Effect of support groups on caregiver’s quality of life

Family Process. 2021;00:1–16. | 1 wileyonlinelibrary.com/journal/famp Received: 13 September 2019 | Revised: 21 April 2021 | Accepted: 23 April 2021 DOI: 10.1111/famp.12684 ORIGINAL ARTICLE Effect of support groups on caregiver’s quality of life ClaudiaBernabéuÁlvarez | JoaquínSalvadorLimaRodríguez | MartaLimaSerrano This is an open access article under the terms of the Creative Commons AttributionNonCommercialNoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is noncommercial and no modifications or adaptations are made. © 2021 The Authors. Family Process published by Wiley Periodicals LLC on behalf of Family Process Institute Faculty of Nursing, Physiotherapy and Podiatry, University of Seville, Seville, Spain Correspondence JoaquínSalvador LimaRodríguez, Faculty of Nursing, Physiotherapy and Podiatry, University of Seville, Street Avenzoar, 6, Postal Code 41009, Seville, Spain. Email: [email protected] Abstract The objective of this study was to evaluate the effect that participating in support groups for caregivers has on the quality of life and psychotropic drug use of family caregivers of adults with limitations in activities of daily living. A controlled quasiexperimental longitudinal design was used with 134 caregivers (64 in the experimental group and 70 in the control group). The outcomes were healthrelated quality of life (EuroQol 5D3L test) and psychotropic drug use (no/yes). The analyses were performed using SPSS and R statistical software. An interaction was observed between the condition and the level of limitations in activities of daily living of the care receiver, having an effect on the caregiver's psychotropic drug use (p=0.003), with this use being lower among caregivers who attend support groups when their relatives present fewer limitations in activities of daily living. Moreover, the quality of life was higher in the posttest in the experimental group (B=8.66, p=0.015). In conclusion, support groups could improve the caregiver's quality of life and decrease psychotropic drug use when the care receiver has low limitations in activities of daily living. KEYWORDS Anxiety, Family caregiver, Quality of life, Psychotropic drug use, Support group 2 | FAMILY PROCESS INTRODUCTION Aging— as well as disabling diseases such as dementia, Parkinson's disease, cardiovascular pathologies, mental disorders, or addictions— increases the need for family care (Pihet et al., 2017; WHS 2020) to provide medication, emotional support, or whatever suits the signs and symptoms and allows to observe the evolution of a disease (LimaRodríguez et al., 2015). Good family care reduces the risk of a patient being hospitalized or the need for emergency services. Moreover, it reduces family anxiety when it is paired with familycentered care and support for the family caregiver (Deek et al., 2016). However, family caregivers usually suffer from unsatisfied needs related to diet, physical exercise, or leisure time (Tatangelo et al., 2018), together with a feeling of guilt. Full dedication to providing care can affect caregivers’ health and quality of life, especially when there is only one person who takes on the caregiver role (Sandstedt et al., 2018; Ullgren et al., 2018). Caregivers are even more prone than the rest of the population to developing anxiety and/or depression symptoms and sleep disorders (CarodArtal et al., 2013; Dahlrup et al., 2015; Jamani et al., 2018; Morimoto et al., 2003) and to using psychotropic drugs such as antidepressants and anxiolytics (Camargos et al., 2012; Sallim et al., 2015). A metaanalysis on the prevalence of mental health disorders among caregivers of people with Alzheimer's disease showed a prevalence of depression of 34.0% (2.51 times more likely in spouse caregivers), anxiety of 43.6%, and psychotropic drug use of 27.2% (Sallim et al., 2015). Moreover, previous studies showed the relationship between psychotropic drug use and the caregiver burden (Treichel et al., 2020), as well as the lack of social support (Camargos et al., 2012; Colell et al., 2014; Mateo et al., 1997; Pérez et al., 2009). Support groups (SGs) are an example of social care provided to caregivers. SGs are made up of people who share a common problem, and they gather voluntarily and freely in order to cope with and overcome their situation through personal and/or social changes (Ahmadi, 2018). In 1994, Monking (1994) showed that attending SGs reduces the burden and physical complaints of those who care for people with mental health disorders. Other authors found that different types of caregiver SGs— especially for people with Alzheimer's and other dementias— had positive effects on psychological wellbeing, depression, burden, and social support (Chien et al., 2011; Guay et al., 2017), as well as on selfefficacy and quality of life (ParkerOliver et al., 2017), although the effects on the latter were not conclusive (Cheng & Zhang, 2020). Regarding psychotropic drug use, previous authors stated that SGs could help to reduce drug prescription (Hunot et al., 2007; Pfeiffer et al., 2011). The integration of psychosocial processes, such as the exchange of emotions and feelings, experiences and information, as well as peer learning, could translate into the benefits mentioned. SGs help caregivers to express repressed emotions, which gives them a sense of relief and protection. The exchange of experiences allows to validate these experiences and produces the feeling of having something in common, but also a feeling of autonomy in caring. In this exchange, there is open discussion, information, and advice. Regarding the information, this involves anything related to community resources and coping and problemsolving strategies. Being part of a new group allows people to increase their social network, reducing feelings of loneliness and promoting continuity of care (BernabéuÁlvarez et al., 2020; Cipolletta et al., 2019; Toseland et al., 1990; Trail et al., 2020). When the SGs are led by professionals, the caregivers can learn new skills for behavioral change, identifying barriers, and problemsolving (Domínguez Orozco, 2012; Guay et al., 2017). We decided to carry out this study because we did not find any studies that assessed the effect of SGs on caregivers’ psychotropic drug use, and neither did we find any controlled experimental or quasiexperimental studies in our field that evaluate the effect of SGs on caregivers’ quality of life. We started from the hypothesis that support groups have a positive impact on | 3 BERNABÉUÁLVAREZ Et AL. healthrelated quality of life and reduce the psychotropic drug use of caregivers of people with limitations in activities of daily living, compared to those who do not participate in SGs. METHOD Study design We used a quasiexperimental longitudinal design with an experimental group (EG) and a control group (CG) and took measurements at the beginning of the study and 6months after. CONSORT guidelines (Montgomery et al., 2018) were followed. The data collection process was carried out between January and September 2016 in Primary Health Care areas in Seville, Huelva, and Cadiz (Spain). Population and sample The study population were caregivers of adult people with limitations in activities of daily living (people who need help to carry out certain basic activities of daily living at least once a day, every day [Ministry of equality and social policies, 2016]). A sample size of 51 participants was estimated for each group (experimental and control groups) in order to perform a onetailed comparison of the averages between two independent groups (ttest), with an alpha error of 5%, a power of 80%, and a mean effect size (d=0.5), taking into account a dropout rate of 40% (BergerHöger et al., 2019; Walters, 2004). The program G*Power 3.1.7. was used. Finally, in the pretest, there were 134 caregivers, 64 in the EG (47.8%), and 70 in the CG (52.2%). The caregivers were selected through intentional sampling, and they were stratified by geographical area, with 47.8% living in urban areas and 52.2% living in rural areas. In the posttest, the attrition rate was 20.9%, with reasons being the death of the relative, discontinued attendance at the SGs (in the EG), and refusal to continue participating in the study (Figure 1, Table 1). The participants in the EG must have been taking part in support groups for less than a year and attending at least 60% of the sessions. Ten SGs for caregivers were recruited, two for caregivers of people with Alzheimer's, one of people with epilepsy, two of people with addictions, one of people with Parkinson's disease, one of people with mental disorders, and three of people with limitations in activities of daily living without a specific pathology. The SGs were led by professionals, consisting of psychologists (5), nurses (3), and social workers (1), except two SGs that were led by the caregivers themselves. The number of participants ranged between 4 and 15. The frequency of meetings was biweekly for half of the groups and weekly for the other half. The duration of sessions oscillated between one hour and a half, and two hours. They were open groups that participants could join or leave without a session limit established over time. Among the support techniques used were free expression of emotions, offering of mutual support, identification of stressful situations and behaviors that could be changed, knowledge of what the illness entails, and acquisition of empowerment tools. Resources such as specific literature, music, relaxation techniques, laughter therapy, or professional support could be used. To avoid selection bias whenever it was possible, the participants in the CG were selected from the same geographical area, and we tried to make it so that the relative of the caregiver in the CG had a similar pathology to that of the caregiver in the EG. However, it was difficult to match both groups, mainly due to difficulties of finding similar caregivers or because some caregivers declined to participate in the study. As a consequence, there were differences in 4 | FAMILY PROCESS some characteristics of the caregivers, as well as in the limitations in activities of daily living of the relatives of both groups (Table 1). Measurement The independent variable was the caregivers’ attendance/nonattendance at the SGs. The dependent variables were the healthrelated quality of life, measured through EuroQol5D, and the use of psychotropic drugs. These variables were measured at the beginning of the study and 6months after. EuroQol5D3L (Brooks, 1996; Herdman et al., 2001) comprises five health dimensions (mobility, selfcare, activity, pain/discomfort, and anxiety/depression), and each one of them has three levels of severity, which are no problems, some or moderate problems, and serious problems. The second part of EQ5D3L is the overall quality of life index, a 20centimeter vertical FIGURE 1 Flow diagram for the sample differences between pretest and posttest Note. Flow diagram for the sample differences between pretest and posttest (after 6months; eligibility, allocation, followup, data analysis) as specified in the Consort Statement (Moher et al., 2005) Data analysis Follow-up Allocation Caregivers evaluated for inclusión in the study (n=134) Experimental group Caregivers (attended SG) (n = 64) Follow up (after 6 months) completed (n=52) -Death of the sick (8) -Non-assistance to SG (4) Data analysis (n=52) -Excluded from analysis (0) Control group Caregivers (no attended SG) (n = 70) Follow up (after 6 months) completed (n=54) -Death of the sick (15) -Refusal to participate (1) Data analysis (n=54) -Excluded from analysis (0) | 5 BERNABÉUÁLVAREZ Et AL. TABLE 1 Baseline characteristics of participants by condition Variable Total Experimental group Control group Baseline difference p Age (25– 80) 58.43 (11.95) 59.19 (13.51) 57.73 (10.38) 0.496c 0.482 Relationship Parents 16 (11.9%) 8 (12.5%) 8 (11.4%) 21.160a 0.001 Spouse 45 (33.6%) 33 (51.6%) 12 (17.1%) Son/daughter 55 (41%) 15 (23.4%) 40 (57.1%) Other 18 (13.4%) 8 (12.5%) 10 (14.3%) Educational level Without studies or Primary education 41 (30.6%) 20 (31.3%) 21 (30%) 5.584a 0.134 Compulsory secondary education 35 (26.1%) 12 (18.8%) 23 (32.9%) Baccalaureate/ Vocational training 37 (27.6%) 18 (28.1%) 19 (27.1%) University studies 21 (15.7%) 14 (21.9%) 7 (10%) Gender Male 25 (18.7%) 19 (29.7%) 6 (8.6%) 9.823a 0.002 Female 109 (81.3%) 45 (70.3%) 64 (91.4%) Area Urban 64 (47.8%) 41 (64.1%) 23 (32.9%) 13.049a 0.001 Rural 70 (52.2%) 23 (35.9%) 47 (67.1%) Years caring 10.01 (10.24) 12.16 (12.7) 8.06 (6.83) 2029b 0.346 Daily hours caring 19.35 (6.63) 18.69 (7.05) 19.96 (6.2) 2001b 0.232 Ill's dependence level (Barthel test) Totally 39 (29.1%) 7 (10.9%) 32 (45.7%) 28.661a 0.000 High 43 (32.1%) 19 (29.7%) 24 (34.3%) Partially 27 (20.1%) 18 (28.1%) 9 (12.9%) Minimally 8 (6.0%) 7 (10.9%) 1 (1.4%) Independent 17 (12.7%) 13 (20.3%) 4 (5.7%) Social support (MOS test) Good 103 (76.9%) 42 (65.6%) 61 (87.1%) 8.705a 0.003 Bad 31 (23.1%) 22 (34.4%) 9 (12.9%) Likelihood of illness (The Social Readjustment Rating Scale) Low 73 (54.5%) 34 (53.1%) 39 (55.7%) 7.663a 0.022 Intermediate 46 (34.3%) 18 (28.1%) 28 (40%) High 15 (11.2%) 12 (18.8%) 3 (4.3%) Caregiver's relative illness Neurological disease 22 (16.4%) 12 (18.8%) 10 (14.3%) 53,719a 0.000 Dementias 30 (22.4%) 22 (34.4%) 8 (11.4%) Pluripathology 53 (39.6%) 7 (10.9%) 46 (65.7%) Mental or addictions 24 (17.9%) 22 (34.4%) 2 (2.9%) Cancer or palliative 5 (3.7%) 1 (1.5%) 4 (5.7%) Note: m (SD)=missing values per variable. aχ2=chisquare. bU=Mann– Whitney test. cANOVA. 6 | FAMILY PROCESS visual analogue scale measured in millimeters. It ranges from 0 (worst imaginable health status) to 100 (best imaginable health status). The use of psychotropic drugs was measured in daily doses, and participants were asked how many psychotropic drugs they consumed per day in order to relax, to improve their mood, to reduce anxiety, or to fall asleep, using as examples lorazepam, alprazolam, fluoxetine, escitalopram, or diazepam. Later, answers were transformed into a qualitative variable (“Yes”/“No”), given the heterogeneity in the daily doses. In addition, characteristics of the caregivers were measured as confounders and consisted of gender, age in years, relationship with the person with limitations in activities of daily living (parent, spouse, son/daughter, other), time dedicated to caregiving (measured in hours of the day), years dedicated to caregiving, educational level (no studies or just primary education [similar to grades 1– 6 in the United States], compulsory secondary education [similar to grades 7– 10], secondary school or professional training [grades 10– 12], and university studies) and area of residence (urban vs. rural, considered rural if the population totaled 30,000 or less), social support measured through the MOS scale (Londoño et al., 2012), and stressful life events measured through the Holmes and Rahe Stress Scale (Holmes & Rahe, 1967; Tescari Quiñones & Poveda De Agustín, 2012). The MOS Social Support Scale measures the quantity (overall social support) and the quality of social support based on nineteen items with five answer options that explore the following domains of social support: emotional/informational support, tangible support, affective support, and positive social interactions. The minimum score is 19, and the maximum is 94. The higher the score, the better the person perceived their social support (Londoño et al., 2012). Cronbach's alpha value for this study was α=0.94. We use the cutoff point of 57 to differentiate good support from bad support. The Stress Scale consists of fortythree items that reflect different stressful life events which are given specific scores. The final impact is obtained by adding the score of the different events that the person suffered in the last year. The risk of getting sick can be considered low risk (<150 points), moderate risk (between 150 and 300 points), and high risk (more than 300 points). Cronbach's alpha value for this study was α=0.67. Furthermore, the level of limitations in activities of daily living of the caregiver's relative was measured using the Barthel Index that assesses performance in different activities of daily living (bathing, grooming, dressing, feeding, toilet use, mobility, climbing stairs, and bowel/ bladder control). The range varies between 0 (greater limitations in activities of daily living) and 100 points (fewer limitations in activities of daily living), with 5 cutoff points: total limitation (0– 20), high limitation (21– 60), partial limitation (61– 90), minimal limitation (91– 99), and no limitation (100 points; Mahoney & Barthel, 1965). This questionnaire showed Cronbach's alpha index of 0.943. Procedure The recruitment for the EG was carried out through the people responsible for/moderators of the SGs. They were found through key informants from the health services or the university and through association finders (LafuenteRobles, 2015). We contacted the person in charge of the SG via e-mail and telephone, and we presented the study to them. The participants in the CG were recruited through family nurses who work in the health centers from the geographical areas in which the participants of the EC were recruited. Data collection was carried out in primary care centers or in the homes of the participants (CG) and in the places where the participants of the EG meet. The questionnaires were hetero administered to avoid information loss (Haneuse, 2016; Pannucci & Wilkins, 2010). | 7 BERNABÉUÁLVAREZ Et AL. Statistical analysis For the descriptive analysis of the data, frequencies and percentages were used for qualitative variables, while means and standard deviations were used for quantitative variables. In the bivariate analysis, to check the differences between the CG and EG in the pretest and posttest, as well as the difference between those who remained and those who dropped out, Pearson's chisquared test was used for qualitative variables, and the F test (or its nonparametric alternatives, the Mann– Whitney's Utest, and the Kruskal– Wallis test) was used for quantitative variables. In order to check the difference between the pretest and the posttest in the outcome variables (healthrelated quality of life and psychotropic drug use dimensions), we used the McNemar test; and the Wilcoxon test was used for the overall quality of life index. Finally, in order to determine those variables associated with the caregiver's healthrelated quality of life (overall quality of life index) and the use of psychotropic drugs, multivariate analyses were carried out using linear regression models and binary logistic models, respectively. At the beginning, the models included all the possible confounders, as well as their interactions with the intervention. Later, for the selection of the final model, the analysis was done based on the theory and statistic criteria. Thus, the nonsignificant interactions (p>0.05) were removed from the model, and the confounders that achieved at least one of the following criteria were maintained: p < 0.10, or that they produced a sufficiently large change (of 20% or greater) in the magnitude of the β coefficient of the other independent variables that remained in the model (Hosmer & Lemeshow, 2000; Katz, 2006; Vetter & Mascha, 2017). The SPSS PASW Statistics 18.0 and R and Rcommander 3.5.2software were used for the analyses. The null hypothesis with risk error of 5% was rejected (Field & Babbie, 2011). RESULTS Comparison of baseline data The initial sample consisted of 134 caregivers with an average age of 58.43years (SD 11.95), the majority of them being women (CG [91.4%] vs. EG [70.3%]; p=0.002). The caregivers were mainly spouses or sons/daughters (with a significant difference between EG and CG; p=0.001) who had been caring for their relative for around 10years, dedicating an average time of 19.35h a day to caregiving (SD 6.63). Almost a third of them had a low level of education (lower than secondary education), without significant differences between the EG and the CG. 64.1% of participants in the EG lived in an urban area versus 32.9% in the CG (p=0.001). Most of them had good social support (CG=87.1% vs. EG=65.6%, p=0.003). Most of the caregivers’ relatives had high limitations in activities of daily living (32.1%), with a statistical difference between CG and EG in the average of the Barthel test (33.86 [SD30.99] vs. 65.62 [SD 31.58]; p<0.001, respectively). The average impact of stressful life events was of 153.3 points (SD 115.1). According to the measurement of stressful life events, the risk of getting ill was higher for the caregivers in the EG (18.8%) than those in the CG (4.3%, p=0.022). Attrition When comparing those who remained and those who dropped out, both in the bivariate analysis and through binary logistic regression, only the level of limitations in the activities of daily living of the caregiver's relative and the area of residence were correlated, with the latter being marginally correlated. The fewer the limitations in activities of daily living of the caregiver's 8 | FAMILY PROCESS relative, the higher the probability of the caregiver remaining in the study (OR 1.02, p=0.011). The caregivers who lived in urban areas dropped out of the study more frequently than those who lived in rural areas (OR 0.45, p=0.076; Table S1). Psychotropic drug use Around 38% of the sample consumed psychotropic drugs at the beginning of the study (34.4% in the EG and 41.4% in the CG). After 6months, the EG reduced the consumption to 30.8%, while the consumption increased up to 46.3% in the CG, even though the differences were not statistically significant (Table S2). Regarding the daily dose, in the pretest, the mean was 0.49 (SD 0.87) in the EG and 0.65 (SD 0.97) in the CG; in the posttest, it was 0.59 (SD 1.29) in the EG and 0.67 (SD 0.93) in the CG. In the bivariate analysis, the use of psychotropic drugs correlated with gender (7.1% in men vs. 92.9% in women, p=0.008), average age (56.11 [SD 11.91] in nonconsumers vs. 62.05 [SD 10.42] in consumers, p=0.010), and the average number of years spent caregiving (8.69 [SD 8.59] in nonconsumers vs. 13.90 [SD 13.67] years of caregiving in consumers, p=0.036). Thus, the consumption of psychotropic drugs was higher in women, the older the age and the greater the number of years they had spent providing care. In the multivariate analysis for the outcome psychotropic drug use, a significant statistical interaction was found (p=0.003) between attending SGs and the level of limitations in activities of daily living of the caregiver's relative (Barthel test). This interaction revealed that belonging to the EG had a statistically significant positive effect on those caregivers whose relative showed lower limitations in activities of daily living. Furthermore, the use of psychotropic drugs was higher in women than in men (OR 8.95, p=0.049). Moreover, the caregiver's age (OR 1.07, p=0.004) and the number of years spent caregiving (OR 1.11, p=0.008) were positively associated with the use of psychotropic drugs and, marginally, with the level of limitations in the activities of daily living of the caregiver's relative (OR 1.02, p=0.094)— that is, a higher score in the Barthel test (Table 2). In Figure 2, the use of psychotropic drugs is compared in the posttest according to the condition (EG vs. CG) and the level of limitations in activities of daily living of the caregiver's relative. It was found that, below percentile 50— that is, with higher limitations in activities of daily living— the use of psychotropic drugs is higher in the EG versus in the CG (50% vs. 42%, X2=0.312 [p=0.577]). However, above percentile 50— lower limitations in activities of daily living— psychotropic drug use is lower in the EG versus in the CG (22.2% vs. 61.1%; X2=7.958 [p=0.005]), this association being statistically significant. Quality of life Regarding the EQ5D3L’s dimensions, the percentage of participants with anxiety problems was reduced in the EG (71.9% with problems in the pretest and 50% with problems in the posttest, p=0.031), whereas in the CG, it remained stable (71.4% with problems in the pretest and 74% in the posttest). Moreover, the difference in the posttest was also statistically significant and favorable to the EG (p=0.001). The posttest also highlights that 61.1% of the caregivers in the CG had extreme problems in the activity dimension, compared to 5.8% in the EG, the difference being significant (p=0.001). However, it is necessary to note that there were already differences in this dimension in the pretest and, using the McNemar test with gathered data (problems vs. no problems in activity), no statistically significant difference was found between the pretest and the posttest in both groups. | 9 BERNABÉUÁLVAREZ Et AL. Regarding the EQ5D3L overall quality of life index, significant differences were found in the posttest, with a score of 71.35 (SD 15.6) in the EG versus 58.70 (SD 20.9) in the CG (p=0.001). Moreover, the index improved over time in the EG (p=0.022), while it worsened in the CG (not significant). In the bivariate analysis, the EQ5D3L overall quality of life index also correlated with gender (75.24 [13.1] for men vs. 62.35 [19.96] for women, p=0.004), use of psychotropic drugs (69.1 [17.84] for nonconsumers vs. 57.69 [20.19] for consumers, p=0.005), and the level of limitations in activities of daily living of the caregiver's relative (Rho Spearman=0.210, p=0.031). Thus, the quality of life index was lower in caregivers who did not attend SGs, in women, in users of psychotropic drugs and in caregivers of adults with significant limitations in activities of daily living. In the multivariate analysis (Table 3), a statistically significant effect of the caregivers’ attendance at SGs on the EQ5D3L overall quality of life index was found, which was 8.66 points higher in the EG (p=0.015). Furthermore, regarding the relationship with the relative, parents obtained 11.49 points less than spouses on the EQ5D3L overall quality of life index (p=0.032). Similarly, sons and daughters had a lower score than spouses, but this association was not significant (p=0.061). DISCUSSION The aim of this study was to find out the effect of attending caregivers’ SGs on psychotropic drug use and on healthrelated quality of life, as well as to understand other sociodemographic variables associated with these outcomes. Attendance at SGs seems to significantly influence the reduction of the use of psychotropic drugs among the caregivers when their relative has fewer limitations in activities of daily living. This interaction could be explained by the notion that the fewer the limitations in activities of daily living the relative has, the more the caregiver can attend the groups and benefit from them (Biegel et al., 2004; Ussher et al., 2008). Previous authors point to online support groups as a strategy to allow caregivers to attend the groups, as well as to provide the relative professional care while the family caregivers attend the SGs (Wynter et al., 2015). It was found that psychotropic drug use was more common among women (almost 9 times more than men), and another associated factor was the number of years spent providing care. The European Monitoring Centre for Drugs and Drug Addiction indicates that Europe is the TABLE 2 Multivariate regression model for dependent variable: psychotropic drugs Estimate Std. error pOR 95% Confidence interval Lower Higher Psychotropic drugs (pretest) [T. Yes] 3.16 0.69 0.000 23.52 6.85 105.77 Attendance to SG [T. Yes] 1.91 1.23 0.119 6.79 0.60 82.31 Ill's dependence (Barthel test) 0.02 0.01 0.094 1.02 0.99 1.05 Gender [T. Female] 2.19 1.11 0.048 8.95 1.23 101.55 Age 0.07 0.03 0.040 1.07 1.01 1.14 Years caring 0.10 0.04 0.008 1.11 1.03 1.20 SG [T. Yes] by Ill's dependence −0.07 0.02 0.003 0.93 0.89 0.97 Note: R2 = 0.492. Adjusted R2 = 0.455. F = 13.55. p = 0.000. 16 | FAMILY PROCESS Tseng, M. H., Chen, K. L., Shieh, J. Y., Lu, L., Huang, C. Y., & Simeonsson, R. J. (2016). Child characteristics, caregiver characteristics, and environmental factors affecting the quality of life of caregivers of children with cerebral palsy. Disability and Rehabilitation, 38, 1– 9. https://doi.org/10.3109/09638 288.2015.1129451 Ullgren, H., Tsitsi, T., Papastavrou, E., & Charalambous, A. (2018). How family caregivers of cancer patients manage symptoms at home: A systematic review. International Journal of Nursing Studies, 85, 68– 79. https://doi. org/10.1016/J.IJNUR STU.2018.05.004 Ussher, J. M., Kirsten, L., Butow, P., & Sandoval, M. (2008). A qualitative analysis of reasons for leaving, or not attending, a cancer support group. Social Work in Health Care, 47(1), 14– 29. https://doi.org/10.1080/00981 38080 1970673 Vetter, T. R., & Mascha, E. J. (2017). Bias, confounding, and interaction: Lions and tigers, and bears, Oh My!. Anesthesia and Analgesia, 125(3), 1042– 1048. https://doi.org/10.1213/ANE.00000 00000 002332 Walters, S. J. (2004). Sample size and power estimation for studies with health related quality of life outcomes: A comparison of four methods using the SF36. Health and Quality of Life Outcomes, 2, 26. https://doi. org/10.1186/14777525226 Wang, L. Q., Chien, W. T., & Lee, I. Y. (2012). An experimental study on the effectiveness of a mutual support group for family caregivers of a relative with dementia in mainland China. Contemporary Nurse, 40(2), 210– 224. https://doi.org/10.5172/conu.2012.40.2.210 World Health Organization - WHO- (November 26, 2020). https://www.who.int/es/newsroom/factsheet s/detai l/ menta ldisor ders Wright, K., Rains, S., & Banas, J. (2010). Weaktie support network preference and perceived life stress among participants in healthrelated, computermediated support groups. Journal of ComputerMediated Communication, 15(4), 606– 624. https://doi.org/10.1111/j.10836101.2009.01505.x Wynter, K., Hammarberg, K., Sartore, G. M., Cann, W., & Fisher, J. (2015). Brief online surveys to monitor and evaluate facilitated peer support groups for caregivers of children with special needs. Evaluation and Program Planning, 49, 70– 75. https://doi.org/10.1016/j.evalp rogpl an.2014.12.019 SUPPORTING INFORMATION Additional supporting information may be found online in the Supporting Information section. How to cite this article: BernabéuÁlvarez, C., LimaRodríguez J.- S., & LimaSerrano M. (2021). Effect of support groups on caregiver’s quality of life. Family Process, 00: 1– 16. https://doi.org/10.1111/famp.12684