The potential and limitations of AI in home care – the users’ view

This week, the English parliament approved a new “health and social care” tax, corresponding to an increase in National Insurance contributions from 12% to 13.25% of salary (i.e., a whopping 10.4% increase!!). This increase is to pay for the home care needs of older people, disabled citizens, and others with high care needs. That is, for carers to visit these citizens in their own homes, and help “with tasks such as washing, dressing, eating and taking medication.”

Home care is not only very costly, but also very difficult to source because there is a shortage of qualified staff for this type of care (made worse with immigration restrictions). Moreover, as we are dealing with vulnerable people, there is also the risk of abuse (physical, emotional, financial…) by unscrupulous staff and agencies.

Homecare seems like an area where technology could be very helpful – not just in terms of being available 24/7 (whereas home care is only for a few hours a week), but also in terms of being cheaper and its performance less variable than human carers. Local authorities could, for instance, use a combination of sensors and motion detectors, coupled with machine learning algorithms, to monitor the movements of vulnerable adults at home, and alert someone when there is an alteration in behavioural patterns, as discussed in the paper “Sensors in Smart Homes for Independent Living of Elderly”, authored by Pireh Pirzada, Neil White and Adriana Wilde.

Image source

Such a scheme is being trialled in a pilot study in the UK. Though, to be fair, the whole idea of assistive technology in the home (also known as “telecare”) is not new. Numerous attempts have been made over the years, with the technology generally proving to do really well… but users proving to be very resistant.  

The key to understanding this resistance is to focus not on what the technology can do, but, instead, on what (or, rather, who), the technology replaces. This is what researchers Martina Čaić, Gaby Odekerken-Schröder and Dominik Mahr found, in their investigation of use of robots in elderly care. The researchers adopted a care network perspective, looking at how the robots fitted in the overall care network of the research participants, including relatives, neighbours and formal caregivers. Čaić and her colleagues reported the findings in a paper entitled “Service robots: value co-creation and co-destruction in elderly care networks”, which was published in the Journal of Service Management, and is available here.

The researchers found that a vulnerable person’s care network offers support in three key areas of that person’s life:

  • Physical health – in the form of safeguarding
  • Psychosocial health – via social contact
  • Cognitive health – through support of cognitive functions

The participants could see that assistive technology might support all of these areas. Namely:

  • Physical health – Assistive technology could enable independent living to the elderly person, while providing relief and reassurance to the carers
  • Psychosocial health – Assistive technology could communicate on behalf of the elderly person if they needed something; and reduce the burden that they place on their relatives (in terms of time and money).
  • Cognitive health – Assistive technology could step in if the elderly person loses their mental abilities.

However, the participants could also see disadvantages of using assistive technology. They were:

  • Physical health – Assistive technology was seen as an intruder, who might expose to their care network things that the elderly person would rather keep private. 
  • Psychosocial health – Assistive technology was deemed to lack tacit knowledge and understanding; and there were fears that the technology would replace – rather than complement – the humans in the care network.
  • Cognitive health – Assistive technology might reduce the elderly person’s cognitive abilities by pre-empting their needs and stepping in to the things on their behalf.
Image source

When it comes to home care, there is a lot more that is provided than doing the shopping, preparing the meal, or giving medication. And, for that, technology is still very much lacking, despite recent developments.

Would you like to have this kind of assistive technology in the home?

3 thoughts on “The potential and limitations of AI in home care – the users’ view

  1. Over the last few months, I have engaged with Social Care services in the city where we live regarding my mother, so I read this blog post with interest. I want to add my thoughts to your discussion based on those experiences.

    I can confirm there is a shortage of staff. My mother was given an Emergency Care Home package. This plan included a person attending my mother four times a day to feed, wash and administer her medicine. In the ten days that the plan was effective, no one, I repeat, no one came because the staff were not available. The plan ended after ten days because my mother fell and broke her collar bone, so she was admitted to A&E at the local hospital. If technology had been used, perhaps this unacceptable level of care could have been avoided.

    The point you make about using technology to help care is a good one. I notice that current ideas are centred on what can technology do or what can it replace. To gain the maximum benefit of technology, one should examine how it can augment human and machine systems. Using my experiences with my mother, I would suggest the following. A hospital bed was installed so that my mother could climb in and out of bed easily. Unfortunately, she fell when returning to her bed. The risk of a fall was identified when Social Care reviewed my mother’s needs, but no technological solution was implemented. If a webcam with a speaker and microphone had been installed, the level of care would have increased. The person monitoring my mother could see if she had fallen, questioned her about the fall, and taken appropriate action. If you multiply this up by everyone needing some form of Social Care, you could have a control room employing one person to monitor several tens of people. In so doing, the length of care would have increased from four visits a day to 24/7 monitoring. You may also need a person to respond to emergencies as not all solutions would involve summoning a paramedic. People may argue that it is better to employ one extra carer than two people to monitor and respond. The question then becomes, does the need of many outweigh the needs of the few? The answer to this question easily forms the topic of another blog post. Also, some people may say that monitoring so many people may result in some delay. Again, technology can assist.

    You mentioned that people were resistant to wearing accelerometers and other technological devices. Yet intelligent clothing exists. Why not provide the people in care with intelligent nightwear as part of the care package. So when my mother fell, this would have raised the alarm at the control centre. The operator could then check the webcam in the flat and take the appropriate action. In this instance, technology is helping and prioritising care needs.

    Not all care is medically related; in my mother’s case, I would still have to shop for food, wash her laundry, and clean the flat. This situation is an opportunity for the Internet of Things (IoT) to help. Robot vacuum cleaners already exist, so that can help with the cleaning. Where the Internet of Things would help the most is in shopping. Apart from the washing machine saying that washing powder and conditioner are needed, food and medicine orders are where significant benefits can be obtained. With help from nutritionists, the system can be programmed with a list of suitable foods to ensure that the patient has a balanced and healthy diet. It would also be programmed with food allergies and any specific information. Using my mother as an example, any pain relief has to be soluble as she cannot swallow tablets. The fridge or cupboard would then order food for delivery as needed.

    Although work has to be carried out to establish the role, nature, and benefits of technology in Social Care, a different approach is needed when it comes to implementation. If you take the study you quoted, they list the advantages and disadvantages as seen by the participants. The first one was:

    ‘Assistive technology was seen as an intruder, who might expose to their care network things that the elderly person would rather keep private.’

    If a health care professional had described how the technology would look after the patient, that is, the benefits, rather than the scientific facts were explained, the resistance to the technology would be less. This point then raises a question about research and medical care. Should the researchers stick to the pure facts that then affect future academic research and thinking, or should it include practical aspects that promote the benefit and adoption of the technology? The answer to that question is probably for another blog post.

    I agree with your conclusion that technology is lacking in this area. I am not sure that to address this issue, the answer is to leave it purely to academics and researchers. When research involves Social Issues, a member of the research team should not be an academic but a practising, experienced and professional practitioner in the area being researched. That way, science and society can work together to produce better, practical and needed solutions.


    1. Thank you for sharing your experience, Tim. I am really sorry to hear that the support programme fell through, eventually leading your mother to end up in A&E (where she is now vulnerable to Covid-19, and other infections).

      I think that the issue with resistance to the technology flagged in this study (and, to be fair, many others, previously) is that they look at people that currently have some level of at-home care. So, when they look at the technology, they are comparing it with their human counter-parts, rather than going from no support to some support enabled by technology. Psychologist call it the anchoring effect. Moreover, we all tend to give more weight to what we stand to lose (in this case, the level of care already in place) than to gains of equivalent magnitude (the endowment effect).

      I had not thought about the “needs of the many vs the few”, as you mentioned. That’s an additional interesting issue, probably compounded by the short-term focus of those in need of care vs the long-term focus of those managing and proposing technological solutions.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s