PropertyValue
is nif:broaderContext of
nif:broaderContext
is schema:hasPart of
schema:isPartOf
nif:isString
  • We carried out a cross-sectional study of 75 patients diagnosed with dementia and their caregivers. Participants were recruited over a one year period through convenience sampling from a memory outpatient clinic and from Alzheimer’s Association support groups. A diagnosis of dementia was confirmed by clinical assessment according to Diagnostic and Statistical Manual of Mental Disorders, 4th Edition, Text Revision (DSM IV TR) clinical diagnostic criteria [31], including administration of the Mini-Mental State Examination (MMSE) [32] and MRI or CT neuroimaging. Informed consent was obtained from all caregivers for provision of their own information and informed consent by proxy was obtained for the investigation of patient characteristics. The study was approved by the Ethics Committee of the Faculty of Medical Sciences, The University of the West Indies, St. Augustine. Demographic characteristics of patients were recorded including age, gender, ethnicity, and marital status. Duration of dementia was determined from medical records, and MMSE score at assessment was considered as a covariate. Caregiver characteristics ascertained were age, gender, relationship to patient, cohabiting status, marital status, level of education, and occupation status. The Zarit Burden Interview (ZBI) [5], used to establish the degree of burden was self -administered for the most part, and caregivers were assisted only if they wanted clarification on the questions. This interview consisted of a 22-item questionnaire with a five-item response set ranging from “never” to “nearly always” graded on a scale from 0 to 4, according to the presence and intensity of an affirmative response. Its questions referred to the caregiver/patient relationship and evaluated the caregiver's health condition, psychological well-being, finances, and social life. Although ZBI categories are often imposed to categorize levels of burden [5], mean values were used in this analysis. In addition, the 28-item General Health Questionnaire (GHQ-28) was also completed by all caregivers [33] and was categorized into a binary variable according to whether scores were below (<47) or above (47+) the sample median. This approach was carried out for pragmatic reasons, in order to define a group at risk of a mental disorder and to compare associations with different elements of caregiver burden. This followed a previously recommended approach for populations in which GHQ screening properties are not known [34]. Stata and SPSS software were used. Having described the sample and caregiver characteristics, mean (SD) ZBI scores were compared between grouped characteristics. The median age was used a priori as a cut-off, as the sample size was too small for defining subgroups. Linear regression models were constructed in Stata and used for testing significance; nonordered exposures with 3 or more groups were entered as fixed covariates applying likelihood ratio tests, and ordered exposures were tested as ordinal covariates on one degree of freedom. Associations between individual ZBI items and common mental disorder symptoms in caregivers were analyzed in logistic regression models, with above-median GHQ-28 score as a binary dependent variable, and individual ZBI item scores entered as 5-point ordinal independent variables. Odds ratios were calculated with 95% confidence intervals and strengths of associations for ZBI items were ranked according to the Nagelkerke R2 statistic derived from the logistic regression model. This R2 statistic is an approximation of the proportion of variance explained–i.e., ranging from 0–1 and with higher values reflecting exposures which have a stronger potential explanatory power. For illustrative purposes, we categorized strong associations on the basis of an R2 >0.20 and weak associations on the basis of an R2 <0.10, with moderate associations in between these limits.
rdf:type