A closed eye on dementia

13 May 2019

Helen Rostill describes how artificial intelligence can monitor people with dementia in the home to reduce hospital admissions and support struggling carers.


People with dementia often feel happier if they can remain independent and in their own homes for as long as possible. Currently, two-thirds of the 850,000 people with dementia in the UK live in their own homes, and a third of these live alone (Alzheimer’s Society, 2014; Mirando-Costillo et al, 2010).

However, the complex needs of a person with dementia combined with the lack of post-diagnostic support available in the community can mean a poor quality of life. People with dementia who are over 65 have on average four comorbidities, while people without dementia have on average two (Poblador-Plou et al, 2014).

The result is that carers and families often find themselves struggling to provide the necessary support. There are 500,000 carers for people with dementia in the UK, and 63.5% say they have either no support or not enough (NHS Digital, 2017). In addition, nearly half of carers (48.4%) have a long-standing illness or disability themselves (NHS Digital, 2017).

Alongside the human cost, dementia has major implications for health and care services. One in four people in acute hospital beds has dementia (Lakey et al, 2012), as do 40% of acute admissions in those over 70 (Sampson, 2009).

However, there is evidence to show that if a system were in place to aid early intervention, the number of unplanned hospital admissions for people with dementia could be reduced. Research shows that four of the five most common comorbidities for which people with dementia are admitted to hospital in the UK are for preventable conditions (Scrutton and Urzì Brancati, 2016).

The key aim of the Technology Integrated Health Management (TIHM) for Dementia trial is to provide a system that improves both early intervention and the quality of life of people with dementia and their carers.

Led by Surrey and Borders Partnership NHS Foundation Trust, and in partnership with the University of Surrey, the Alzheimer’s Society and technology provider Howz, TIHM uses ‘Internet of Things’ (IoT) technology by combining a range of digital devices with artificial intelligence to remotely monitor the health of people with dementia around the clock to help them stay well in their homes for as long as possible. TIHM has won many plaudits, including a 2018 HSJ award for improving care with technology, and an NHS70 parliamentary award.


How does TIHM work?

A network of internet-enabled devices installed in the home continuously collects data about a person’s health and environment. Sensors monitor sleep and movement inside the home; smart plugs record usage of electrical appliances such as fridges, kettles and toasters; interactive monitors register body temperature, blood pressure, heart rate, weight and hydration; and a GPS device tracks a person’s whereabouts if they go outside. In addition, people with dementia are asked to answer three questions each day, sent to them via phone, about their health and how well they slept the night before.

Data streamed from all of these devices is integrated and analysed, and machine-learning algorithms are used to identify important or unusual changes in a person’s health or behaviour that could signal they are becoming unwell (see Architecture of the TIHM for Dementia system). For example, their blood pressure reading may be out of their normal range, they may be dehydrated or have walked too far from home, or they may be developing a urinary tract infection (UTI) or showing signs of agitation.

If the technology identifies a physiological, environmental or technical problem, an alert is flagged and prioritised on a digital dashboard known as the ‘integrated view’ and immediately followed up by a monitoring team. The team is guided by algorithms and their own judgement to determine what action is needed. This may involve contacting the carer, asking the person with dementia to take a second reading of a vital sign, advising them to see their GP or, if necessary, contacting the emergency services. All interactions are recorded and followed up by the monitoring team (Enshaeifar et al, 2018).

The design of TIHM for Dementia was influenced by a co-design group of 20 people with dementia and their carers, known as the ‘trusted users’. The group were the early adopters of the TIHM system and provided ongoing feedback about the technological devices and the deployment and monitoring processes throughout the study.


Trial results

The first phase of TIHM for Dementia, a randomised controlled trial, involved more than 400 people with dementia and their carers from across Surrey and north-east Hampshire and was completed at the end of March 2018. People were involved in the study for six months, with half receiving the technologies and half continuing with treatment as usual.

An independent evaluation of the first phase of TIHM by the University of Surrey found that TIHM for Dementia was successfully deployed in people’s homes and that the technology, combined with the monitoring team, was overwhelmingly welcomed by participants, with more than 70% saying they would recommend TIHM to others.

Crucially, the study found that the people with dementia in the technology arm reported a sustained and statistically significant (p<0.05) reduction in neuropsychiatric symptoms, such as depression, agitation, anxiety and irritability. Given the link between the behavioural and psychological symptoms of dementia and transfer to long-term care, this finding is important. A reduction in these symptoms could ease pressures on carers and family and help them to carry on with their caring role rather than see a loved one admitted to hospital or even a care home.

The evaluation report also highlighted the significance of the development of prototype algorithms during the study to detect UTI and agitation, irritability and anxiety (Enshaeifar et al, 2019). UTIs are a top-five cause of hospital admission among this group.

An analysis of 11,487 alerts in the period between July 2017 and March 2018 found that the most common were linked to blood pressure (2483), blood oxygen (1894), agitation (1769) and pulse (1669). The monitoring team responded to alerts by connecting with the person with dementia, and their carer and the majority of the alerts were managed through a simple phone call. A UTI algorithm was designed towards the end of the study and associated with 124 alerts. We also developed a predictive algorithm to detect early stages of agitation.


Architecture of the TIHM for dementia system
The future and impact on CPs

A second phase of the TIHM for Dementia study has just been launched, involving up to 150 people with dementia and their carers from across the same region. The aim of this shorter, more agile study, which will be completed at the end of July, is to refine the range of devices needed in the home to effectively monitor the health of a person with dementia. It will also be used to further develop the UTI and agitation, irritability and anxiety algorithms.

The more refined version of the TIHM system, recently approved as a CE marked device, will then be ready to be adopted by other NHS trusts, general practices and integrated care systems and primary care networks. Pilots with a number of these organisations are already planned.

The development of TIHM in this context is in line with NHS England’s primary care home model, which brings together a range of health and social care professionals to work together to provide enhanced personalised and bespoke care for their local community. The key is to focus on local population needs and to provide care closer to patients’ homes, with a particular focus on the frail elderly.

For the community practitioner, having access to the TIHM for Dementia system could have a hugely positive impact on their work in supporting both people with dementia and also the families caring for someone with this condition.

Community practitioners would benefit from having access to up-to-date health and wellbeing data, at any point in time, about the person they were supporting. This could help to target resources and facilitate more informed and faster decision-making and potentially aid early intervention to keep someone out of hospital or a care home. In addition, the recording of vital sign data by people who have the TIHM technology could also potentially save community practitioners from having to carry out some of these readings themselves.

The extra layer of support provided by the TIHM for Dementia monitoring system could also reduce some of the pressure on families caring for a person with this condition and this may in turn help the community practitioner to better support that family.

Case Studies

  • A 74-year-old woman, discharged from hospital following a 12-week stay for treatment for pancreatitis, triggered a urinary tract infection (UTI) alert within a few days of being home.
  • The GP was immediately informed and visited the lady to confirm the diagnosis.
  • As a result, antibiotics were prescribed and being taken by the woman within four hours of the TIHM UTI alert.
  • A 78-year-old woman triggered a blood pressure alert.
  • TIHM data enabled the GP to prescribe medication immediately. TIHM then identified that the medication was causing frequent urination and dehydration.
  • The TIHM data enabled the GP to promptly optimise the medication.


TIHM for Dementia is part of the NHS Test Beds Programme and is the first programme of work in the UK, or internationally, that has developed and installed an IoT-based system to support management of dementia in the community.

It has proved it has the potential to provide the NHS with a new digital and data-driven type of intervention to support people with long-term and complex health conditions. It could also transform anticipatory care by providing clinicians with up-to-date information indicating when and where support is most needed. By providing timely and targeted support, it should make it possible for people with dementia to remain in their own homes for longer with confidence.

Professor Helen Rostill is director of innovation, development and therapies, Surrey and Borders Partnership NHS Foundation Trust, and senior responsible officer for the TIHM for Dementia clinical trial. 


For more information about TIHM, go to sabp.nhs.uk/tihm

To find out more about assistive technologies, visit bit.ly/AD_assistive_technology

Time to reflect

How could technology help clinicians provide better care in your service? Share any insights and join the conversation at @CommPrac using #TIHMindementia


Alzheimer’s Society. (2014) Dementia UK update. See: https://www.alzheimers.org.uk/sites/default/files/migrate/downloads/dementia_uk_update.pdf (accessed 24 April 2019).

Miranda-Castillo C, Woods B, Orrell M. (2010) People with dementia living alone: what are their needs and what kind of support are they receiving? International Psychogeriatrics 22(4): 607-17.

Poblador-Plou B, Calderón-Larrañaga A, Marta-Moreno J, Hancco-Saavedra J, Sicras-Mainar A, Soljak M, Prados-Torres A. (2014) Comorbidity of dementia: a cross-sectional study of primary care older patients. BMC Psychiatry 14(1): 84.

NHS Digital. (2017) Personal social services survey of adult carers in England, 2016-17. See: tinyurl.com/y6whqnx3 (accessed 24 April 2019).

Lakey L, Chandaria K, Quince C, Kane M, Saunders T. (2012) Dementia 2012: a national challenge. See: tinyurl. com/y8hqfrjr (accessed 24 April 2019).

Sampson EL, Blanchard MR, Jones L, Tookman A, King M. (2009) Dementia in acute hospital: prospective cohort study of preva­lence and mortality. British Journal of Psychiatry 195(1): 61-6.

Scrutton J, Urzì Brancati C. (2016) Dementia and comorbidities: ensuring parity of care. See: https://ilcuk.org.uk/wp-content/uploads/2018/10/Dementia-and-Comorbidities-Ensuring-Parity-of-Care.pdf (accessed 24 April 2019).

Enshaeifar S, Zoha A, Markides A, Skillman S, Acton ST, Elsaleh T, Hassanpour M, Ahrabian A, Kenny M, Klein S, Rostill H, Nilforooshan, Barnaghi P. (2018) Health management and pattern analysis of daily living activities of people with demen­tia using in-home sensors and machine learning techniques. PLoS ONE 13(5): 1-20. 

Enshaeifar S, Zoha A, Skillman S, Markides A, Acton ST, Elsaleh T, Kenny M, Rostill H, Nilforooshan R, Barnaghi P. (2019) Machine learning methods for detecting urinary tract infection and analysing daily living activities in people with dementia. PLoS ONE 14(1): e209909.