Participants and Procedure
This cross-sectional study was conducted in Ningxiang County of Hunan province from November 2015 to January 2016. Our sample was primary caregivers of PLS that were registered in the “Central Government Support for the Local Management and Treatment of Severe Mental Illnesses project”, also named as “686 program”, which is China's largest demonstration project for mental health services that seeks to integrate hospital and community services for serious mental illness [33, 34]. The 686 Program provides multiple comprehensive mental health services including: patient registration and initial assessment, free medication and regular follow-up in the community, management of community emergencies, and free emergency hospitalization [33, 34].
A one-stage cluster-sampling method was used to recruit primary caregivers of PLS from the 686 Program. First, three Towns and 1 Xiang (an administrative unit similar to a town but with lower socio-economic development) were randomly selected from Ningxiang County, and then followed by whole sampling of all communities within each Town and all villages within each Xiang, leading to a total sampling frame of 55 representative communities/villages. This sampling method was used to obtain a sample as representative as possible and minimize sampling bias. A local primary care physician within each community/village assisted in identifying eligible primary caregivers based on a list of PLS registered in the 686 Program. Eligibility criteria for primary family caregivers included: 1) caring for a family member with schizophrenia who is registered in the 686 Program; 2) the family member being cared for meets criteria of the Chinese Classification of Mental Disorders-3 (CCMD-3) or the International Classification of Diseases-10 (ICD-10) for schizophrenia; 3) the primary caregiver is living with the PLS at least for the past two years and assumes major responsibilities for caregiving; 4) the primary caregiver is older than 16 years of age, since by Chinese civil law, people at the ages of 16 and older who are engaged in labor (here referred to as “caregiving”) and have a source of economic income are considered to have full capacity for civil conduct; and 5) the primary caregiver speaks Chinese, is literate, and not seriously disabled, thus enabling the person to understand and communicate in face-to-face interviews with the caregivers. Exclusion criteria included: 1) caring for family member with diagnosis other than schizophrenia such as depression and epilepsy; 2) primary family caregivers who do not speak Chinese or having serious physical or mental illness that makes them unable to communicate effectively in a face-to-face interview. A total of 352 primary family caregivers of PLS were identified in the final sample.
The community/village primary care physician visited each study household with a member of the research team to obtain informed consent from the caregivers, who then completed a series of questionnaires (see measures below) in face-to-face interviews, which were all conducted in Chinese. All caregivers were reimbursed with small gifts equivalent to RMB ¥ 10 (approximately USD$1.5) in return for their participation. Details of the subject recruitment process have been described elsewhere [35, 36].
Of the 352 caregivers approached, 327 participated in the interview (response rate: 93%). Among the 25 caregivers that refused the interview, reasons for refusal included no interest in the study (n=11), concern about stigma (n=9), too painful to share caregiving experiences (n=3) and other reasons not disclosed (n=2). No significant differences in socio-demographic background were observed between those who did or did not agree to participate based on age, gender, marital status, employment and education (available as supplementary material, S1). In the current study, only parents (n=151) and spouses (n=113) were included in the analyses, resulting in a final sample of 264. This sample size of 264 detected an effect size of 0.4 between parent and spouse caregivers by two-tailed t-test, assuming α = 0.05, β = 0.10.
Measures
Socio-demographic information. Demographic information of both the PLS and the primary caregiver was collected by asking the primary caregivers on a questionnaire designed for this study, which included gender, age, marital status, occupation, education, family financial status, and kinship between PLS and the primary caregiver. For the primary caregivers, we also asked if there were any co-caregivers, dependents, physical illnesses, and how care had been provided for the PLS. This information was collected through face-to-face interviews between the research team and primary caregivers.
PLS functioning. PLS functioning was assessed using the Global Assessment of Functioning scale (GAF) [37]. The GAF is an axis of the DSM-IV and assessed the person’s overall social, occupational and psychological functioning in the past month from 1 (lowest) to 100 (highest), with examples provided for each ten-point interval. The GAF was first translated in Chinese by Zhang in 1984 [38] and proved to be a reliable and valid measure of psychiatric functioning [37, 38]. For the current study, PLS GAF was assessed by the research team based on a combination of caregivers’ report, direct assessment of the PLS, and a review of their clinical records, whenever available.
Caregiving activities. Specific caregiving activities were assessed by four “Yes-No” questions asking if the respondent was involved with the following four aspects of caregiving activities for the PLS: daily activities (e.g., eating, drinking, getting dressed), medication management (e.g., monitoring medication, buying medicine), hospital visits (e.g., registration, hospitalization), and financial help (e.g., spending and giving money). Responses of “no” were scored as 0, and “yes” as 1, and then followed by asking how often the respondent was involved with each caregiving activity, ranging from “occasionally” (scored as 1) to “always” (scored as 4). Detailed information about the questions and optional answers can be found in the Appendix S2. For the current study, questions on caregiving activities were administered by the research team to primary caregivers.
Objective burden. Objective burden was assessed using the Family Burden Interview Schedule (FBIS) [39] classified into six categories: financial burden, disruption of routine family activities, family leisure, family interactions, and effect on physical and mental health of others. The scale consists of 24 items rated on a 3-point Likert scale from 0 (no burden) to 2 (serious burden). The total score ranges from 0 to 48 with higher scores showing higher family burden. The FBIS was first translated into Chinese by Chien et al.[40] in 2004 and proved to be a reliable and valid measure of objective family burden. For the present study, FBIS was administered by the research team to primary caregivers and the Chinese version of FBIS showed acceptable internal consistency with a Cronbach’s α of 0.86 for the total scale and ranged from 0.63 to 0.86 for the subscales.
Subjective burden. Subjective burden was assessed using the Zarit Burden Interview (ZBI) [41]. The ZBI consists of 22 items scored on a 5-point Likert scale from 0 (never) to 4 (nearly always), except for the final item on global burden, rated from 0 (not at all) to 4 (extremely). The total score ranges from 0 to 88 with higher scores indicating higher perceived burden [42, 43]. The ZBI was first translated into Chinese by Lu et al.[44] in 2009 and proved to be a reliable and valid measure of subjective caregiver burden. For the present study, ZBI was administered by the research team to primary caregivers and the Chinese version of ZBI showed acceptable internal consistency with a Cronbach’s α coefficient of 0.89.
Caregiver depression. Caregiver depression was measured using the 9-item Patient Health Questionnaire (PHQ-9) [45], which consists of 9 items scored in 4-point Likert scale from 0 (not at all) to 3 (nearly every day). The total score ranges from 0 to 27, with higher scores implying more depressive symptoms, and a cut-off point of 10 differentiating depression and non-depression [46, 47]. The PHQ-9 was first translated into Chinese by Yeung et al. [48] in 2008 and proved to be a reliable and valid measure of depression. For the present study, PHQ-9 was administered by the research team to primary caregivers and the Chinese version of the PHQ-9 demonstrated good internal consistency with a Cronbach’s α coefficient of 0.89.
Caregiver anxiety. Caregiver anxiety was measured by the 7-item Generalized Anxiety Disorder Scale (GAD-7) [49] which detects the primary caregiver’s anxiety symptoms during the past two weeks. Each item is rated on a 4-point Likert scale from 0 (not at all) to 3 (nearly every day). The total score ranges from 0 to 21, with a cut-off point of 10 differentiating anxiety and non-anxiety [50]. The GAD-7 was first translated into Chinese by He et al.[51] in 2010 and proved to be a reliable and valid measure of anxiety. For the present study, GAD-7 was administered by the research team to primary caregivers and the Chinese version of the GAD-7 demonstrated good internal consistency in the current study with a Cronbach’s α coefficient of 0.91.
Caregiving rewarding feelings. Positive feelings about caregiver were assessed using the caregiving rewarding feelings (CRF) scale which was designed for this study. CRF was initially developed based on qualitative interviews on a convenience sample of 30 pairs of PLS and their primary caregivers, then revised based on pre-testing and Delphi’s method, and validated in a larger sample. The detailed development and validation process of the CRF has been described elsewhere [52]. The CRF was developed in Chinese and consists of 12 items asking about a range of possible rewarding feelings that caregivers may have during caregiving. Each “yes” response was scored 1 and “no” responses 0, with total scores ranging from 0 to 12; higher scores indicate more positive feelings in caregiving. For the present study, CRF was administered by the research team to primary caregivers and the Chinese version of CRF showed acceptable reliability with a Cronbach's alpha of 0.77.
Caregiver perceived family functioning. Caregiver perceived family functioning was assessed using the Family Adaptation, Partnership, Growth, Affection and Resolve Index scale (APGAR) [53]. It consists of 5 items scored in 3-point Likert scale from 0 (hardly ever) to 2 (almost always). The total score ranges from 0 to 10 with higher scores indicating higher satisfaction with family functioning. The APGAR was first translated into Chinese by Chang et al.[54] in 1993 and proved to be a reliable and valid measure of family functioning. For the present study, APGAR was administered by the research team to primary caregivers and the Chinese version of APGAR showed acceptable internal consistency with a Cronbach’s α coefficient of 0.91.
Statistical Analysis
Data analyses examined missing values, influential values and outliers, skewness, and kurtosis. Frequencies and percentages were displayed for categorical variables, and means with standard deviations or medians with interquartile ranges (IQR) displayed for continuous variables. We compared socio-demographic characteristics, caregiving activities, and six types of caregiving impacts (objective family burden, subjective caregiver burden, depression, anxiety, caregiving rewarding feelings, and family functioning) between spouse and parent caregivers using two-group tests. Depending on variable type and data distribution, we conducted various two-group tests including: independent two sample t-tests for normally distributed continuous data, Mann-Whitney U test for non-normally distributed continuous data, and Pearson’s chi-square tests for categorical variables. In order to examine the predictive effect of kinship (spouse vs parent) on the six types of caregiving impacts, we further conducted six separate multivariate linear regressions, with kinship (spouse vs parent) as independent variables, and the six types of caregiving impacts as dependent variables, while controlling for all potential confounders (PLS functioning, gender, and education, as well as caregiver age, gender, marriage, employment, education, family income, whether have co-caregivers, whether there were additional dependents, whether caregivers have physical illness, and length of caring) to avoid analysis bias (Table 4). Data were analyzed using STATA software version 15.0, with p values smaller than 0.05 in two-tailed tests considered as statistically significant.
Ethical Considerations
This study was approved by the Ethics Review Committee of the Xiangya School of Public Health of Central South University in China. The interviewed participants were informed verbally and in writing of the study’s purpose, their right to refuse to participate, and the voluntary nature of their participation. All participants provided written informed consent before the interviews.