Caregiving burden and depression in paid caregivers of hospitalized patients: a pilot study in China

This was a cross-sectional survey study in a tertiary referral hospital (Chengdu, China) performed with a self-developed questionnaire that is described in the “Instruments” section. Nurses participated in the design, distribution and return of the questionnaire.

Considering that caregiving burden could be affected by the working environment and patients of paid caregivers, the subjects were all paid caregivers who worked in four medical and four surgical wards (i.e., cardiology vs. cardiac surgery, respiratory medicine vs. thoracic surgery, gastrointestinal medicine vs. gastrointestinal surgery, and neurology vs. neurosurgery). The subjects were enrolled if they met the following criteria: (1) agreed to participate and signed the informed consent form and (2) could read and complete the questionnaire. Consequently, 120 of the 124 (96.8%) registered paid caregivers from the aforementioned wards were enrolled; four people did not fulfil the inclusion criteria as they refused to participate.

The questionnaire was written in Chinese and consisted of three parts. In the first part, the baseline characteristics of the study subjects were collected; this included proposed factors associated with the caregiving burden or depression that were categorized into general, work, income, and family information. General information included the age, gender, education, and marital status. Work information included the ward in which they worked, number of years as a paid caregiver, and number of vacation days per month. For the income information, “income satisfaction” was a required field, while “income per month before and after becoming a paid caregiver” was optional. The number of dependents, frequency of meeting his/her own family, and housing status in Chengdu (owned or rental) were categorized as family information.

The second part assessed the caregiving burden by adopting the Chinese version of the Caregiver Burden Inventory (CBI), which has been validated and shown to be sufficiently reliable in a Chinese population. [13, 14] The CBI has 24 items, and each item is answered on a scale ranging from 0 (not at all disruptive) to 4 (very disruptive). The CBI describes the caregiving burden in 5 different domains (factors): time-dependence (items 1 to 5), developmental (items 6 to 10), physical (items 11 to 14), social (items 15 to 19), and emotional (items 20 to 24) burden. The time-dependence burden is associated with the time restrictions of caregivers. The developmental burden describes the perception of being “taken out” of their development or being excluded from the expectations and opportunities enjoyed by their peers. The physical burden is associated with feelings of chronic fatigue and their own physical health problems. The social burden describes the poor balance between many roles. The emotional burden is associated with negative feelings towards patients and is possibly induced by the unpredictable and bizarre behaviours of the patients. [15, 16] Consequently, the score in the physical burden factor should range from 0 to 16 (only 4 items), and each of the scores in the other factors should range from 0 to 20 (5 items for each factor). The total score of the CBI ranges from 0 to 96, and a higher score indicates a higher caregiving burden. To compare between the 5 aforementioned factors (Fig. 1), the physical burden score was multiplied by 1.25 to give an equivalent score out of 20. [15].

The third part of the questionnaire assessed the depression symptoms of paid caregivers with the Chinese version of the Center for Epidemiologic Studies Depression Scale (CES-D), which has been validated and showed good reliability in a Chinese population. [17‐19] The CES-D includes 20 items that cover affective, cognitive, behavioural, and somatic symptoms associated with depression. The subjects were asked to indicate the frequency of symptoms on the following 4-point scale from 0 (rarely or none of the time) to 3 (most or all of the time). The 4 positively formulated items (items 4, 8, 12, and 16) were recorded in reverse. Therefore, the total score of the CES-D ranged from 0 to 60, and a score of 16 or higher indicates the likely presence of clinically significant depression. [18, 20].

According to the study design, the copies of the questionnaire with inadequate quality were not entered into the final analysis, including those with an incomplete CBI or CES-D section, and a consistent lowest score (i.e. “0” score) or highest score (i.e. “4” score in the CBI section and “3” score in the CES-D section) for each item in either section. This extreme and unchangeable scoring situation was considered to indicate that the participant had not understood the survey or had not seriously considered it.

Data for the 4 domains are presented in Table 1. All paid caregivers were middle-aged, and most were female. The majority of the subjects achieved an educational level below high school. The number of years that each participant worked as a paid caregiver was categorized into five groups in the questionnaire (i.e., less than 1 year, 1 ~ 3 years, 3 ~ 5 years, 5 ~ 10 years, and more than 10 years); however, 73 (66.7%) paid caregivers worked in this field for more than 10 years.

The total CBI score was normally distributed among the study population with a mean value of 29.7 ± 12.5. There were some differences among the 5 factors of the CBI that were detected by the ANOVA (F ₍₄₅₃₅₎ = 194.207, P < 0.001) and post hoc analysis. The time-dependence burden factor had the highest score of 15.3 among the 5 factors; it was significantly higher than those of the remaining factors (all P < 0.001). It was followed by the physical burden score of 6.5, which was significantly higher than the emotional, developmental, and social burden scores (all P < 0.001). The scores of the last 3 factors were similar. (Fig. 1).

The CES-D score of the study participants also had a normal distribution with a mean value of 11.9 ± 8.7. Using the CES-D score ≥ 16 cut-off, depression was found in 28 (25.9%) paid caregivers. Moreover, the CBI score of the depression group (CES-D score ≥ 16) was significantly higher than that of the non-depression group (36.3 ± 14.3 vs. 27.4 ± 11.0, P = 0.001). The CBI score was positively correlated with the CES-D score (r = 0.452, P < 0.001). (Fig. 2).

A multiple linear regression analysis was used to identify the factors that were related to higher CBI (Table 2) and CES-D (Table 3) scores. The parameters in Table 1 were all included as covariates in the model, where either the CBI or CES-D score was the dependent variable. Three independent variables were associated with a higher CBI score, which included a longer time as a paid caregiver (P = 0.009), lower income satisfaction (P = 0.001), and more frequent meeting with his/her own family (P = 0.006). (Table 2) Similarly, a lower income satisfaction was found to be associated with a higher CES-D score (P = 0.008). (Table 3).

The finding that paid caregivers suffered from a high caregiving burden was corroborated by the findings of previous reports; [21, 22] however, our study included a more diverse population of paid caregivers who worked in a hospital setting rather than a population of paid caregivers who worked for patients with a specific severe disease condition, such as dementia, cancer, and heart failure. Compared with family caregivers, paid caregivers showed a similar [23, 24] or higher [25, 26] CBI score.

The hired caregivers usually worked around the clock, except for short breaks at lunch and dinner or when the patients’ relatives came to visit. In particular, they always had to do everything by themselves at night, sleeping next to their patients in portable camping beds and observing the needs and movements of the patients; such sleep interruptions were frequent, inevitable and unpredictable. Although they could take a few vacation days per month, usually after fulfilling an existing contract and before starting a new one, they preferred not to do so because they were paid per working day. Therefore, this prolonged period of caregiving explains why the time-dependence burden was the most prominent among the 5 factors of the CBI; this finding was also observed in family caregivers. [16, 27, 28] Therefore, interventions to tackle this time-dependence burden could be suggested, e.g., caregiver companies could enforce a change from the current 24-h, 7-day working schedule to a better arrangement with day and night shifts. In addition, chronic sleep disturbances, tedious caregiving work and contact with infectious patients result in concerns about their own physical health, [12] which was the second most prominent burden among paid caregivers. This was different from that observed in family caregivers, who were more affected by a developmental burden than physical burden. [27, 29].

Previous studies showed that the risk factors for caregiving burden included female gender, [30] a low education, [31, 32] sleep deprivation, financial stress, [33] depression, [34] social isolation, [35] and time restrictions [36]. Because the paid caregivers in the current study worked on a 24-h schedule, we gathered the number of years they spent as paid caregivers rather than the number of hours per day spent performing their job and found the former to be associated with the CBI score. This association might be attributed to the self-perceived long-term time restrictions and deteriorating physical health of the paid caregivers. As previously reported, financial stress was a risk factor for caregiving burden [33], and poor payment for paid caregivers was a common issue. [37, 38] In China, most tertiary referral hospitals (as the study site) provide high-quality, specialized medical health care services to a number of districts that are located in urban areas; however, these areas also have the largest demand for paid caregivers. [39] In contrast, most paid caregivers are from the countryside and have moved to a city for a better living. Not surprisingly, lower income satisfaction was significantly associated with a higher caregiving burden, indicating a conflict among their actual payment, self-expectations and financial circumstances. However, this conflict cannot be resolved only by raising their salaries. It is possible that organized lectures and regular meetings could identify what paid caregivers need and think. More interestingly, it was observed in the study that paid caregivers who met with their own families more frequently had a higher CBI score. This phenomenon has not been previously identified, and its cause remains unexplored in the current study.

Using the CES-D score ≥ 16 cut-off, we found that the prevalence of depression among paid caregivers reached 25.9%; this was higher than that of the general population in China, which was reported to be between 5.9% and 12.4%. [40, 41] However, other studies showed that more than 50% of family caregivers were at risk of depression according to the same CES-D score cut-off. [42, 43] This difference in the prevalence of depression might be explained by the “nature” of the caregiving work performed by paid caregivers compared to family caregivers. Paid caregivers usually consider the caregiving work as a job from which they are paid for their countable time and effort, while family caregivers treat it as a serious responsibility in which their emotions, time and money are involved. It is also possible that family caregivers are less likely to be gainfully employed. Consistent with previous studies on family caregivers [44, 45], our study showed that the caregiving burden of paid caregivers was positively correlated with depression symptoms.

The major limitations of this study include its cross-sectional nature and the absence of health consequence data, i.e. the result of coupling between the caregiving burden and the physical/emotional health of paid caregivers. Although some practical interventions to reduce the caregiving burden could be developed based on the results of the current study, their impact on the professional health of paid caregivers remains unclear. The relatively small sample size is another limitation.