Pragmatic RCT to Assess the Effectiveness of an Online Self-help Programme for Older Adults After Spousal Bereavement

Introduction:

Theoretical background:

The death of a partner is a frequent and very stressful critical life event in later life. It implies a dissolution of social and emotional ties. This deeply affects the attachment system, requires the acceptance of the loss, as well as the formation of a new identity and a new perspective for the future. It involves the adaptation of daily routines which can be even more challenging when social, physical, and financial resources decline in later life. Grief and psychological distress after the loss of a partner are normative reactions. For most people, grief intensity weakens to a manageable degree within several weeks or months. After the most intensive period, grief is still present, but the loss becomes gradually integrated and no longer hinders the way of ongoing life. However, some individuals are less able to cope with bereavement and show symptoms of disturbed or prolonged grief, or adaptation problems. About 10% of older adults develop a prolonged grief disorder or persistent complex bereavement disorder, characterised by separation distress, frequent or disabling cognitive, emotional and behavioural symptoms, such as avoidance of reminders of the loved one, difficulties moving on with life and functional impairment. The development of prolonged grief or adaptation problems often goes unnoticed by the widow(er) who then fail to realize that help may be needed. This leads to a situation in which older adults are unaware of their own needs, do not seek help, and do not receive the care needed. Previous focus groups and stakeholder's feedback have noted that Portuguese older adults who lost their partner (and their family members) avoid and downplay the grief process, not seeking help when needed. This may lead to a high usage of primary care and, later on, to mental health services, with complaints of insomnia, physical and depressive symptoms.

A wide range of interventions are available for either preventing or treating prolonged grief disorder including support groups, writing exercises, and individual psychological counselling, and, more recently, internet-based options. Such interventions mostly consist of writing exercises with minimal therapist involvement. Previously, Brodbeck and colleagues found that a 12-week intervention with a bereavement online support service (LIVIA) in Swiss adults 40+ years old had a medium to large effect on grief, depression and loneliness over 3 months as well as increasing life satisfaction.

Adapted from LIVIA and combining features from the "Before you leave" app, LEAVES is an online self-help interactive programme founded in the task model of mourning and the dual-process model of coping with bereavement that supports older adults who have lost their partner in dealing with and preventing prolonged grief following a user-centred, iterative and agile approach. In LEAVES-PT the investigators will carry out a 10-week pragmatic randomised two-armed parallel-group controlled trial to compare the effectiveness of LEAVES vs. usual care in reducing grief in community-dwelling Portuguese adults over 55 years registered with the local primary care unit. Besides the clinical evaluation, a qualitative study of implementation and engagement with the LEAVES online service and a cost-effectiveness analysis will be carried out. This trial (LEAVES-PT) is part of a larger study from an international consortium (LEAVES) with two more trial centres but with distinct design and methodology (The Netherlands and Switzerland).

Context on Lower Alentejo's Population and ULSBA:

LEAVES-PT is coordinated by the Local Health Unit of Lower Alentejo's (ULSBA) Psychiatry Department. ULSBA comprises several primary care services across the region, a public health unit, a Hospital as well as a palliative care community support team. ULSBA is located in the largest district of Portugal (district of Beja), covering 10.4% of Portugal's total area, while serving the second least populated regional district area of Portugal with a population density 9 times lower than the national average. Lower Alentejo (and the district of Beja) is a predominantly rural and relatively poor area, with a high percentage of older adults, declining population, underdeveloped public transportation system and, up until recently, underserved in terms of mental healthcare assistance. ULSBA has a long-standing history of community health prevention and promotion. Specifically, ULSBA has implemented a community-based mental health promotion program for older adults in social isolation that included strategies on how to cope with grief. Leveraging on this expertise and in the context of LEAVES-PT, ULSBA will contact community-dwelling older adults registered in primary care services in Lower Alentejo.

Hypothesis:

The investigators hypothesize that:

1. a dialogue-based online self-help programme (LEAVES) intervention significantly decreases grief depression symptoms and perceived loneliness (small to medium effect) compared to the usual care control arm over 20 weeks in a clinical sample of Portuguese adults over 55;
2. participants where the loss occurred more recently (especially < 6 months) and/or have more severe grief symptoms at baseline will benefit more from the LEAVES intervention.

Objectives:

To test these hypotheses, the investigators aim to:

1. determine the effectiveness of a 10-week intervention with LEAVES vs usual care control over 20 weeks in a pragmatic randomised two-armed parallel-group controlled trial on: 1) Grief symptoms (primary endpoint) and, 2) Depression symptoms and Loneliness (secondary endpoints);
2. determine if time since loss and severity of grief symptoms at baseline are moderators for the effectiveness of the intervention.

Sample size:

The investigators aim to recruit and randomise 100 participants where 50 participants will be allocated to each arm. The intervention will last 10 weeks, and participants will be followed for a further 10 weeks (20 weeks since baseline).

Previously, Brodbeck and colleagues found that an intervention with LIVIA (a bereavement online support service) in Swiss adults 40+ years old had a large effect on grief (d = 0.81) compared to the control arm. In Alentejo, Portugal and in adults 55 years and over the investigators expect the LEAVES intervention to have a small to medium effect on grief (d > 0.30) (as measured by TRIG) vs usual care (control arm). To find a significant difference between intervention and control at 80% power and with an α of 0.05, assuming a correlation of 0.65 among repeated measurements and using ANOVA, a total of 52 participants are necessary. The investigators anticipate a dropout of 20% between recruitment and the start of the trial (10 weeks) and a 30% dropout over the study duration (20 weeks). In view of the sample size calculation, the investigators aim to recruit at least 100 participants with 50 participants per arm. Power calculations were done with G*Power.

Recruitment:

Recruitment will be coordinated by ULSBA located in Beja, Portugal by 4-6 clinical psychologists and/or mental health nurses. Participants registered with ULSBA's primary care services, who are 55 years and over and widowed will be randomly contacted by clinical staff through telephone following a six-step model for optimal recruitment: (1) explain the condition, (2) reassure patients about receiving treatment, (3) establish uncertainty, (4) explain the study purpose, (5) give a balanced view of treatments, and (6) Explain study procedure. The first 100 participants to accept to participate in LEAVES-PT will be screened according to the eligibility criteria and if included, assigned a treatment. If the investigators do not manage to recruit 100 participants, the recruitment will be extended to other districts in Portugal besides Beja. Participants will also be asked for the preferred place of assessment, if in-person and at the hospital/ primary care centre or in-person and at their usual place of residence. To minimize dropout and increase adherence, participants will be contacted regularly to address troubleshooting or other technological issues.

Data Collection and management:

Besides the initial telephone risk assessment data on most variables will be collected through face-to-face interviews at the participant's usual residency or at ULSBA, however is more convenient for the participant. Data on module completion, duration and intensity of the LEAVES support service will be collected automatically within the online platform. In accordance with the data minimisation principle, the investigators certify that only variables indispensable for the trial will be collected. All data will be stored and analysed in an anonymous and unidentifiable format in such a way that the database entries cannot be linked to the personal identity of the study participants. Servers are protected by high-end firewall systems and only researchers directly involved in the study have access to the data.

Statistical analysis:

Data cleaning and quality control will be performed prior to analysis. Normality will be assessed by Q-Q plots: normally distributed variables will be presented as means and SDs, non-Gaussian distributed variables as medians and interquartile ranges (IQR) and categorical data as percentages (with corresponding frequency). Missing value analysis will be carried out to explore the patterns, the extent and the type of missing data. Assuming that data will be missing at random, linear mixed models will be used to determine the effectiveness of the intervention vs the control on grief symptoms over the 20 weeks of follow-up and using the intention-to-treat principle. Models for the primary and secondary outcomes will have terms for time (pre vs post intervention measures and post intervention vs follow-up measures), group (usual care vs. LEAVES + usual care), possible non-randomised confounders and interaction terms. As a sensitivity analysis, LEAVES-PT will explore moderators such as sex, age, time since loss and severity of grief symptoms. Cohen's d will be calculated for all models. Point estimates and confidence intervals will be used to assess statistical and clinical significance. Results of the linear mixed models will be presented as βs and 95% confidence intervals (CI).

Cost-effectiveness analysis:

The investigators will estimate the incremental cost-effectiveness ratio (ICER) of the intervention, in terms of cost per point-improvement in the grief scale. The investigators will also look at depression and loneliness (secondary outcomes) and estimate ICERs using improvements in the depression or loneliness scales in the denominator according to the following: ICER = cost of delivering the intervention/ total improvement in grief symptoms. The investigators will adopt a provider perspective, taking into account, in the numerator, the costs borne by the provider of the intervention. The investigators will consider both technological (equipment costs, cost of using the platform) and human costs (training, time spent by staff delivering the intervention) associated with providing the intervention.