This paper presents the research protocol according to the SPIRIT (Standard Protocol Items: Recommendations for Interventional Trials) guideline [
A multiple baseline single case study will be conducted in everyday care practice. Multiple baseline single case studies are based on the principle that evidence of a causal effect can be strengthened by assigning participants to baselines of varying lengths. The staggered introduction of the intervention helps to rule out alternative explanations such as maturation or testing and session experience [
The study period consists of 3 phases (within 13 weeks) and a follow-up phase (
Study planning overview. CAU: care as usual; CiSR: care involving a social robot.
Baseline phase: participants will be randomly allocated to 1 of 4 baseline lengths (2, 3, 4, and 5 weeks, respectively), resulting in 4 different clusters. In the baseline phase, participants will receive care as usual.
Intervention adjustment phase: a 6-week phase in which each participant starts to receive compare care involving a social robot (CiSR), that is, using and adjusting to the social robot.
Effect measurement phase: this period varies in length based on the cluster to which the participant is assigned, from 2 to 5 weeks. Participants will continue CiSR.
Follow-up phase: to observe whether the results are maintained while participants continue CiSR, a follow-up measurement will take place 6 months after the start of the intervention adjustment phase.
Study planning.
| Cluster | Baseline | Intervention | Effect | Follow-up |
|
| 1 | 2 weeks of CAUa | 6 weeks of CiSRb | 5 weeks of CiSR | Weeks 34-35 | |
| 2 | 3 weeks of CAU | 6 weeks of CiSR | 4 weeks of CiSR | Weeks 35-36 | |
| 3 | 4 weeks of CAU | 6 weeks of CiSR | 3 weeks of CiSR | Weeks 36-37 | |
| 4 | 5 weeks of CAU | 6 weeks of CiSR | 2 weeks of CiSR | Weeks 37-38 | |
aCAU: care as usual.
bCiSR: care involving a social robot.
This study was approved by the Ethics Review Board Tilburg School of Social and Behavioral Sciences of Tilburg University (TSB_RP1461) and was exempted from ethics approval by the Medical Ethical Review Committee of Radboudumc, Nijmegen (2023-16798). Both the participants and care professionals will be provided with verbal and written information about the study, and their informed consent will be obtained before randomization. To promote full participation in the project, residential care organizations will receive a financial contribution for the research activities, up to a maximum amount of €10,000 (US $10.845,69; currency rate as per October 18, 2024) when different targets are met (such as enrolling 10 participants and completing data collection). Individuals participating in the study will receive no financial compensation. Pseudonymized data will be stored at a secure site with limited access and separated from personal data, such as names. The World Health Organization trial registration dataset is presented in
Individuals with challenges in daily structure and planning will be recruited from 3 Dutch residential disability care organizations. Facilities will have to meet the criteria detailed in
Have an intention to sustainably implement the social robot
Have an IT infrastructure to support the use of the social robot
Show commitment to implementation and participation in the research, characterized by the following:
release funding for the purchase of the social robot
provide human resources for implementation and training
provide a project leader to coordinate the project
Individuals will be eligible for inclusion in the study if they meet the predefined inclusion and exclusion criteria, as detailed in
One coresearcher is recruited from each of the 3 participating care organizations. A coresearcher is an individual who receives care within the context of this care organization and can represent the perspective of the participants. The 3 coresearchers are each coupled with a researcher and will function as a duo. The duos will receive training on how to collaborate and perform coresearch. During the study, coresearch activities will be logged, and the impact of the coresearcher’s involvement on the participants will be evaluated.
Aged ≥18 years (no upper age limit)
Currently receive residential care
Struggle with daily structure and planning, for example, have trouble remembering appointments and activities, or require assistance to initiate them
Capable of understanding and following verbal instructions
Possess the cognitive capacity to set 3 to 5 personalized goals with the help of a care professional and are motivated to work on these goals with the support of the social robot; these are goals for which the participant usually receives daily or weekly support by a care professional at the start of the study
Supported by motivated care professionals to guide them in using the social robot
Prone to convert stress into physical aggressiveness
Overly sensitive to stimuli
The SEAR (Screened, Eligible, Approached, Randomized) framework [21 will be used to systematically monitor and report the recruitment of participnats, according to the CONSORT (Consolidated Standards of Reporting Trials) extension for N-of-1 flowchart [22.
Eligible participants will receive an information letter about the study. The purpose of the research will be explained in a conversation and questions will be answered. The participant and care professional will be given sufficient time to consider the information. When they decide to participate in the study, participants will be asked to sign an informed consent form.
The researcher (KNvD) will randomly assign each location within the participating care organizations to a cluster. Due to practical considerations, randomization will be applied at the location level rather than the participant level, as this ensures a consistent and manageable implementation for the care team. This cluster determines the duration of the baseline length (2, 3, 4, or 5 weeks). The first 4 locations will be randomly assigned to one of the 4 clusters using a random sequence generator. Subsequent locations will be assigned to the cluster with the fewest participants at that moment. If multiple clusters have the same number of participants, a random sequence generator will be used to decide the assignment. The risk of unequal exposure to external influences is limited due to the initial randomization of the first 4 locations and the balanced allocation strategy of the following locations, which aims to maintain the balance between the clusters.
Within each participating care organization, the research period for all locations will begin simultaneously.
In this study, the social robot Tessa will be used to support the participants. Tessa was developed by Tinybots [
Each participant will tailor the use of the social robot to their own goals. Before the start of the study, the type, frequency, and duration of support provided by the care professionals, specifically regarding daily structure and planning, will be assessed. In consultation with the care professional, 3 to 5 goals will be selected to work on with the support of the social robot. These can involve (1) activities of daily living (such as brushing teeth, shaving, drinking enough water, and getting up and going to bed on time); (2) household activities (such as doing the dishes, cleaning, changing the bed, and making a shopping list); and (3) planning (such as being ready for an appointment on time or being aware of the daily schedule).
The selected goals will be formulated using the goal attainment scaling (GAS) method [
To integrate the social robot into the care process, care professionals need to be trained through a demonstration and guided practice in using the technology. This will be carried out during a training session, in which an implementation expert will install the social robot together with the care professional or the care team. After this, the care professional will introduce the social robot to the participant. Besides learning how to use the technology, care professionals will need to change their working routines, as the social robot will handle reminders that they previously provided. Thus, they will need to deimplement CAU for these specific goals. This will be discussed with the care professionals during weekly phone calls.
The social robot will be integrated into the care process, with the research team monitoring its implementation by logging and resolving support requests related to Tessa. During weekly phone calls with the care professionals, a researcher will check whether they have encountered any prevailing problems and directly help to solve these problems (eg, technical problems, problems with programming the messages, and communication issues between the various parties involved).
During the baseline phase, the participant will receive CAU from care professionals. The care for individuals with challenges in structure and daily planning usually consists of several components, as described in
Visual aid: the care professional provides instructions on paper or in digital format
Hints: the care professional provides verbal instructions before the participant performs an activity
Supervision: the care professional checks and provides general instructions during or after the performance of an activity
Verbal instructions: the participant performs an activity while the care professional instructs the participant on what to do
Demonstration: the care professional shows how to perform the activity, which the participant then replicates
Doing tasks together: the care professional provides physical assistance or takes over part of the activity, while the participant performs the rest
Taking over: the care professional performs the entire activity on behalf of the participant, who is not required to be involved.
The study uses mixed methods, yielding quantitative as well as qualitative data, to answer the RQs.
The primary outcome variable is the support provided by care professionals to the participant per week. These support moments will be registered daily in a “support diary” for each participant throughout the research period. Each day, the care professionals involved in the support of the participant will log the frequency and duration of each support moment provided to the participant during their shift. On the basis of the intake information provided by the care professionals, the diaries will be standardized and personalized by prefilling the typical daily prevalent types of support relevant to each participant. In this way, the caregivers only need to log frequency and duration of these predefined types of support and provide content descriptions only in cases of nonroutine support. To ensure consistency, the care professionals will receive detailed instructions on how to fill in the diaries by a standardized procedure, which includes participation of the researchers in team meetings and a “test week” before the study period to practice diary logging. Throughout the study, the researcher will hold weekly meetings with a designated care professional at each location to monitor compliance. In addition, to assess interrater reliability of the diary entries, on-site observations will be performed by the researchers.
The care professionals will log the content of the support they provided to the participants in the support diary. During weekly calls with the care professional, a researcher will also inquire whether there are changes in the content of the provided support and what these changes are.
The Personal Wellbeing Index–Intellectual Disability will be used to measure well-being. The scale is designed specifically for individuals who have an intellectual disability or other form of cognitive impairment and is considered a state-of-the-art instrument to measure subjective well-being in this population with acceptable psychometric properties [
The 10-item version of the Psychosocial Impact of Assistive Device Scale-Adults measures the psychosocial impact of CiSR, on items such as happiness, independence, self-esteem, and sense of control [
Both questionnaires will be filled in as part of an in-depth interview. During this interview, participants will also be asked about their experiences with the social robot and its added value, with questions such as “What do you find most helpful about Tessa?” or “What would you miss the most if Tessa would be no longer available to you?” The interviews will be guided by a specific topic list, which is available upon request from the corresponding author.
To measure work engagement, the care professionals and involved care teams will be asked to complete a short questionnaire, consisting of eight 5-point Likert scale questions, such as “Working with the social robot brings me more enjoyment in my work” or “Working with the social robot helps me save time.”
Complementary in-depth interviews with care professionals will be held to explore their experiences with the implementation of the social robot and how it affects their work engagement. In addition, care professionals will be asked about the perceived added value of CiSR for each client and how it influences the support provided to this participant, with questions such as “How do you perceive the impact of using Tessa on the participant’s progress toward their personal goal?” or “How do you incorporate the use of Tessa into the support you provide to the participant?” The interviews will be guided by a specific topic list, which is available upon request from the corresponding author.
Together with the intervention costs, such as the purchase of the social robot and subscription costs, medical consumption and productivity costs will be considered. To measure the medical consumption and productivity costs, a questionnaire containing relevant items from the iMTA Medical Cost Questionnaire and iMTA Assessment Productivity Cost Questionnaire will be conducted at baseline, effect, and follow-up [
To measure the participants’ quality of life, the Icepop Capability Measure for Adults (ICECAP-A) will be administered [
Study measurements overview.
| Measurement | Baseline | Intervention adjustment | Effect | Follow-up | Used to answer RQa |
| Daily registration of support (diary)b | ✓ | ✓ | ✓ | ✓ | RQ1 and RQ2 |
| Weekly evaluation of goals (GASc)b |
|
✓ | ✓ | ✓ | RQ3 |
| Interview: well-being (PWI-IDd)b | ✓ | ✓ | ✓ | ✓ | RQ4 |
| Interview: psychosocial impact (PIADS-10e)f |
|
|
✓ | ✓ | RQ4 |
| Interview: added value of CiSRf,g |
|
|
✓ | ✓ | RQ4 |
| Interview: work engagement and added value of CiSRh |
|
|
✓ | ✓ | RQ3, RQ4, and RQ5 |
| Questionnaire: work engagementh |
|
|
✓ | ✓ | RQ5 |
| Questionnaire: medical consumption and productivity cost (iMCQi+iPCQj)b | ✓ |
|
✓ | ✓ | RQ6 and RQ7 |
| Questionnaire: quality of life (ICECAP-Ak)b | ✓ |
|
✓ | ✓ | RQ6 and RQ7 |
aRQ: research question.
bInformation on individuals with challenges in daily structure and planning, provided by the care professional.
cGAS: goal attainment scaling.
dPWI-ID: Personal Wellbeing Index–Intellectual Disability.
ePIADS-10: 10-item Psychosocial Impact of Assistive Devices Scale.
fInformation on individuals with challenges in daily structure and planning provided by the individuals themselves.
gCiSR: care involving a social robot.
hInformation on care professional.
iiMCQ: iMTA Assessment Medical Consumption Questionnaire.
jiPCQ: iMTA Assessment Productivity Cost Questionnaire.
kICECAP-A: Icepop Capability Measure for Adults.
Due to the staggered design, power was determined through simulation using the R (R Foundation for Statistical Computing) packages simstudy and lme [
Primary outcome data will be analyzed using a multilevel model (R, version 4.0+; package lme4) to investigate whether there is a significant difference between the level of professional care support moments (frequency or duration) per week in the baseline phase and the effect phase. The dependent continuous variable is the frequency and duration of professional care support moments per week. We will analyze all professional care support moments provided, as well as a selection of the specific professional care support moments connected to the goals for which the social robot is deployed.
The independent binary variable is the moment of measurement (baseline phase or effect phase). Because there are multiple measurements per participant at both phases, the model will contain a random intercept for participants. Time will be included in the model as a covariate. Measurements are consequently averaged and compared within participants per phase. Following the intention-to-treat principle, all participants will be included in the analysis, regardless of whether the participant has used the social robot in all weeks during the intervention adjustment phase and effect phase. Because multilevel models handle missing values well, no imputation is needed [
Clinical relevance will be determined per participant. Here, two conditions must be met as follows: (1) a Reliable Change Index of >1.96 for the level of support provided by care professionals [
The follow-up data will be used to determine how many participants continue to receive CiSR, whether the effect is sustained in these participants, and what reasons there are to discontinue CiSR.
As recommended by recent single-case intervention research design standards [
Qualitative data will be coded and analyzed thematically using Atlas.ti (version 8.4.20; Lumivero, LLC).
A model-based economic evaluation with a follow-up time equal to the life span of the social robot will be performed. This study will be reported by following the Consolidated Health Economic Evaluation Reporting Standards 2022 guidelines for reporting economic evaluations in health care [
If the collected data suffices, the cost-effectiveness of CiSR will be examined by a model-based economic evaluation. For this, a Markov model, following the established guidelines [
For the analyses, R (version 4.0+; package lme4) will be used. Missing measurements will be handled using multiple imputations
If, in practice, it appears that the data are of insufficient quality, making it impossible to create a model, the costs and effects will be described with a cost-consequence analysis.