Effectiveness of a support intervention for family caregivers and stroke survivors

Objective: to analyze the effectiveness of a support intervention on the burden and stress of family caregivers and on the stroke survivors’ independence level, compared to the Control Group. Method: a quasi-experimental study conducted with 37 participants (Intervention Group, n=20; and Control Group, n=17). The intervention lasted 8 months. The outcomes of the caregivers (burden and stress) and of the survivors (independence level) were measured by the Zarit, Perceived Stress and Katz scales, at the following moments: pre-intervention, the fourth month of the intervention and post-intervention. The differences of these outcomes between groups and intra-group and the effect size were calculated using the Mann-Whitney and Friedman tests (Bonferroni adjustment by Wilcoxon) and the Kendall’s W coefficient. Results: the Intervention Group reduced burden (p=0.039) and stress (p=0.009), mainly, after 8 months of intervention, which was not observed in the Control Group. The independence level did not change between the groups or moments (p>0.05). The intervention presented moderate effect size (p=0.45 and p=0.54). Conclusion: the intervention was effective to reduce the burden and stress of family caregivers, but did not alter the stroke survivors’ independence level, when compared to the Control Group.


Introduction
Stroke is one of the main causes of chronic cognitive and functional impairments in adults and aged individuals worldwide. The rehabilitation of stroke survivors is generally long; the family, or a single family caregiver, assumes home care, with the execution of complex tasks and without due preparation, which results in physical, economic, emotional and psychosocial burden (1) .
In the absence of a source of social support, especially health services and health, this burden can increase and jeopardize both the health of the caregiver and the quality of care provided to the patient. This can result in the emergence or worsening of stress in the caregiver and in the difficulty recovering the autonomy of the stroke survivors to carry out the everyday basic activities (2) .
These negative consequences for the caregiverstroke survivor dyad show the need for reducing such burden. Thus, the literature recommends enabling to these caregivers multicomponent interventions that emphasize care for their own health, while providing continuous training for the care of the stroke survivor (3) .
Nurses can be at the forefront among the health professionals that are part of formal social support (4) , establishing partnerships with other members of the multidisciplinary team in the implementation of strategies or multicomponent interventions to provide support, education and advice to the caregivers, either individually or collectively.
These interventions must be developed considering the caregivers' needs (5) and the recovery phase of the stroke survivors, which is divided into acute phase (occurring between the transition period from hospital discharge to the first six months in the community) and the chronic phase (starting from six months after returning to the community) (3) .
Although the caregiver's burden remains throughout the care period (6) , there are more publications of intervention studies in the acute phase, which reveals a need for more interventionist research studies in the chronic phase (3) , since the caregivers need constant guidelines on elementary and complex care, knowing that problems can emerge during the care period.
Most of the research studies on support interventions after stroke, focusing on the caregivers, come from highincome countries. In Brazil, the interventions to reduce the burden on the family caregiver family caregivers of survivors of this disease based on scientific evidence are scarce (5) , especially in the chronic phase of stroke. In this sense, interventions with low financial cost and easy development can be viable in health care, especially in the public sector, so as to supplement routine care or as an alternative to it, usually provided by the public and private health services.
As the burden generally has relation with the caregiver's stress and with the stroke survivor's independence level (7)(8) , the following hypothesis will be tested: There will be differences in the burden and stress of family caregivers and in the independence level of stroke survivors in the Intervention Group, when compared to the Control Group. Therefore, the objective of the study was to analyze the effectiveness of a support intervention on the burden and stress of family caregivers and on the stroke survivors' independence level compared to the Control Group.

Study design
A prospective and quasi-experimental study, based on the Transparent Reporting of Evaluations with Nonrandomized Designs (TREND) statement, which represents a supplement to the Consolidated Standards of Reporting Trials (CONSORT) statement (9) .
Randomization is a process that distributes a "random" number of participants in different groups (experimental and control) in order to preserve the characteristics that similar between them. For a study to be classified as randomized, it is necessary that all the randomization stages (allocation sequence generation, concealment and implementation mechanism) are rigorously performed (10) .
Thus, this study was not classified as randomized because allocation sequence generation occurred by means of the only alternative and acceptable method called minimization, which, although not eliminating bias in all the known and unknown factors, allows for the balance between the Intervention and Control groups, by selecting the participants' factors (10) . The choice of this method is recommended for smaller groups and was necessary due to the lethality of stroke, which could further reduce the sample.

Participants and sample size
The sample was for convenience and the study took place in the city of Guanambi, Bahia (BA), Brazil. The www.eerp.usp.br/rlae 3 Silva JK, Boery RNSO.
family caregivers were recruited between September 2017 and March 2018, in their homes. Therefore, it was necessary to identify the stroke survivors in the chronic phase of the disease that had an attendance register from 2014, either in the regional hospital unit (n=152) or in the 17 Basic Health Units (BHUs) (n=69) managed by the

Intervention
The intervention consisted in supporting the family caregivers. Therefore, it was divided into two parts (individual and group) and lasted 8 months, combining the terms "acquisition of skills" and "education".
Before starting the intervention, the lead researcher (first author) conducted a previous study that investigated the needs of all the family caregivers, regardless of their future allocations in the Intervention Group (IG) or Control Group (CG). The results showed that these caregivers needed education in health about stroke and about daily care to the survivors, having free time and getting assistance for their physical and mental health (11) .
Knowing these needs supported the organization of the group and individual components of the intervention. adapted to the study context (12) .
An action plan was developed for all the sessions containing information relating to their respective stages: problematization (description of the problem to be discussed), topic, date, time, locus, objectives, methodological strategies (presentation of the proposal, video, dynamics, group discussion, etc.), food for the participants, human and material resources, evaluation of each session and references used for each theme.
Each session was recorded with a video camera by a research assistant and the following themes were covered:

Instruments and outcomes
A form containing variables about the caregiver and the stroke survivor was developed for this study, based on previous surveys (1)(2)4,(7)(8) . The primary outcome was the caregiver's burden in relation to the care provided to the stroke survivor. To assess it, the Zarit Burden Interview scale was used, consisting in 22 items that assess health condition, psychological and financial situation, interpersonal relationships and social and personal life.
The total score of the scale varied from 0 to 88, where the higher the score, the greater the burden. Although it was developed for caregivers of older adults and of people with dementia, this scale can be applied to caregivers of people with various mental and physical ailments. In Brazil, its internal consistency and validity were tested in caregivers of older adults with depression and the Cronbach's alpha value was 0.87 (13) .  (14) .
To identify the stroke survivor's independence level, the Basic Activities of Daily Living scale was applied to the caregivers. This instrument assesses the functional independence of older adults and of other dependents in six respective functions: feeding, sphincter control, transfer, going to the bathroom, ability to dress and bathe.
The score varies from 0 to 6 points, where one point is assigned to each "yes" answer, the individual being considered as follows: independent in all functions (zero score) or dependent in one to six functions. This scale was cross-culturally adapted Brazil in older adults and their caregivers, having its internal consistency tested and approved by Cronbach's alpha values ranging from 0.80 to 0.92. Its validity was not assessed directly, but the Cronbach's alpha coefficient levels provided empirical evidence of its validity (15) .
All four instruments were applied in the homes of the family caregivers. To verify the suitability of these instruments to the study, a pilot test was performed with 8 caregivers who were not part of the sample (n=37).

Allocation and blinding
After collecting data at the pre-intervention moment, a statistician (external to the research) conducted the generation of the participants' allocation sequence, through the minimization process according to the outcomes, the sociodemographic characteristics of the caregiver and the stroke survivor, seeking homogeneity of these variables in both groups. In this allocation process (minimization), the first individual was randomly allocated and, for each subsequent participant, the treatment allocation which minimized the imbalance of the factors selected from the groups was identified (10) .
Blinding (blinding of the participants' allocation in the Control and Intervention groups) was possible only for the trained nurse who applied all the instruments and for the statistician who conducted the minimization process and data analysis, since the caregivers and professionals were aware of the intervention. Kendall's W coefficient assumes that values equal to or near to 0 indicate a negligible effect size and that values close or equal to 1 indicate large effect (16) . The Statistical Package for Social Sciences (SPSS) software, version 21.0, was used for all the statistical analyses, considering a significance level of 5%.

Results
The results are presented according to the TREND statement. Of the 44 participants eligible for the study, 37 consented to participate, being assigned to the Intervention (n=20) and Control (n=17) groups.  Table 1. which was not the case in the Control Group (Table 2).  community resources and applying knowledge about the care techniques (18)(19) .
Unlike the aforementioned studies that were developed in the chronic phase of stroke, this support intervention used two groups (Intervention and Control), which reinforces the greater effect of the group strategy in reducing the caregivers' burden and stress during home care, especially in the psychosocial aspects.
Generally, coexistence in a support group awakens the perception of not being alone and is preferable to individual care. Social support between the group members and the professionals improves the feeling of emotional well-being (20) , mainly in the sharing of successful experiences for problem solving. The participants of this study also made these findings explicit. Similarly to this support intervention, another family-centered study, which used a short message service and cell phone calls for 8 weeks, with the objective of improving the basic daily activities of stroke survivors, also failed to increase the participants' independence in the chronic phase of this disease (21) . It is probable that the focus of this support intervention, in responding only the unmet needs of the caregivers, influenced this result. Stroke survivors may also have reached the apex of their motor recovery prior to the study, as it is during the acute phase that the professionals intensify the use of therapies for musculoskeletal strengthening (22) and, therefore, after hospital discharge, the survivors can present functional decline. In addition to that, the way in which the family caregivers stimulated the stroke survivors' mobility and independence at home after receiving guidelines in the group component was not verified in this research.
Verifying these care practices can ease understanding of the effect of the intervention on the independence of these survivors for the daily activities.
Most of the recovery of the motor functions occurs until the sixth month post-stroke and, after this period, functional capacity decreases progressively until it stabilizes (23) . Some factors can interfere with the recovery of these functions, such as the early use of antithrombotic therapy, type of stroke (23) , its severity at admission and advanced age of the survivors (24) , its severity at hospital admission and older age of survivors  (20) , as well as to their effect towards the stroke survivors. In Brazil, this scenario is even more critical, since these interventions in health are still scarce. Therefore, this support intervention that has been developed represents In intervention studies, it is not always possible to blind researchers and participants due to the interaction between them, which can generate some bias in the result. However, to reduce this risk, the evaluator and data collector were blinded and a Control Group was used, as recommended by the TREND statement, which represent strengths of this study. In addition to these points, the minimization process was adopted for the participants allocated in the Intervention and Control groups, since it was not possible to perform randomization.
The limitations of the study refer to the sample size and to the non-randomization of the sample, which precludes generalizing the results to other groups and, therefore, the data must be interpreted with caution.
Therefore, in future research studies, a larger sample selected by randomization can be used, especially when analyzing sustainability over time. Even so, the findings in this study may help in the design of such research studies.

Conclusion
The intervention was effective to directly reduce the family caregivers' burden and stress in the post-stroke chronic phase. However, the repercussion of its effects on stroke survivors was not identified, as they did not present changes in their independence level for the performance of the basic daily activities, when compared to the Control Group.
Such findings reinforce that care to the caregivers is essential to improve their physical and psychological health and should not be limited to the early months of their role. In addition to assessing and incorporating the stroke survivors' needs, in this intervention, may generate more benefits for these participants.