When older adults lose their partner, they often lose the most important person in their life. Most mourners grieve for a period and then find a way to cope with their loss. Some mourners, however, develop severe or persistent grief symptoms, with a clinical diagnosis of a prolonged grief disorder (as defined in International Classification of Diseases [ICD]-11) or a persistent complex bereavement disorder (as defined in Diagnostic and Statistical Manual of Mental Disorders [DSM]-5). Within the context of LEAVES (optimizing the mentaL health and resiliencE of older Adults that haVe lost thEir spouSe via blended, online therapy), we will use the term prolonged grief (PG). PG is the “failure to return to pre-loss levels of performance or states of emotional wellbeing” [1] and commences when grief-related symptoms are still present after 6 months. These symptoms consist of separation distress (eg, yearning, intensive sorrow, and emotional pain), reactive distress (eg, difficulties accepting the loss, self-blame, and avoidance of reminders of the loss), and social/identity disruption (eg, loneliness, meaninglessness, and role confusion), which can lead to impairments in important life domains. Comorbidity of PG with depression, post-traumatic stress disorder, and suicidal ideation is high [2]. Bereavement in later life has also been associated with physical problems and risk behaviors including involuntary weight loss, poor sleep, change in smoking and alcohol habits, chronic pain, inflammation and cardiovascular risks, as well as increased mortality [3-6]. PG has an estimated prevalence of 10% among the general elderly population [7], with the estimated rates differing per gender. For example, a German population-based study found prevalence rates of complicated grief of 9.6% for female and 2.7% for male older adults [8]. A study conducted on a large sample of Dutch older adults found that of those experiencing grief, 25.4% develop PG [9].

The course of the grieving period and whether or not a person develops PG depend on several factors, such as the mourner’s resilience and their social network [10]. Although it is possible to predict the risk for one to develop PG [11], screening is not a common practice. Furthermore, the development of PG often goes unnoticed by the widow/widower, who then also does not realize that she or he can benefit from help [12]. This has led to a situation in which older adults are unaware of their own needs, do not seek help, and do not receive the care they need.

A wide range of interventions are available for either preventing or treating PG including support groups, writing exercises, and individual psychological counseling. Internet-based options have also started to emerge (eg, [13-16]). Such interventions mostly consist of writing exercises with minimal therapist involvement. A recent evaluation of an internet-based, preventive intervention shows a moderate to large reduction in the duration of grief, depression, anxiety, and psychopathological distress [17]. In general, however, the evidence on the effect of online preventive interventions remains inconclusive.

In this proposal and the upcoming project LEAVES, we will develop, implement, and evaluate an online bereavement support program that will support the prevention and treatment of PG. In parallel, we will develop an exploitation strategy for the service, so that it can persist on a sustainable model following project completion.

The LEAVES service will provide support to older adults who have lost their spouse, helping them cope with their grief, by preventing PG or treating PG once detected. Ultimately, the aim is to prevent depression or social isolation, strengthen resilience and well-being, and accelerate one’s return to society. LEAVES helps older adults to process the loss of a spouse in an empathic and caring online environment. It can detect persons at risk for complications following loss of a spouse, reveal negative trends in the emotional life of the widow or widower, and intervene to counter such trends. The service will rely on online treatment whenever possible but will be blended with telephone or face-to-face counseling when necessary.

The LEAVES value chain (Figure 1) begins when a person at the end of their life is presented with the LEAVES service, for example, by an elderly association (as a service to their members) or a funeral home (as part of their service package). Once the person using the service passes away, the mourning spouse is automatically invited to the service. Alternatively, those in mourning can register for the service without prior participation of the deceased spouse. After registering with the service via the LEAVES website, a user profile is created for each individual. This profile includes demographics (for personalizing the service) and a risk assessment for developing PG based on an online survey founded upon scientific risk prediction models. The mourner then enters the treatment pipeline, which is based on the LIVIA program [17,18]. LIVIA focuses on guiding mourners during marital bereavement via a self-service, online program. Online treatment is supported by stories and memories, provided by the person who passed away via the Before You Leave service [19]. To facilitate the interaction between technology and mourner, the LEAVES service will make use of a conversational agent-based interaction paradigm, in which the user interacts through natural language with a virtual on-screen character, an approach that is gaining popularity in health-related applications (eg, [20,21]). Using this technology, a caring, empathic, and individually tailored online environment will be developed to foster trust and compliance. The mourner’s emotional state will be constantly monitored both by assessing the user’s responses during dialogue interactions with the virtual agent and by experience sampling methods (ie, short, periodic Q&A modules), offered via tablet or mobile platforms. If any negative trend is observed indicating potential risk for PG, the mourner will be directed toward LEAVES’ offline services (eg, phone helpdesk, face-to-face meetings). The service will be developed around the notion that the user will be helped online when possible, but offline when necessary.

The LEAVES service focusses on the following target groups:

The project consortium consists of a collection of organizations including research, end user, and clinical organizations; small- and mid-sized enterprises; and a large enterprise. The project is led by Roessingh Research and Development (RRD). Service model design and visual design are mainly being taken care of by The Dutch National Foundation for the Elderly (NFE) and the Swiss SME Nothing AG (NTH). Technology development is being led by RRD. RRD has created a solid foundation for an electronic health (eHealth) platform and AI dialogue systems in numerous European and commercial projects. NTH will complement these skills with their functional front-end and graphical user interface design skills. The University of Bern, Switzerland (UoB), the School of Social Work of the University of Applied Sciences and Arts in Olten, Switzerland (SSW), and the Psychiatric Department at the Health Unit of Baixo Alentejo, Portugal (ULSBA), bring in their vast and extensive experience and knowledge of treating PG, and their skills in leading clinical evaluations. NOVA University of Lisbon (UNL) will bring in their expertise on testing and validating technological tools for health intervention, as well as economic evaluation and cost-effectiveness analysis. The Dutch insurer and funeral undertaker DELA Natura- en levensverzekeringen (DELA) will contribute with expertise from the innovation department, where they continuously seek new ways to support their clients around end of life, both online and offline. Finally, the Portuguese SME Sensing Future Technologies (SFT) will utilize their experience in marketing eHealth technologies for the Portuguese and international markets.

The core technology components in LEAVES are the mobile app Before You Leave (technology readiness level [TRL] 9, developed by DELA), the Livia Online Grief Program (TRL 7, developed by the UoB), and the natural language-based Virtual Agent Platform (TRL 7, developed by RRD). All modules will receive an accessible design and a clean user experience given the subject matter of coping with grief (furnished by NTH).

The workplan of the LEAVES project is divided into the 5 work packages illustrated in Figure 6: (1) iterative design, (2) technology development, (3) real-life evaluation, (4) business modeling and exploitation, and (5) management and dissemination.

The first 2 years of the project will focus on the design and development of the LEAVES platform and its service model. These activities are mostly handled in work packages 1 and 2.

The objective of the first work package (ie, iterative design) is to make sure that the service, tools, and user interface design are in line with the expectations of all end-user groups of the LEAVES service. Several multidisciplinary design workshops will take place in which a wide array of techniques from user-centered design will be employed, including user story mapping, interviews, definition of “jobs to be done,” focus group discussions, information architecting, persona definitions, and card sorting for content aggregation. The results of these procedures will lead to first sketches and wireframes of the interface and its interactive experience. The user interface will be crafted with careful attention to accessibility standards, considering the specific needs of the elderly in user experience and user interaction. Iterative evaluations with different prototypes (with low, medium, and high fidelity) will be performed with example members of the target group. For the low- and mid-fidelity prototype evaluation, a series of fairly detailed mock-ups will be available, providing simulated interactive functionality and navigation. The prototypes will be presented to the users in a controlled laboratory environment, and the evaluations will be conducted in different iterative cycles with usability experts. Before the final pilot, an extensive usability study will be performed to make sure that all aspects are usable within the evaluation.

The second work package (ie, technology development) deals with all technology developments needed to realize a stable platform to be used in the real-life evaluation. To do so, the technology partners will work together setting working prototypes to maximize the impact of the iterative design work in the first work package. Technical and clinical partners will work together to translate the plain text content of LIVIA and Before You Leave into dialogue-based text. This work package will also deal with the personalization of the content to specific end users based on user profiles. A continuous personalization of the service will be performed as a response to the emotional assessment performed in the platform. Last but not least, this work package will ensure that the content system will be designed with a “privacy by default” approach. In the LEAVES project, we will go beyond the state of the art in privacy design and will include a personal data control interface that allows older adults (with relatively undeveloped technological skills and low awareness of privacy control) to control the access and storage of their personal data. Given the sensitive nature of the treatment and the data that are collected about the individual who uses the LEAVES service, ensuring privacy and control over personal data has top priority.

The third and last year of the project will be mostly dedicated to real-life evaluation, refinement of the business model, and exploitation. The third work package (ie, real-life evaluation) focuses on the evaluation of the clinical and economic potential of the LEAVES service by conducting a longitudinal real-life evaluation across 3 countries: Portugal, the Netherlands, and Switzerland. This work package concerns the development of the study protocol as well as the clinical evaluation (in Portugal and Switzerland) and acceptance of technology and service design (in the Netherlands). Finally, an economic evaluation will be performed by determining the differences in costs and the health benefits that result from the use of LEAVES.

The clinical evaluation will involve older adults (65+) who express the need for help in mourning their spouse and will focus on well-being (eg, a decrease in grief symptoms measured with the Texas Revised Grief Inventory [27]; loneliness, as assessed with the De Jong Gierveld Short Scale for Emotional and Social Loneliness [28]; and improved quality of life measured with the Satisfaction with Life Scale [29]) and end-user satisfaction. Results will be compared with those obtained in the same timeframe from people that are “regular mourners” (in The Netherlands), and in a waiting control group (in Switzerland and Portugal). Next, close ones to older adults and care professionals will be consulted on potential effects of the service, to gather a comprehensive overview. Exit strategies will be put in place for participants who need additional care or want to stop using the service. The objective of the economic evaluation is fourfold and will aim to assess the financial consequences from implementing the LEAVES service. It will (1) estimate the change in symptoms averted with the implementation of LEAVES; (2) evaluate the costs associated with the implementation of LEAVES; (3) evaluate the avoided costs related to minimization of the grief symptoms; and (4) compare the costs of LEAVES with the benefits of implementing it, in terms of improved health.

The fourth work package (ie, business modeling and exploitation) concerns the development of a business model, integrated and supported on the evidence from the field, as well as a go-to-market strategy and commercial exploitation of the project output.

The fifth and final work package (ie, management and dissemination) concerns project management and ensures that the project will be executed on time, within budget, and with outstanding quality. Within this work package a dissemination and communication strategy is also defined and executed. Finally, the ethical guidelines for the project will be established, including procedures for medical ethical permission and informed consent, as well as project-wide methods for data handling (in a General Data Protection Regulation [GDPR]-compliant manner).