Predictors of Caregiver Burden in Caregivers of Bedridden Patients

ABSTRACT Background Caregivers are at risk of experiencing caregiver burden. It is therefore important to determine the caregiver burden of caregivers who provide care to bedridden patients and related factors. Purpose The aim of this study was to determine the caregiver burden of caregivers who provide care to bedridden patients and the factors that impact this burden. Methods This cross-sectional study was executed at a state hospital in Istanbul, Turkey, on bedridden patients registered in the home healthcare unit and their caregivers. During study period, the researchers made 312 visits to patients and their caregivers. A sociodemographic questionnaire, the Burden Interview, and the Katz Index of Independence in Activities of Daily Living were used to collect data. Descriptive statistics, an independent sample t test, one-way analysis of variance, and stepwise multiple regression analysis were used for data analysis. Results The participants reported a moderate level of caregiver burden. Existing caregiver health problems, caregiver employment status, the ability of the caregiver to maintain his or her own good health, type of home, and the degree of patient dependence in terms of activities of daily living were each found to be significant predictors of caregiver burden. Conclusions/Implications of Practice The support provided to caregivers by home healthcare units is important in terms of protecting the physical, mental, and social health conditions of caregivers and preventing the exacerbation of caregiver burden.

. Furthermore, caregiver burden affects the level of well-being of the caregiver and, as a consequence, reduces the caregiver's ability to provide good care (Collins & Swartz, 2011;Sanuade & Boatemaa, 2015). Chang et al. (2010) found that the mental health and caregiver burden of care providers were related to the health problems experienced by caregivers. Women make up the large majority of caregivers in family settings (Jeong et al., 2015;Rha et al., 2015;Yıkılkan, Aypak, & Görpelioğlu, 2014). Whereas some research has identified gender as a significant predictor of caregiver burden (Brodaty et al., 2014;Etters et al., 2008;Sousa et al., 2016), some have found no correlation between gender and caregiver burden (Orak & Sezgin, 2015;Roopchand-Martin & Creary-Yan, 2014;Sanuade & Boatemaa, 2015). Chiou et al. (2009) reported that caregivers with poor social support and family functions have a higher level of caregiver burden and that perceived social support is a better indicator of caregiver burden than the social support actually received. Furthermore, studies have shown that living in the same house with a patient (Adelman et al., 2014;Conde-Sala et al., 2010) and having a direct role in that patient's physical care (Sanuade & Boatemaa, 2015) increase caregiver burden.
Professional healthcare providers have an important impact on the health and well-being of caregivers (Yıkılkan et al., 2014). Nurses may engage in training primary caregivers and support caregivers by aiding in care-related activities. Thus, nurses have an important role to play in lessening the caregiver burden of care providers (Schulz & Sherwood, 2008). Nurses are best positioned to make early diagnoses of caregiver burden and to help caregivers avoid/minimize the adverse effects of caregiving (Etters et al., 2008). Nurses should observe caregivers during their home visits and evaluate them in terms of caregiver burden risk (Etters et al., 2008;Yıkılkan et al., 2014). When caregiver burden is identified, the perceptions of caregivers regarding the burden that they are taking on may be eased with appropriate interventions (Etters et al., 2008). Thus, for all of the above-stated reasons, it is important to identify caregiver burden and its related factors, to ensure that caregivers receive support, and to develop, organize, and implement caregiver-burden prevention programs. Therefore, the purpose of this study was to determine the caregiver burden of individuals who provide care to bedridden patients and the factors that impact this burden. This study is significant in that it was conducted in Turkey with a large sample and because many factors affecting caregiver burden were examined as variables. The study sought answers to the following research questions: • What is the relationship between the sociodemographic characteristics of caregivers and their caregiver burden? • What is the relationship between the way caregivers provide care and their caregiver burden? • What is the relationship between the degree of dependency of bedridden patients and the caregiver burden of their caregivers? • Are caregiving-related factors predictive of caregiver burden?

Design and Data Collection
This study used a cross-sectional design and was executed at a state hospital in Istanbul, Turkey, between January and April 2014. The participants were bedridden patients who were registered in the hospital's home healthcare unit and their caregivers.

Participants
Home healthcare services in Turkey are given by the Ministry of Health, municipalities and private organizations. Home healthcare units of Ministry of Health work in affiliation with the hospitals. Individuals who require services from these units at home are provided examination, testing, treatment, medical care, rehabilitation, and social and psychological support services during home visits. The costs of these services are covered by the social security administration. This study targeted the caregivers of bedridden patients registered in the home healthcare unit of a state hospital in Anatolian Istanbul (N = 4,500). The formula that was used to determine the necessary sample size on the basis of a finite population determined that a minimum sample size of 251 bedridden patients and caregivers was required (Sümbüloğlu & Sümbüloğlu, 2002). The study was conducted on 312 bedridden patients and their caregivers who met the criteria for inclusion. The criteria for inclusion included being responsible for the care of a bedridden patient, providing care for at least 1 month, willing to participate in this research, aged 18 years or older, able to read and write, and able to understand and answer questions. The questionnaires were filled out during face-to-face interviews that were conducted with the participants during a home visit by the researcher. Patient information was obtained from patients and caregivers. Questionnaires with missing data were excluded from the study.

Ethical Considerations
Verbal and written permission was obtained from the institution in advance. The university's ethics committee granted its approval for this study (September 9, 2013, No. 47). The purpose of this study was explained to the patients and their caregivers, and their written and verbal consent was obtained.

Demographic variables
The descriptive characteristics of the caregivers (gender, age, educational status, civil status, type of residence, employment status, and income status), their personal health situation, their ability to attend to their own health, their relationship to the patient, the duration that they had been caring for their patient, the areas of caregiving, and the patient's gender, age, and educational status were collected using a standard questionnaire.

Caregiver burden
The caregiver burden of the caregivers was measured using the Zarit Burden Interview (BI). In 1980, Zarit, Reever, and Bach-Peterson developed BI as a scale to evaluate the level of stress experienced in providing care to the sick and older adults. The scale questionnaire, which may be filled out either by the caregiver or a researcher, consists of 22 statements on the effect of caregiving on the respondent caregiver's life. Each of the statements is answered using a Likert-type scale, with scores ranging from 0 to 4 (never, rarely, sometimes, frequently, and always). Studies have indicated an internal consistency coefficient of .87-.94 for the scale and a test-retest reliability of .71. The total possible scores for the BI range from 0 to 88, with 0-20 indicating "no burden," 21-40 indicating "mild burden," 41-60 indicating "moderate burden," and 61-88 indicating "severe burden." The scale items generally address the social and emotional domains, with a higher total score associated with a greater burden experienced (Zarit et al., 1980). İnci and Erdem (2008), who carried out the validity and reliability studies for the Turkish version of this scale, found a Cronbach's alpha value of .95. The Cronbach's alpha value for the BI found in this study was .90.

Functional status
The dependency status of participants was assessed using the Katz ADL was used as Katz Index of Independence in Activities of Daily Living. In 1963, Katz, Ford, Moskowitz, Jackson, and Jaffe developed Katz ADL as a tool to assess basic ADL, and Yardımcı (1995) completed the Turkish translation of the index. The index contains six headings, including bathing, dressing, toileting, transferring, continence, and feeding. Each heading has three possible responses: "dependent," "partially dependent," and "independent," which are assigned scores of 1, 2, and 3, respectively. ADL index score totals are assessed as follows: 0-6 = dependent, 7-12 = partial dependence, and 13-18 = independent. The ADL Cronbach's alpha value for this scale found in this study was .92.

Data Analysis
The study data were analyzed using SPSS 16.0 (SPSS, Inc., Chicago, IL, USA). Descriptive statistics (numbers, percentages, means, and standard deviations) were used in the analysis. The one-sample Kolmogorov-Smirnov test was used to assess normal distribution, and the p value was found to be > .05. As the data displayed normal distribution, parametric tests were employed in the advanced analysis. The parametric tests used included the independent sample t test for two independent variables and the one-way analysis of variance for more than two independent variables. Stepwise multiple regression analysis was used to determine the factors affecting caregiver burden. The Durbin-Watson statistic was used to measure autocorrelation. The Durbin-Watson statistic was found to be 2.086. Tolerance was found to be between .85 and .99, and the variance inflation factor was 1.00-1.17. Tolerance and variance inflation factor values were both found to be within acceptable limits. The results were found to be within the 95% confidence interval, and significance was assessed as p < .05.

Sample Characteristics
Two thirds (65.1%) of the caregiver participants were women, 57.4% were 36-55 years old, and 60.6% had received over 8 years of training. In addition, 72.8% were married, 30.4% were employed, 64.1% earned an income that was roughly equal to their expenditures, and a large majority (80.1%) lived in apartment residences. Slightly over two fifths (42.6%) self-reported as having some kind of health problem, and 57.1% stated that they did not tend to their health needs. It was found that 71.8% were taking care of their parent or sibling, 56.7% had been a caregiver for 2 years or more, and 85.9% said they tended to every need of their patient. In terms of care recipients, 56.7% were women, the large majority (82.4%) were older than 65 years, and 79.5% had an education of 8 years or less (Table 1). The ADL item mean scores of the care recipients varied between 1.16 AE 0.47 and 1.52 AE 0.70, and their ADL total mean score was 7.64 AE 2.94 (Table 2).

Factors Associated With the Burden of Caregiving
The BI total mean score in this study was 43.56 AE 15.05. The BI mean score of the female caregivers was significantly higher than that of the male caregivers (p < .01). The BI mean score of the caregivers who were 56 years old or more was significantly higher than those of caregivers in the other age Caregiver Burden in Caregivers VOL. 27, NO. 3, JUNE 2019  groups (p < .001). The BI mean score of the unemployed caregivers was significantly higher than that of the employed caregivers (p < .001). The BI mean score of caregivers living in squatters' homes and apartments was significantly higher than that of caregivers living in private homes (p < .001). The BI mean score of caregivers who had health problems was significantly higher than that of caregivers with no health problems (p < .001). Those caregivers who did not tend to their health needs had a BI mean score significantly higher than that of caregivers who did take care of their health needs (p < .001). Caregivers who had been tending to their patients for 2 years or more had a BI mean score significantly higher than that of caregivers who had been tending to their patients for 1-2 years (p < .05). The BI mean score of caregivers who tended to all of the patients' needs was significantly higher than that of caregivers who provided only financial or psychological support (p < .05). The BI mean scores did not display any statistically significant differences based on the level of education, marital status, or family income level of the caregivers or on their relationship to the care recipient (p > .05). The BI mean score of the caregivers of patients with a level of education over 8 years was significantly higher than that of caregivers of patients with a level of education of 8 years or less   (p < .05). No statistically significant differences were found between the BI mean scores of caregivers categorized, respectively, by care recipient gender and age (p > .05; Table 1). In comparing the BI mean scores of patients according to their ADL, the BI mean scores for caregivers of patients who were partially dependent on the caregiver for dressing, continence, and feeding issues were significantly higher than the mean scores for those caregivers who were caring for completely dependent patients (p < .01). No statistically significant differences were found between the BI mean scores of caregivers based on the bathing, toileting, and transfer dependency needs of patients (p > .05; Table 3).

Determinants of the Burden of Caregiving
A stepwise multiple regression analysis was performed to determine the predictors of the burden of caregiving. In the univariate analysis, the variables that had a significant effect on the burden of caregiving were considered as independent variables. Thus, gender, age, employment status, the health problems of the caregiver, the ability of the caregiver to tend to personal health, type of home, duration of care, the areas of care, and the patient ADL score were adopted as the independent variables. It was determined that gender, age, duration of care, and the areas of care of caregivers were not significant predictors of caregiver burden. However, significant relationships were found between burden of caregiving and the caregivers' health problems, employment status, ability to tend to personal health, type of home, and ADL (R 2 = .25, p < .001), with 25% of the total variance explained by these variables. The order of significance for the variables ranged from the caregiver's health problems (β = .257, p < .001) to employment status (β = −.225, p < .001), type of home (β = −.182, p < .001), ability to tend to personal health (β = −.170. p < .001), and the dependency of the patient in ADL (β = .147, p < .001; Table 4).
The burden of the caregivers in this study was found to be higher in those who had health problems, who were unable to manage their own health, who were not employed, who lived in a squatter home or in an apartment, and who were less dependent on their patients.

Discussion
The aim of this study was to determine the caregiver burden of individuals who were caring for bedridden patients and the factors that impact this burden. This study was conducted on a sample of bedridden patients who were receiving services from a home healthcare unit and their caregivers. It was determined that the caregivers experienced a moderate level of caregiver burden. Using univariate analysis, a significant relationship was found between caregiver burden and the caregiver's gender, age, employment status, type of home, health problems, ability to tend to personal health, and duration of caregiving and the areas where caregiving was needed as well as the patient's dependency in terms of ADL. Multiple regression analysis showed that caregiver health problems, their employment status, their ability to tend to their own health, their type of home, and the degree of dependence of patients in terms of ADL were each significant predictors of caregiving burden.
The caregiver burden scores for caregivers with health problems were higher than those of caregivers who had no health problems, indicating that caregiver burden increases when caregivers have personal health problems. One study pointed to correlations between health problems and the mental health of caregivers and caregiver burden (Chang et al., 2010). Half of the caregivers in one prior study had at least one chronic health problem (Collins & Swartz, 2011). Caregiving creates a physical and mental burden that adversely affects the health of caregivers (Bauer & Sousa-Poza, 2015;Etters et al., 2008;Kim et al., 2012), and poor personal health disrupts personal quality of life (Jeong et al., 2015). Furthermore, poor caregiver health lowers the quality of care given, increases caregivers' formal demands for healthcare (Bauer & Sousa-Poza, 2015;Collins & Swartz, 2011), and increases healthcare costs (Bauer & Sousa-Poza, 2015). It is important that caregivers are supported by home healthcare units so that they do not feel alone and helpless, do not develop health issues, and do not experience a deterioration in their existing health problems.
The caregiver burden of caregivers who do not tend to their own health is higher than that of caregivers who Note. R 2 = .254, adjusted R 2 = .242, F = 20.84.

The Journal of Nursing Research
Aysun BEKDEMİR et al.
do. Paying appropriate attention to personal health reduces caregiver burden. The health of caregivers is known to impact caregiver burden strongly (Rha et al., 2015). Chiou et al. (2009) observed that caregivers with a low level of social and functional family support experienced higher levels of caregiver burden. The fact that caregivers allocate time to deal with their own health problems indicates that they receive social support from friends and family when needed. The literature supports that social support reduces the caregiver burden and increases the quality of life of caregivers who provide care for patients with chronic diseases (Atagün, Balaban, Atagün, Elagöz, & Özpolat, 2011). This study found that the caregiving burden of unemployed caregivers was higher than that of their employed counterparts, indicating that outside employment reduces caregiver burden. A prior study reported that the caregiver burden of employed caregivers is of a lower level compared with that of unemployed caregivers (Sanuade & Boatemaa, 2015). Another study found that self-employed caregivers had lower caregiver burden scores (Roopchand-Martin & Creary-Yan, 2014). As caregivers who hold jobs outside the home cannot serve as primary, full-time caregivers, their caregiving time is shorter than that of caregivers who are not employed, which may explain why employed caregivers have a lower caregiver burden. Chiou et al. (2009) found that caregiver burden increases as the duration of caregiving. In Yeşil, Uslusoy, and Korkmaz (2016), no difference was found in caregiver burden based on employment status.
The caregiver burden of caregivers who live in squatter houses or apartments was shown to be higher than those who live in private houses, indicating that living in a private house reduces caregiver burden. This result indicates that physical circumstances may affect caregiver burden. Istanbul is one of Turkey's most densely populated and most expensive cities. Accordingly, it is likely that people who live in private homes in Istanbul are of a relatively high socioeconomic status. As caregivers at higher economic levels care for their patients in more comfortable physical conditions and are more likely to employ outside help, their caregiver burden may subsequently be less.
Caregivers of patients with higher levels of education were found to have higher burdens of care than those of patients with low levels of education. As patients with higher levels of education have higher life expectancies, they likely have greater expectations from their caregivers. This may increase the burden of care of their caregivers.
This study found that patients were most commonly dependent on their caregivers for bathing and least dependent for feeding, with results showing that the degree of patient dependency was a significant predictor of caregiving burden. As patient dependency lessened, caregiver burden increased. Concurrently, the caregiver burden of individuals caring for patients who were partially dependent because of continence or feeding issues was higher than the burden of those caring for dependent patients. This outcome suggests that the caregivers of partially dependent patients may not have been able to accept the additional dependency-related burdens. As patient dependency increases, caregivers tend to feel that the patients actually do need them and therefore accept the situation, leading to lower levels of perceived caregiver burden. In a study by Taşdelen and Ateş (2012), as patient dependency grew in terms of ADL, the caregiver's emotional burden lessened, which is consistent with the results of this study. However, contrary to the results of this study, other studies have shown that caregiving burden increases as patient dependency rises (Abdollahpour et al., 2012;Kim et al., 2012;Zaybak et al., 2012). It is important that further qualitative studies be conducted to discover the reasons for these outcomes, so that appropriate interventions may be designed to lessen the burden of caregiving.
Women comprised a large majority of caregivers in this study. Other studies have also found that most caregivers are women (Rha et al., 2015;Unver, Basak, Tosun, Aslan, & Akbayrak, 2016;Yeşil et al., 2016). Whereas the univariate analysis revealed a higher level of caregiver burden in female caregivers compared with male caregivers, the multiple regression analysis did not reveal a significant relationship between gender and caregiver burden. Similar to the results of this study, other studies have not detected any significant relationship between gender and caregiver burden (Orak & Sezgin, 2015;Roopchand-Martin & Creary-Yan, 2014;Sanuade & Boatemaa, 2015). However, other studies still have pointed to female caregivers having greater levels of caregiver burden than their male counterparts (Sanuade & Boatemaa, 2015;Sousa et al., 2016;Unver et al., 2016). These results reveal that the burden of caregiving in women may be affected not only by gender but also by normal responsibilities such as housework and childcare, by personal characteristics, by employment status, and by other relevant factors.
Although the burden of caregiving was found to be higher in caregivers aged 56 years and above, age was found not to be a significant determinant of this burden. Similarly, other studies have shown age not to significantly impact the burden of caregiving (Abdollahpour et al., 2012;Roopchand-Martin & Creary-Yan, 2014). Despite this, it is still believed that the older a caregiver is, the higher the caregiving burden may be because of age-related health problems and physical limitations.
It was found that caregivers with 2 or more years of caregiving experience had a higher level of caregiver burden than those with durations of care of 1-6 months and 1-2 years. Similar to the results of this study, Çetinkaya and Karadakovan (2012) found that longer caregiving durations were positively associated with higher caregiver burden. In a study by Yıkılkan et al. (2014), caregivers who cared for their patients for more than 3 years had higher levels of depression and anxiety than caregivers with shorter caregiving durations. Regression analysis showed that duration of caregiving was not a significant predictor of the burden of caregiving. This finding is an important outcome, as it shows that caregiver burden is associated with more than only the duration of the caregiving. It may be that long-term caregiving results in higher caregiver burden because of the increases in frequency and severity of physical, mental, and social problems.
In this study, the caregiver burden of caregivers who were required to meet all of the needs of their patients was higher than that of caregivers who met only the financial needs or provided psychological support to their care recipients. Multiple regression analysis found no significant relationship between areas of care and caregiver burden. Mollaoğlu, Özkan Tuncay, and Kars Fertelli (2011) found that those caregivers who met all of the needs of their patients had a higher level of caregiver burden. Moreover, Sanuade and Boatemaa (2015) found that caregivers who provided only financial support and caregivers who received outside financial and physical support had lower levels of caregiver burden. Caregivers who provide patients with only financial or psychological support are not primary caregivers. In these types of cases, the main caregiver is usually another member of the family, and the caregiver's burden is less compared with those who must meet all of their patients' needs.
The cross-sectional approach used in this study limits its generalizability to similar populations only. In addition, the self-report nature of data collection potentially limits the accuracy and generalization of results. The relatively large sample size is a strength of this study. Future studies should consider more complex variables dealing with caregiving as predictive variables (e.g., caregiver-perceived social support, coping strategies, daily care hours, having help available at home, number of caregivers). Furthermore, in line with the results of this study, it is recommended that experimental studies be carried out to evaluate the effectiveness of nursing interventions that are carried out to reduce the caregiver burden of individuals who provide care to bedridden patients.

Conclusions
The caregivers in this group reported a moderate overall level of caregiver burden. Furthermore, the health status, employment status, ability to tend to personal health matters, and type of home of the caregiver and the degree of dependence of the patient in terms of ADL were all found to be significant predictors of caregiver burden.