Introduction
Caregiver burden refers to the many physical, psychological, and socio-economic problems that arise from caring for an individual with a chronic and disabling disease (Zarit et al., Reference Zarit, Todd and Zarit1986). Caring for people with PD or dementia places a significant level of burden on the carer, which can lead to mental health problems and declining physical health. Being a carer in a setting such as sub-Saharan Africa (SSA), where little is known about these conditions and where there are few healthcare workers to provide formal support or education may be even more demanding. Conversely, multigenerational living may provide significant family support and allow sharing of care among many caregivers. Such living arrangements are common in rural SSA.
Both PD and dementia are associated with certain health beliefs in SSA, which might lead to further stress for caregivers. For example, we have previously reported that patients with PD experienced stigma within the community as many patients, carers, and laypersons believed that their symptoms were caused by witchcraft (Mshana et al., Reference Mshana, Dotchin and Walker2011).
Caregiver burden of dementia
Strain on carers for people with dementia in developing countries measured by the Zarit Burden Interview (ZBI) has previously been reported (Prince, Reference Prince2004; Prince et al., Reference Prince2012), yet few of these data came from SSA. In a large multicenter international study, 20 people with dementia and their carers in Nigeria were the only African participants (Prince et al, 2004). Carers in Nigeria had high mean ZBI scores (50.3) and a high prevalence of psychiatric morbidity (95%) as measured by the 12-item General Health Questionnaire (GHQ-12). While the majority of carers lived in extended family households, levels of caregiver strain were at least as high as those seen in developed countries. A multicenter study of 673 carers by Prince et al. (Reference Prince2012) noted that higher carer strain was associated with more severe dementia, behavioral and psychological symptoms and greater time spent caring. Median ZBI scores ranged from 16 (rural China and urban Peru) to 29 (rural Peru).
Two previous studies from Africa have examined levels of stress for carergivers of people with dementia, although neither used the ZBI. Ukpong and Makanjuola (Reference Ukpong and Makanjuola2003) reported caregiver emotional distress as a result of caring for people with dementia (n = 20) or schizophrenia (n = 24) in Nigeria. The two groups differed in age (as would be expected), with the dementia patients being significantly older than the schizophrenic patients (mean 71.5 years vs. 34.8 years). There was also a difference in location at the time of the questionnaire being completed: all schizophrenic patients were inpatients, whereas all dementia patients were outpatients. GHQ-30 scores were reported as a measure of carer emotional distress. The GHQ-30 is related to the GHQ-12, and a score of 5 or greater indicates possible “caseness,” i.e. the possibility of being a case with a minor mental health disorder. The mean score for carers of patients with schizophrenia was higher than carers of dementia patients (6.08 and 5.4, respectively), but this difference did not reach significance. Both levels, however, are above the “caseness” threshold.
Baiyewu et al. (Reference Baiyewu2003) reported level of caregiver distress for carers of people with dementia in Nigeria, using the Neuropsychiatric Inventory (NPI). From an original community-based cohort of 98 people with dementia, 40 were still alive and they and their carers were interviewed at home. Of these, 39/40 patients had probable or possible Alzheimer's disease (AD). Symptoms on the NPI that correlated with the most caregiver distress were reported. In 37/40 situations the carer lived with the patient. Most of the carers were female (30/40), with daughter being the most common relationship to the patient (14/40). The mean age of participants in this study was 84.7 years. The most common symptoms reported in the people with dementia were appetite change, depression, irritability, and apathy. The most distressing symptoms were: depression, irritability, anxiety, and nighttime behavior. It should be noted that 58 people were not followed up from the original study, of whom 91% had died or were unable to be examined. These patients may well have had the most severely disabling disease, and therefore, highest levels of carer strain, leading to an underestimation of strain in this study.
Caregiver burden of PD
The ZBI has previously been used to assess carer strain in people with PD in the USA, Japan and Europe (Wallhagen and Brod, Reference Wallhagen and Brod1997; Edwards and Ruettiger, Reference Edwards and Ruettiger2002; Edwards and Scheetz, Reference Edwards and Scheetz2002; Marsh et al., Reference Marsh, Williams, Rocco, Grill, Munro and Dawson2004; Martinez-Martin et al., Reference Martinez-Martin, Arroyo, Rojo-Abuin, Rodriguez-Blazquez, Frades and de Pedro Cuesta2008). The largest study of this type was conducted in Spain, with 286 carers interviewed in a longitudinal study (Martinez-Martin et al., Reference Martinez-Martin, Arroyo, Rojo-Abuin, Rodriguez-Blazquez, Frades and de Pedro Cuesta2008). Greater carer burden was associated with greater disease severity, disease duration, disability, anxiety, depression, and psychosis in the person with PD. These findings are broadly supported by other smaller studies (Wallhagen and Brod, Reference Wallhagen and Brod1997; Edwards and Ruettiger, Reference Edwards and Ruettiger2002; Edwards and Scheetz, Reference Edwards and Scheetz2002; Marsh et al., Reference Marsh, Williams, Rocco, Grill, Munro and Dawson2004). There are no previous studies of care burden in PD from African cohorts.
Aims and objectives
We have previously conducted community-based prevalence studies of dementia (Longdon et al., Reference Longdon2013) and PD (Dotchin et al., Reference Dotchin2008) in the Hai district demographic surveillance site in northern Tanzania (population 161,119 in 2009). Our previous experience of research work in the Hai area had suggested that dementia was not a particularly big cause of burden among caregivers. Many relatives did not believe that there was any medical problem, and that the person was simply “getting old.” As intergenerational living is the norm in Hai, younger family members are expected (and would expect) to look after older relatives living in the family home. We hypothesized that, in contrast to what one might expected in a developed world setting, persons with dementia, particularly in the early stages, would not present a substantial burden to carers. In contrast, our experience of people with PD in Hai was that, as the majority of patients were not on drug treatment, they tended to be physically quite disabled, requiring significant caregiver input. Furthermore, PD was also frequently not understood to be a medical problem and a great deal of stigma was associated with it in this setting (Mshana et al., Reference Mshana, Dotchin and Walker2011).
The aim of the current study was to document the burden of caregiving for people with PD and dementia and to compare and contrast the difficulties associated with being a caregiver for two distinct conditions: one with predominantly physical symptoms and one predominantly cognitive symptoms. In order to compare the burden of PD and dementia to that experienced by carers of the general elderly in Hai, we also collected data from carers of a cohort of age-matched elderly people living in Hai, randomly selected from the background population, who did not have PD or dementia.
Methods
Participants
All participants were living in the community at the time of assessment.
Data from three groups of people were included in this study:
-
(1) All surviving people with PD from the prevalence study (Dotchin et al., Reference Dotchin2008), along with several people newly diagnosed with PD as part of a study of neurological disorders in those aged 70 years and over (Dewhurst et al., Reference Dewhurst2013). All participants had to have an identifiable formal carer.
-
(2) All people with dementia from the dementia prevalence study (Longdon et al., Reference Longdon2013) with an identifiable regular main caregiver.
-
(3) A control group of general elderly people with an identifiable main carer, but without a diagnosis of PD or dementia. The control group was selected from all people interviewed as part of the dementia study but found to be cognitively normal (Longdon et al., Reference Longdon2013). Selection was by a random number generator with stratification for age, such that there was overlap in the interquartile range (IQR) for age between the control group and those with PD or dementia. The size of the control group was selected to be the same as the combined number of people with dementia and PD included in the study. All controls were assessed for neurological disorders; however, we have no data relating to other possible diagnoses, such as cardiovascular disease or musculoskeletal disorders.
Assessment and diagnosis
PD diagnosis was according to the UK PD Society Brain Bank Criteria (Gibb and Lees, Reference Gibb and Lees1988) and dementia diagnosis was based on DSM IV criteria (American Psychiatric Association, 1994). Diagnoses were made by the research doctors who performed the clinical examination. These diagnoses were subsequently validated independently by a UK-based specialist in movement disorders (by video recording review, RW) or dementia (by case history and CT brain scan review; Longdon et al., Reference Longdon2013).
The ZBI is the most commonly used tool to define caregiver strain in dementia (Zarit et al., Reference Zarit, Todd and Zarit1986). The 22-item questionnaire gives the respondent a list of statements describing how they might feel in their role as a carer. Carers report for each statement how often they feel this way, from “never” (score = 0), to “nearly always” (score = 4). The ZBI scores can range from 0 to 88, with higher scores reflecting higher burden on carers. Hérbert et al. (Reference Hérbert, Bravo and Préville2000) have suggested cut-offs for low (0–8), moderate (9–17), high (18–32), and severe (33 and over) levels of strain. Schreiner et al. (Reference Schreiner, Morimoto, Arai and Zarit2006) report a cut-off of 24, above which 72% of carers with probable depression were identified. The ZBI has been used as part of the 10/66 research group protocol for assessment of carers of people with dementia in developing countries, thus, we felt it was the most appropriate tool to use here (Prince et al., Reference Prince2012).
The ZBI was administered by the dementia research nurse (AK, fluent in English and Swahili) assisted by one of the study doctors (S-MP, AL, RW) if necessary. People who had PD with dementia (PDD) had their data analyzed with people with PD, since PD was the primary diagnosis. In addition, age, gender, and disease duration were recorded for person with PD and gender for each carer. Disease severity was also assessed using the Hoehn and Yahr (Hoehn and Yahr, Reference Hoehn and Yahr1967) scale for people with PD and the Clinical Dementia Rating (CDR) scale for dementia (Morris, Reference Morris1993).
Statistics
The data were quantitative in nature and were analyzed using standard statistical software, PASW-18 for windows (SPSS, Chicago, IL, USA). All variables were found to be non-normally distributed (Kolmogorov-Smirnov test) and so did not meet parametric assumptions. Therefore, the data are described in terms of the median and IQR. Linear regression modeling was used to identify independent predictors of ZBI score. Stepwise methods were used to construct models and the robustness checked by examination of residuals, Cook's distances, and collinearity diagnostic tests.
Ethics
Ethical approval was obtained from the Tanzanian National Institute for Medical Research.
Results
Of the original PD prevalence cohort (n = 32), 13 were still alive. ZBI was recorded for 11 people; the remaining two had no identifiable formal carer and were largely independent. Of the 12 people with PD identified from the neurological disorders study, ZBI was completed for 9. Three had no identifiable formal carer as they were independent. Of the people with dementia (n = 78), three were excluded due to a diagnosis of PDD and for 32 people no ZBI was completed because the relatives did not believe that the care they provided for their relative was out of the ordinary intergenerational expectation in Hai. Thus, five of 25 (20.0%) people with PD were excluded since they had no formal carer, compared to 32 of 75 (42.7%) people with dementia.
Of 218 eligible controls, 120 had an identifiable main carer who completed the ZBI. From these, 63 people were selected using the methods described. Six of the selected controls had a neurological diagnosis: two had essential tremor, two had spinal cord pathology, one had chronic headaches, and one had muscle wasting. Demographic and clinical characteristics of all participants are detailed in Table 1. For controls, only ten (15.9%) of the carers did not have a total ZBI score of zero.
Demographic and clinical characteristics of participants
H&Y = Hoehn and Yahr scale; CDR = Clinical Dementia Rating scale; IQR = interquartile range.
Comparison of carer burden in people with PD and people with dementia
In linear regression models, disease type (PD or dementia), patient age and gender, carer gender, and disease duration were investigated as univariate and multivariate predictors of ZBI score. Only disease duration (β = 1.054, 95% CI for β 0.164 to 1.965, p = 0.021) and disease type (β = 14.911, 95% CI for β 7.937 to 21.884, p < 0.001) emerged as univariate predictors of ZBI score; however, disease duration was not an independent predictor of ZBI score on multivariate regression, with disease type as the only independent predictor.
Discussion
Caring for a relative with a progressive neurological condition in a resource-poor setting is associated with a moderate to high level of caregiver strain as determined by the ZBI. Prior to the prevalence studies in this area, the majority of the people with PD and/or dementia had not been diagnosed and was untreated (Dotchin et al., Reference Dotchin2008; Longdon et al., Reference Longdon2013). No formal support systems, either financial, educational or emotional, exist in this setting. Within the Hai district, there are no institutions for older persons requiring a high level of care. All such people would be cared for long-term at home, with their family providing the bulk of the informal care. Very few patients or families could afford to pay someone to help care for them. Many carers would have a “day job” that they had had to cut back or stop to enable them to care for their relative, meaning a loss of productivity for the family.
ZBI scores were higher in the PD caregiver group than the dementia or control group, although this was partly explained by the longer disease duration for people with PD, greater strain on carers for an illness with predominantly physical symptoms, rather than cognitive symptoms, is likely to be one of the main reasons for this. A number of carers and relatives of the people with dementia reported thinking that the person's symptoms were a normal sign of aging and not having sought any medical help for the problem. We have previously published data on the stigma associated with having PD (Mshana et al., Reference Mshana, Dotchin and Walker2011) in Hai and are currently conducting a similar study around dementia. It will be interesting to note whether stigma is generally lower in relation to having dementia (or a relative with dementia) than it is in relation to PD. This may in turn account for some of the difference in the level of burden appreciated by the carer.
There may be other explanations for the relatively low burden for carers of people with dementia in this setting. Many elderly people in Hai live a day-to-day life that is less cognitively demanding that might be experienced by someone living in a high-income country (Paddick et al., Reference Paddick2013). Furthermore, there are often numerous potential caregivers to assist in activities of daily living (ADLs), such that someone with moderate dementia would still be able to function relatively well, with little carer input. ZBI scores for carers of people with dementia were generally lower in Hai than seen in other low- and middle-income countries, including India, Latin America, and China (Prince et al., Reference Prince2012). Kinship and multigenerational living are likely to be factors in the lower burden. In our experience, multigenerational living is much more common in rural Tanzania than seen in many other world regions and in urban SSA (Aboderin, Reference Aboderin2004). Younger family members expect, and are expected to, care for older family members. It may be that only when someone becomes physically disabled by his or her condition does the carer perceive a significant burden in this setting. In societies where multigenerational living is less common and feeling of kinship may be eroding due to demographic transition, the burden of dementia on carers may well be greater.
Conversely, the physical symptoms of PD, which can be successfully treated by medication, at least at the beginning of the disease course, are not generally as well controlled in Hai as in high-income countries (Dotchin et al., Reference Dotchin2008). The reasons for this are complex, but include low rates of diagnosis, difficulty sourcing medication, poor access to multidisciplinary team input, and lack of local nurses and doctors to deal with follow up of the condition. Carers often find the symptoms very distressing, as the visible signs of PD (tremor, stooped posture, mask like face, and so on) can sometimes be misinterpreted as being due to witchcraft or evil spirits (Mshana et al., Reference Mshana, Dotchin and Walker2011). The physical environment in Hai can also be challenging with uneven floors and roads, and few devices and adaptations to assist people with physical disability to carry out ADLs. In general, our research staff found that physical caregiving was more burdensome to carers in this setting than the psychological problems. Hence, those with milder dementia, with few or no physical needs, did not engender carers feeling as burdened as those with PD.
Limitations
Limitations of the study include the relatively small sample size (although larger than previously published reports from SSA) and the small number of people in the more severe disease stage sub-groups for both PD and dementia. This reflects reduced survival of those in the most advanced stages, but may also have served to underestimate carer burden in dementia, as those with the most advanced disease would be most likely to require increased levels of care from carers and have more physical needs to be met. We also acknowledge that by not including the substantial number of people with PD and dementia, who did not have a recognized carer, we may have biased our results. We could have included those whose carer did not believe they were doing anything more than culturally expected for their elders, and therefore that there was no care burden, by scoring them as zero for the ZBI. However, we wanted to focus on those who had an identifiable carer. Moreover, including these individuals would only have increased the difference between those with PD and those with dementia, since only 20% of those with PD had no carer, compared to over 40% of those with dementia. Anyone wishing to interpret our results by including those with no identifiable carer will be able to do so.
Conclusion
Caring for an individual with PD appears to be more burdensome than caring for an individual with dementia in Hai, Tanzania. This may in part be due to how the diseases are perceived by the general population, with dementia being considered a normal part of aging and societal roles for younger family members in Hai dictating that caring for their elderly people is expected. People with PD attract greater stigma and require more “heavy” care than people with dementia. Age alone is not a predictor of higher caregiver strain.
Our results are likely to have relevance in other regions of SSA and our study may have some important implications for health service provision in this resource-poor setting. While dementia may be much more frequently seen than PD in SSA, the impact PD has on patients affected, and their carers, appears to be much greater. The lack of public awareness, and the stigma associated with PD, heightens this burden (Mshana et al., Reference Mshana, Dotchin and Walker2011). If they were diagnosed earlier, and treated more effectively, the burden could be significantly lessened. Training for health professionals and education for people with PD and their carers may make a substantial difference to the quality of life of carers. Despite the higher carer burden of PD, the burden of dementia in many families should not be ignored, and similar strategies to those outlined above are also likely to be of benefit.
Conflict of interest
None.
Description of authors’ roles
Catherine Dotchin, Richard Walker, and Paul Chaote designed this study. Catherine Dotchin, William Gray, Stella-Maria Paddick, and Anna Longdon carried out background literature searches. Aloyce Kisoli, Stella-Maria Paddick, Anna Longdon, Felicity Dewhurst, and Matthew Dewhurst conducted data collection. William Gray, Catherine Dotchin, Stella-Maria Paddick, and Richard Walker conducted data analysis and interpretation. Catherine Dotchin and William Gray wrote the first draft of the paper. All authors reviewed and critically appraised the final paper prior to submission.
Acknowledgments
The Hai Dementia Prevalence Study was part funded by a British Geriatric Society SpR start up grant and an Academy of Medical Sciences (UK) Clinical Lecturer start up grant. The PD prevalence study was funded by the UK Parkinson's Disease Society. Everyone who contributed to the writing of this paper is listed as an author. We thank all patients and their carers who took part in the study and all the members of staff in Hai who assisted in the project.
