Predictors of caregiver burden after stroke in Nigeria: Effect on psychosocial well-being

Abstract

Background:

Family caregivers (CGs) are critical to the care and recovery of stroke survivors (SSs), particularly in the community. However, little is known about their psychosocial well-being, especially in developing countries. In this study, we assessed CG burden, psychiatric morbidity, quality of life (QoL), and predictors of burden.

Materials and Methods:

We recruited 150 dyads of CGs/SSs from two outpatient clinics in Nigeria. Data were collected through sociodemographic/clinical questionnaire, the Zarit caregiver burden interview, the World Health Organization QoL-BREF, and the Mini-International Neuropsychiatric Interview. Disability in SSs was graded with the Modified Rankin Scale.

Results:

Compared to CGs who were mostly females (86.7%) with a mean age of 42.8 (±12.6) years, SSs were older with a mean age of 61.6 (±12.5) years and were mainly males (62.7%). The mean CG burden score was 31.6 (±10.5) and eight in 10 CGs, 124 (82.7%) reported moderate-to-severe burden scores. Mental disorders in the form of depression and/or anxiety disorders were diagnosed in 26 (17.3%) CGs. Significant predictors of high-burden scores (P < 0.05) were incontinence in SSs, psychological symptoms in SS, worse poststroke disability, and performance of more CG tasks (B = 8.3, 4.5, 3.9, and 3.5, respectively). CGs QoL scores correlated negatively with burden scores, with medium-to-large effect sizes (r = −0.4–−0.6) across QoL spheres.

Conclusion:

In this study, psychosocial burden in CGs was indexed by various determinants and impacted QoL negatively. Given the integral role of CGs to the care of SSs in the community, mental health services, psychosocial support programs, and stratification based on known vulnerability factors seems viable options for intervention, assessment, and planning.

INTRODUCTION

Stroke is a leading cause of death and adult disability worldwide with recent trends that further indicates increased stroke morbidity and mortality in low- and middle-income regions.[1,2] Emphasis of stroke rehabilitation has recently shifted from a stroke survivor (SS) focused to SS/caregiver (CG) dyad approaches, in recognition that CGs play an essential role in preserving rehabilitation gains and the long-term well-being of SSs.[3] Resources for stroke rehabilitation are deficient in the developing countries, whereas caregiving still remains the backbone of services provided for SSs by spouses, children, or other relatives.[4,5,6] The wide range of disabilities faced by SSs predisposes them to variable degrees of dependence on CGs to live a normal life.[7]

During the period leading to hospital discharge, family CGs of SSs may face immense uncertainties and new challenges.[8] The CGs must learn to cope with their concurrent grief and an uncertain future.[8] Thus, the sudden need to care for a SS can be emotionally difficult and physically challenging.[8] Caregivers (CGs) are often unaware of their new role, unfamiliar with the disease, its progression, and the kinds of community support options available.[9]

Caregiver burden refers to alterations in CGs emotional and physical health which can occur when care demands outweigh available resources,[10] the extent to which CGs feels there emotional, physical, social life, and financial status has suffered as a result of caregiving,[11] which indeed affects their overall health and quality of life (QoL).[12] Caregiver burden has been shown to have cumulative negative effects on SSs treatment outcomes.[12]

The levels of psychosocial burden on CGs of SSs is most likely influenced by a variety of factors which could be directly linked to the length of time spent caregiving and levels of assistance devoted to dealing with the consequences of disability.[12] Other CG factors such as age, gender, disturbed sleep, financial stress, little leisure time, depression, lack of social support, the presence of incontinence, and significant neurological deficits in SSs have also been found to significantly increased levels of burden.[2,12,13,14,15]

There is a dearth of studies devoted toward understanding the complex and multi-layered phenomenon of caregiving after stroke.[16,17,18] This study evaluated CG burden, psychiatric morbidity (depression and/or anxiety disorder), QoL, and predictors of burden among principle family CGs of SSs at outpatient clinics of two teaching multispecialty hospitals. We posit that certain variables of interest will predict CG psychosocial burden and also that CG burden negatively impacts QoL.

MATERIALS AND METHODS

Study setting and design

The study was carried out at two University Teaching Hospitals in Nigeria. These two health institutions are the main teaching and multispecialty tertiary hospitals in Lagos, Nigeria. The state of Lagos in South Western Nigeria is the most densely populated and commercial nerve center of the country. The hospitals both serve as referral centers for the metropolis of Lagos and its environs.

This study is a cross-sectional survey. Data collection was done over 6 months between October 2014 and March 2015. The sample size was calculated using the Public Domain Statistical Software Java-Stat with power set at 80%.[19] Recruitment of participants was done using convenience sampling of every eligible participant who met the study criteria until the minimum estimated sample size was completed. Overall, a total of 164 participants were recruited from both study centers, but 14 had incomplete data and were not included in the final analysis. The study participants consisted of 150 adult principle family CG-SS dyads recruited at neurology and physiotherapy clinics at both hospitals.

The CGs for this study were spouses, direct relatives, or in-laws of SSs. These family CGs were also the principal people who, without being a health professional or belonging to a social support network, live with the SSs and in some way, were directly implicated or affected by the SS's health problem.[20] Where a SS had more than one CG, they or their family members are usually asked to nominate the principal CG[21] who if not present initially was invited for interview at the next available SS's clinic visit. Also included were principle family CG who on an average spent at least 20–25 h a week (for this study equated to 5 h a day for at least 5 days in a week) attending to the SS throughout the care period, CG who have been giving care for at least 1 month and care recipients with current stroke duration of at least 1 month at interview.[22,23,24]

Excluded from the study were CGs not up to the age of 18 years, CGs with a history of mental and severe chronic debilitating physical illnesses predating the commencement of caregiving. Also excluded were CG of SS providing care services for other chronically ill people during the study period.

Ethical consideration

Ethical approval was obtained from the Health Research and Ethics Committee of both Lagos University Teaching Hospital and Lagos State University Teaching Hospital before the commencement of the study. Informed consent was given by all participants following explanation of the purpose of study, study procedure, and expected duration of involvement. Other issues such as confidentiality and voluntariness regarding participation were also explained.

Instruments and procedure

The instruments used included: sociodemographic/Clinical Questionnaires (SDCQ), the Zarit Burden Interview (ZBI),[25] and the World Health Organization QoL-BREF scale (WHOQoL-BREF).[26] The Mini-International Neuropsychiatric Interview (MINI)[27] and the Modified Rankin Scale (MRS).[28]

The ZBI is a CG self-report measure. The 29-item questionnaire version of the ZBI was used for this study.[25] Each item in the ZBI is a statement which the CG is asked to endorse using a 5-point scale.[25] The ZBI measures caregiving burden as it relates to time, physical health, emotional health, and social relationships.[25] The WHOQoL-BREF is a 26-item self-administered instrument that measures health-related QoL.[26] The WHOQoL-BREF is a shorter version of the WHOQoL-100 scale which measures the subjective response of respondents in the past 2 weeks[26] The WHOQoL-BREF has cross-cultural applications having been designed for diverse cultural groups including sub-Saharan Africa, and has been validated in Nigeria.[29] The MINI is a structured diagnostic tool patterned along the Diagnostic and Statistical Manual Text Revision IV (DSM IV-TR) diagnostic criteria of the American Psychiatric Association.[27] The MINI is interviewer administered and used as a diagnostic tool for neuropsychiatric disorders.[27] The 6^th version of the MINI used for this study was chosen because of its reliability, validity, availability, ease of use, and shortened time of administration.[27] The MRS is a quick measure of the degree of disability or dependence in daily activities in people who have suffered strokes or neurological sequel of other medical conditions.[28] The MRS is a widely used clinical outcome measure for stroke research.[28] The MRS scale describes six grades of disability after stroke.[28] All the study instruments have been validated and/or used in Nigeria.[29,30,31,32,33]

Principal family CG/SS dyads were recruited consecutively from neurology and physiotherapy outpatient clinics of the hospitals on alternate clinic weekdays. Written consent was obtained from the study participants after details of the study were explained to them. Principal family CGs were initially administered the CG sections of a designed SDCQ. Those that meet the study criteria of inclusion were then administered; the ZBI, the WHOQoL-BREF, and the MINI. SSs were administered relevant aspects of the SDCQ. This included portions detailing symptoms of psychological distress in them. SSs that subsequently meet the study inclusion criteria were also administered the MRS. Sometimes, the assistance of proxies which was usually their principal family CG was required to obtain clinical data of SSs with significant cognitive deficits. Other relevant information needed to be inputted into the SDCQ was also obtained from retrieved individual case folders of SSs.

Data analysis

All data collected were coded and entered for analysis. This was done with the Statistical Package for Social Sciences for windows (SPSS) version 17.0 (SPSS Inc. Chicago, USA).[34] Descriptive statistics such as means, frequencies, and percentages were used to analyze the data. Chi-square and Fisher's exact tests was used to test for significance between categorical variables. Continuous variables were analyzed using the Independent t-tests and ANOVA. Correlation and regression analysis was also performed to predict associations between significant variables. A confidence level of 95% will be used which allows for 5% sampling error with significance ≤0.05.

RESULTS

Clinicodemographic/burden profile of caregivers and its associations with Zarit caregiver burden scores

A total of 150 CG-SS dyads were recruited for the study. The mean (±standard deviation) age of CGs was 42.8 (±12.6) years. Majority of CGs were female 130 (86.7%), married 100 (66.7%) and employed 94 (62.7). Higher proportions of CGs 72 (48.0%) were spouses to SSs. The prevalence of mental disorders was 26 (17.3%) that consisted of anxiety disorders – 6 (4%), depression – 19 (12.7%), and anxiety/depression – 1 (0.6%). The mean Zarit burden score was 31.6 (±10.5) with 124 (82.7%) experiencing moderate-to-severe burden. Majority of CGs spent an average of >10 h daily rendering care assistance 50 (68.0) and performed on an average >5 care duties daily 113 (75.3). High mean Zarit burden scores were significantly associated with CGs who were, female (P = 0.041), CGs who usually performed caregiving tasks on an average >10 h daily (P < 0.001), CGs averaging >5 care-related tasks daily (P < 0.001), CGs cohabiting in the same room with SS (P < 0.001), and CGs with positive mental diagnoses (P < 0.001) [Table 1].

Table 1.

Clinico-demographic/burden profiles of stroke survivor caregivers and its associations with Zarit Caregiver Burden Scores

Parameter	Mean (SD)/n (%)	Zarit, mean (SD)	TOS	df	P
Zarit Burden Mean Score		31.6 (10.5)
Zarit Burden Grade
Mild	26 (17.3)
Moderate	78 (52.0)
Severe	46 (30.7)
Caregiver mean age (years)	42.8 (12.6)
Duration of caregiving (years)
<10	85 (56.7)	32.80 (10.5)	1.609	148	0.110
≥10	65 (43.3)	30.03 (10.4)
Gender
Male	94 (62.7)	27.15 (10.4)	−2.057	148	0.041
Female	130 (86.7)	32.28 (10.4)
Marital status
Unmarried	50 (33.3)	29.74 (10.7)	1.541	148	0.126
Married	100 (66.7)	32.53 (10.4)
Employment status
Employed	94 (62.7)	30.35 (10.7)	−1.903	148	0.059
Unemployed	56 (37.3)	33.70 (10.0)
Relationship with SS
First degree	62 (41.3)	29.48 (10.6)	2.180	2	0.117
Spouse	72 (48.0)	33.07 (10.5)
Others	16 (10.7)	33.19 (9.4)
Caregiver training
Yes	2 (1.3)	21.00 (11.3)	−1.442	148	0.151
No	148 (98.7)	31.74 (10.5)
Living arrangement
Same room with SS	113 (75.3)	33.92 (10.3)	−5.112	148	<0.001
Separate room with SS	37 (24.7)	24.51 (7.8)
Caregiver assistance
Yes	125 (83.3)	31.34 (10.3)	−0.687	148	0.493
No	25 (16.7)	32.92 (11.5)
Average caregiving hours
≤10	48 (32.0)	24.40 (7.8)	−6.515	148	<0.001
>10	50 (68.0)	34.99 (10.0)
Average caregiver daily tasks
Usually ≤5	37 (24.7)	22.74 (7.7)	−6.897	148	<0.001
Usually >5	113 (75.3)	34.61 (9.7)
Mental diagnoses
Positive	26 (17.3)	41.04 (6.4)	5.515	148	<0.001
Negative	124 (82.7)	29.62 (10.1)
Total	150 (100.0)

SD – Standard deviation; df – Difference of means; P – Statistical significance; TOS – Test of significance; In bold – Statistically significant, SS – Stroke survivor; P – level of significance. Mental diagnosis – Anxiety disorder - 6, depression - 19, anxiety and depression - 1

Clinico-demographic profile of stroke survivors and its associations with Zarit caregiver burden scores

The mean (±SD) age of SSs was 61.63 (±12.5) years. Majority of SSs were male 94 (62.7%) and unemployed 129 (86.0%). A high proportion of SSs also had moderate-to-moderately severe levels of poststroke disabilities 80 (53.3%). High Zarit burden scores were significantly associated with SSs that were unemployed (P < 0.001), incontinent (P < 0.001), that had dysarthria/aphasia (P < 0.001), and SS with worse poststroke disabilities (P < 0.001). Significantly high-burden scores were also associated with SSs with medical comorbidities (P = 0.008) and symptoms of psychological distress (P < 0.001). SSs paying >10,000 Naira monthly as health costs also had significantly higher burden scores (P < 0.001). Finally, where health-care costs were borne by CGs and/or SSs, significantly high-burden scores (P = 0.015) were also found [Table 2].

Table 2.

Clinico-demographic profile of stroke survivors and its associations with Zarit Caregiver Burden Scores

Variable	Mean (SD)/n (%)	Zarit, mean (SD)	TOS	df	P
Zarit Mean Burden Score		31.6 (10.5)
SS mean age (years)	61.63 (12.5)
Gender
Male	94 (62.7)	31.63 (10.6)	0.042	148	0.967
Female	56 (37.3)	31.55 (10.5)
Marital status
Unmarried	52 (34.7)	31.79 (9.8)	−0.160	148	0.873
Married	98 (65.3)	31.50 (10.9)
Employment status
Employed	21 (14.0)	24.14 (9.6)	−3.651	148	<0.001
Unemployed	129 (86.0)	32.81 (10.2)
Stoke duration (months)
≤12	84 (56.0)	32.20 (10.6)	0.791	148	0.430
>12	66 (44.0)	30.83 (10.5)
Stroke lateralization
Right	76 (50.7)	30.62 (10.5)	1.297®	2	0.277
Left	71 (43.7)	32.89 (10.4)
Indeterminate/brain stem	3 (2.0)	26.00 (14.4)
Dysarthria/aphasia
Yes	79 (52.7)	35.36 (10.2)	4.972	148	<0.001
No	71 (47.3)	27.42 (9.2)
SS incontinence (urine and/or fecal)
Yes	44 (29.3)	42.39 (6.0)	10.800	148	<0.001
No	106 (70.7)	27.12 (8.5)
Modified Rankin Scale
None/slight disability	70 (46.7)	24.80 (8.1)	9.308	148	<0.001
Moderate-moderately severe disability	80 (53.3)	37.55 (8.6)
Monthly health bill (Naira)
≤10,000	60 (40.0)	26.35 (9.3)	−5.461	148	<0.001
>10,000	90 (60.0)	35.10 (9.9)
Health bill paid by
CG + CG/SS	35 (23.3)	35.37 (10.2)	2.467	148	0.015
Others	115 (76.7)	30.45 (10.4)
SS medical comorbidity
Yes	19 (12.7)	37.53 (9.0)	2.686	148	0.008
No	131 (87.3)	30.74 (10.5)
SS psychological symptoms
Yes	21 (14.0)	40.05 (7.6)	4.190	148	<0.001
No	129 (86.0)	30.22 (10.3)
Total	150 (100.0)

SD – Standard deviation; df – Difference of means; TOS – Test of significance; In bold - Statistically significant; CG – Caregiver; SS – Stroke survivor; P – Level of significance

Multiple linear regression analysis: Predictors of caregiver burden scores and clinico-demographic significant parameters in caregivers/care recipients

The multiple linear regression model predicted that Zarit caregiver burden scores were statistically significant (F = 20.759, R²= 0.665) for the presence of incontinence in SSs (B = 8.348, P < 0.001), presence of psychological distress symptoms in SSs (B = 4.544 P < 0.007), more severe poststroke disability (B = 3.974, P = 0.006), and CGs with usual daily care duties of >5 (B = 3.504, P = 0.026) [Table 3].

Table 3.

Multiple linear regression analysis: Predictors of caregiver burden scores and clinico-demographic parameters in stroke caregivers/care recipients

Parameters	Unstandardised coefficient (B)	Standardized coefficients (β)	t	P	Lower CI	Upper CI
Constant	17.939		10.226	0.000	14.470	21.408
CG gender	1.042	0.034	0.641	0.523	−2.173	4.257
Living arrangement	1.234	0.051	0.818	0.415	−1.749	4.217
CG tasks	3.504	0.144	2.253	0.026	0.429	6.580
CG mental diagnosis	2.372	0.086	1.439	0.152	−0.888	5.632
SS employment status	2.330	0.106	1.878	0.063	−0.123	4.783
SS dysarthria/aphasia	−0.229	−0.011	−0.182	0.856	−2.722	2.264
SS incontinence	8.348	0.363	5.773	0.000	5.488	11.207
SS Modified Rankin Score	3.974	0.189	2.808	0.006	1.175	6.772
SS health bills paid by	0.396	0.019	0.354	0.724	−1.815	2.608
SS monthly health cost	1.362	0.064	1.083	0.281	−1.126	3.850
SS medical comorbidity	2.046	0.065	1.222	0.224	−1.265	5.356
SS psychological symptoms	4.544	0.151	2.755	0.007	1.282	7.806
CG care giving hours	2.418	0.108	1.747	0.083	0.319	5.154

Dependent variable – Zarit Score. Model F=20.759; R²=0.665. In bold – Significant at P<0.05. CG – Caregiver; SS – Stroke survivor; t – t-test; P – Level of significance; CI – Confidence interval

Stroke survivor caregiver's quality of life mean scores and correlations between mean Zarit caregiver burden scores and quality of life scores

Correlation analysis showed statistically significant (P < 0.05) negative relationships between Zarit burden scores and across all six QoL spheres scores with large effect sizes (ES) seen in psychological health (r = −0.639), social health (r = −0.567), overall health (r = −0.551), and physical health (r = −0.533) QoL spheres. Moderate ES was, however, seen in environmental health (r = −0.485) and health satisfaction (r = −0.414) QoL spheres [Table 4].

Table 4.

Stroke survivor caregiver’s quality-of-life mean scores and correlations between mean Zarit Caregiver Burden Scores and quality-of-life scores

Variable (n=150)	Mean (SD)	Mean Zarit Score	r	P
Overall health	66.4 (15.07)	31.6 (10.5)	−0.551	<0.001
Health satisfaction	69.2 (13.032)		−0.414	<0.001
Physical health	67.76 (9.098)		−0.533	<0.001
Psychological health	61.99 (9.536)		−0.639	<0.001
Social health	56.64 (10.093)		−0.567	<0.001
Environmental health	56.83 (7.960)		−0.485	<0.001

r – Parsons correlation coefficient; P – Level of significance; SD – Standard deviation; In bold – Statistically significant. r <±0.3 – Mild; r±0.3–±0.5 – Moderate; r >±0.5 – Large

DISCUSSION

This study investigated caregiver (CG) burden in principle family CGs of SSs. The study also evaluated predictors of CG psychosocial burden and correlations between burden and QoL.

Clinico-demographic profile of caregivers/stroke survivors and its predictors of caregiver burden

The mean age of CGs in this study was 42.8 (±12.6) years, which is in keeping with other studies reviewed.[4,6,11] Majority of CGs were female, married spouses of SSs also in agreement with other studies reviewed.[4,6,7,18,23] These findings are not unusual because, in most African communities, family CG roles are mostly ascribed to female spouses.[35] The mean CG burden score in this study was indicative of moderately severe burden.

High-burden scores were significantly associated with CGs being female. Female CGs having significantly higher burden scores was replicated in other studies reviewed.[7,13,15,36] Other studies reviewed, however, found no association between CG burden and gender.[4,37] Reason for this inconsistency is unclear. It has, however, been argued that in Africa, sociocultural factors may limit male CGs ability to express burden as it may indicate incompetence.[7]

In this study, CG parameters such as increased CG tasks, more CG contacts hours, and closer CG cohabitation with SSs was associated with higher caregiver burden. These findings above have been replicated in other studies reviewed.[36,37,38,39] Regression analyses also predicted that for every point increase in Zarit burden score there was an almost four-fold increase in CGs performing more daily care tasks.

Financial factors such as high monthly health-care bills and where the responsibility of paying such bills rested on CGs and/or SSs was also significantly associated with high CG burden, findings which were replicated in other studies.[5,13,24,36] These findings above are not out of place. In Nigeria, like a majority of developing countries, health-care financing is mostly out of pocket with scarce or poor health insurance coverage. The added financial health-care bills payment burden by CGs and/or SSs is likely to compound other pre-existing factors accentuating CG burden.

The prevalence of mental disorders in the form depression and/or anxiety disorders among CGs was 26%. The combined prevalence of anxiety and depression in CGs of SSs from studies reviewed ranged from 21.0% to 30.8%.[40,41,42] A recent meta-analyses by Loh et al.,[41] found a pooled global mean combined prevalence of 30.8%, slightly higher than findings of this study. Methodological differences such as, differences in defining who a CG is; differences in sample sizes, differences in timing of data collection in reference to period after stroke and differences in diagnostic criteria for assessment may have been responsible, i.e., use of symptom rating scales versus structured instruments for diagnoses as was done in our study. The study, furthermore, found that CGs with positive mental diagnoses were significantly more likely to have a higher CG burden. This finding was replicated in other studies reviewed.[13,38,39]

Our findings of worse post stroke disabilities/complications, being positively associated with higher CG burden were replicated in other studies reviewed.[4,5,15,23,38] This study also predicted that for every point increase in Zarit burden score, there were, eight-fold increases in caring for SSs that were incontinent and almost four-fold increases in caring for SSs with worse poststroke disabilities as measured using the MRS.[28] Furthermore, we also predicted that for every unit rise in burden score there was an almost five-fold rise in caring for SSs with psychological symptoms in agreement with other studies reviewed.[9,43]

Correlates/predictors of caregivers World Health Organization quality of life and Zarit Care burden

QoL is perceived as the result of a complex process of interaction between personality traits, medical outcomes, coping behavior, support, and quality of healthcare received.[44] Health-related QoL measures are important in quantifying burden of diseases and evaluating the effects of therapies.[45] This study found significant negative correlations between Zarit CG burden scores and all spheres of the WHOQoL-BREF scores with ES ranging from moderate to large. This finding was replicated in other studies reviewed.[4,18,36,46,47] Boakye et al.,[23] however, found no significant association between burden scores and any of the eight domains of the QoL measure used. The authors, however, recruited relatively younger CGs compared to this study that may have had significantly better baseline QoL and possibly coped well in physically strenuous care tasks. In this study, the social health QoL domain had the lowest mean score, whereas the psychological health QoL domain had the largest ES negative correlation. The above infers that besides psychological factors, social support, sexual activity, and personal relationships are negatively affected by CG burden and should be focal points for CG/SS dyad interventions to improve their psychosocial well-being.

The hospital-based nature of the study and cross-sectional design limits making causal inferences. The exclusions of certain categories of caregiver/care recipients, limit generalizability of findings. The occasional use of proxy responses may also limit the accuracy of findings.

Implications of the study

Reports show that SS characteristics correlate with the range and intensity of their CGs' burden.[15,48,49] Conversely, there is evidence to suggest that CGs burden adversely impacts the care rendered to those with stroke.[12,50] This implies that the quality of care SSs receive relates to the well-being of their carers. It is imperative, therefore, for studies such as this to identify those factors that predict burden in CGs of SSs. Understanding these factors is a key step and a valuable tool toward the design of early intervention strategies for carers who are at risk of distress from caregiving. This would improve CGs health and SSs' care outcomes if successfully implemented.

CONCLUSION

This study found that CGs of SSs faced significant psychosocial burden. This burden was influenced by varied clinico-demographic factors in CG-SS dyads and negatively correlated with all spheres of CGs QoL There is need to streamline intervention programs based on research findings supporting CG-SS dyad approaches for better outcomes.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.