, such as involving carers/family members in the monitoring, offering home visits instead of patients coming to the clinic for training, improving the electronic reminders to avoid spam filters, and to consider restricting the study to evaluating only one technology rather than multiple technologies at the same time.

One way that might help simplify it, and also perhaps help identify the key research question would be to just test one thing rather than testing visual fields and pressure. I don’t know which would be best. The technology for pressure is more developed but visual fields are probably more important.

P036, PI, Site 1

Exploring the potential for utilising participants’ social support within both the training and use of the devices to be more inclusive of those at greatest risk of vision loss was also suggested.

The other thing I suppose is that we focused on self-measurement, but I think that looking at carers and how they can help and support the patient is really important too because I find it hard to put eye drops in my own eyes, now if I had glaucoma I’d be asking my partner to do it for me, I’m terrible, and I think, well patients do that too, and I think having someone to help you and support you really to take the measurements at home would help because one of my worries about this sort of technology is it’s the vulnerable groups, the people who are probably at highest risk of losing vision, the people who can’t put eye drops in themselves, and they’re the people who might need it most and they might not be able to do it. So I think looking at, if there was further work with this, I think looking at carers as well would be useful.

P036, PI, Site 1

Within I-TRAC, digital reminders for home monitoring devices (such as e-mails and text messages) were provided to enhance adherence. Site staff highlighted the need to ensure these reminders are received as intended and not filtered out by spam filters or blocked by mobile phone providers. One site staff member reported that the need for patient participants to travel to the clinic for the training and provision of equipment acted as a barrier for several participants, undermining the potential benefit for patients who cannot travel. For this site staff participant, they suggested that providing study visits at home may improve recruitment of those who are often less represented in research.

I think it would be entirely possible. I think you… I think definitely within the capacity of going and visiting patients there’s no reason it couldn’t be done at the patient’s home, it’s just whether that’s a reasonable use of time I think.

P031, Research Assistant, Site 2

The importance of selecting research sites and staff purposively was also commented upon, due to the importance of confidence with technologies and the patience required to help teach others. A further recommendation in relation to design was the potential for a further study to determine the duration of follow-up.

I wonder if you almost need another… a pilot study to see how long do you need to… do patients need to keep them for to get full results, because that would be a key question in terms of how many devices you need and then the cost of the study….

P036, PI, Site 1

1. Recruitment of participants

When asked about their expectations for recruiting a larger number of participants for a future trial, there were mixed views. Although site staff felt recruitment for this feasibility study, with very broad criteria for participant selection, had been easy to recruit to, there were concerns about how easy recruitment of a more tightly defined eligible population might be.

I think the fact that we have a virtual clinic where there is a large pool of what would’ve been considered suitable patients for this all coming to the same place in a non-stressful environment made a massive difference to our ability to recruit patients, and that has two real implications for me. Going ahead to a randomised controlled trial, which I guess is what you’re hoping to do, I would be very much looking at recruiting centres which run virtual clinic services because the patients are quite easy to access or more easy to access. But it also creates a barrier as well because in a way they are generally more straightforward patients and it depends who you’re trying to evaluate these devices on. If you want more advanced patients then you’re going to have to look elsewhere and then it will be harder to recruit. This was particularly in relation to the typical age demographic of glaucoma patients, and assumptions about older patients abilities to use the technologies.

P035, PI, Site 2

Especially at the moment, we didn’t have any restrictions, the availability was anyone with glaucoma, so that was basically… if in the context of a trial, you have a particular population, it might be more difficult… I can imagine from my impression, this technology might not be for everybody. I guess if there is a trial, there will be an inclusion and exclusion criteria, there would be a particular population of glaucoma people, people with glaucoma. I guess, you know, I’m just thinking that perhaps recruitment will be difficult.

P034, PI, Site 3

Site staff discussed how there were uncertainties about which glaucoma patients home monitoring would best meet the needs of in terms of usefulness for the patient, usefulness for the clinician, and from a cost-effectiveness perspective.

Then yeah, maybe those ones who find it difficult, who are stable but also find it difficult to come to appointments, and I agree with [other participant] that those younger ones that are familiar with technology and have busy lives.

P032, Research Optometrist, Site 3

That’s the biggest thing because I think… yeah, it’s a bit… it’s a chicken and egg situation isn’t it because you don’t really know what sort of patients it might benefit… But yeah, I think it’s really hard because it’s just speculating and guessing because we don’t really know what kind of patients or what scenarios it would be of use.

P036, PI, Site 1

For one PI, the difficulty arose because the two technologies being utilised together as the home monitoring intervention in the I-TRAC study are measuring different outcomes. For them, each of those outcomes independently would help very different subgroups of glaucoma patients, and those who would benefit from both were a very exclusive group. Clarifying who this intervention would be most appropriate for was indicated to be an important priority prior to any full-size trial.

But I think separating the pressure and the visual fields and looking at the different indications for those is really important because it’s very different situations where you might do one or the other. So yeah, I think there’s almost three situations, there’s three questions, how can we use home glaucoma monitoring, but how can we use home fields and home IOP because they’re different, different groups.

P036, PI, Site 1

1. Improvements to intervention training (for staff and patient participants)

Many site staff participants reported suggestions as to how the training for patient participants could be improved. The first was to identify and develop ‘good trainers’ at research sites – those who are patient and able to promote confidence and reassurance among patient participants. It was considered helpful if the person delivering the training had some awareness of the patients’ clinical and disease status, as this provided insightful information as to how best to help the patient learn to use the devices.

I think also the other thing that I noticed was that it’s helpful if the person teaching the patient has an understanding of the patient’s condition, because we had a patient where they had poor vision on one eye and it was really hard to get them to measure the pressure in that eye, but they could do the other eye, and we didn’t realise initially, and then once the notes were opened and you’re like, ‘Oh, the vision’s quite poor in that eye’, it became apparent why there was that big difference. I think just having some… I think if you just dump a patient with a technician who knows nothing about that patient, that might not be the best way to teach them.

P036, PI, Site 1

In terms of the specific tips and stuff, it’s generally been stuff like knowing that some patients find it far easier with one eye closed or knowing that actually sometimes I just need to go in and adjust the device a bit myself and help them get it in the right position, and then once they know how that should feel, a lot of them just… they can remember that and it helps a lot.

P031, Research Assistant, Site 2

A number of site staff wondered if group training for patient participants would be helpful.

For me, they’d be in an environment where the equipment would be there for them to play with. As I say, internally here, there was issues about getting space so I didn’t think that initially helped. If I had a conference room for them, get the videos up where they can see it, huge, not make it too long because they’ll maybe get a wee bit tired, provide a wee provision of some fluids and what-have-you, cups of tea, they’d like that, it would be a social thing for them. Have the equipment there that they can play with at the time, each one get a wee shot themselves, see how they feel with it, ‘How do you feel? Go to that table, have a wee shot’. Somebody like myself always there on hand to explain to them and also demonstrate. I think a huge presentation behind us, like the videos you get on YouTube showing them, big enough for them to see, take questions, play with it, I’m always there to help them.

P039, Research Nurse, Site 1

This group approach was perceived to be a potentially efficient approach of delivering training for sites, but the additional social aspect of the training environment was also expected to be helpful for patient participants.

It would probably be easier if we just had an I-TRAC demo and just brought everybody on the same day to do training, or maybe did group training with the patients rather than individually. That would be something that would save time, so one person can maybe teach three or four people at the same time and maybe those people have a bit of support as well because they’re all part of the same study… People with conditions can feel isolated sometimes if they think they’re alone, but if they have a group where maybe they come to the clinic once a month, like you say, it’s a social thing. I know older patients love coming, getting tea and biscuits and a chat, that’s why it takes so much time!

P038, Research Optometrist, Site 3

Several aspects of training for site staff, directly related to the intervention devices themselves, were also suggested as areas of improvement for the future. Increased provision of more detailed technical information on the devices, what they are measuring and how they work, were suggested as ways of improving site staff confidence not just in the devices but in their ability to answer patient participant questions about the technologies being explored: ‘I think more clarity in what that test is actually doing would be good, there maybe wasn’t that much information on that’ (P032, Research Optometrist, Site 3).

Site staff were generally in agreement that they would like to see more training opportunities to increase the time they could spend practising using the devices and practising training another person to use the device, so as to increase their familiarity and confidence with the devices. Site staff also suggested several benefits of having spare equipment on site, such as that staff can continue to practise with the device while participants are home monitoring, and it would be helpful in supporting participants when they contact for help with their device over the phone.

I don’t know if this is doable in terms of resources, but to have a device, one of the [tonometers] on site that we get to keep to solely use as a practice device, and then if we’ve got new staff coming on board or whatever we’ve always got something there that we can do in-house training for people.

P033, Research Optometrist, Site 3

Pragmatic considerations for the delivery of site training also included ensuring a suitable training environment is made available at site for the training session; as was reflected on by one site, space can be limited at NHS sites.

There wasn’t much space for them [a suitable room for the commercial trainers to deliver training at site] here. It didn’t really hit off very well initially. Do you know what I would have liked? If we could have all gone to a wee training course, you know, meeting up with the other people. I know the budget is only so much but if we’d done a training course with the other people where we could all ask questions, take our time and been shown how to use it, we could take notes and things like that as well and then try and start a wee bit earlier.

P039, Research Nurse, Site 1

Multisite training days, where staff from across sites get together off-site to complete their device training, was mentioned as a preference by several site staff participants.

I think it’s something I wouldn’t necessarily think would need to be an organised event, but it’s something I’d recommend for future sites is familiarising themselves with the devices, maybe just letting them use some of the probes, just actually go through, ‘This is how the device works. If I’m going to use this, this is how I use it’, show it, get all of the members of the team used to how it works so that then they can explain it to patients quickly, because when I learnt how to use it, it was just the one meeting, I think was the only person who had it demonstrated on, and then from then it was just, ‘Now you’ve got to show patients how this works, and actually having a bit more experimentation in the meantime might’ve been useful, yeah’.

P031, Research Assistant, Site 2

Impact and influence of COVID-19 pandemic on perceptions of acceptability to patients and clinicians

Site staff were asked to discuss whether they believed the COVID-19 pandemic had any influence on recruitment to I-TRAC. No instances were reported where the pandemic appeared to influence a participant’s decision to participate. It was noted, however, that the pandemic may have made some patients more amenable to the use of technologies, through increased familiarity gained by using them during the pandemic to stay in touch with friends and family and use of remote health consultation services. Our PI in site 2 reported that more patients were now being seen in virtual clinics post pandemic, as patient care pathways have changed due to the pandemic, and as a result this could be said to have made recruitment easier as the more stable patients were available to be recruited from one clinic, as opposed to searching multiple clinics.

I wonder, just add to that, whether patients might be more happy to try technology and might be… I think there were certainly some people that had never used an iPad and Zoom before and now they do to talk to relatives and maybe, so maybe it helped, and people might’ve been worried about coming to clinics and, so the idea to monitor disease at home might, they might’ve been more open to that than they would’ve been before the pandemic.

P036, PI, Site 1

Table 19 provides a summary of the strengths and weaknesses of I-TRAC perceived by site staff. Table 20 reports a summary of site staff suggestions regarding a future evaluation of digital technology for home monitoring glaucoma.

Patient acceptability

Low digital literacy skills, anxieties about using technologies, difficulties understanding how to use technologies, and no or low support at home to assist participation were commonly reported by researchers as reasons for patients choosing not to engage with DHT studies or trials.

So we got quite a lot of, ‘No, sorry, I can’t’, or, ‘Sorry, I tried but I can’t manage this, it’s just giving me too much stress and hassle’, or, ‘Sorry, I need my son to help me but he’s not around because I can’t use a computer’. So there was a lot of… I would say probably, you could probably say about a third of them were very happy and capable and able, a third were willing and needed assistance and the other third were probably like you don’t even ask.

P028, Trial Manager

Researchers described how these issues reflected low acceptability of DHT from the patients’ perspective, resulting in a lack of willingness to participate in studies or trials of this type. Several researchers reported how DHT study or trial populations were over-represented by younger, more educated, white participants – a finding that, while not dissimilar to issues across research fields, researchers felt was more difficult to overcome in evaluating DHTs due to generational differences in ability and perceptions of technologies. A further impediment to recruiting patients was the assumptions made by clinicians or site staff about patients’ willingness and/or ability to take part in DHT studies or trials, often making judgements based upon age (i.e. perceiving their patient group as too old to use technologies and anticipating barriers to participation prior to discussing with patients). Researchers frequently reported how sites varied in their attempts to offer DHT study/trial participation to as a great a number of participants as possible. Some researchers highlighted selective approaches and offers, stating that many of their patients would not be able to use technologies or would not be interested.

. . . well identifying patients that were eligible wasn’t necessarily a problem because that’s irrelevant whether or not they wanted to take part in the study or not, it’s I think the sites, some sites were more selective with who they approached depending on how they felt the patient would cope with it, and some sites were… would approach anyone who was eligible and try get them on board. In a lot of ways the latter was better because you get a better picture, but we still had a lot of rejection because of the technology.

P023, Trial Manager

Clinician and site staff acceptability

It also became apparent that there were a number of barriers put up by clinicians in agreeing to host DHT research at their site. Researchers discussed how a lack of supportive evidence on the effectiveness and benefits of the technology being tested contributed to the lack of staff/clinician buy-in. In some circumstances, clinicians challenged the benefit of the intervention, questioning what value the digital intervention had for them personally, and at times for their patients.

And the other beauty was that we were able to search the records to look at the number of face to face consultations, and we were able to demonstrate that there was a significant fall in face to face consultations in the group that were being managed that way, a 25% reduction in all face to face consultations. Once you have this sort of data and you can show this to people, they start to get interested…’

P026, Chief Investigator

If clinicians did not see that the digital technology would make their life easier – that is, if they perceived the technology to be cumbersome, time-consuming and not leading to a clear improvement for the patient – they would be reluctant to participate.

. . . the main thing is that the equipment works. That’s really, really important that it works and is reliable. If you’re a busy ophthalmologist, a busy GP and it just doesn’t work a couple of times, you’ll think, ‘Stuff this’, and you just won’t use it, you know? Even an enthusiast, after a while, they might like you and say, ‘We’re doing our best here but I’ve got a busy life and I can’t be bothered with this’. It’s really important that it works.

P026, Chief Investigator

Researchers reported that many of these acceptability issues, believed by the researchers to inhibit recruitment and retention, could be enhanced in DHT studies or trials through several mechanisms. Ensuring the technology has a clear purpose and is as simple to use as possible, with minimal steps and actions, and that this is communicated clearly to clinicians, can aid buy-in. The importance of simplicity was also echoed in relation to improving the acceptability of DHT for patients.

For the [study name] the patient population, the average age was, I think it was 74 and, so that hugely impacts how that patient population should be approached for the trial because the main reason for not being interested in taking part was being put off by the technology. So that… you will need to consider how best to approach the patient population without putting them off. We had training sessions for an hour long where the site would demonstrate the technology, go through it, but it’s still a lot to take in and obviously we weren’t just testing one device, we had a paper journal plus two apps, plus a MiFi device, it was a lot to take on. So I feel like yeah, considering the patient population is really important, and also not to overwhelm them. I appreciate if a trial’s being set up and you want to measure as much as possible, but keeping it to one device or one test is the easiest way from a patient perspective, yeah.

P023, Trial Manager

Exploring the needs of stakeholders early in the development of the technology, and building these into the design, was considered important. Many researchers felt using technologies they (clinicians and patients) are already familiar with aids acceptability. However, this issue may be a double-edged sword, linking with ‘digital exclusion’, discussed later in this chapter. Also, many believed that recruiting patients face to face in a clinical setting, demonstrating the technologies, helps with training coherence. In relation to the training to be delivered, multiple pilots of training materials were considered a good approach to ensuring the final training programme is effective and acceptable to patients and clinicians. Contextually, many considered that the COVID-19 pandemic helped with both clinician and patient acceptability through increased interest, motivation, and confidence in using technologies. The value of digital technologies was also perceived by researchers to have increased for patients and clinicians; DHTs have offered a lifeline for remote monitoring when hospital visits were not possible or were undesirable due to COVID-19 risk: ‘So basically have a system where you only have one click option so they can’t click on the wrong thing, that’s ultimate necessity’ (P028, Trial Manager).

Some researchers reported how selecting a research site experienced in running DHT studies/trials may be beneficial. One participant described this as identifying research champions, those who are motivated and enthusiastic enough to see new DHTs through the early testing period where many problems are anticipated. These champions were less likely to give up at the first hurdle and set the path for future implementation. This was considered good practice for facilitating feasibility or pilot studies of DHT. As was stated, busy clinicians often drop challenging studies at the first hurdle. Additionally, researchers reported that the involvement of researchers from communities and ethnicities that were under-represented in recruitment helped to balance the representation.

It’s really good if you can get champions, so if you can get a few people who are tech enthusiasts who will get this thing going, who don’t mind if it’s going to be difficult in the first – there’ll be bumps in the road. We always did this, we would pick a few practices, for example, where we knew the people there were reasonable folk that understood technology, that understood that this was not perfect, you know? We would say that to them at the start, ‘You are helping us develop this. We know there are going to be bumps in the road and the reason we have chosen you is because you’ve got that approach and that you will be prepared to put up with this and you will feed back. You know that we will take what you say and make changes as a result of that’. I think that’s really, really important. Try to as much as possible involve the early people in this as much as you can in the actual design and running of this, continually feed back to them. We always kept feeding back to people, ‘Has anyone come up with an idea? We’ve heard that this works, this is another way of recruiting people’, you know, this sort of stuff. ‘Here are the numbers, these people have managed to recruit this number of people. Let us know what trouble you’re having, we’ll see if we can sort it out’, and respond quickly every time… .

P026, Chief Investigator

Expert glaucoma clinicians not involved in In-home Tracking of glaucoma: Reliability, Acceptability, and Cost

While some clinicians believed that home monitoring could free up staff time, others stressed a number of situations where additional hospital resources would be needed. First, extra staff time will be needed to review the device data and take relevant actions. Second, unreliable data from the tonometer or data incompatibility with the existing medical records can increase the burden on NHS staff. Third, signs of disease progression from the device data may trigger contacts to HES if device readings are visible to home-monitored patients.

Regarding data collected from home monitoring, as the device data are currently not well integrated with the existing medical records, software development may be needed to resolve the issue. Clinicians also mentioned the potential use of AI, for example to interpret data and only highlight to clinicians any abnormal findings. AI may help to conduct some primary review of home monitoring data, making clinical decision-making more efficient.

Clinicians viewed the current home monitoring device used in the I-TRAC study as an expensive technology due to the equipment cost and the maintenance and patient training needed; however, technology advancement might help to reduce these costs. Home monitoring was also seen as a way to manage overtreatment and thus reduce unnecessary costs of treatment. In the current post-pandemic NHS environment, clinicians are uncertain on the time of the next patient assessment. Therefore, clinicians might opt to treat patients instead of maintaining them under no-treatment and observation.

In-home Tracking of glaucoma: Reliability, Acceptability, and Cost site staff

Several issues which may trigger further resource use were mentioned in the interviews with site staff. First, issues with logistics were reported, including challenges with the equipment return and booking appointments with patients after they finished their home monitoring period. Moreover, additional devices might be needed if these get lost or delays are experienced during the return process (e.g. due to the time needed to get the devices ready for the next patient). While these issues are important for any home monitoring service provision, they are particularly relevant to a future study.

Second, the resource use implications on patient training were mixed. On the one hand, site staff reported a decrease in staff time on patient training due to increased experience. In addition, site staff brought in the idea of group training, which can be more time-efficient and reduce staff time devoted to training. On the other hand, as participants might not fully use the provided support at the site, additional technical support might be needed during the monitoring period. In addition to the devices being used by patients, extra devices are required to allow site staff to train the next cohort of patients using the monitoring devices. These requirements imply additional equipment and staff cost.

Eligibility criteria

Intervention

We included studies conducting economic evaluation of any glaucoma monitoring strategies in any setting; that is, in or out of hospital.

Comparator

Eligible studies must state the group comparator (e.g. standard care or usual care) against the intervention group. Eligible studies can also include a situation in which no active or organised monitoring service is available.

Outcomes

The primary outcome of interest was HRQoL outcomes, costs and quality-adjusted life-years (QALYs).

Study type

The qualified studies must be full economic evaluations [i.e. cost-minimisation analysis (CMA), cost–consequence analysis (CCA), cost-effectiveness analysis (CEA), cost–utility analysis (CUA) or cost–benefit analysis (CBA)]; that is to say, both health outcomes and costs for the intervention and comparator should be considered.

Eligibility criteria

General characteristics of included studies

We identified 27 records from MEDLINE and EMBASE. After initial title/abstract screening and deduplication, 22 were excluded, leaving 5 for full-text screening. 65–69 A variety of health conditions were included in these systematic reviews. Following advice from the clinicians in the project team and with support of the SSC, we decided to focus the review only on home monitoring of hypertension due to the comparability between hypertension and glaucoma in disease progression, that is, slow, with no immediate action required on the part of patients post monitoring. For this reason, four systematic review studies that focused on non-hypertension diseases were excluded,65–68 leaving the study by De Guzman et al. as the only systematic review deemed eligible for inclusion (Figure 6 for the PRISMA chart). This is a recent systematic review that includes economic evaluation studies of remote monitoring for several chronic diseases (including hypertension) published from inception until September 2021. 69 Studies on hypertension remote monitoring from this systematic review were extracted and analysed. A list of initially excluded studies and further excluded studies in full-text review, and reasons for exclusion, is presented in Appendix 9.

FIGURE 6.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses diagram for the search of systematic reviews of economic evaluation studies on home monitoring.

Table 25 provides the general characteristics of the primary studies deemed eligible from the identified systematic review. In total, eight primary studies were included for data extraction,70–77 seven70–74,76,77 of which were published within the latest 10 years. Five studies targeted patients with hypertension (with or without age limits) and were conducted in Europe,71–73,75,76 two were conducted in North America,74,77 and one70 was conducted in Asia. In all included studies, the standard care of BP monitoring for hypertension patients was reported as a primary care setting. For the interventions, telemonitoring using a RPM device was the main intervention component reported in all studies, but some interventions also included remote counselling services. 72,74,77

Most studies attempted to explore the cost-effectiveness of the interventions from the perspective of the healthcare system or payer, while one study also analysed this issue from the perspective of personal social services. 73 In terms of the analytical model used, four studies conducted a trial-based economic evaluation only,70,71,75,77 three were extrapolation studies using a Markov model,73,74,76 and one72 conducted a trial-based economic evaluation together with an extrapolation analysis. However, the trial periods were generally not long (i.e. all ≤ 1 year), implying that a trial extrapolation is needed in conjunction with the trial to estimate the long-term effect of the intervention. No study used a system-based model for home monitoring for hypertension. Regarding the functionality of the RPM device used in the interventions, the majority of studies reported that the devices allowed remote BP reading by health professionals, while patients in one study19 had to write down the readings from a self-monitoring BP device and post them to the doctors for review. The functionality of the device used by Padwal et al. 74 was not detailed. Four studies reported that the BP devices used in their interventions enabled remote communication with doctors,70–72,75 yet the device used by Dehmer et al. 77 did not allow this functionality. Information is not available in the other three studies. 73,74,76

Selection of most appropriate outcomes

The objectives of I-TRAC were to assess the outcomes of feasibility and acceptability regarding digital technology for home monitoring glaucoma. The clinical outcomes collected at baseline and 3-month follow-up centred on IOP and VFs. We collected IOP and visual function data weekly during the home monitoring period. Clinical outcome data were collected and reported for descriptive purposes only, with no statistical testing undertaken. One of the site staff interviews felt that the outcomes being collected are of benefit to different subgroups of glaucoma patients and those who would benefit from both were in the minority. This links back to eligibility and who the patient population are. It is worth noting that IOP is the most frequently selected primary outcome in glaucoma trials, with over 90% of all glaucoma trials registered reporting this outcome as the primary end point, and is considered a core outcome. 82,85

It is also worth considering the secondary outcomes required for a future evaluative study. The qualitative research generated data from clinicians that identified a range of important additional outcomes, including: improved management from increased data collected, detecting progression quicker and preventing sight loss (either directly from monitoring or indirectly due to clinicians spending more time with high-risk patients in clinic); patient anxiety; and reduced travel. In addition to the resource use categories such as time for patient training, ongoing patient support, and healthcare contacts triggered by information from the home monitoring devices, a future study should consider health-related (e.g. QALYs) and non-health-related QoL outcome measures. Patient participants stated convenience as a benefit of home monitoring and this is a possible source of non-HRQoL utility; however, they also mentioned concerns about the potential lack of reassurance from clinicians under home monitoring. These are examples of possible sources of utility that should be explored further when deciding the outcome measures for an economic evaluation within an evaluative study.