This systematic literature review explores the emerging field of remote-based deep learning predictive algorithms for depression, focusing on addressing the limitations of traditional diagnostic methods and examining the current state of research in this novel area. A systematic search was conducted in Embase, Medline, Web of Science Core Collection, CINAHL, and PsycINFO in June 2023. To capture relevant studies, titles and abstracts of the papers were reviewed against predefined inclusion and exclusion criteria using four groups of keywords addressing prediction, depression, validity, and deep learning. Eligible studies were systematically reviewed based on the Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling Studies (CHARMS) checklist. The risk of bias was assessed using the Prediction Model Risk of Bias Assessment (PROBAST) Tool for methodological quality. The synthesis of data was conducted using the Synthesis Without Meta-Analysis (SWiM) framework. From 286 studies initially identified, 6 studies met all inclusion criteria, published between 2020 and 2023. Performance metrics revealed the potential of deep learning models, with accuracy rates reaching as high as 98.23

This paper addresses debates concerning the ‘participatory turn’ in healthcare. It focuses on the case of blood pressure self-monitoring, understanding this as a form of patient participation at the level of individual care. Drawing and expanding on the work of Marres and Wynne and their notions of material participation and of uninvited engagement, we examine how patients’ home blood pressure self-monitoring is incorporated into clinical care, how the materials of blood pressure self-monitoring mediate participation and how we might characterise the practices of participation found within everyday clinical care. Our analysis makes new conceptual links, suggesting that, in this context, invited participation appears to align with participation made easy, while uninvited participation involves more invested, more engaged participation. We offer two further developments of these concepts. First, we trouble characterisations of invited and uninvited participation as distinct and separate, observing movement between these. Second, through applying the logics of material participation in a new context, everyday clinical practice, we suggest that the logic of participation made easy might be extended beyond lay people, to apply to professionals as well. Our analysis illustrates how materials are mobilised to facilitate invited and uninvited participation within the context of the on-going asymmetries in doctor patient relationships.

Background:Multiple sclerosis (MS), epilepsy, and depression are chronic central nervous system conditions in which remote measurement technology (RMT) may offer benefits compared with usual assessment. We previously worked with clinicians, patients, and researchers to develop 13 use cases for RMT: 5 in epilepsy (seizure alert, seizure counting, risk scoring, triage support, and trend analysis), 3 in MS (detecting silent progression, detecting depression in MS, and donating data to a biobank), and 5 in depression (detecting trends, reviewing treatment, self-management, comorbid monitoring, and carer alert). Objective:In this study, we aimed to evaluate the use cases and related implementation issues with an expert panel of clinicians external to our project consortium. Methods:We used a Delphi exercise to validate the use cases and suggest a prioritization among them and to ascertain the importance of a variety of implementation issues related to RMT. The expert panel included clinicians from across Europe who were external to the project consortium. The study had 2 survey rounds (n=23 and n=17) and a follow-up interview round (n=9). Data were analyzed for consensus between participants and for stability between survey rounds. The interviews explored the reasons for answers given in the survey. Results:The findings showed high stability between rounds on questions related to specific use cases but lower stability on questions relating to wider issues around the implementation of RMT. Overall, questions on wider issues also had less consensus. All 5 use cases for epilepsy (seizure alert, seizure counting, risk scoring, triage support, and trend analysis) were considered beneficial, with consensus among participants above the a priori threshold for most questions, although use case 3 (risk scoring) was considered less likely to facilitate or catalyze care. There was very little consensus on the benefits of the use cases in MS, although this may have resulted from a higher dropout rate of MS clinicians (50%). Participants agreed that there would be benefits for all 5 of the depression use cases, although fewer questions on use case 4 (triage support) reached consensus agreement than for depression use cases 1 (detecting trends), 2 (reviewing treatment), 3 (self-management), and 5 (carer alert). The qualitative analysis revealed further insights into each use case and generated 8 themes on practical issues related to implementation. Conclusions:Overall, these findings inform the prioritization of use cases for RMT that could be developed in future work, which may include clinical trials, cost-effectiveness studies, and the commercial development of RMT products and services. Priorities for further development include the use of RMT to provide more accurate records of symptoms and treatment response than is currently possible and to provide data that could help inform patient triage and generate timely alerts for patients and carers.

BACKGROUND:A variety of smartphone apps and wearables are available both to help patients monitor their health and to support health care professionals (HCPs) in providing clinical care. As part of the RADAR-CNS consortium, we have conducted research into the application of wearables and smartphone apps in the care of people with multiple sclerosis, epilepsy, or depression.METHODS:We conducted a large online survey study to explore the experiences of HCPs working with patients who have one or more of these conditions. The survey covered smartphone apps and wearables used by clinicians and their patients, and how data from these technologies impacted on the respondents' clinical practice. The survey was conducted between February 2019 and March 2020 via a web-based platform. Detailed statistical analysis was performed on the answers.RESULTS:Of 1009 survey responses from HCPs, 1006 were included in the analysis after data cleaning. Smartphone apps are used by more than half of responding HCPs and more than three quarters of their patients use smartphone apps or wearable devices for health-related purposes. HCPs widely believe the data that patients collect using these devices impacts their clinical practice. Subgroup analyses show that views on the impact of this data on different aspects of clinical work varies according to whether respondents use apps themselves, and, to a lesser extent, according to their clinical setting and job role.CONCLUSIONS:Use of smartphone apps is widespread among HCPs participating in this large European survey and caring for people with epilepsy, multiple sclerosis and depression. The majority of respondents indicate that they treat patients who use wearables and other devices for health-related purposes and that data from these devices has an impact on clinical practice.

An increasing number of psychological interventions are shifting to online modes of delivery. One such intervention is peer-to-peer support, which in this context may provide internet users living with mental health disorders an opportunity to connect with and support others living with similar conditions. This paper presents a call for further research into how platforms such as Facebook could be used as channels for peer support and the mechanisms that may underlie their effectiveness. We discuss the background of peer support, how it has transitioned online, and consider theories and models that may have relevance. We also consider the importance of moderation within online peer support and the development of specific social network–based online interventions. We conclude that for social network sites to be used as peer-to-peer support interventions, more research is needed to understand their effectiveness, the role of moderation in these communities, and the mechanisms that produce the benefits experienced by users.

Background Patient activation is defined as a patient’s confidence and perceived ability to manage their own health. Patient activation has been a consistent predictor of long-term health and care costs, particularly for people with multiple long-term health conditions. However, there is currently no means of measuring patient activation from what is said in health care consultations. This may be particularly important for psychological therapy because most current methods for evaluating therapy content cannot be used routinely due to time and cost restraints. Natural language processing (NLP) has been used increasingly to classify and evaluate the contents of psychological therapy. This aims to make the routine, systematic evaluation of psychological therapy contents more accessible in terms of time and cost restraints. However, comparatively little attention has been paid to algorithmic trust and interpretability, with few studies in the field involving end users or stakeholders in algorithm development. Objective This study applied a responsible design to use NLP in the development of an artificial intelligence model to automate the ratings assigned by a psychological therapy process measure: the consultation interactions coding scheme (CICS). The CICS assesses the level of patient activation observable from turn-by-turn psychological therapy interactions. Methods With consent, 128 sessions of remotely delivered cognitive behavioral therapy from 53 participants experiencing multiple physical and mental health problems were anonymously transcribed and rated by trained human CICS coders. Using participatory methodology, a multidisciplinary team proposed candidate language features that they thought would discriminate between high and low patient activation. The team included service-user researchers, psychological therapists, applied linguists, digital research experts, artificial intelligence ethics researchers, and NLP researchers. Identified language features were extracted from the transcripts alongside demographic features, and machine learning was applied using k-nearest neighbors and bagged trees algorithms to assess whether in-session patient activation and interaction types could be accurately classified. Results The k-nearest neighbors classifier obtained 73% accuracy (82% precision and 80% recall) in a test data set. The bagged trees classifier obtained 81% accuracy for test data (87% precision and 75% recall) in differentiating between interactions rated high in patient activation and those rated low or neutral. Conclusions Coproduced language features identified through a multidisciplinary collaboration can be used to discriminate among psychological therapy session contents based on patient activation among patients experiencing multiple long-term physical and mental health conditions.

Understanding stakeholders' views on novel autonomous systems in healthcare is essential to ensure these are not abandoned after substantial investment has been made. The ExTRAPPOLATE project applied the principles of Responsible Research and Innovation (RRI) in the development of an automated feedback system for psychological therapists, 'AutoCICS'. A Patient and Practitioner Reference Group (PPRG) was convened over three online workshops to inform the system's development. Iterative workshops allowed proposed changes to the system (based on stakeholder comments) to be scrutinized. The PPRG reference group provided valuable insights, differentiated by role, including concerns and suggestions related to the applicability and acceptability of the system to different patients, as well as ethical considerations. The RRI approach enabled the anticipation of barriers to use, reflection on stakeholders' views, effective engagement with stakeholders, and action to revise the design and proposed use of the system prior to testing in future planned feasibility and effectiveness studies. Many best practices and learnings can be taken from the application of RRI in the development of the AutoCICS system.

Background Epilepsy, multiple sclerosis (MS) and depression are long term, central nervous system disorders which have a significant impact on everyday life. Evaluating symptoms of these conditions is problematic and typically involves repeated visits to a clinic. Remote measurement technology (RMT), consisting of smartphone apps and wearables, may offer a way to improve upon existing methods of managing these conditions. The present study aimed to establish the practical requirements that would enable clinical integration of data from patients’ RMT, according to healthcare professionals. Methods This paper reports findings from an online survey of 1006 healthcare professionals currently working in the care of people with epilepsy, MS or depression. The survey included questions on types of data considered useful, how often data should be collected, the value of RMT data, preferred methods of accessing the data, benefits and challenges to RMT implementation, impact of RMT data on clinical practice, and requirement for technical support. The survey was presented on the JISC online surveys platform. Results Among this sample of 1006 healthcare professionals, respondents were positive about the benefits of RMT, with 73.2% indicating their service would be likely or highly likely to benefit from the implementation of RMT in patient care plans. The data from patients’ RMT devices should be made available to all nursing and medical team members and could be reviewed between consultations where flagged by the system. However, results suggest it is also likely that RMT data would be reviewed in preparation for and during a consultation with a patient. Time to review information is likely to be one of the greatest barriers to successful implementation of RMT in clinical practice. Conclusions While further work would be required to quantify the benefits of RMT in clinical practice, the findings from this survey suggest that a wide array of clinical team members treating epilepsy, MS and depression would find benefit from RMT data in the care of their patients. Findings presented could inform the implementation of RMT and other digital interventions in the clinical management of a range of neurological and mental health conditions.

Methods to facilitate co-production in mental health are important for engaging end users. As part of the Technology for Healthy Aging and Wellbeing (THAW) initiative we organized two interactive co-production workshops, to bring together older adults, health and social care professionals, non-governmental organizations, and researchers. In the first workshop, we used two activities: Technology Interaction and Scavenger Hunt, to explore the potential for different stakeholders to discuss late life mental health and existing technology. In the second workshop, we used Vignettes, Scavenger Hunt, and Invention Test to examine how older adults and other stakeholders might co-produce solutions to support mental wellbeing in later life using new and emerging technologies. In this paper, we share the interactive materials and activities and consider their value for co-production. Overall, the interactive methods were successful in engaging stakeholders with a broad range of technologies to support mental health and wellbeing and in co-producing ideas for how they could be leveraged and incorporated into older people’s lives and support services. We offer this example of using interactive methods to facilitate co-production to encourage greater involvement of older adults and other under-represented groups in co-producing mental health technologies and services.

In this paper we report progress on a novel explainable artificial intelligence (XAI) initiative applying Natural Language Processing (NLP) with elements of codesign to develop a text classifier for application in psychotherapy training. The task is to produce a tool that will facilitate therapists to review their sessions by automatically labelling transcript text with levels of interaction for patient activation in known psychological processes, using XAI to increase their trust in the model’s suggestions and client trajectory predictions. After pre-processing of the language features extracted from professionally annotated therapy session transcripts, we apply a supervised machine learning approach (CHAID) to classify interaction labels (negative, neutral, positive). Weighted samples are used to overcome class imbalanced data. The results show this initial model can make useful distinctions among the three labels of patient activation with 74% accuracy and provide insight into its reasoning. This ongoing project will additionally evaluate which XAI approaches can be used to increase the transparency of the tool to end users, exploring whether direct involvement of stakeholders improves usability of the XAI interface and therefore trust in the solution.

Background Home self-monitoring of blood pressure is widely used in primary care to assist in the diagnosis of hypertension, as well as to improve clinical outcomes and support adherence to medication. The National Institute for Health and Care Excellence (NICE) care pathways for hypertension recommend specific guidelines, although they lack detail on supporting patients to self-monitor. Aim To elicit primary care practitioners’ experiences of managing patients’ home blood pressure self-monitoring, across surgeries located in different socioeconomic areas. Design & setting A qualitative focus group study was conducted with a total of 21 primary care professionals. Method Participants were GPs and practice nurses (PNs), purposively recruited from surgeries in areas of low and high deprivation, according to the English indices of multiple deprivation. Six vignettes were developed featuring data from interviews with people who self-monitor and these were used in five focus groups. Results were thematically analysed. Results Themes derived in the thematic analysis largely reflected topics covered by the vignettes. These included: advice on purchase of a device; supporting home monitoring; mitigating patient anxiety experienced as a result of home monitoring; valuing patients’ data; and effect of socioeconomic factors. Conclusion The work provides an account of methods used by primary care practitioners in the management of home blood pressure self-monitoring, where guidance may be lacking and primary care practitioners act on their own judgement. Findings complement recent policy documentation, which recognises the need to adopt new ways of working to empower patients (for example, additional support from healthcare assistants), but lacks detail on how this should be done.

Background: Remote measurement technologies (RMT) can be used to collect data on a variety of bio-behavioral variables, which may improve the care of patients with central nervous system disorders. Although various studies have explored their potential, prior work has highlighted a knowledge gap in health care professionals' (HCPs) perceptions of the value of RMT in clinical practice. Objective: This study aims to understand HCPs' perspectives on using RMT in health care practice for the care of patients with depression, epilepsy, or multiple sclerosis (MS). Methods: Semistructured interviews were conducted with 26 multidisciplinary primary and secondary care HCPs who care for patients with epilepsy, depression, or MS. Interviews were transcribed verbatim and analyzed using thematic analysis. Results: A total of 8 main themes emerged from the analysis: (1) potential clinical value of RMT data; (2) when to use RMT in care pathways; (3) roles of health care staff who may use RMT data; (4) presentation and accessibility of data; (5) obstacles to successful use of RMT; (6) limits to the role of RMT; (7) empowering patients; and (8) considerations around alert-based systems. Conclusions: RMT could add value to the system of care for patients with central nervous system disorders by providing clinicians with graphic summaries of data in the patient record. Barriers of both technical and human nature should be considered when using these technologies, as should the limits to the benefits they can offer.

BACKGROUND:Epilepsy, multiple sclerosis (MS), and depression are chronic conditions where technology holds potential in clinical monitoring and self-management. Over 5 years, the Remote Assessment of Disease and Relapse - Central Nervous System (RADAR-CNS) consortium has explored the application of remote measurement technology (RMT) to the management and self-management of patients in these clinical areas. The consortium is large and includes clinical and nonclinical researchers as well as a patient advisory board. OBJECTIVE:This formative development study aimed to understand how consortium members viewed the potential of RMT in epilepsy, MS, and depression. METHODS:In this qualitative survey study, we developed a methodological tool, universal points of care (UPOC), to gather views on the potential use, acceptance, and value of a novel RMT platform across 3 chronic conditions (MS, epilepsy, and depression). UPOC builds upon use case scenario methodology, using expert elicitation and analysis of care pathways to develop scenarios applicable across multiple conditions. After developing scenarios, we elicited views on the potential of RMT in these different scenarios through a survey administered to 28 subject matter experts, consisting of 16 health care practitioners; 5 health care services researchers; and 7 people with lived experience of MS, epilepsy, or depression. Survey results were analyzed thematically and using an existing framework of factors describing links between design and context. RESULTS:The survey elicited potential beneficial applications of the RADAR-CNS RMT system as well as patient, clinical, and nonclinical requirements of RMT across the 3 conditions of interest. Potential applications included recognition of early warning signs of relapse from subclinical signals for MS, seizure precipitant signals for epilepsy, and behavior change in depression. RMT was also thought to have the potential to overcome the problem of underreporting, which is especially problematic in epilepsy, and to allow the capture of secondary symptoms that are not generally collected in MS, such as mood. CONCLUSIONS:Respondents suggested novel and unanticipated uses of RMT, including the use of RMT to detect emerging side effects of treatment, enable behavior change for sleep regulation and activity, and offer a way to include family and other carers in a care network, which could assist with goal setting. These suggestions, together with others from this and related work, will inform the development of the system for its eventual application in research and clinical practice. The UPOC methodology was effective in directing respondents to consider the value of health care technologies in condition-specific experiences of everyday life and working practice.

Background A growing number of apps to support good mental health and well-being are available on digital platforms. However, very few studies have examined older adults’ attitudes toward the use of these apps, despite increasing uptake of digital technologies by this demographic. Objective This study sought to explore older adults’ perspectives on technology to support good mental health. Methods A total of 15 older adults aged 50 years or older, in two groups, participated in sessions to explore the use of digital technologies to support mental health. Interactive activities were designed to capture participants’ immediate reactions to apps and websites designed to support mental health and to explore their experiences of using technology for these purposes in their own lives. Template analysis was used to analyze transcripts of the group discussions. Results Older adults were motivated to turn to technology to improve mood through mechanisms of distraction, normalization, and facilitated expression of mental states, while aiming to reduce burden on others. Perceived barriers to use included fear of consequences and the impact of low mood on readiness to engage with technology, as well as a lack of prior knowledge applicable to digital technologies. Participants were aware of websites available to support mental health, but awareness alone did not motivate use. Conclusions Older adults are motivated to use digital technologies to improve their mental health, but barriers remain that developers need to address for this population to access them.

Under-diagnosis of depression and anxiety is common in older adults. This project took a mixed methods approach to explore the application of machine learning and technology for early detection of these conditions. Mood measures collected with digital technologies were used to predict depression and anxiety status according to the Geriatric Depression Scale (GDS) and the Hospital Anxiety and Depression Scale (HADS). Interactive group activities and interviews were used to explore views of older adults and healthcare professionals on this approach respectively. The results show good potential for using a machine learning approach with mood data to predict later depression, though prospective results are preliminary. Qualitative findings highlight motivators and barriers to use of mental health technologies, as well as usability issues. If consideration is given to these issues, this approach could allow alerts to be provided to healthcare staff to draw attention to service users who may go on to experience depression.

Background: Depression is currently underdiagnosed among older adults. As part of the Novel Assessment of Nutrition and Aging (NANA) validation study, 40 older adults self-reported their mood using a touchscreen computer over three, one-week periods. Here, we demonstrate the potential of these data to predict future depression status.Methods: We analysed data from the NANA validation study using a machine learning approach. We applied the least absolute shrinkage and selection operator with a logistic model to averages of six measures of mood, with depression status according to the Geriatric Depression Scale 10 weeks later as the outcome variable. We tested multiple values of the selection parameter in order to produce a model with low deviance. We used a cross-validation framework to avoid overspecialisation, and receiver operating characteristic (ROC) curve analysis to determine the quality of the fitted model.Results: The model we report contained coefficients for two variables: sadness and tiredness, as well as a constant. The cross-validated area under the ROC curve for this model was 0.88 (CI: 0.69-0.97).Limitations: While results are based on a small sample, the methodology for the selection of variables appears suitable for the problem at hand, suggesting promise for a wider study and ultimate deployment with older adults at increased risk of depression.Conclusions: We have identified self-reported scales of sadness and tiredness as sensitive measures which have the potential to predict future depression status in older adults, partially addressing the problem of underdiagnosis.