Health-Related Quality of Life and Related Factors among Primary Caregivers of Children with Disabilities in Shanghai, China: A Cross-Sectional Study

Health-related quality of life (HRQOL) of caregivers of children with disabilities (CWD) is important for both children’s rehabilitation and caregivers’ life, but the corresponding attention is far from enough in mainland China. Thus, we investigated the HRQOL of 170 caregivers and related factors in Shanghai. The 12-item Short Form Health Survey (SF-12) was used to measure HRQOL. The potential factors were collected, including child characteristics, caregiver characteristics, and environmental factors. Univariate analysis and multiple linear regression were performed to identify the key factors that could be intervened. Compared with the general population, caregivers of CWD had a slightly higher score on the physical component summary (PCS, 52.57 ± 8.41), but the score of mental component summary (MCS, 31.58 ± 7.72) was extremely low. Caregiver’s illness condition, family size, and household income were significant factors of physical HRQOL. Caregivers with illness and caregivers living in an extended family were associated with higher mental HRQOL. Whereas these two factors had opposite effects on physical HRQOL. This finding indicated poor mental HRQOL among caregivers of CWD in Shanghai and thus requiring urgent attention and intervention. Improving physical fitness, maintaining family integration, and providing financial support should be considered when developing intervention for this population.


Introduction
A caregiver is defined as a person who provides care to those who have difficulties in completing the tasks of daily living and thus need supervision or assistance due to some form of illness or disability [1]. They may provide the care in an institution or an organization, which is called formal caregiving. The concept also refers to an informal caregiver who has a social relationship with the care recipient including family member, relative, friend, or neighbor. With the rapid growth of the global economy, institutionalized long-term care may be an option for children with disabilities (CWD), but the informal caregivers are preferable because of children's high dependency and cognitive deficiency. Therefore, the caregivers of CWD generally refer to informal caregivers especially family caregivers including parents, grandparents, siblings, or other family members. In recent years, there has been a tremendous change in the health care system which exerts a shift toward family-centered services.

Participants and Procedure
We recruited the caregivers of CWD from one district of Shanghai. This district has a well-constructed rehabilitation service system and high-level government attention to CWD. Since the exact number of CWD is not yet available, we determined the sample size based on the number of children who received the Sunshine Baby Cards. The Sunshine Baby Card is the most important project of rehabilitation service system provided for CWD by the Shanghai Health Bureau, Shanghai Municipal Education Commission, Shanghai Finance Bureau, and the Shanghai Disabled Person's Federation. It can be applied for any child who has been certified as disabled by a professional organization. According to the regulation, every child who holds a Sunshine Baby Card can receive rehabilitation training in designated rehabilitation institutions and enjoy the corresponding subsidies. Thus, the number of children with a Sunshine Baby Card approximates the actual number of CWD. In 2019, there were 900 children with the Sunshine Baby Card in the district, 15% of the total number of cardholders were selected for the survey. Then the sample size was expanded by 20% to take into account the dropout rate, which resulted in a sample size of 900 × 15% × (1 + 20%) = 162. Finally, 170 cases were included in the survey.
We collected data in both rehabilitation institutions and community-based settings, considering that children aged 0-6 years are more accessible in rehabilitation institutions, while children aged ≥7 years need to go to school. Since there are private rehabilitation institutions, public hospitals, and community health centers in Shanghai, we selected all types of rehabilitation institutions for investigation. For community-based settings, we covered all communities of this district.
Caregivers were included if (1) they were caring for a child diagnosed with disabilities by the responsible departments; (2) their children were aged 0-16 years; and (3) they were the primary caregivers. Caregivers who did not identify themselves as primary caregivers were excluded from the study. In each rehabilitation institution, the children's basic information and rehabilitation schedule was first obtained from their health records to determine whether they met the criteria or not. Then, we created the investigation list. As for the communities, we first collected the basic information of CWD from the staff of the communities. If children met the sample requirements, their name would be fed back to the staff. The purpose and procedure of the research project were explained to caregivers. If they agreed to participate in the study, the investigators would conduct a one-to-one investigation. Ultimately, we recruited 97 caregivers from the rehabilitation institutions and 73 caregivers from the communities.
Ethics approval was obtained from the Ethics Committees of School of Public Health of Fudan University (Grant No. IRB#2019-10-0782).

Measure of Health-Related Quality of Life
HRQOL was measured using the 12-item Short Form Health Survey (SF-12), an abbreviated form of MOS 36-item Short Form Health Survey (SF-36). SF-12 was designed to take less time to complete and was widely used due to its high degree of acceptability and data quality. It contains 8 domains: Physical Functioning (PF), Role-Physical (RP), Bodily Pain (BP), General Health (GH), Energy/Fatigue (VT), Social Functioning (SF), Role-Emotional (RE), and Mental Health (MH) [28]. Two summary scores are reported: Physical Component Summary (PCS) score and Mental Component Summary (MCS) score which can be used to measure physical HRQOL and mental HRQOL, respectively. For comparison, scores were calculated using the US norm-based scoring algorithm. The test-retest reliability (0.89) and reliability coefficients (0.76) showed the scale was reliable and valid [28]. In 2005, a study conducted in Hong Kong using SF-12 showed that this instrument was valid and equivalent for the Chinese population as well as the items and scoring algorithm. Therefore, SF-12 can be applied to the Chinese population for further health-related research [29].

Measure of Determinants of HRQOL
Child characteristics: Child characteristics were designed under the framework of ICF-CY. For body functions and structures, sleep time and emotional stability were taken into consideration. For activities and participation, we measured the frequency of physical activities and group games during the last four weeks. For personal factors, we collected age, gender, cause of disability, disability type, disability severity, and whether they had the Sunshine Baby Card or disability certificate.
According to the People with Disabilities Act of the Peoples Republic of China, disability is divided into visual disability, hearing disability, speech disability, physical disability, intellectual disability, mental disability, multiple disability, and other disabilities. Since our inclusion criteria did not strictly limit the type of disability, all types above were involved in our objects. Additionally, we specially listed cerebral palsy as one specific type of disability because of its high prevalence among children. Disability severity was graded with China's legal criteria of disability which divided all kinds of disabilities into four levels, ranging from level I (extremely severe) to IV (mild severe).
Caregiver characteristics: Caregiver characteristics were collected under the guidance of ICF. The dimension of body function and structure was measured with the question of "whether you have the following disease or not (e.g., hypertension, diabetes, Parkinson's disease)". The feature of activity and participation was identified by caregivers' employment status and caring time. Furthermore, we collected the personal factors including relationship with the child, gender, household register, marital status, and education. Previous studies mostly focused on the caregiving of parents or merely mothers, rarely involving other families. In this study, we took fathers, mothers, paternal grandfathers, paternal grandmothers, maternal grandfathers, and maternal grandmothers into consideration.
Environmental factors: Environmental factors included number of children, family size, household income, policy satisfaction, and social friendliness. Among these factors, policy satisfaction and social friendliness were reported by subjective perception of caregivers and were evaluated in the form of the Likert scale, ranging from 1 (not friendly at all or dissatisfaction) to 5 (extremely friendly or satisfaction).

Statistical Analysis
Descriptive statistical methods were applied in presenting basic characteristics for child, caregiver, and environment, and the measures (mean, standard deviation [SD], frequencies, and percentages) were respectively reported. Mean scores of the SF-12 of caregivers of CWD were compared with norm scores of the general population using the one-sample t-test. In order to preliminarily examine the associations between participants' characteristics and HRQOL, univariate analyses including t-test and one-way analysis of variance (ANOVA) were conducted. To further investigate the potential predictors, two multiple linear regressions were performed with PCS and MCS scores as dependent variables. Any covariates that achieved p < 0.10 for the bivariate analyses were included in the multivariate linear regression models. We examined the multicollinearity among predictors in each model using Variance Inflation Factor (VIF). We removed the predictors which lead to multicollinearity. All statistical analyses were performed using SPSS Statistics version 20.0 (IBM Corp., Armonk, New York, NY, USA).

Descriptive Statistics
A total of 170 caregivers completed the survey. The mean age of children was 7.16 years (SD: 4.04 years), 60.6% of the children were male, and most of them (62.4%) got disabled because of acquired disease, trauma, or unknown reason. Both children with cerebral palsy and children with physical disability had the highest percent (20.0%). Moreover, the disability severity was mostly unrated (70.6%) ( Table 1). Tables 2 and 3 illustrate the caregiver characteristics and environmental factors. The mean age of the caregivers was 45.10 years (SD: 12.16 years). The relationship with the child was mother (48.2%), father (25.3%), and grandparents (25.9%). A total of 65.9% of the caregivers were female and most of them were currently married (91.2%). In addition, 78.8% of the disabled children were only children and most of the families were small families with no more than four persons.    Table 4 shows the scores on HRQOL. The mean PCS score was 52.57 (SD: 8.41) and the mean MCS score was 31.58 (SD: 7.72). Compared to mothers, fathers had the highest scores on PCS (55.59 ± 5.89), but the lowest scores on MCS (30.67 ± 6.29). For disability types, caregivers of children with hearing disabilities had the lowest scores on PCS (45.77 ± 13.23), and caregivers of children with cerebral palsy had the lowest scores on MCS (29.25 ± 6.43). Moreover, compared with the Chinese (HK) and the US norm of the general population, caregivers of CWD had significantly higher mean scores on PCS and lower mean scores on MCS (p < 0.001) ( Table 5) [28,29].

Univariate Analysis
Results of the univariate analysis using the PCS and MCS of the caregivers as the dependent variables and 24 factors are shown in Table 6. In summary, among 11 factors associated with CWD, the physical HRQOL of caregivers were significantly different in age, disability certificate, disability severity, sleep time, and emotional stability (p < 0.10). However, the mental HRQOL of caregivers was found as the only statistically significant difference in disability type (p < 0.10). For the eight factors associated with caregivers, statistically significant differences in the physical HRQOL were found in role, gender, education, employment status, and accompanying disease (p < 0.10). Statistically significant difference in the mental HRQOL was found in caregivers' accompanying disease (p < 0.10). For the five environmental factors, the physical HRQOL of caregivers were significantly different in family size, household income, and policy satisfaction. Whereas statistically significant differences in mental HRQOL were found in number of children, family size, and policy satisfaction (p < 0.10).

Multiple Linear Regression
The two fitted multiple regression models identified predictors of caregivers HRQOL as measured with the SF-12 (Tables 7 and 8). PCS (or MCS) score was used as the dependent variable, and the significant factors in univariate analysis were included as independent variables. We examined the severity of multicollinearity in the multiple linear regression models using Variance Inflation Factor (VIF). Therefore, we removed the disability certificate because of the high VIF (7.937 > 5). The PCS regression model showed that caregiver's illness condition, family size, and household income had a statistically significant effect on the physical HRQOL of caregivers of CWD. Specifically, caregivers with illness (β = −0.291, p < 0.001) and caregivers living in an extended family with more than six persons (β = −0.165, p = 0.032) had lower physical HRQOL. Household income was positively correlated with the physical HRQOL of caregivers. The MCS regression model showed that caregiver's illness condition and family size were statistically significant factors, indicating that caregivers had higher mental HRQOL when they had an accompanying illness (β = 0.286, p = 0.005) or the household's resident population was six or more (β = 0.216, p = 0.003).

Exploring Factors Affecting the HRQOL of Caregivers in Shanghai has Important Value for the Rehabilitation of CWD
Our research group was dedicated to the disability prevention, rehabilitation, and health of CWD for a long time. HRQOL, as the health aspect of quality of life, focuses on people's level of ability, daily functioning, and ability to experience a fulfilling life. Whereas caregivers' HRQOL may not only affect their own living standard, but also has significant impact on the work productivity which is the guarantee for adequate care and caregiving quality. Moreover, previous studies also demonstrated significant correlations between caregivers' HRQOL and the rehabilitation effect as well as community reintegration of CWD [3,30]. Therefore, exploring the facilitators and barriers affecting the caregivers' HRQOL and identifying the key factors can provide guidance for implementing targeted interventions. They are also valuable for promoting the rehabilitation of CWD and improving their health.

Parents Are the Primary Caregivers of CWD, but the Caregiving Role of Grandparents Cannot be Ignored
Among the 170 caregivers of CWD, most of them were parents (73.5%) with the highest proportion of mothers (48.2%), which is consistent with most studies [31,32]. It is worth noting that the proportion of grandparents in this study is 25.9%, reflecting the social phenomenon of the elderly bringing up their grandchildren. In fact, raising grandchildren by the elderly is not a unique scenery in China, as the elderly in many countries are on the road to bring up grandchildren, and even become the main force. In Europe, 58% of grandmothers and 49% of grandfathers took care of at least one of their grandchildren in the preceding year [33]. In the United States, 5.7 million grandparents lived with their grandchildren in 2000, and the numbers seemed to be on the rise [33]. In 2002, 60-70% of Taiwanese elderly lived with their married children or with grandchildren [34]. For families with CWD, it is important to have two earners that maintain the functioning of the families. Neither mothers nor fathers may be in a position to provide the desired amount of childcare inside the nuclear household. According to a carer survey in urban Britain, 79% of the respondents were predominantly grandparents [35]. In this study, we found that grandparents (51.81 ± 9.92) had lower scores on physical HRQOL than parents (52.77 ± 7.83) because of the physical degeneration of elders. Thus, intervention strategies should cover not only parents but also grandparents.

Physical HRQOL of Caregivers of CWD is Moderate, But the Mental HRQOL Needs to Be Urgently Concerned
The results of the normative comparison showed that the physical HRQOL of caregivers was better than that of the general population [28,29]. Possible explanations for this phenomenon are that 73.53% of the study population were parents, while the normative model consisted of the entire population. In addition, there is an adverse selection tendency to choose the person with better HRQOL as the primary caregiver in order to ensure the caregiving quality for the disabled children.
Meanwhile, we found that the mental HRQOL of caregivers of CWD in this study (31.58 ± 7.72) was worse than that of the general population. This finding is consistent with other studies [36], but the mental condition among caregivers in our study was extremely poor. Actually, this result validates our previous hypothesis. In developing countries, there are substantial shortfalls in the availability of rehabilitation services, and the effectiveness of rehabilitation might not be satisfactory [37,38]. Insufficient social inclusion in these places also makes it difficult for CWD to attend school [39]. The United Nations Children's Fund estimates that 90% of CWD in Africa do not attend school and are thus less likely to engage in other opportunities for social participation [40,41]. Furthermore, living in developed cities like Shanghai, the high living expense often puts great financial pressure on caregivers, especially those who have CWD. In addition, poor mental HRQOL may be also associated with China's one-child policy, which had been implemented for 30 years and caused the prevalence of the only-child family [42]. On one hand, if parents were the only child, their capability of adapting themselves to the adversity was relatively weak. On the other hand, if the child with a disability is also an only child, the psychological trauma to the parents is conceivable. All of these issues could increase parental concern and aggravate caregivers' psychological status. Therefore, relevant government departments and community committees should pay more attention to this group and provide corresponding social support.

The Disease Condition of Caregiver Has a Negative Effect on Physical HRQOL But a Protective Effect on Mental HRQOL
Disease condition was found to have a significant impact on both the physical and mental HRQOL of caregivers. Specifically, caregivers with a disease had a worse physical HRQOL than those without illness, which is consistent with some previous research [36]. However, several studies found no statistical difference in physical and mental HRQOL between caregivers with and without disease [43]. In this study, 34.7% of the caregivers were found to have one or more diseases, including hypertension (15.3%), arthritis/rheumatic disease (5.3%), diabetes (4.7%), and coronary heart disease (3.5%). This is consistent with a Canadian study [30] that found that caregivers of children with cerebral palsy had significantly higher rates of hypertension, diabetes, and migraines than ordinary caregivers. Both findings suggest that caregivers should be concerned about their disease condition and be provided with timely and targeted professional support, for example, physical examination and medical assistance service. Additionally, the psychological HRQOL of the caregivers with disease was better than those without disease. This difference is beyond our expectation, but in fact is a common psychological phenomenon. Caregivers with a disease have lower caregiving expectations which increases the feasibility of achieving the goal. Therefore, there is rising satisfaction and higher mental HRQOL.

Family Size Expansion Benefits the Mental HRQOL of Caregivers, But Often Means a Greater Need for Physical Assistance and Support
Different from our common sense, this study found that caregivers had better mental HRQOL but worse physical HRQOL when they lived in a family with the permanent residents of six or more, compared with those who lived with less than two persons. With the development of urbanization, the family size tends to be smaller and simpler, with the nuclear family becoming the mainstream family form in modern society [44]. It is a well-documented phenomenon that women's fertility intention is low or falling rapidly everywhere except in parts of Sub-Saharan Africa [45,46]. The more developed the region, the more pronounced this phenomenon is [47]. In these places, older people emphasize fostering independence in their children, family integration is unlikely to occur if the nuclear family could withstand the caregiving pressure. However, if the parents are unable to cope with the corresponding stress because of poor physical HRQOL, they would likely turn to the child's grandparents in order to increase the overall resilience [48]. Meanwhile, there was high consistency in independent views of mothers and fathers of grandparents identified by both as the most available and supportive [49]. Therefore, the family size expands.
In addition, it has been confirmed by several studies [50][51][52] that social support (including family support, and support from friends) can positively influence the mental HRQOL of caregivers. The corresponding plausible explanation is that other family members are able to provide emotional support to caregivers when there are more persons in the family. Furthermore, the psychological stress of caregivers is reduced. Therefore, relevant departments should strengthen humanistic care for caregivers with small families, providing psychological support such as door-to-door communication and psychological counseling.

Caregivers of Children with Disabilities Face Tremendous Financial Pressures, Providing Financial Support Could Improve Their HRQOL
In this study, we found that household income was significantly positive associated with the physical HRQOL of caregivers. Higher income could improve caregivers' parental physical environment and increase the access to health services [53]. This finding is consistent with most previous studies. In fact, for all families of CWD, caregivers are under tremendous financial pressure [23]. On the one hand, providing the basic necessities and rehabilitation for these children is costly. On the other hand, caregivers often have to quit their jobs to take care of the child full time, which reduces their source of income [54]. However, we did not find a statistically significant effect of household income on the mental HRQOL of caregivers. This is mainly because the household income of families with CWD is always insufficient compared with the child's rehabilitation expense, and financial need is always the most pressing. This finding suggests that although some financial subsidies are already available, they are far from enough and need to be further adjusted. Therefore, relevant government departments could provide additional financial support depending on families' economic status, and this might be useful in enhancing the HRQOL of caregivers and their children.

Implications and Limitations
Results from this study have practical implications for health professionals and policymakers to help improve the HRQOL of caregivers of CWD. We provide three suggestions as follows. First, rehabilitation physicians should pay attention to the mental health of caregivers and regularly conduct psychological counseling to reduce their psychological pressure; second, local governments should understand the comprehensive and dynamic information about the caregivers, their children, and living environment, so as to provide targeted physical examination, medical assistance service, humanistic care, and financial support; finally, government departments should improve the effectiveness of services for CWD and provide coverage for them. After all, as long as the child is healthy, the caregiver is better off.
There are several limitations in this study. Firstly, the study design was cross-sectional. Thus, it was difficult to establish causal relationships between caregivers' HRQOL and related factors. Secondly, caregivers of children with all types of disabilities were included in this study, therefore it was hard to identity the determinants of caregivers' HRQOL of children with specific disability. Thirdly, the sample size is only 170, which may affect the accuracy of the HRQOL levels. Finally, we include many factors that could influence caregivers HRQOL, but these factors were measured crudely, such as using just one question. Actually, this is a preliminary exploration to discover significant factors, and more in-depth studies will be followed up based on this research.

Conclusions
In summary, this study found poor mental HRQOL among caregivers of CWD in Shanghai, and thus requiring urgent attention and intervention. The factors we identified provide an intervention framework for promoting caregivers' HRQOL from the perspective of children, caregivers, and the environment. Significant factors indicate the key role of caregiver's illness condition, family size, and household income. Therefore, interventions should be targeted at these key factors to improve HRQOL of caregivers and further facilitate children's rehabilitation and health. Additionally, it is worth noting for local health care policymakers and public health researchers that relevant health management strategies should not only consider parents but also grandparents.
Author Contributions: C.X. contributed to the design and conception of the study, analysis of data, and wrote the manuscript. M.S. contributed to the design of the investigation questionnaire, the conception of the study, and manuscript revision. Material preparation and data collection were performed by X.L. and X.G. M.S., J.L., and G.C. participated in the design and coordination of the study, assisted with data acquisition, and finalized the manuscript. C.L. participated in the manuscript revision and English language editing. All authors have read and agreed to the published version of the manuscript.