Background
In a context of de-institutionalization and community-based mental health services, the family caregiver’s role is becoming increasingly important at a sociological, economic, and political level [
1,
2]. Caregivers serve as main providers of care for people with mental illness and thus shoulder a substantial part of the burden [
3]. The burden of caregiving is often described as consisting of two parts: objective burden (such as completion of manual tasks and household duties) and subjective burden (such as caregiver perceptions of emotional distress and stigma) [
4‐
6]. In combination, objective and subjective burden encompass physical, mental, financial and social aspects of caregiving [
4]. Identifying a measure to assess burden for caregivers of people living with schizophrenia (PLS) is not only a vital step in understanding the types of support that caregivers need, but also critical to developing effective intervention programs that target specific caregiver needs [
7,
8]. Although a plethora of measures have been developed to assess caregiving burden, two have been validated in international studies and used most often among caregivers of PLS: The Family Burden Interview Schedule (FBIS) [
9] and the Zarit Caregiver Burden Interview (ZBI) [
10].
The Family Burden Interview Schedule (FBIS) was initially developed by Pai and Kapur among Indian caregivers of psychiatric patients [
9]. The original FBIS contains 26 items, with the first 24 items grouped under 6 categories to measure objective burden: financial burden, disruption of routine family activities, disruption of family leisure, disruption of family interaction, effect on physical health of others, and effect on mental health of others. Each item of objective burden is rated on a 3-point Likert scale (0 = no burden, 1 = moderate burden, 2 = severe burden). The 25th item comes as a supplementary question asking about any other family burden not mentioned in the first 24 items, and the 26th item serves as a subjective burden assessment by asking one standard question “How much would you say you have suffered owing to the patient’s illness?” on a 3-point Likert scale (0 = not at all, 1 = a little, 2 = severely). Most studies only use the first 24 items to assess objective family burden for convenient calculation and easy interpretation, which we refer to as FBIS-24. The total score of the FBIS-24 ranges from 0 to 48, with higher scores indicating higher objective burden. A recent study showed an optimal cut-off score of 23 to distinguish lower and higher burden for risk of psychological distress, with sensitivity being 76% and specificity being 68% for depression as measured by PHQ-9 [
11]. The FBIS-24 has now been used among caregivers of a wide variety of health conditions as a standardized measure for assessing caregiving burden of both hospitalized and community living care-recipients [
12‐
14].
The Zarit Caregiver Burden Interview (ZBI) was initially developed by Zarit and his colleagues among US caregivers of dementia seniors [
15]. The original ZBI contains 29 items on a four-point Likert scale, which was later revised as 22 items on a five-point Likert scale and commonly referred to as ZBI-22 [
10]. Each item of the ZBI-22 assess the respondent’s subjective burden by asking “do you feel or do you wish….” with optional answers scored from 0 to 4 (0 = never, 1 = rarely, 2 = sometimes, 3 = quite frequently, 4 = nearly always) except for the item 22 (0 = not at all, 1 = a little, 2 = moderately, 3 = quite a bit, 4 = extremely). The total score of the ZBI-22 ranges from 0 to 88, with higher scores indicating higher subjective burden [
16]. A recent study showed an optimal cut-off score of 48 to distinguish lower and higher burden for risk of psychological distress, with sensitivity being 73% and specificity being 80% for depression as measured by PHQ-9 [
17]. The ZBI-22 is the most widely used tool for measuring the level of subjective burden among caregivers, has been validated across many populations of caregivers (i.e. spouses/partners, children, and parents) and care-recipients (i.e. AD/dementia, physical illness, and mental illness), and is available in most languages (e.g., French, Spanish, and Chinese) [
18,
19].
Although both measures have been used extensively in international studies to assess caregiving burden, data for the FBIS-24 and ZBI-22 have not been reported in the same study. A careful comparison between the two scales have shown that apart from differences in item number, Likert scale rating, score range, cut-off value, sensitivity, specificity and dimensionality, they also differ in their contents and emphasis (as shown in Table
1). FBIS-24 focuses on objective burden on the family level, while ZBI-22 targets at measuring subjective burden on individual level. Yet there is lack of evidence on the psychometric performance comparison between the two scales for burden assessment in the same study. The purpose of the present study is to address this issue by reporting for each measure: 1) sociodemographic and clinical differences for lower and higher levels of burden, 2) reliability (Cronbach’s alpha), and 3) validity in terms of convergent validity, and known groups’ validity for lower and higher levels of burden. Our sample is a Chinese rural community sample of caregivers of PLS in which all measures were collected simultaneously, thus allowing for the examination of unique and overlapping characteristics of objective and subjective burden using the FBIS-24 and ZBI-22, respectively.
Table 1
Simple comparison between FBIS-24 and ZBI-22
Item number | 24 | 22 |
Likert scale rating | 3-point (0-no,1-moderate, 2-serious) | 5-point (0-never, 1-seldom,2-sometimes, 3-often, 4-always) |
Score range | 0–48 | 0–88 |
Cut-off value& | 23 [11] | 48 [17] |
Sensitivity& | 76% | 73% |
Specificity& | 68% | 80% |
Dimensionality | 6 (financial burden, disruption of routine family activities, family leisure, family interactions, and effect on physical and mental health of family members) | 1–5 (different studies proposed different factors) |
Results
Group comparisons of socio-demographic characteristics using two burden measures
Table
2 shows the socio-demographic characteristics of the sample and their comparisons between lower and higher burden groups. The median (interquartile range) age of the caregivers was 59 (49–67) years, most of them were married (82.26%) and with a primary grade education (59.94%). Slightly more than half of the caregivers were female (53.82%) and employed (52.91%). The two largest family relationships to the PLS were as a parent (46.18%) or spouse (34.56%). The caregivers spent a median (q1–q3) of 15 years (9–25) caregiving.
Table 2
Demographic characteristics of caregivers categorized by validated cutoff values of FBIS-24 and ZBI-22 (n = 327)
Age (years) | median (IQR, min-max) | 59 (49–67; 17–81) | 59 (48–67; 17–81) | 58 (50–66; 25–81) | 0.345 | 0.730 | 59 (48–67; 17–81) | 59 (50–66; 29–77) | 0.103 | 0.918 |
Gender | Male | 151 (46.18) | 74 (52.48) | 59 (38.06) | 6.203 | 0.013 | 87 (48.60) | 46 (38.33) | 3.068 | 0.080 |
Female | 176 (53.82) | 67 (47.52) | 96 (61.94) | | | 92 (51.40) | 74 (61.67) | | |
Marriage* | Married | 269 (82.26) | 119 (84.40) | 130 (83.87) | 0.0153 | 0.902 | 153 (85.47) | 98 (81.67) | 0.773 | 0.379 |
Unmarried | 58 (17.74) | 22 (15.60) | 25 (16.13) | | | 26 (14.53) | 22 (18.33) | | |
Occupation | Employed | 173 (52.91) | 78 (55.32) | 78 (50.32) | 0.740 | 0.390 | 93 (51.96) | 64 (53.33) | 0.0547 | 0.815 |
Unemployed | 154 (47.09) | 63 (44.68) | 77 (49.68) | | | 86 (48.04) | 56 (46.67) | | |
Education | Primary | 196 (59.94) | 82 (58.16) | 93 (60.00) | 6.320 | 0.042 | 98 (54.75) | 78 (65.00) | 7.411 | 0.025 |
Middle | 87 (26.61) | 32 (22.70) | 47 (30.32) | | | 48 (26.82) | 33 (27.50) | | |
High | 44 (13.46) | 27 (19.15) | 15 (9.68) | | | 33 (18.44) | 9 (7.50) | | |
Kinship | Spouses | 113 (34.56) | 54 (38.30) | 45 (29.03) | 9.794 | 0.007 | 66 (36.87) | 34 (28.33) | 12.917 | 0.002 |
Parents | 151 (46.18) | 52 (36.88) | 85 (54.84) | | | 68 (37.99) | 70 (58.33) | | |
Other | 63 (19.27) | 35 (24.82) | 25 (16.13) | | | 45 (25.14) | 16 (13.33) | | |
Care duration (years) | 15 (9–25; 1–49) | 16 (8–25, 1–49) | 14 (9–23; 2–44) | 0.630 | 0.529 | 15 (8–25; 1–49) | 15 (9.5–23; 2–44) | −0.173 | 0.863 |
Using a score of 23 and 48 as the cut-offs for FBIS-24 (objective burden) and ZBI-22 (subjective burden) to distinguish those with higher and lower burden, we compared these groups on demographic characteristics. Overall, the demographic characteristics of these two groups were comparable. Relative to the lower burden group, caregivers in the higher burden group were more likely to be parents (54.84% vs 36.88% for FBIS-24, 58.33% vs 37.99% for ZBI-22) and less likely to have a high school education (9.68% vs 19.15% for FBIS-24, 7.50% vs 18.44% for ZBI-22). For either measure, no significant differences were found between lower burden and higher burden groups based on caregiver age, marital status, occupation, and number of years caregiving. However, there were more female caregivers with higher objective burden (FBIS-24) than with lower burden (61.94% vs 47.52%, p < 0.05), but this gender difference was only observed at a trend level of significance for subjective burden with the ZBI-22 (61.67% vs 51.40%, p = 0.080).
Group comparisons of clinical characteristics
Table
3 shows the clinical characteristics and comparisons between lower and higher burden groups for the two measures. The median score of depression, anxiety, and family functioning of caregivers as measured by PHQ-9, GAD-7, and APGAR were 9 (4–15), 9 (4–15) and 7 (4–9), respectively. The median score of PLS functioning as measured by GAF was 41 (20–61).
Table 3
Clinical characteristics of caregivers categorized by validated cutoff values of FBIS-24 and ZBI-22 (n = 292)
Depression (PHQ-9) | 9 (4–15; 0–27) | 6 (2–9; 0–27) | 14 (8–18.5; 0–27) | −8.30 | < 0.001 | 6 (2–9; 0–26) | 15 (10–20; 0–27) | −9.58 | < 0.001 |
Anxiety (GAD-7) | 9 (4–15; 0–21) | 6 (1–9; 0–21) | 13 (7–18; 0–21) | −7.99 | < 0.001 | 6 (1–10; 0–21) | 14 (9–18; 0–21) | −8.95 | < 0.001 |
Family functioning (APGAR) | 7 (4–9; 0–10) | 8 (5–10;0–10) | 6 (3–9; 0–10) | 3.30 | 0.001 | 8 (5–10; 0–10) | 5 (2–8; 0–10) | 5.12 | < 0.001 |
PLS functioning (GAF) | 41 (20–61; 1–99) | 55 (30–68;4–90) | 30 (15–55; 1–99) | 5.21 | < 0.001 | 55 (30–65;4–99) | 23 (13–50.5; 1–85) | 6.45 | < 0.001 |
Clinical comparisons between higher and lower burden using cut-off scores of 23 and 48 for FBIS-24 and ZBI-22, respectively, showed comparable patterns between the two measures, regardless of whether the assessment was of objective or subjective burden. Relative to caregivers with lower burden, those with higher burden also reported higher levels of depression (14 vs 6 for FBIS-24, 15 vs 6 for ZBI-22) and anxiety (13 vs 6 for FBIS-24, 14 vs 6 for ZBI-22) scores, lower family functioning scores (6 vs 8 for FBIS-24, 5 vs 8 for ZBI-22), and lower PLS functioning (30 vs 55 for FBIS-24, 23 vs 55 for ZBI-22).
Psychometric comparisons
Table
4 shows comparisons of reliability and validity results between the FBIS-24 and ZBI-22. Overall, reliability for each measure was generally good, with Cronbach’s alpha at 0.86 for the FBIS-24 and 0.89 for the ZBI-22. Convergent validity was also in general accordance with expectations and similar between the FBIS-24 and ZBI-22, with significant positive correlations with PHQ-9 score (
r = 0.51 and 0.64,
p < 0.01) and GAD-7 score (
r = 0.49 and 0.56,
p < 0.01), and significant negative correlations with APGAR score (
r = − 0.26 and − 0.31,
p < 0.01) and GAF score (
r = − 0.43 and − 0.46,
p < 0.01), respectively. All correlations were of medium to large effect size, except for the APGAR score, indicating family functioning may be less strongly correlated to caregiver burden than PLS functioning and caregiver psychological distress. For known groups’ validity, both the FBIS-24 and ZBI-22 showed significantly higher burden scores in caregivers with a physical illness than for those without a physical illness (26 vs 22 and 47 vs 36.5,
p < 0.05).
Table 4
Reliability and Validity of the FBIS-24 and ZBI-22(n = 292)
Internal consistency |
Cronbach’s α | 0.864 | 0.888 |
Convergent validity |
Caregiver depression (PHQ-9) | 0.505** | 0.639** |
Caregiver anxiety (GAD-7) | 0.494** | 0.558** |
Family functioning (APGAR) | −0.261** | − 0.307** |
PLS functioning (GAF) | −0.427** | − 0.455** |
Known groups’ validity |
Caregivers with physical illness | 26 (18–31; 0–44) | 47 (33–56; 1–84) |
Caregivers without physical illness | 22 (12–29; 2–44) | 36.5 (20–55.5; 0–88) |
Z | −1.967 | − 2.508 |
P | 0.049 | 0.012 |
Cohen’s kappa | 0.523 | < 0.001 |
Among the 295 caregivers with non-missing data in both FBIS-24 and ZBI-22, most of caregivers (52.54%) were identified as having higher family burden using the FBIS-24 cut-off of 23, whereas the prevalence of higher burden was 40.0% in the same group using the ZBI-22 cut-off of 48, with moderate chance-corrected agreement between the two instruments (Cohen’s kappa: 0.52). The overlap of higher burden prevalence is illustrated as following: 101 of the 295 caregivers (34.2%) were captured by both FBIS-24 > 23 and the ZBI-22 > 48.
Discussion
This study is the first to compare the psychometric properties of two widely used measures of caregiving burden, the FBIS-24 and the ZBI-22, as indicators of objective and subjective burden for caregivers of PLS. We examine sociodemographic and clinical differences, reliability and validity for lower and higher burden scores on each measure among Chinese caregivers. Our findings indicate that both scales demonstrate acceptable psychometric performance with respect to the internal consistency, convergent validity and know group’s validity, and moderate chance-corrected agreement between them (Cohen’s kappa: 0.52).
What is most striking is how comparable both measures are psychometrically, despite measuring slightly different aspects of the same construct, caregiver burden. Both measures demonstrate strong internal consistency and are significantly associated with expected clinical indicators of caregiver depression, anxiety, and family functioning, and PLS functioning. In addition, both measures reveal a gender difference in caregiving burden in which women report significantly higher levels of objective burden on the FBIS-24 and higher subjective burden at a trend level of significance on the ZBI-22. Higher burden among women is consistent with gender role theory and is consistent with the caregiving literatures [
46‐
48]. Some recent research has also noted that women may have fewer resources available to them for caregiving and use less effective ways of coping [
49‐
52].
The comparable psychometric properties of the FBIS-24 and ZBI-22 indicate that the two scales may be used interchangeably in assessment of caregiver burden. The choice between FBIS-24 and ZBI-22 may depend on your research question and interest based on the observed differences such as sensitivity, specificity and contents, as shown in Table
1. For instance, FBIS-24 focuses on objective burden at family level, with higher sensitivity at cut-off value of 23 than ZBI-22 (76% vs 73%) [
11]. Meanwhile, ZBI-22 mainly targets at subjective burden at individual level, with higher specificity at cut-off value of 48 than FBIS-24 (80% vs 68%) [
17]. If the primary interest is addressing objective indicators of burden related to manual tasks and duties at family level, the FBIS-24 may be particularly useful in identifying a broad spectrum of caregivers for assessment or intervention purposes. In contrast, if the primary interest is excluding individuals for assessment or intervention with lower levels of subjective burden at personal level, in terms of emotional distress and experiences of stigma, then use of the ZBI-22 to do so may be more appropriate.
Another implication of the present findings is that the use of both measures in the same study, may be warranted as the FBIS-24 and ZBI-22 complement one another in our understanding of caregiver burden, and offer consistent correspondence to one another to related sociodemographic and clinical measures of a study population [
53]. Although it may not be feasible to use both measures in real-world contexts, future research should examine ways that both measures offer a more precise and nuanced understanding of caregiver burden that facilitates assessment and subsequent intervention.
This study has a few limitations. First, the data reported is cross-sectional, and so no longitudinal psychometric data is reported for each measure. However, this limitation is mitigated somewhat because the purpose of this study was to examine psychometric properties of each measure and their relationship to various concurrent sociodemographic and clinical indicators. Second, the sample came from a single study population, a relatively rural county of Hunan province, which may limit the study’s representativeness. Future research should examine our findings in additional populations.
Conclusions
In conclusion, the present study provides empirical support for the psychometric comparability of two measures of caregiver burden, the FBIS-24 and the ZBI-22, for the reliable and valid assessment of caregiver objective and subjective burden, respectively. The choice between the FBIS-24 and the ZBI-22 for assessment of caregiver burden depends on your research question and interest.
Acknowledgements
The authors are grateful to the families that participated in this study for sharing their experiences. The authors also appreciate the assistance of the health and family planning bureau of Ningxiang County, the government of the Towns of Liushahe, Shungfupu, and Yutan as well as the Xiang of Chengjiao, for their administrative support, and the village/community physicians that assisted in home visits to each household in the rural areas of Ningxiang County, Hunan province.
Open AccessThis article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit
http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (
http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.