Box 8 provides a summary of Chapter 4 of this report.
Summary of key points
- Perhaps unsurprisingly with new and emerging technology, there was a lack of understanding of exactly what AI-based technology can provide, which led to the emergence of a broad range of anticipated benefits.
- The expected outcomes anticipated by multiple stakeholders can be summarised within five domains: increasing preventative care; improving assessments and diagnoses; supporting independent living; providing reassurance for those who draw on care and support and their carers; and the conservation of resources and reduction in costs.
- While some participants could cite anecdotal examples of ways in which some of their aspirations had been met for some people, others provided caveats to these observations or offered alternative perspectives. Overall, participants felt that the sensors were not sufficiently stable or effective to collect reliable data over the necessary period of time.
This chapter explores the first six of our seven research questions, with the final chapter (our Discussion and conclusion) placing our findings in the context of Greenhalgh et al.’s (2017) ‘NASSS framework’37 and the ‘rational model’ of implementation, and addressing our final question: how can the process of implementing new technology be improved?
Desired outcomes
Adult social care faces a series of financial, demographic and service pressures, and sites recognised that new and emerging technology could have a significant role to play. However, there was a lack of clarity as to exactly what such technology might deliver (or how), and different people often had different assumptions about what might be possible (even in the same site). In several cases (and with the benefit of hindsight), the ‘shininess’ of new technology seemed very seductive for sites looking for solutions to a series of different problems and hoping that technology could help. Others reflected that it was hard to know what new and emerging technology might achieve in advance of giving it a go.
In particular, different stakeholders seemed to be trying to achieve five different outcomes, sometimes all at the same time:
- Preventative care: Some people hoped that the digital technology (home-based sensors with AI capabilities) would help understand patterns of behaviour so that changes could be identified and interventions implemented earlier to prevent a crisis or a deterioration in someone’s condition: Part of the selling point of IndependencePlus was that, you know, the machine learning would pick up when somebody’s daily routine had changed and would alert you to that fact. So, you know, the kit would send a text message to a carer saying ‘Usually your mum has five cups of tea by this point and today she’s only had one’, you know. ‘Do you want to check this out?’ or, you know, ‘Your mum’s usually out of bed by this time; she hasn’t got up yet, might be worth going round’.(CS2 P05)
- Improve the accuracy of assessments and diagnoses: Other people suggested that the increased detail captured by a broader range of more sensitive sensors might enable better informed and more accurate assessments, leading to care packages that more fully met people’s needs: We were hoping it would help them with the assessment process and help them to genuinely understand how people use their homes and therefore what their needs were, so if that person wasn’t really showing, for example, that they were making or seeming to be making themselves regular drinks, that’s something that we would be able to factor in, because sometimes when you speak to someone and they say ‘oh yes I eat very regularly and oh yes I’ve had no trouble at all making a cup of tea’, but they’re either telling you what they think you want to hear or forgetting or fibbing or something, so we felt it would be a useful kind of tool to help a professional really understand the holistic needs of someone in not too intrusive a way.(CS2 P04)
- Promote independence: The more nuanced data produced by the digital technology might enable a more accurate determination of the potential risks faced by those who draw on care and support and better inform mitigating interventions: [Our] ethos is to help people … to live and work as independently as they choose and that’s very much what we’re trying to do in the assistive technology project.(CS1 P02)
- Increase reassurance for those who draw on care and support and their carers: The flow of real-time data could be monitored by family members (in situations where people had family to take on this role), thereby providing reassurance as to their relative’s safety. Some people also hoped that people drawing on care and support would feel reassured and more confident going about their daily lives if they knew that the digital technology was monitoring them and looking out for potential problems: We had a few young people… moving on from living at home but parents were very concerned but wanted to have paid support in the environment, they wanted people in the house all the time, whereas we were convinced, I mean, part of the enablement service is to set people up in their homes, give them daily living skills, set them up into routines so they can manage on their own – we know what they could do – we knew that they would be safe – so this was just a reassuring piece of kit that would hopefully support the parents a bit more so they weren’t so anxious.(CS2 P06)
- Conserve resources and reduce costs: It was hoped that the use of a unified system (containing sensors and an incorporated AI element to support data analysis) would reduce costs. If support could be better tailored to people’s actual needs, then there may also be scope to reduce the amount of staff time spent on providing direct care: We wanted to change the way we did things in order that we can make our money go further basically, and our way of doing that is to take a strength-based approach and to promote independence and we feel that the technology influence and potential is something that really aligns with that vision.(CS2 P03)
Sometimes, different people in the same site identified various of these different outcomes, raising the question of whether any one new way of working could ever deliver all these different aspirations at once. If different stakeholders wanted the technology to achieve different things, it was highly likely that many people – perhaps everyone – would end up disappointed. Table 8 sets out these different outcomes in more detail, alongside often anecdotal examples of where individual participants felt that some of these benefits had (at least partially or for a few individuals) been realised. However, other participants would often offer caveats or alternative perspectives, and perceived responses by people drawing on care and support and carers were largely those reported by senior and operational leads, rather than by these groups themselves.
| Outcomes | Anticipated benefits | Were anticipated benefits met? (Perceptions of different/individual participants) |
|---|---|---|
| 1. Preventative care | Observe changes over time in behaviour patterns | There was enough data captured for one individual (including time spent active, hours in bed asleep, the number of occasions they left the house, etc.) to confirm there was no deterioration over time in their levels of activity. |
| Observe changes over time in physiological functions | ||
| 2. Improve the accuracy of assessments and diagnoses | Increased understanding of daily routines | Sensors demonstrated that the kettle was being used regularly and that the fridge was being accessed, leading the care provider to infer that the individual was eating and drinking frequently. |
| Short-term changes in the data could be used in diagnosis | A number of restless nights recorded by the bed sensor were considered an indication of the onset of vertigo for one person. For another person, the data captured by the bed sensor identified petit-mal seizures at night. | |
| As a source of additional information for health-care professionals | A range of data including heart rate was usefully provided for paramedics called out to attend an individual. | |
| 3. Promote independence | Determine appropriate level of care support | The data demonstrated that one person was capable of preparing their own breakfast, so staff did not need to do this for them. |
| Inform calls to response service (alerts) | The sensors showed that an individual was leaving the house in the early hours, which necessitated a call to response services. Such alerts might mean that the person could continue living in their current accommodation (which might not ultimately be possible if the person went out late at night with no alarm raised). However, as mentioned, in most cases the sensors were not stable enough to collect the reliable data needed to inform calls to response services. As the technology evolved to include more health data, there were also challenges associated with care staff not having the necessary medical training to interpret whether the individual needed medical attention (if data seemed to suggest risk of a deterioration in someone’s health). | |
| 4. Increase reassurance for those whose draw on care and support and their carers | Reassure families as to the safety and well-being of older relatives | There were several examples where staff felt that sensor data were providing reassurance to families – these data were accessed directly by families or used by care providers as evidence of the safety of the individual. However, as mentioned, in most cases the sensors were not stable enough to collect the reliable data needed to reassure families. This also raised issues of what happens in situations where a person does not have family available to take on this role. |
| Reassure people drawing on care and support | Several individuals were felt to be drawing reassurance from the fact they were being monitored, albeit remotely. | |
| 5. Conserve resources and costs | Financial savings from using a unified system | Purchasing the combined AI and sensor technology was felt to be cheaper than purchasing two systems. |
| Involve families in monitoring | A family was successfully engaged in monitoring the data, though baulked at paying for the technology when the pilot had concluded. | |
| Safely reduce staff contact | One site identified a reduction in hours of direct care needed for those with the technology installed. |
Reflecting back on this with the benefit of hindsight, some participants wondered whether this lack of clarity about desired outcomes was connected to the myriad of different possibilities that new and emerging digital technology seemed to offer.
For one person:
What we should be doing is going on an individual level – what is it that this person can do, what is it that they could do with a little bit more support, maybe through technology… – and then going out to look at a range of different options and technology being a part of that option. So, it might be that there’s people out there that have care needs, so you have to have in-person support, but they might want to control their environment a little bit more, so you might look at environmental control devices or commercially based devices like [Alexa] and smart homes and generating things like that. There might be a cohort of people that you are identifying where there’s a specific need, so then you might want to go ‘OK, so we have this cohort, because they keep telling us that this is a problem that we need to fix, here’s now a technology that can suit that cohort’, so it might be a remote monitoring device or something similar and then you’d go out to procure a lot of those different devices for that cohort.
For another participant:
The problem is people probably want all of it and they can’t have all of it, so you need to decide what’s your highest priority. Is our highest priority to know when somebody’s fallen over so we can go and pick them up and maybe get them to hospital, is that our highest priority? I don’t know… Or is our highest priority to have lots of data about people so when we come to review them or assess them we make better decisions?... Or is our priority something as simple and practical as I want a very good automated meds dispenser for people who are able to take their own medication because overnight that saves me about 500 hours of care a week and, what’s that, a lot of money?
Framing this in a slightly different way, others felt that there was a risk of being seduced by the apparent lure of new technology (especially given the quality of the IndependencePlus ‘pitch’ discussed in Chapter 5):
Where we’d got it wrong was we’d got this shiny bit of kit and we were trying to fit it in somewhere and make it work somewhere. So the big lesson we learnt was actually don’t start off with the kit or the technology, you start off with the people and the need. And so the question that we now ask ourselves when we look at technology in social care is: right, what is the need this person might have; where our traditional social care package is not meeting that need, and in between that gap, is there a piece of technology that could fit that gap? So that’s the big lesson that we learnt in our approach to technology in adult social care which you know, it sounds obvious now, but yes, we’ve learnt not to be taken in by the shininess of a kit or how good a sales pitch might be.
Responses also seemed to suggest that sites were struggling with whether to experiment with new and emerging technology to see what it could do, or whether to start with people’s needs and reflect as to whether new technology might have a role to play. Thus, one site was motivated in part by a curiosity to see what new and emerging technology was capable of, rather than having a pre-planned sense of what might be achieved:
We were all kind of, like, ‘I don’t know!’, like, ‘we don’t know what things will happen, we don’t know what the possibilities are until we have it’. So I think that was part of our interest, was actually like we’re not sure what we’re looking for but we think we’ll know if we find it!
However, another site expressed the opposite view:
I often explain it that we used to, when we were going out for like IndependencePlus and other technology projects in the past, we used to think what technology’s out there and let’s go and buy it and now let’s look at people that we provide care for and fit them to that technology. That is probably the biggest learning, that that’s a mistake. We shouldn’t be doing that.
For several participants, the lack of clarity around outcomes was made worse when the focus of the digital technology changed during lockdown. During the early stages of the COVID pandemic, installation was paused and IndependencePlus made a number of changes to the system before it was reintroduced, so that the focus became monitoring and recording health-related data:
They took away the behavioural notifications that we had started to receive, so we’re no longer getting prompting on this person’s not hydrating as much as they usually do or that they’re not going into the fridge as much as they usually do, all those kind of things. Instead we were getting notifications around heart rate spikes or heart rate drops, with notifications set up for blood SpO2 and temperature.
Hardly surprisingly, this evolution or pivot in the focus of the technology made it difficult to achieve desired outcomes, as the nature of the technology was different at the end of some of the pilots than at the start.