The TECH@HOME study, a technological intervention to reduce caregiver burden for informal caregivers of people with dementia: study protocol for a randomized controlled trial

TECH@HOME is a randomized controlled trial to test the effectiveness of an innovative technological intervention among people with dementia and their primary informal caregivers in Sweden. The intervention will last 12 months. A total of 320 dyads, including people with dementia and their primary informal caregivers (640 participants in total), will be recruited. These dyads will be randomized either into an intervention or a control group (see Fig. 1 for the Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) figure and for the complete SPIRIT Checklist see Additional file 1). All dyads, both the intervention and control groups, will receive three home visits (at baseline and 3 and 12 months after enrollment) from a dementia nurse in charge of data collection.

In the intervention group, besides usual care, the people with dementia will have a technological home monitoring kit installed in their home (as described under the “Intervention tested” section) while dyads in the control group will only receive usual care. Usual care for people with dementia in Southern Sweden can vary. In the target area, people with dementia usually receive comparable pharmaceutical treatment depending on the type of dementia, as prescribed by a general practitioner or a specialist at a memory clinic. Patients of memory clinics receive regular follow-ups depending on the type of diagnosis and on-going treatment response, while the majority of people with dementia are treated and monitored at a primary health care center and by the municipal health care services. Social workers from the municipality (“Biståndshandläggaren”) where the person resides, together with district nurses, have a meaningful role in tailoring the care plan by mediating access to other care services such as respite care homes, home help, and (dementia) nurse home visits. Use of such services depends on the specific needs of the person with dementia, which can also be unrelated to dementia, but rather dependent upon concomitant health and social issues.

The intervention includes the installation and use of a technical monitoring kit in the home for 12 months. The home monitoring kit includes devices that are easy to use and do not require significant technical expertise for installation and maintenance. The kits will be composed of a control unit and a set of sensors that immediately notify caregivers, through their phones, of any potential risks for the person with dementia. The kit will have home-leaving sensors, bed-occupancy sensors, smoke and water leak sensors, automatic lights, and other interactive functions. These devices will be connected to a single-board microcontroller that will transmit alarm messages to the caregivers in case of need. The monitoring kits will be assembled by an external contractor and installed by a handyman who will also train the caregivers about their use (see Fig. 2). The handyman will be trained and receive a manual on how to provide a standardized introduction to all caregivers, who will be probed to ask for additional information. Written information for users will be developed in collaboration with caregivers, followed by a subsequent field test.

The primary outcome of the study is the amount of informal care support provided to the individual with dementia by the primary informal caregiver.

Secondary outcomes of the study include:

People with dementia living in the area of Region Skåne (Southern Sweden), together with their primary informal caregiver will be invited to participate in the study.

The inclusion criteria for the study that are applied to the person with dementia are:

Those not meeting the inclusion criteria will not be eligible for the trial. In addition, specific exclusion criteria will be:

Informal caregivers are defined as those people who provide care and support to an individual with dementia on a regular basis without any form of financial compensation. Exclusion criteria applied to the informal caregiver are:

The person with dementia and their caregiver can either be cohabitating or live in different homes.

Informed consent will be actively sought from both parties. Participation in the study is completely voluntary and participants will receive verbal and written information about the study prior to their participation and during the study implementation.

Participants will be recruited at the Memory Clinic at Ängelholm’s Hospital in Sweden, which is coordinated by PJ. The neurologists and nurses working at the Memory Clinic will review their list of patients to identify those meeting the inclusion criteria and send an invitation letter to participate that also describes the aims and characteristics of the study. We estimated a participation rate of 50% based on previous studies in this area [25]. In addition, we plan to distribute a brochure describing the characteristics of the project among the primary health care centers in the surrounding areas, informing potential candidates about the study. The potential participants (both people with dementia and their caregivers) will then be assessed by the dementia nurse coordinating the field activities (CDB) to double-check whether the inclusion criteria are met and to further assess the exclusion criteria. Eligible dyads will then be asked to sign the informed consent for the study. Basic information (age gender and town of residence) about those participating in the study will be compared with that of the people meeting the inclusion criteria who are enlisted in the Memory Clinic database, only on aggregate level, for the purpose to assess potential participation bias.

Following written informed consent, names of the dyads enrolled will be sent directly to the principal investigator of the study (CC) who will perform the randomization. Dyads will be randomized (according to a 1:1 ratio) either to the experimental arm (receiving the technological intervention), or to the control arm (receiving usual care). The allocation sequence will be generated using the statistical software STATA (StataCorp, College Station, TX, USA).

All participants signing the informed consent to participate will be enrolled in the study and will receive three home visits by a dementia nurse trained in the methodology used in the project. Informed consent will be actively sought both from people with dementia and their caregivers. In case of people declared legally incompetent, informed consent will be requested from a family member or from another person legally appointed by a judge to act as substitute. Home visits will occur at enrollment and after 3 and 12 months. Before the home visit, the self-administered sections of the questionnaires will be mailed to the caregivers participating in the study. The home visits will take approximately 1 h, during which the remaining sections of the questionnaire will be administered to the study participants and a general check of the self-administered forms will be made by the dementia nurse. Those in the experimental arm will receive the installation of the home monitoring kit at latest 1 month following the baseline home visit.

A 12-section study-specific questionnaire has been designed. The questionnaire sections addressing the people with dementia will be administered by the dementia nurse and cover the following domains: (A) sociodemographic information, (B) cognitive function, (C) ADL, (D) health and quality of life, and (E) use of health care resources.

The sections addressing the informal caregiver will be self-administered and cover the following domains: (F) sociodemographic information, (G) lifestyle, (H) health and quality of life, (I) mental health and wellbeing, (J) caregiver burden, (K) caregiving time, and (L) use of health care resources.

The amount of informal care provided by caregivers to the people with dementia, considered as a proxy of the caregiver burden, will be measured in h/week and assessed using a specific section of the Resource Utilization in Dementia (RUD) instrument [26]. The change from baseline to months 3 and 12 will be considered as primary outcome for the analysis.

Quality of life of the person with dementia will be measured using the Quality Of Life in Alzheimer’s Disease (QOL-AD) instrument [27], which has been developed in collaboration with caretakers, caregivers and experts in dementia care to ensure validity. Previous studies suggest that the instrument’s validity and reliability are satisfactory in a study population of people with dementia: Cronbach’s α ranges from 0.84 to 0.88 for people with dementia and caregivers and 1-week test-retest reliability is also acceptable (intraclass correlation coefficient (ICC) = 0.76 for people with dementia and 0.92 for caregivers) [28].

Fear of falling of the person with dementia will be measured using the 16-item Falls Efficacy Scale-International (FES-I) [29‐31]. The evaluation of the Swedish version of the instrument showed high internal reliability (Cronbach’s α = 0.95) and an interitem correlation averaging 0.55 [32]. Quality of life of the caregiver will be assessed using the EuroQol, 5 dimensions, 3 levels (EQ-5D-3 L) health survey. This standardized instrument to measure health outcomes [33] has been used and validated in previous Swedish studies [34, 35], showing both a good internal reliability (Cronbach’s α = 0.73) and validity [36]. Caregiver anxiety will be assessed using the anxiety component of the Hospital Anxiety and Depression Scale (HADS) [37]. HADS is a 14-item scale; seven of the items relate to anxiety and seven relate to depression; each item is a Likert scale rated from 0 to 3, and this means that the overall score for either anxiety or depression will range from 0 to 21. Cronbach’s α for HADS-Anxiety varies from 0.68 to 0.93 (mean 0.83) and for HADS-Depression from 0.67 to 0.90 (mean 0.82) [38]. The Zarit Burden Inventory (ZBI) will be used to measure the level of caregiver burden [38]. This is the instrument most consistently used in dementia caregiving research [39], and it is often used to measure the change of caregiver burden over time, resulting from the progression of the disease severity of the care recipient or from interventions aimed at reducing burden. The revised version of the ZBI with 22 items will be used, as this has shown both a high reliability (Cronbach’s α from 0.88 to 0.91) and validity (it is highly correlated with a single global burden rating, r = 0.71) [40]. Domestic accidents are defined as accidental falls, bruises, cuts, episodes of wandering outside the home, burns and fires, flooding, or other events which potentially or concretely harm the people with dementia. The incidence of domestic accidents will be assessed using an ad-hoc form.

Basic social and demographic data, such as age, sex, marital status, place of living, housing characteristics, income, education and current and previous employments, will be collected from both the people with dementia and their informal caregivers. Information regarding alcohol use, smoking, physical exercise and diet of the caregiver will be retrieved using an ad-hoc module based on the forms of the Swedish National Health Survey.

The questionnaire developed for the study includes other instruments that have already been tested in studies addressing people with dementia. To assess the functional status of the people with dementia (dependency in both personal and instrumental activities of daily living (IADL)), the instruments of the interRAI suite will be used. As the full form of the RUD questionnaire will be administered, full information about use of health care services, medications, and other characteristics of the caregiving relationship will be available.

Although TECH@HOME is a nonpharmacological trial, the research team will actively monitor adverse events, i.e., any unfavorable and unintended sign, symptom, or disease temporally associated with the intervention, whether or not it is causally related to the study (e.g., falls, acute episodes of illness, etc.). This will mainly be done using two sources of information: (1) people with dementia and their caregivers will be instructed to refer to the clinical study coordinator in case of serious adverse events (requiring general practitioner support or hospital care) and (2) similarly, the clinical staff at the memory clinic will report in case any adverse event is acknowledged. Participants (both people with dementia and their caregivers) may be asked to withdraw if it is determined that participation in the trial does not represent the best interests of the participants themselves.

According to Nordberg and colleagues [5], the mean number of hours spent in caregiving by informal caregivers of people with dementia ranges from 7.7 h per week (people with clinical dementia rating (CDR) 0.5) to 46.9 h per week (people with CDR 2.0). Wimo and colleagues [41] suggest that a standard deviation (SD) of 20 can be considered acceptable when using informal caregiving hours as an outcome for formal sample size calculation. Given these numbers, a sample size of 320 dyads randomized in a 1:1 ratio (160 per arm) is sufficient to detect a mean decrease in caregiving time of 1 h per day with a statistical power of 0.8 and a dropout rate of 20%, aiming at a final number of 128 dyads in the experimental arm and 128 dyads in the control arm.

The final database will be checked and cleaned using statistical routines after completion of each wave of assessments. All modifications made during this phase will be recorded in a specific file in order to allow for replicability of all operations made on the raw database.

All collected data will be made anonymous in accordance with the Swedish Data Protection Act (1998: 204) and coded in a database with access protection. The database will be stored on a server with continuous daily safety backup. Researchers will analyze data containing serial number encoded data without access to names or social security numbers. All information relating to the participants, the sampling framework, written consent, and the results from questionnaires are handled so that no unauthorized access to them can occur, and kept locked up. Only current researchers in the project will get access to the materials. Materials will be archived for 10 years.

The first step of the analysis will be exploratory in nature. A descriptive analysis of the sample will be conducted using univariate and bivariate statistical analyses with the aim of verifying the comparability of the study groups. Significant differences between exposures and outcomes will be compared using the chi-square test, Fisher’s exact test (in the case of categorical variables), the Mann-Whitney U test (for ordinal data such as the FES-I), the t test, or analysis of variance (ANOVA) test (for comparisons of continuous variables between groups according to a normal or non-normal distribution). Additionally, to check possible selection bias, the characteristics of the subjects in the sample will be compared, only on aggregate level, to those of the population with neurocognitive disorders included in the memory clinic register of the study area (e.g., age, sex, and place of living).

The primary outcome “time of informal care provided” will be compared between the two groups using analysis of covariance (ANCOVA). The analyses will be adjusted for variables which are nonhomogenously distributed in the two groups and controlling for other factors that the medical literature suggests are associated with the two outcomes (e.g., sex and age of caregivers, severity of dementia, and use of private care workers). The analyses of outcomes will be intention-to-treat (ITT).

Among the secondary outcomes of the study, we will investigate the cost-effectiveness of the technological intervention, in case the overall intervention does not yield statistically significant improvements in the other primary and secondary study outcomes. Therefore, cost of illness, budget impact and cost-utility analyses will be performed using the perspective of the National Health Service [42] and in agreement with the ITT principle per the CHEERS statement [43].

Probabilistic sensitivity analyses (PSA) will also be performed on the health economics models using the software TreeAge Pro 2009. Results from the PSA will also be presented as an acceptability curve, graphically illustrating the probability of the intervention being cost effective over a range of willingness-to-pay values.

The implementation of a technological intervention can be considered “complex” according to the definition of the UK Medical Research Council [44]. This means that the study researchers need to identify and appraise all the intertwining components which are interacting and determining the outcomes themselves. Therefore, specific subanalyses are planned such as intervention integrity, dyad adherence to the study, and acceptance of the technology by the users. The research group plans to administer a further ad-hoc questionnaire at the end of the trial to both the people with dementia and the caregivers to assess their level of satisfaction with the technologies used. A specific ethical approval for this aspect of the study will be sought in the future.

Risk of selection bias will be assessed by looking at the results of the enrollment process. Clinical staff will be instructed about how to explain, in very simple terms, what participation in the trial will entail. This is expected to reduce refusals among those people who are not very familiar with new technology. Considering the high level of “digital” education among the older population in Sweden, even compared to other Western countries, the likelihood of a selection bias should be minimized. The involvement of expert dementia nurses in the study also addresses the need to use research staff with a high level of know-how on how to relate with people with dementia and their families. This is also expected to reduce the dropouts and withdrawals during the study, thus increasing overall retention.

Other mediating and confounding variables might be associated with study outcomes, and these could lead to wrong/false results at the end of the trial. The use of a multidimensional questionnaire and the randomization process are expected to reduce such risks by providing several covariates that will be available for the statistical analyses to ensure the two study arms are comparable under several social, clinical and functional profiles.

Last but not least, the lack of adherence to the technology can constitute a major flaw for the study implementation and results. This risk will be minimized by using a standardized but thorough approach to the training of users (both people with dementia and their caregivers) with technologies, by using a very simple technological device, and by having the possibility to remotely monitor in real time the actual use of the devices. The study coordinator will be able to detect when a person is not using the technology and will promptly react in order to understand (and possibly address) the reasons behind the lack of use.

The trial started recruiting in March 2016 and the assessments are scheduled to continue until December 2017.