SAMS - Software Architecture for Mental health Self management

Lead Research Organisation: Lancaster University
Department Name: Computing & Communications


SAMS is a proposed 3-year project to that will investigate the potential for novel data and text mining techniques for detecting subtle signs of Cognitive Dysfunction that may indicate the early stages of Alzheimer's disease. Promoting self-awareness of change in cognitive function is will investigate the potential for novel data and text mining techniques for detecting subtle signs of change in cognition that may indicate the early stages of Alzheimer's disease. Promoting self-awareness of change in cognitive function is a key step in encouraging people to self-refer for clinical evaluation. A key motivation for SAMS, therefore, is to provide a non-invasive tool that helps develop such self-awareness.
An increasing number of older people, the group most at risk of cognitive dysfunction and dementia, regularly use the Internet to keep in touch with their families, particularly grandchildren. This Internet activity presents an opportunity to harness rich, routinely available information that may contain indications of changes in the linguistic, executive and motor speed abilities in older people.
Development work is needed to develop the software to harness this opportunity, to establish the optimal thresholds for flagging up important changes in cognition and the optimal methods for feeding this information back to individuals. SAMS will validate thresholds by examining changes in performance in people with established cognitive dysfunction and mild Alzheimer's disease and begin to explore the potential for technology-enhanced detection of early cognitive dysfunction. Patterns of computer use and content analysis of e-mails, such as forgetting topics, expressions of concern, emotion, etc., will be analysed and coupled to feedback mechanisms to enhance users' cognitive self awareness, empowering them self administer follow up tests and decide when to self refer themselves for expert medical advice.
Tackling cognitive change detection requires the novel pooling of knowledge and integration of techniques from different sub-disciplines within a Computer Science. In addition to developing techniques for MCI detection and supporting self-referral, an explicit goal of the research is to develop a generic sense making and user-centred feedback architecture. This could be applied to a wide range of problems where interpreting computer use may be appropriate, e.g. mental health, social loneliness, privacy and social exploitation.

Planned Impact

From strand 1, the elderly and other potential patients of Alzheimer's disease will be the main beneficiaries as the prototype software will be released as an operational product by the Alzheimer Society, and NHS-related organisations such as e-Heath NW and the Institute for Health Sciences in Manchester and the MRC Centre in Kings. The health impact will be a decrease in the number of people suffering from the debilitating effects of Alzheimer's disease because early interventions will delay progress of symptoms and future research may eventually prevent onset. More widely, the SAMS system will improve detection of, and intervention for a wide range of ageing and mental health problems. The tools will enable planning of home visits to encourage follow-up diagnosis as well as long-term management of dementia-related illness in the community, and coordinating with family and close friends of the patient. Wellbeing among the elderly will be improved by the reassurance afforded by awareness and self management of cognitive age-related problems.

Wider economic impacts will accrue from of reduction in healthcare costs resulting from delaying the onset of Alzheimer's disease. Given the projected increase in the ageing population and hence exposure to cognitive impairment, there will be savings for the NHS, with fewer patients hospitalised after dementia-related accidents and reduced community care; with even greater savings for local authority home care budgets as full-time care home/home visit support will be reduced. Taking a very conservative estimate of delaying the onset of incapacity related to dementia by one year of 10 % within the current UK patient population, assuming care cost savings of £20K per annum, the saving is £1,720 million recurrent per annum. Other economic impacts will arise from exploitation of the generic software architecture (strand 2) in a range of health management, education and related people centric tracking applications. The potentialorldwide impact is obvious with ageing populations in all OECD countries and life long education being an international priority.

From strand 2, companies in the general IT industry with interests in monitoring, tracking and feedback/management applications based on interaction data and text document context will benefit from the generic architecture and reusing its components in products ranging from marketing recommenders, to education and training progress trackers
Description We undertook a requirements elicitation exercise that has helped us understand the features that are essential of the SAMS software is to be accepted and used by its target user group.

As a side-effect, we had to develop a novel mix-mode requirements elicitation technique for home healthcare-monitoring systems for senior users.

We also developed experience in eliciting clinical expertise to inform the requirements for an analytical, diagnostic software tool, in the challenging situation where (a) consensus among the clinicians was imperfect, and (b) the envisioned software tool was outside of the experts' previous experience. The experience we gained has been developed into a journal paper which is currently (as of March 2018) in its second round of revisions.

The SAMS software was been installed on 32 consenting older adults' home computers and logged encrypted and anonymised data collected from computer use to a server running in Lancaster University.

The SAMS software was engineered to work across a range of operating systems, applications and web browsers. A great deal of effort was devoted to making it robust and resilient and benign with respect to its affect on users and their use of their computers.

Analysis of this data continues as resources become available post-project. It is being informed by the results of an earlier (also in 2015) cross-sectional study in which cohorts of healthy adults and people with Mild Cognitive Impairment or early Alzheimer's Disease performed a set series of tasks on a computer. Early analysis of the results suggests marked differences in performance between the two cohorts in terms of (e.g.) speed of completing tasks.

The data from the longitudinal study is still in progress. There is a large volume of data to analyse.

The text analysis too completed near the end of the project (SNOWCAT - see Software) has now been validated on two previous studies in the literature; one on novels and one on press conferences. This has broadly confirmed the findings of these earlier studies but also revealed that propositional density is not a reliable indicator of language decline in dementia, as was suggested in a third earlier study.

This work has been taken forward by an Aston University-funded PhD student who is investigating the potential of machine learning to infer cognitive decline consistent with transitioning to dementia by people with a diagnosis of MCI through the analysis of spoken text.
Exploitation Route The requirements elicitation method is likely to have applications to other discretionary use software systems, particularly those that are in affect-laden domains such as mental health monitoring.

The SAMS monitoring software and the overall systems architecture of which it forms a part could serve as a reusable resource for other projects. Some of the architectural lessons have informed the of logging eye-movement data in EP/M006255/1.
Sectors Digital/Communication/Information Technologies (including Software),Healthcare

Description Early Detection of MCI that progresses to dementia - An Interdisciplinary Approach
Amount £60,000 (GBP)
Organisation Aston University 
Sector Academic/University
Country United Kingdom
Start 04/2018 
End 03/2021
Title The SAMS system 
Description SAMS monitors computer interaction to identify early signs of the onset of cognitive decline that may signal the development of dementia. At the end of the SAMS project, the SAMS monitoring system (See under Software) had been deployed to over 30 older adult users's Windows computers, and data collected over a c. 8-month period. This data is still being analysed, but the data itself forms a potentially valuable resource for other researchers. 
Type Diagnostic Tool - Non-Imaging
Current Stage Of Development Initial development
Year Development Stage Completed 2016
Development Status On hold
Impact The impacts to date relate to the development of monitoring software, interns of: - software architecture - privacy and security - software reliability - user requirements The understanding that has been developed is likely to be of benefit to developers of ambient healthcare monitoring systems. 
Title SAMS monitoring system 
Description The SAMS monitoring system runs as a background application on Windows 8 and Windows 10 machines, collecting a variety of data on user interactions with the operating system and with a range of Windows applications. Any users on whose computer the SAMS software runs are assigned a unique identifier by a designated member of the SAMS clinical research team. The collected data is tagged with this user ID and stripped of all other information that could be used to identify the user. It is then encrypted and periodically uploaded to a server. The server stores the data securely for use by the SAMS research team. Only the designated members of the clinical research team is able to associate the user IDs with which the data is tagged with the actual user identity. 
Type Of Technology Software 
Year Produced 2016 
Impact To date impact is listed to research publications as the software is tailored to the goals of the SAMS project. There have been discussions about using it for new mental and brain health research projects but these are still at an early stage. 
Description SNOWCAT is a tailored integration of several text analysis techniques to detect signs of decline in language consistent with dementia. Given a set of documents (e.g. diary entries over a period of months), a range of metrics is produced that summarise the author's use of language. 
Type Of Technology Software 
Year Produced 2016 
Impact To date, replication of published analyses of novels and press conferences. 
Description Attendance of Dementia Catalyst Event 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Professional Practitioners
Results and Impact Prof Burns and prof. Sawyer attended a one-day Dementia Catalyst Event organised by the Alzheimer's Society in May 2016 at Aintree racecourse. Prof Burns gave the keynote address, while Prof. Sawyer participated in discussion and working group break-out sessions on the role of technology in dementia diagnosis and care.
Year(s) Of Engagement Activity 2016
Description Presentation at International Psychogeriatrics Association International Congress 2016 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Professional Practitioners
Results and Impact This was a talk, selected by submission of an abstract, by members of SAMS' clinical team on the project. The audience was a mix of researchers and practitioners. The talk generated interest from other researchers and stimulated early plans for future international collaboration, which are in progress now (2017).
Year(s) Of Engagement Activity 2016
Description Project poster at Alzheimer's Society annual conference 2014 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Professional Practitioners
Results and Impact generated significant interest from the audience who were attending the poster session

The poster raised awareness about the project and anticipation of the final results.
Year(s) Of Engagement Activity 2014
Description Public Dementia Futures Event, Lancaster Town Hall, 18th September 2015 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach Regional
Primary Audience Public/other audiences
Results and Impact The focus of the event was on Dementia Futures, bringing together academics from Lancaster University, clinicians and practitioners, city councillors and members of the public to hear about some of the latest developments in dementia research and practice. A room was set aside for demonstrating a range of interactive activities used in research that proved a huge success with those attending.
Year(s) Of Engagement Activity 2015
Description Royal College of Physicians mature healthcare IT principles workshop 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Professional Practitioners
Results and Impact The workshop was in part a response to Professor Robert Watcher's report "Making IT Work: Harnessing the Power of Health Information Technology to Improve Care in England", which recommended principles for the adoption of digital technology by the NHS. However, Watcher took a traditional approach: fixing the issues that were presented to it with the tools currently in use. It assumed the benefits we experience from digital in our daily lives could simply transfer to healthcare.
Rather than critique Watcher, it was realized the report could be used as a platform to strengthen systemic and strategic thinking not just for the next few years, but to navigate forward:

• What were the critical questions that Watcher did not ask?
• Thinking about possible futures for the NHS, what are the biggest questions we should be asking?
• How can we devise powerful and enduring guiding principles for the development and adoption of digital healthcare that would enable us to be more strategic and discerning in digitisation or in developing "IT strategies"?
Year(s) Of Engagement Activity 2017