Study Design
This study is a cross-sectional observational study. This study is part of a broader study to identify indicators that determine the effectiveness of home care for patients with chronic CVD, and to identify variables that determine effective support systems for their home carers. The study involved 350 patients with CVD. In order to define indicators specific to home care, 193 patients remained in the home environment under the care of primary care family nurses, while 157 patients went out to see their GP for follow-up visits. The study also included caregivers of patients under home care of primary care family nurses. The study involved 161 caregivers. This article presents a partial analysis of the results of this study on the level of needs and severity of burnout in home care for people with CVD.
Setting
The study was conducted in Polish CVD patients and their informal caregivers. These patients received home care from a family nurse working in basic health care in the Opolskie, Dolnośląskie, Mazowieckie, Lubelskie and Podlaskie provinces. Eight primary care institutions took part in the study. Patients and caregivers were encouraged to take part in the study by a family nurse during planned home visits. Respondents completed questionnaires in their home environment. Patients and caregivers were provided with one set of questionnaires each, and nurses completed an additional questionnaire concerning the patient (i.e. paired questionnaires with respect to the same patient). Data was collected from March 2016 to January 2017.
Participants
180 informal tutors (i.e. caregivers) were invited to participate in the survey, but the final sample of participants was determined on the basis of their temporal availability. Finally, 161 people took part in the survey. The criteria for inclusion in the study were as follows: 18 years of age, and taking care of a patient with chronic CVD under home care at least 12 months before the study. The exclusion criteria (disqualification as determined by the family nurse) were cognitive disorders and other intensified mental disorders, and/or other difficulties preventing active participation in the study (e.g., visual disorders, a patient not speaking Polish as the first language, etc.).
Instruments And Data Collection
In this study, 36 variables were analysed, ranging from sociodemographic variables (e.g., sex, age, marital status, education, place of residence, the degree of relation to the patient, employment, period of care given, etc.), as well as variables relating to the degree of burden experienced by home care providers.
Needs of informal home care providers of CVD patients were assessed with the CANSAS. The CANSAS consists of 24 questions that cover 22 different problem areas related to the somatically and chronically ill. Of note, the questionnaire is for individuals who do not display severe mental disorders. Participants rate the different needs as either unmet (coded 0) or met (coded 1). We computed the total number of needs (N) that participants reported as unmet and met. Of note, missing data (i.e., participant nonresponses) were omitted from analysis. In addition, we computed the Camberwell index, which was calculated according to the formula 1-M/N. The consistency for the CANSAS questionnaire (i.e., Cronbach’s alpha) is 0.82 [17].
The present study also administered the Maslach Burnout Inventory (MBI), which was developed by Maslach and Jackson in 1981. The MBI provide insight into three components of the burnout syndrome, which are divided into subscales, namely: (1) emotional exhaustion (EE), (2) depersonalization (DE), and (3) a decreased level of personal achievement (PA). The MBI questionnaire contains 22 test items that are divided into the three burnout subscales, listed above. Nine of the 22 items correspond to the emotional exhaustion subscale, five items to the DE subscale, and eight items correspond to decreased PA. The test items in emotional exhaustion and depersonalization subscales are formulated negatively, whereas items from the decreased PA subscale are positive. Each question is structured affirmatively and relates to the attitude and feelings that a participant may experience and rates on a 7-point scale (where 0 means never, and 6 means every day). The results are calculated separately for each subscale. A high level of burnout is characterized by high scores on the EE and DE subscales and low scores on the decreased level of PA subscale [18]. The MBI questionnaire is frequently applied in research on occupational burnout. According to Maslach and Jackson, the MBI can be used in professions that require contacting other people. It was, however, indicated, that after the adaptation, the questionnaire might be applied more broadly by changing some words, but the meaning of the scale remains the same [19].
To assess the sociodemographic aspects of the care providers, an interview questionnaire was administered that assessed sex, age, marital status, education, place of residence, degree of relation to the patient, employment, and the period of care provided.
Study Size
161 informal care providers accepted the invitation to participate and ultimately completed the study. Of note, there were missing data from care providers for the following variables: age, marital status, employment, and period of caretaking. Thus, the numbers provided in the columns do not sum up to 161. Of note, for three cases, a non-family caregiver was both a neighbour and an informal partner.
Ethics Approval And Consent To Participate
The study was approved by the Bioethical Commission at Medical University in Wroclaw (No KB -86/2016). Participation in the study was voluntary and anonymous. All participants were informed about the study aims, methods, and the ability to withdrawal participation at any stage of the examination.
Statistical Methods
The results of the study were subject to statistical analysis with the use of R statistical package (version 3.4.0).
For the quantitative variables, the arithmetic mean, standard deviation, first quartile (Q.25%), second quartile (Q.50%) of the median, third quartile (Q.75%), minimum, and maximum were calculated. For qualitative variables, frequency (i.e., percent) was determined. The Shapiro-Wilk test showed that only two qualitative variables (i.e., age, decreased level of PA on the MBI) showed a normal distribution. The other included variables diverged from the normal distribution. Chi-square test was used to assess the qualitative variables.
The relationship between sociodemographic variables and the assessment of needs and severity of burnout was analysed using Spearman’s rank correlation coefficient, which does not require normal distribution of the variables. The null hypothesis (H0) is tested wherein the Spearman’s rank correlation coefficient equals 0. The alternative hypothesis that the correlation coefficient differs from 0. The null hypothesis (H0) was rejected if the p-value was < 0.05 (α = 0.05).
Logistic regression analysis was used to describe the relationship between demographic data of care providers and their needs and severity of burnout, via an odds ratio. Separate regression analyses were carried out for each of the following dichotomous outcome variables:
- unmet needs: 0, if median scores on the Camberwell ≤ 0.875; 1, if Camberwell > 875,
- EE: 0, if median scores on the MBI-EE ≤ 20; 1 if MBI-EE > 20,
- DE: 0, if median scores on the MBI-D ≤ 6; 1, if MBI-D > 6,
- decreased level of PA: 0, if MBI-PA ≤ 29 (median); 1 if MBI-PA > 29.
The variables used in logistic regression analysis models were selected from the following 29 variables (Table 1):
Table 1
Descriptive variables for logistic regression analysis models
| Variables | Coding |
x1 | Gender | 1 – Women 2 – Men |
x2 | Age (in years) | Number of years |
x3 | Marital status | 1 – Single 2 – Married 3 – Widowed 4 – Divorced |
x4 | Place of residence | 1 – Big city < 100 thousand inhabitants 2 – Middle town from 20–100 thousand inhabitants 3 – Town small > 20 thousand inhabitants 4 – Village |
x5 | Education | 1 – Primary 2 – Vocational 3 – Secondary without Matura Exam 4 – Secondary with Matura Exam 5 – Post-secondary 6 – BA 7 – MA |
x6 | Family-related carer | 1 – Wife/husband 2 – Brother/sister 3 – Mother/father 4 – Uncle/aunt 5 – Cousin 6 – Other |
x7 | Non-family carer | 1 – Neighbour 2 – Informal partner 3 – Other |
x8 | Employment | 1 – Full time 2 – Part-time 3 – Sick leave-child care 4 – Sick leave 5 – Unemployment benefit 6 – Unemployment |
x9 | Period of homecare | Number of years |
Next, for each explained variable, a separate logistic regression analysis, for at least 9 different explanatory variables, was carried out to examine all possible models. For further analysis, only specific models that demonstrated significance were chosen. All variables in the model had to be statistically significant and included the largest number of explanatory variables, in the smallest number of models. Using the models selected, the odds ratio for the events examined were calculated and conclusions formulated on their basis. This approach did not require the use of model-matching procedures. The significance level was established at 0.05.