Burden of informal caregivers of people without natural speech: a mixed-methods intervention study

This study focuses on examining self-perceived care-related burden, resources, and coping among a group of informal caregivers that has been little studied to date. These are caregivers of people who have no or no intelligible natural speech due to a congenital or acquired disability and who use augmentative and alternative communication (AAC). It is estimated that worldwide, about 97 million people have either severe limitations in using natural speech or no natural speech at all [13]. For Germany, there are no reliable data on the prevalence of people who use AAC. The group of people who use AAC is heterogeneous in terms of age, disability, and the extent of physical, intellectual, and communication limitations [14]. In a cross-sectional study, we have already found significant interrelations between caregiver burden and caregiver assessment of the health-related quality of life and functioning of people without natural speech [15]. Although it is unclear whether (1) higher burden leads to poorer above-mentioned outcomes for people without natural speech, (2) whether poor outcomes lead to higher burden, or (3) only the assessment of these outcomes is affected, the results can be seen as an indication of the major importance of burden in this care context. To the best of our knowledge, longitudinal studies examining relationships regarding care-related burden in this group do not yet exist.

Different caregiver and care recipient characteristics such as caregiver’s education level, family income, mental impairments, and multiple disabilities of care recipients are known as significant predictors of burden of informal caregivers of people with disabilities [16]. Nevertheless, perceptions of caregiving as well as coping with challenges can vary widely between individuals [5].

To better examine and understand this variance and the caregiving situation of caregivers of people without natural speech, the transactional model of stress and coping (TMSC) by Lazarus and Folkman [17] will be used as a theoretical framework for the present study. The TMSC explains stress responses by two essential reactions, cognitive appraisal and coping. In cognitive appraisal, a stressor is initially classified as positive, harmful, or irrelevant. If a stressor is classified as harmful, secondary appraisal occurs, in which the availability of resources for handling the stressor is being assessed. As a result, the situation is classified as challenging, harmful, or threatening. Subsequently, the person makes cognitive (emotion-focused) and behavioural (problem-focused) efforts to cope with the stressor that has been classified as harmful or threatening. Further coping efforts, if needed, occur after the reappraisal of the stressor [18, 19]. The focus of the present study will be primarily on identifying and examining stressors, resources, and coping strategies. In our study, we use the term “stressor” not in the meaning of a neutral environmental stimulus before the appraisal stage, but as a stimulus that has already been classified as harmful or threatening and leads to the experience of stress.

Data were collected within the research project “New Service Delivery Model for Augmentative and Alternative Communication (AAC) Devices and Intervention” funded by the Federal Joint Committee’s Innovation Fund in Germany [20]. This publication represents one part of the larger study. The primary goal of the developed complex intervention was to improve the communication skills and consequently the health-related quality of life and functioning of people without natural speech (results on these main outcomes will be presented elsewhere). The complex intervention was evaluated in three AAC counselling centres in Germany. The new service delivery model (nSD) that represents the intervention studied includes

Another component of the intervention is the accompanying case management, through which the AAC intervention process is coordinated and the collaboration between the involved stakeholders, is strengthened [20]. The recommendation for an AAC system is based on the participants' skills and needs and includes both aided AAC systems (e.g., voice output devices or paper-based communication books) and non-aided AAC systems (e.g., gestures). The intervention implies the mandatory participation of an informal caregiver in the initial counselling session. This allows a detailed assessment of the different activities and settings that are important for the care recipient, with the aim of selecting the most appropriate AAC system. The participation of caregivers in the following AAC training and therapy sessions is not mandatory, but desirable at any time according to the needs and possibilities of participants and their caregivers. It is important to mention that within the intervention it is intended and encouraged that AAC training and therapy sessions can take place in different settings (e.g., school, kindergarten, residential home). The aim is to ensure the successful use of the AAC system in many contexts and with various stakeholders. This flexibility already implies that it is not necessarily required for informal caregivers to be present at all AAC training and therapy sessions. Case management aims to ensure that informal caregivers are continuously involved in the exchange with other stakeholders about the AAC training and therapy process and the use of the AAC system in other contexts. In most cases, informal caregivers are the main contact persons of the AAC counselling centres and therefore take a central role in the intervention. The present study addresses the caregiving situation of informal caregivers. Although the intervention does not primarily address the caregivers’ situation, it is anticipated that output such as improved AAC system supply, optimised collaboration with stakeholders, and improved communication with the cared-for person can lead to a reduction in burden in the care context. The reduction in burden can therefore be defined as an unintended, but expected positive consequence or “side effect” of the intervention.

Within the quasi-experimental intervention design, the control group consists of people without natural speech and their caregivers receiving service delivery under the existing contract (SDeC). This group receives only an initial counselling session not followed by training, therapy or case management. Since only insured persons of a specific health insurance company can participate in the nSD, health insurance affiliation represents the only anticipated difference between the intervention group nSD and the control group SDeC. Quantitative data were collected by means of a standardised postal survey of informal caregivers after the initial consultation (T0), 4 weeks after AAC system receipt (T1) and 4 months after AAC system receipt (T2).

This study is an interventional mixed-methods study with an embedded design. Parallel to the AAC intervention and the long period of quantitative data collection, qualitative data on stressors, resources, and coping were collected and analysed. The findings of both methods are integrated in the results section and jointly considered in the discussion [21, 22]. While the quantitative data focus on burden level, the qualitative data address specific stressors, resources, and coping strategies. Figure 1 illustrates the mixed-methods study design used.

The inclusion criterion was the presence of a congenital or acquired disability associated with loss of natural speech. The need for AAC can occur at any age and be associated with a variety of disabilities; therefore, people of all ages and disabilities were included to address this heterogeneity. Participants in the two groups were recruited during the initial consultation at one of the three participating AAC counselling centres. The potential participants received all information about the study in written form. Their affiliation with a health insurance company and thus their group affiliation was inquired in advance by telephone. Participation in the study required the naming of two caregivers (formal and informal) to whom the questionnaires would be sent. Only informal caregivers are included in the present study. For the qualitative interviews, caregivers of participants in the nSD group were recruited. Only caregivers of participants who have completed at least 4 AAC training sessions and 10 AAC therapy sessions at the time of the interview were recruited. For participation in the interviews, it was not crucial whether the caregivers attended the AAC training and therapy sessions with the participants or not. The method of purposeful sampling [23] was used to ensure a diverse sample in terms of the relation of the interviewed person to the cared-for person, age, and sex (see Table 1).

Within the quasi-experimental survey study, quantitative data were collected between June 2018 and April 2021. The questionnaires were checked for comprehensibility and practicability in cognitive pretest interviews (n = 16) [24] and were further adapted if needed (adaptations were made, but not in relation to the instrument used to measure burden). Caregiver burden was measured with the validated German short version of the Burden Scale for Family Caregivers (BSFC-s; Cronbach's α = 0.92; see supplementary material) [25]. Caregivers were asked to provide information on their emotional and physical burden using 10 items to be answered on a 4-point Likert scale from 0 strongly disagree to 3 strongly agree. Item scores were added up and divided by the number of items. The T0 questionnaire contained various sociodemographic questions (see the data analysis section). Semi-structured interviews with n = 16 informal caregivers from the nSD group were performed. The first interviews with n = 8 caregivers (2 individual and 2 focus group interviews) took place at the three AAC counselling centres involved in the project. Due to the COVID-19 pandemic, the following interviews with n = 8 caregivers were conducted by telephone. All participants were informed verbally and in writing about the interview’s aims and the data protection measures and were asked to give their written informed consent prior to the interviews. The interview guideline was developed according to the underlying theoretical model and the defined research questions. The semi-structured interview guideline [26] contained the following guiding questions:

The duration of the interviews was between 8 and 26 min.

The qualitative study allowed data to be collected from n = 16 informal caregivers. The sociodemographic characteristics of this group are presented in Table 1.

Data from n = 154 informal caregivers (n = 88 in the nSD group and n = 66 in the SDeC group) could be analysed. The sociodemographic characteristics of informal caregivers and of the people without natural speech to whom they provide care are presented in Table 2.

The recorded material was transcribed verbatim and pseudonymised. The interview data were analysed using the software MAXQDA Analytics Pro 2020 (version 20.3.0) by structured qualitative content analysis following Kuckartz [27]. Initially, a priori main categories were defined according to the interview guideline and the underlying theoretical model (TMSC). During the coding process of the first interview, subcategories were inductively formed. Two researchers discussed the developed category system in a consensus-building process before conducting the remaining analysis. The remaining coding process was carried out by two researchers independently with subsequent consensus finding.

The paper questionnaires were electronically imported using the Electric Paper TeleForm software. The data were then exported to the IBM SPSS V.27 software and underwent extensive plausibility checks. Missing values of the BSFC-s scale were imputed using the expectation-maximization algorithm [28]. For n = 13 cases with T0 and T2 data, but no T1 data, last observation carried forward (LOCF) analysis was performed, and thus, missing T1 values on the BSFC-s scale were imputed with data from T0. To address potential group differences, certain sociodemographic variables were used to form the propensity score (PS). The PS is defined as the probability of being assigned to a certain group, in this case intervention vs. control group [29]. The PS was calculated based on caregiver characteristics (sex, age group, native language, and education) and on characteristics of persons without natural speech (sex, age, employment status, and congenital vs. acquired disability). To determine whether there is a significant difference in burden between the nSD and SDeC groups at different time points, first a mixed ANCOVA with repeated measures controlled for T0 and after that a mixed ANCOVA with repeated measures controlled for T0 and the calculated PS were performed [30].

The stressors identified in the qualitative analysis can be divided into two main areas: on the one hand, there are primary or care-related stressors that are directly related to daily care activities, such as limited communication with the cared-for person or the feeling of being overwhelmed by certain behaviours. On the other hand, there are secondary or psychosocial stressors that result from the primary stressors (e.g., giving up employment in favour of care tasks and the resulting financial difficulties). A detailed explanation of all subcategories as well as sample quotations can be found in Table 3.

The resources available for dealing with the stressors reported in qualitative interviews can be divided into two main categories. First, this entails internal resources, such as feeling a strong bond with the care recipient. Secondly, resources could be identified that originate from the environment, such as support with daily caregiving tasks from family and friends (see Table 3).

The coping strategies reported by caregivers in qualitative interviews could be divided into two main categories. Emotion-focused strategies, such as relativising one's own problems by comparing them with other families' problems, aim to change one's own perception of the (caregiving) situation for the better. Problem-focused strategies, such as detailed organisation of daily routines or utilisation of professional services are aimed at changing the situation itself (see Table 3).

The mean burden scores of the two groups (Table 4) are similar at baseline T0 (nSD mean = 1.03; SDeC mean = 1.04). By the T1 time point, burden has decreased slightly in the intervention group (nSD mean = 0.97) and increased slightly in the control group (SDeC mean = 1.07). This development is maintained from T1 to the T2 time point, showing a decrease in burden in the intervention group (nSD mean = 0.92) and an increase in burden in the control group (SDeC mean = 1.20). The raw sum scores of BSFC-s necessary to categorise the burden level are in the category "moderate" (range between 9.193 and 12.000) in both groups, nSD and SDeC, at all 3 time points [31]. Tests for pairwise mean differences within each group and between groups were performed and the results are shown in Table 4.

The results of the mixed ANCOVA, controlling only for T0 as a covariate, are shown in Table 4. The significant effect for the group confirms the second research hypothesis: after controlling the baseline values at T0, the measured values in the intervention group nSD are significantly lower than in the control group (F(1, 151) = 4.386, p = 0.038). The statistical effect is small to moderate (partial η² = 0.028). However, if PS is also controlled for, the intervention effect no longer proves to be statistically significant, and the effect size decreases considerably to T2 η² = 0.009. Accordingly, the measured main effect for the groups is not to be regarded as being due to the intervention, but as resulting from confounding effects due to the group composition (F(1, 150) = 1.405, p = 0.238, partial η² = 0.009). The significant interaction of time and group, however, turns out to be dependent on a confounding effect: after control for T0: F(1, 150) = 4.574*, p = 0.034, partial η² = 0.029; after additional control for PS: F(1, 150) = 4.050*, p = 0.046, partial η² = 0.026. Accordingly, the decrease in values from T1 to T2 in the intervention group can be considered stable in relation to the increase in values in the control group from T1 to T2. The positive change in the sense of an improvement in the intervention group nSD, in contrast to the deterioration in the control group SDeC, is to be regarded as a systematic intervention effect in accordance with the research hypothesis.

The caregivers interviewed consistently rated the intervention as beneficial, both in terms of improving the cared-for person's communication skills and in terms of reducing their own burden in the caregiving situation. It became evident that these two aspects are strongly interrelated. The reduction in burden could be primarily attributed to the following two aspects that were perceived as (more) burdensome prior to the intervention: first, improved communication of needs and desires of the cared-for person led to an increase in frustration tolerance, reduction in behavioural problems and thus to a facilitation in the management of daily care tasks. Second, the support received through the intervention in communicating with stakeholders also led to a reduction in burden.