J Clin Psychol Med Settings DOI 10.1007/s10880-014-9399-y
Caregiving for Patients in Vegetative and Minimally Conscious States: Perceived Burden as a Mediator in Caregivers’ Expression of Needs and Symptoms of Depression and Anxiety M. Pagani • A. M. Giovannetti • V. Covelli D. Sattin • M. Leonardi
•
Springer Science+Business Media New York 2014
Abstract Involvement of family members is crucial to provide daily informal caring to patients in vegetative state and minimally conscious state. Previous studies showed that perceived burden is a risk factor for informal caregivers as it increases psychophysical distress. This research further investigated the relationship between these factors and aimed at providing a model that thoroughly describes this mechanism of functioning. In the frame of a national survey on people with disorders of consciousness, 487 informal caregivers of children and adult patients in vegetative and minimally conscious state were administered measures of depression, anxiety, caregiver needs, and family strain. Regression models proposed by Baron and Kenny and the Sobel test were adopted to investigate the relationship between depressive and anxiety symptoms, perceived burden and needs expressed. Our study shows that the relation between those symptoms and needs is mediated by burden, where higher burden accentuates and lower burden mitigates the needs expressed by caregivers. Our findings demonstrate that psychosocial components of the burden perceived by caregivers of patients with disorders of consciousness play a key role in shaping those caregivers’ needs, especially their needs for information and communication. We recommend implementation of comprehensive steps to meet the needs of these caregivers, steps that incorporate improved economic and public health programs, social support, and use of psychological
M. Pagani (&) A. M. Giovannetti V. Covelli D. Sattin M. Leonardi Neurology, Public Health and Disability Unit, Scientific Directorate, Neurological Institute Carlo Besta IRCCS Foundation, Via Celoria 11, 20133 Milan, Italy e-mail: [email protected]
interventions to ameliorate caregivers’ psychological distress and decrease their burden. Keywords Informal caregivers Vegetative state Minimally conscious state Burden Needs Depression Anxiety
Introduction Vegetative state (VS) and minimally conscious state (MCS) are clinical conditions grouped as disorders of consciousness (DOC) that represents possible outcomes following both traumatic and non-traumatic acquired severe brain injury. Despite clinical differences between diagnosis (Royal College of Physicians, 2003; Giacino et al., 2002) and in functioning and disability (Leonardi, Sattin et al., 2012), both patients in VS and MCS show a low to none interaction with the environment and necessitate prolonged assistance from health care professionals and informal caregivers. Recent progresses in medicine and intensive care techniques, including cardiopulmonary resuscitation (Daubin et al., 2011) and decompressive craniectomy (Cooper et al., 2011), has led to a gradual increase of patients that survive acquired severe brain injury (Pagani, Leonardi, Covelli, Giovannetti, & Sattin, 2014). In line with data in other Western countries, Italian prevalence and incidence of DOC has increased to 10/100,000 and 4/100,000 inhabitants, respectively (Zampolini, 2003). Higher numbers of patients live longer and family caregivers dedicate a larger amount of time for their relatives in VS or MCS (Leonardi, Giovannetti, et al. 2012). Tireless involvement and unpaid nature of informal caregiving have been acknowledged as important in the pathway of care, and informal assistance provided by family
123
J Clin Psychol Med Settings
members of patients with DOC has been slowly gaining attention from governments, policy makers and health professionals. Patients with DOC require continuous care and relatives’ participation is crucial to plan the process of caring efficiently, especially when patients are discharged from hospitals and become outpatients at home or in long term care (Lippert-Gru¨ner, Wedekind, & Klug, 2003; Pagani, Giovannetti, et al. 2013; Pagani, Covelli, et al. 2013). Previous studies on caregivers of patients with DOC reported that prolonged assistance, not only when patients are at home but also when patients are hosted in long term care institutions, may be perceived as stressful and overwhelming, and may contribute to the increase of physical, psychosocial and financial distress as well as overall perceived burden (Giovannetti, Leonardi, et al. 2012; Giovannetti, Pagani, et al. 2012). As for many other chronic diseases, having a family member with a prolonged disabling disorder such as DOC has several financial implications reflected in direct and indirect costs that increase the overall caregivers’ disease burden, as reported in a recent study by Gustavsson et al. (2011). In addition, as a reaction to the responsibilities related to the continuous care of patients with DOC, caregivers experienced severe emotional burden and reported both poor physical and mental health, expressed needs for social involvement and high level of depressive and anxiety symptoms (Leonardi, Giovannetti, et al. 2012). In line with findings on adults, caregivers of children in VS and MCS also reported poor outcomes in terms of burden, mental health, anxiety and depressive symptoms regardless of children’s disability and level of functioning (Giovannetti, Leonardi, et al. 2012). Furthermore, as reported in a recent study, caregivers reported that is not always easy to understand their real needs from the health and social professionals that work with their impaired family members (Giovannetti, Pagani, et al. 2012). Caregivers of patients in VS and MCS experience high needs for information and social support. In fact they want to be informed by physicians and health professionals about what is done to their relative, as well as to cooperate and be involved in decisions that affect the patient, to effectively communicate with all the professionals of the medical team, to be able to easily contact the staff, to know what problems may possibly arise in the course of the disease, and to be enabled to perform all the task related to the care of their relative. Level of social support, financial resources and patient’s condition may play a crucial role in determining caregivers’ expression of needs. Subjective psychological characteristics of caregivers also impact their needs, e.g., their level of emotional and psychophysical distress, anxiety and depression, as well as their use of situation-focused coping strategies. The impact of such factors
123
is likely to continue over time long after the acute event (Chiambretto & Vanoli, 2006). In a biopsychosocial framework (Engel, 1977), family members are considered an environmental factor that interacts with a patient’s health condition. In line with this approach caregivers’ daily efforts and informal care act as a facilitator, i.e., as a factor that improves the patient’s functioning and decreases disability. As the health of caregivers, themselves, is affected by caregiving, this study focuses on the relationship between patients’ need for care and the caregivers’ perceived burden, needs and emotions. The complex two-way interaction between patients and caregivers is investigated within the framework of the biopsychosocial model. Therefore, investigating caregivers burden (Leonardi, Giovannetti, et al. 2012) will improve understanding of caregivers’ needs, which in turn, suggests interventions that can reduce caregiver burden and promote better caregiver health, which in turn, strengthens caregivers’ ability to act as facilitators for their disabled family member. The relationship between caregivers’ psychological factors, expressed needs and perceived psychosocial burden has been investigated in other diseases. For instance, higher prevalence of anxiety and depression symptoms in primary caregivers of children and adolescents with myelomeningocele was found to be related to higher burden and impaired quality of life for caregivers (Valenc¸a, de Menezes, Calado, & de Aguiar Cavalcanti, 2012). Caregivers’ burden in the immediate aftermath of stroke was mainly determined by patient and caregivers’ anxiety rather than the level of disability, age, gender, or support from family or social services (McCullagh, Brigstocke, Donaldson, & Kalra 2005). Caregivers of terminally ill patients who expressed needs for social assistance complained about severe health problems and overall burden (Bialon & Coke, 2012). A study on needs of caregivers of patients with severe brain injury (McPherson, McNaughton, & Pentland, 2000) reported that they demanded more information immediately after the medical consultation, and the need for information was unrelated to either the severity of injury or level of functional deficit. Additionally, level of depression was significantly correlated with unmet needs in areas such as activities of daily living, instrumental activities of daily living, and transportation. Previous findings of family needs after a person’s traumatic brain injury also reported an association between informal caregivers’ needs and diminished quality of caregivers’ life over time. Family support is also highlighted as a direct predictor of emotional well-being in the long-term in caregivers of patients with traumatic brain injury (Kendall & Terry, 2009). There is extensive theoretical and empirical literature describing needs, burden, anxiety and depressive symptoms of caregivers. Little attention has been given to informal caregivers of patients in VS and MCS. The
J Clin Psychol Med Settings
understanding of the process underlying the relationship between perceived burden, depressive and anxiety symptoms, as well as expressed needs, is essential to the development of evidence-based methods to provide clinicians and service providers with useful insights on factors related to caregivers’ needs. The present paper seeks to achieve a more comprehensive understanding of mechanisms by which DOC patients’ informal caregivers’ ways of expressing their own needs is influenced by their perceived burden and symptoms of depression and anxiety.
Materials and Methods This was an observational cross-sectional multi-centre study, conducted in 78 Italian centres located in northern, central, or southern Italy. The centres specialized either in post-acute treatment, or in long-term assistance and care of patients with disorders of consciousness. Sixty-nine out of the 78 centres collected data, and the project was coordinated by the Neurological Institute ‘‘C. Besta’’ Foundation. The final aim of this project was to highlight crucial issues related to the complex situation of patients and caregivers for development of better care management strategies and inclusive health care programmes (Sattin, Pagani, Covelli, Giovannetti, & Leonardi, 2013). Neurological Institute ‘‘C. Besta’’ Foundation is a regional and national centre of excellence for the treatment of neurological diseases. The fundamental aims of the Foundation are to provide high quality, ethical and efficient care and treatment to its patients, making use of the result of its own research and that of others, and always putting the interests of patients first. In fulfilment of these aims, the Foundation provides care and treatments to patients, carries out clinical and basic research and promotes research in the area of public health. The group of centres that provide services to DOC patients was built using formal contacts with the Ministry of Health and informal contacts with the association of patients’ family members. All known and recognized centers were invited to participate in the study. The Ministry of Health supervised every phase of the project, including: analyzing the reports of the study; coordination of staff; and organizing two meetings during the project with associations of DOC patients’ family members that focused on ethical and psychological implications of the project. Only the main caregivers were recruited as participants, i.e., the person who was mainly involved in terms of time spent with the patient for informal caring, who felt responsible for the patient, and who most frequently interacted with doctors and clinical staff. Main caregivers were contacted and held face to face meetings with psychologists, or other members of the project team, at the centre that provided services to the patient. Participation in
the study was voluntary. The protocol consisted of administering a battery of self-report questionnaires. Sixhundred caregivers were contacted of which 487 (81.2 %) agreed to participate in the study. Standardized training on the assessment protocol was provided to the health professionals who administered the protocol at various centres. The training included a brief introduction on instrumentation, research goals, as well as instruction on data storage and sharing with the coordinating institution. Research data were stored in a database. Data quality control was performed after the enrolment of the first 60 cases: data were checked by the coordinating institute, and a report with corrective recommendations was sent to each centre. Common errors were the absence of demographic information (e.g. age, sex, marital status, hours spend at the bedside). The coordinating centre recommended that steps be taken to complete the missing values, and in the final analysis, the percentage of missing values was very low. Informed consent was signed and confidentiality was preserved by recording the data in de-identified form. It was hoped this procedure would encourage participants to express honest opinions and feelings. The study was approved by the ethical committee of the coordinating centre and the approval was ratified by the participating institutions. Assessment Protocol Neurologists and neuropsychologists with experience in DOC administered to caregivers the following self-report tests: the beck depression inventory (BDI-II) (Ghisi, Flebus, Montano, Sanavio, & Sica, 2006), the Spielberger state trait anxiety inventory-Y (STAI-Y), the caregiver needs assessment (CNA) (Moroni, Sguazzin, & Filipponi, 2008), the family strain questionnaire (FSQ) (Rossi Ferrario, Bacchetta, Omarini & Zotti, 1998) and a socio-demographic questionnaire. The assessment lasted approximately 30–45 min. BDI-II is a 21-item self-report inventory that evaluates presence and severity of depressive symptoms. It measures cognitive manifestations such as pessimism, guilt, selfcriticism and self-esteem and somatic-affective symptoms, including loss of interest, loss of energy, changes in appetite and sleep, agitation and crying. An overall index, ranging from 0 to 63, is provided and higher scores indicate more severe symptoms. Two typical sample item of BDI-II assess ‘‘sadness’’ and ‘‘nervousness’’. Scores are also grouped in categories of minimal, mild, moderate and severe symptoms. STAI-Y is an inventory that measures feelings of apprehension, tension, nervousness and worry. Two main factors are independently calculated to differentiate between the temporary condition of anxiety or ‘‘State Anxiety’’, and the more general and longstanding property
123
J Clin Psychol Med Settings
of anxiety or ‘‘Trait Anxiety’’. STAI-Y is composed of 40 items ranging from 1 to 4 and both ‘‘State Anxiety’’ and ‘‘Trait Anxiety’’ scores range from 20 to 80. Higher scores suggest higher level of anxiety. Only ‘‘State Anxiety’’ was used in this study and two typical sample items are ‘‘I feel good’’ and ‘‘I feel calm’’. The CNA is a 17-item questionnaire that assesses needs expressed by caregivers of chronic and severely impaired patients, including social, relational, emotional, physical, cognitive, behavioural, and spiritual needs. A sample item for each type of caregiver need is as follows: for social needs, ‘‘I need to easily contact the medical staff of my relative;’’ for relational needs, ‘‘I need to exchange views with other caregivers of patients with the same clinical condition;’’ for emotional needs, ‘‘I need to be reassured;’’ for physical needs, ‘‘I need to know what is done and will be done to treat my relative;’’ for cognitive needs, ‘‘I need to collaborate with and be involved in the decisions related to my relative;’’ for behavioural needs, ‘‘I need to learn how to cope with behavioural problems of my relative;’’ and for spiritual needs, ‘‘I need spiritual support.’’. Scores range from 0 to 51 and higher values indicate stronger intensity of expressed needs. The FSQ is a semi-structured interview of 30 dichotomous items that investigates psychological problems of caregivers, including emotional burden, problems in social involvement, need for knowledge about the disease, satisfaction with family relationships and thoughts about death. Typical sample items for each content area are as follows: for emotional burden, ‘‘I suffer from insomnia, digestive problems, headache and tiredness’’ and ‘‘I am pessimistic about the future;’’ for problems in social involvement, ‘‘I gave up my personal interests’’ and ‘‘The clinical condition of my relative makes me tired;’’ for need for knowledge about the disease, ‘‘I feel I need suggestions on how to assist my relative’’ and ‘‘I need to talk with an expert (e.g., doctor, clinical psychologist);’’ for satisfaction with family relationships, ‘‘In case there is a need, I know that I can rely on somebody else in the family’’ and ‘‘In this moment, I feel very close to relatives and friends;’’ for thoughts about death, ‘‘I feel helpless about the disease’’ and ‘‘I sometimes think of the death of my relative.’’. An index of overall burden is provided, with higher score indicating higher presence of perceived burden. Statistical Analyses Prior to analyses, data were screened for univariate and multivariate outliers, non-normal distribution responses and missing values following the procedures outlined by Tabachnick and Fidell (2007). Frequencies and percentages were used to report descriptive statistics of categorical variables. Cronbach’s alpha for each test was calculated as
123
a measure of internal consistency and alpha values higher than 0.80 indicated a good consistency. Regression models were adopted to investigate two alternative models of the combined influence of depressive symptoms and perceived burden as predictors of DOC caregivers expression of needs. The models consider two different ways to chain the three variables together. As shown in Table 3, Model 1 examines a depression–burden– needs chain, i.e., the model treats perceived burden as a mediating variable that influences the relationship between caregivers’ depressive symptoms and their expression of needs. Model 2 reverses the order of the two predictors in the chain and examines a burden–depression–needs chain, i.e., the model treats depressive symptoms as a mediating variable that influences the relationship between caregivers’ perceived burden and their expression of needs. Age, gender, disease duration and STAI-Y Trait Anxiety scores were entered as control variables in both Model 1 and Model 2. Additional regression models were used to investigate two further models that include state anxiety as a variable that influences DOC caregivers expression of needs. As shown in Table 4, Model 3 examines an anxiety–burden– needs chain, i.e., the model treats perceived burden as a mediating variable that influences the relationship between caregivers’ anxiety and their expression of needs. Model 4 reverses the role of the two predictors in the chain and examines a burden–anxiety–needs chain, i.e., the model treats state anxiety as a mediating variable that influences the relationship between caregivers’ perceived burden and their expression of needs. Age, gender, disease duration and BDI-II scores were entered as control variables in both Model 3 and Model 4. For a variable to be considered a mediator four criteria need to be met, as proposed by Baron & Kenny (1986): (1) Independent variable should significantly predict mediator (path a); (2) mediator should significantly predict dependent variable (path b); (3) independent variable should predict dependent variable (path c); and (4) the relationship between dependent and independent variable should be reduced or no longer significant in the presence of mediator (path c’). In order to corroborate this effect, the significance of each mediation model was assessed by testing the joint significance of the independent variable to the mediator (path a) and of the mediator to the dependent variable (path b), following the suggestion of MacKinnon et al. (2002). The Sobel test, which yields a z-value, was used to test the significance of the joint effect of the a and b paths. All path coefficients represents standardised betas (regression coefficients). Caregivers’ gender and age as well as patient’s disease duration were entered as covariates in all regression models. Significance level a = 0.05 was adopted, and all statistical analyses were performed using SPSS v. 18.0.
J Clin Psychol Med Settings Table 1 Patient-related socio-demographic characteristics Patient diagnosis—n (%) SV
272 (68.5)
SMC
125 (31.5)
Patient’s disease duration—mean (SD), in years
3.4 (3.3)
Patient’s place of residence—n (%) With family
51 (12.8)
Post-acute rehabilitation centre
233 (58.7)
Long term care centre
113 (28.5)
Relationship to the patient—n (%) Partner
154 (38.8)
Relative
84 (21.2)
Son/Daughter
72 (18.1)
Grandparent, grandson/daughter, mother-in-law, sisterin-law and others
87 (21.9)
Table 2 Caregiver’s scores obtained at protocol assessment BDI-II score—mean (SD)
15.4 (10.1)
Minimal
165 (41.6)
Mild Moderate
59 (14.9) 32 (8.1)
Severe
141 (35.5)
STAI-Y trait anxiety score—mean (SD)
46.1 (10.4)
FSQ perceived burden score—mean (SD)
13.2 (6.1)
CAN needs expressed score—mean (SD)
33.8 (9.57)
Results A total of 487 caregivers of children and adult patients in VS or MCS were invited to participate, and 397 (81.5 %) questionnaires were completed and considered for analyses. Participants were mainly females (n = 266; 67 %), married (n = 301; 75.8 %), with a mean age of 51.5 years (SD = 12.9) and 10.6 years (SD = 4.5) of education. Patient-related socio-demographic characteristics and scores obtained by caregivers at protocol assessment are displayed in Tables 1 and 2, respectively. Procedures for screening raw data for completeness prior to statistical analysis identified few missing values, less than 5 % for all the variables. Missing values where replaced by the median value of the variable in order not to inflate correlation between variables. FSQ score was positively skewed and so a square-root transformation was applied. No cases were found to be a univariate (z-score = ± 3 or
beyond) or multivariate outlier (Mahalanobis distance at p \ 0.01). Regarding Cronbach’s alpha, internal consistency was good for STAI-Y Anxiety (a = 0.87), CNA (a = 0.87), BDI-II (a = 0.89), and FSQ (a = 0.92). In Model 1, greater depressive symptoms were significantly associated with greater perceived burden (path a) which, in turn, was positively associated to needs expressed (path b). Greater depressive symptoms were also significantly associated to greater needs expressed (path c) and this correlation was no longer significant when perceived burden was entered as mediator variable (path c’ in Fig. 1). Tests of the indirect or joint effect of the paths a and b revealed that perceived burden was a significant intervening variable between depressive symptoms and needs expressed (Sobel test, z = 4.639, p \ .001), as shown in Table 3. In Model 2, paths a, b and c were all significant and path c’ was still significant when depressive symptoms were entered as mediator (p \ .001), and the Sobel test was far from being significant (z = .214, p = .830). In Model 3, similar to Model 1, greater anxiety symptoms were significantly associated with greater perceived burden (path a) which, in turn, was positively associated to needs expressed (path b). Greater anxiety symptoms were also significantly associated to greater needs expressed (path c) and this correlation was no longer significant when perceived burden was entered as mediator variable (path c’ in Fig. 2). The Sobel test showed that perceived burden was a significant intervening variable between anxiety symptoms and needs expressed (z = 5.232, p \ .001). In Model 4, similar to Model 2, paths a, b and c were all significant and path c’ was still significant when anxiety symptoms were entered as mediator (p \ .001), and the Sobel test was far from being significant (z = .385, p = .700), as shown in Table 4. Although the relationship between variables in Tables 3 and 4 are referred to as ‘‘paths,’’ they are not the same as ‘‘paths’’ calculated in Path Analysis. In the current paper, the term ‘‘path’’ is used to refer to a regression coefficient that quantifies the association between two variables. The same holds true in Figs. 1 and 2, i.e. the arrow vectors shown in those Figures are referred to as ‘‘paths,’’ but they should not be confused with ‘‘paths’’ as calculated in Path Analysis. They are regression coefficients.
Discussion To the best of our knowledge, this is the first study that shows the important role of perceived psychosocial burden in determining the effects of depression and anxiety on DOC caregivers expression of needs. Our study shows that the relation between these symptoms and caregivers’ needs is mediated by burden, where higher burden accentuates and lower burden mitigates the needs expressed by
123
J Clin Psychol Med Settings Fig. 1 Depression-burdenneeds chain. This model treats perceived burden as a mediating variable that influences the relationship between caregivers’ depressive symptoms and their expression of needs
Perceived burden (FSQ) Path a .333 (p
Caregiving for Patients in Vegetative and Minimally Conscious States: Perceived Burden as a Mediator in Caregivers’ Expression of Needs and Symptoms of Depression and Anxiety M. Pagani • A. M. Giovannetti • V. Covelli D. Sattin • M. Leonardi
•
Springer Science+Business Media New York 2014
Abstract Involvement of family members is crucial to provide daily informal caring to patients in vegetative state and minimally conscious state. Previous studies showed that perceived burden is a risk factor for informal caregivers as it increases psychophysical distress. This research further investigated the relationship between these factors and aimed at providing a model that thoroughly describes this mechanism of functioning. In the frame of a national survey on people with disorders of consciousness, 487 informal caregivers of children and adult patients in vegetative and minimally conscious state were administered measures of depression, anxiety, caregiver needs, and family strain. Regression models proposed by Baron and Kenny and the Sobel test were adopted to investigate the relationship between depressive and anxiety symptoms, perceived burden and needs expressed. Our study shows that the relation between those symptoms and needs is mediated by burden, where higher burden accentuates and lower burden mitigates the needs expressed by caregivers. Our findings demonstrate that psychosocial components of the burden perceived by caregivers of patients with disorders of consciousness play a key role in shaping those caregivers’ needs, especially their needs for information and communication. We recommend implementation of comprehensive steps to meet the needs of these caregivers, steps that incorporate improved economic and public health programs, social support, and use of psychological
M. Pagani (&) A. M. Giovannetti V. Covelli D. Sattin M. Leonardi Neurology, Public Health and Disability Unit, Scientific Directorate, Neurological Institute Carlo Besta IRCCS Foundation, Via Celoria 11, 20133 Milan, Italy e-mail: [email protected]
interventions to ameliorate caregivers’ psychological distress and decrease their burden. Keywords Informal caregivers Vegetative state Minimally conscious state Burden Needs Depression Anxiety
Introduction Vegetative state (VS) and minimally conscious state (MCS) are clinical conditions grouped as disorders of consciousness (DOC) that represents possible outcomes following both traumatic and non-traumatic acquired severe brain injury. Despite clinical differences between diagnosis (Royal College of Physicians, 2003; Giacino et al., 2002) and in functioning and disability (Leonardi, Sattin et al., 2012), both patients in VS and MCS show a low to none interaction with the environment and necessitate prolonged assistance from health care professionals and informal caregivers. Recent progresses in medicine and intensive care techniques, including cardiopulmonary resuscitation (Daubin et al., 2011) and decompressive craniectomy (Cooper et al., 2011), has led to a gradual increase of patients that survive acquired severe brain injury (Pagani, Leonardi, Covelli, Giovannetti, & Sattin, 2014). In line with data in other Western countries, Italian prevalence and incidence of DOC has increased to 10/100,000 and 4/100,000 inhabitants, respectively (Zampolini, 2003). Higher numbers of patients live longer and family caregivers dedicate a larger amount of time for their relatives in VS or MCS (Leonardi, Giovannetti, et al. 2012). Tireless involvement and unpaid nature of informal caregiving have been acknowledged as important in the pathway of care, and informal assistance provided by family
123
J Clin Psychol Med Settings
members of patients with DOC has been slowly gaining attention from governments, policy makers and health professionals. Patients with DOC require continuous care and relatives’ participation is crucial to plan the process of caring efficiently, especially when patients are discharged from hospitals and become outpatients at home or in long term care (Lippert-Gru¨ner, Wedekind, & Klug, 2003; Pagani, Giovannetti, et al. 2013; Pagani, Covelli, et al. 2013). Previous studies on caregivers of patients with DOC reported that prolonged assistance, not only when patients are at home but also when patients are hosted in long term care institutions, may be perceived as stressful and overwhelming, and may contribute to the increase of physical, psychosocial and financial distress as well as overall perceived burden (Giovannetti, Leonardi, et al. 2012; Giovannetti, Pagani, et al. 2012). As for many other chronic diseases, having a family member with a prolonged disabling disorder such as DOC has several financial implications reflected in direct and indirect costs that increase the overall caregivers’ disease burden, as reported in a recent study by Gustavsson et al. (2011). In addition, as a reaction to the responsibilities related to the continuous care of patients with DOC, caregivers experienced severe emotional burden and reported both poor physical and mental health, expressed needs for social involvement and high level of depressive and anxiety symptoms (Leonardi, Giovannetti, et al. 2012). In line with findings on adults, caregivers of children in VS and MCS also reported poor outcomes in terms of burden, mental health, anxiety and depressive symptoms regardless of children’s disability and level of functioning (Giovannetti, Leonardi, et al. 2012). Furthermore, as reported in a recent study, caregivers reported that is not always easy to understand their real needs from the health and social professionals that work with their impaired family members (Giovannetti, Pagani, et al. 2012). Caregivers of patients in VS and MCS experience high needs for information and social support. In fact they want to be informed by physicians and health professionals about what is done to their relative, as well as to cooperate and be involved in decisions that affect the patient, to effectively communicate with all the professionals of the medical team, to be able to easily contact the staff, to know what problems may possibly arise in the course of the disease, and to be enabled to perform all the task related to the care of their relative. Level of social support, financial resources and patient’s condition may play a crucial role in determining caregivers’ expression of needs. Subjective psychological characteristics of caregivers also impact their needs, e.g., their level of emotional and psychophysical distress, anxiety and depression, as well as their use of situation-focused coping strategies. The impact of such factors
123
is likely to continue over time long after the acute event (Chiambretto & Vanoli, 2006). In a biopsychosocial framework (Engel, 1977), family members are considered an environmental factor that interacts with a patient’s health condition. In line with this approach caregivers’ daily efforts and informal care act as a facilitator, i.e., as a factor that improves the patient’s functioning and decreases disability. As the health of caregivers, themselves, is affected by caregiving, this study focuses on the relationship between patients’ need for care and the caregivers’ perceived burden, needs and emotions. The complex two-way interaction between patients and caregivers is investigated within the framework of the biopsychosocial model. Therefore, investigating caregivers burden (Leonardi, Giovannetti, et al. 2012) will improve understanding of caregivers’ needs, which in turn, suggests interventions that can reduce caregiver burden and promote better caregiver health, which in turn, strengthens caregivers’ ability to act as facilitators for their disabled family member. The relationship between caregivers’ psychological factors, expressed needs and perceived psychosocial burden has been investigated in other diseases. For instance, higher prevalence of anxiety and depression symptoms in primary caregivers of children and adolescents with myelomeningocele was found to be related to higher burden and impaired quality of life for caregivers (Valenc¸a, de Menezes, Calado, & de Aguiar Cavalcanti, 2012). Caregivers’ burden in the immediate aftermath of stroke was mainly determined by patient and caregivers’ anxiety rather than the level of disability, age, gender, or support from family or social services (McCullagh, Brigstocke, Donaldson, & Kalra 2005). Caregivers of terminally ill patients who expressed needs for social assistance complained about severe health problems and overall burden (Bialon & Coke, 2012). A study on needs of caregivers of patients with severe brain injury (McPherson, McNaughton, & Pentland, 2000) reported that they demanded more information immediately after the medical consultation, and the need for information was unrelated to either the severity of injury or level of functional deficit. Additionally, level of depression was significantly correlated with unmet needs in areas such as activities of daily living, instrumental activities of daily living, and transportation. Previous findings of family needs after a person’s traumatic brain injury also reported an association between informal caregivers’ needs and diminished quality of caregivers’ life over time. Family support is also highlighted as a direct predictor of emotional well-being in the long-term in caregivers of patients with traumatic brain injury (Kendall & Terry, 2009). There is extensive theoretical and empirical literature describing needs, burden, anxiety and depressive symptoms of caregivers. Little attention has been given to informal caregivers of patients in VS and MCS. The
J Clin Psychol Med Settings
understanding of the process underlying the relationship between perceived burden, depressive and anxiety symptoms, as well as expressed needs, is essential to the development of evidence-based methods to provide clinicians and service providers with useful insights on factors related to caregivers’ needs. The present paper seeks to achieve a more comprehensive understanding of mechanisms by which DOC patients’ informal caregivers’ ways of expressing their own needs is influenced by their perceived burden and symptoms of depression and anxiety.
Materials and Methods This was an observational cross-sectional multi-centre study, conducted in 78 Italian centres located in northern, central, or southern Italy. The centres specialized either in post-acute treatment, or in long-term assistance and care of patients with disorders of consciousness. Sixty-nine out of the 78 centres collected data, and the project was coordinated by the Neurological Institute ‘‘C. Besta’’ Foundation. The final aim of this project was to highlight crucial issues related to the complex situation of patients and caregivers for development of better care management strategies and inclusive health care programmes (Sattin, Pagani, Covelli, Giovannetti, & Leonardi, 2013). Neurological Institute ‘‘C. Besta’’ Foundation is a regional and national centre of excellence for the treatment of neurological diseases. The fundamental aims of the Foundation are to provide high quality, ethical and efficient care and treatment to its patients, making use of the result of its own research and that of others, and always putting the interests of patients first. In fulfilment of these aims, the Foundation provides care and treatments to patients, carries out clinical and basic research and promotes research in the area of public health. The group of centres that provide services to DOC patients was built using formal contacts with the Ministry of Health and informal contacts with the association of patients’ family members. All known and recognized centers were invited to participate in the study. The Ministry of Health supervised every phase of the project, including: analyzing the reports of the study; coordination of staff; and organizing two meetings during the project with associations of DOC patients’ family members that focused on ethical and psychological implications of the project. Only the main caregivers were recruited as participants, i.e., the person who was mainly involved in terms of time spent with the patient for informal caring, who felt responsible for the patient, and who most frequently interacted with doctors and clinical staff. Main caregivers were contacted and held face to face meetings with psychologists, or other members of the project team, at the centre that provided services to the patient. Participation in
the study was voluntary. The protocol consisted of administering a battery of self-report questionnaires. Sixhundred caregivers were contacted of which 487 (81.2 %) agreed to participate in the study. Standardized training on the assessment protocol was provided to the health professionals who administered the protocol at various centres. The training included a brief introduction on instrumentation, research goals, as well as instruction on data storage and sharing with the coordinating institution. Research data were stored in a database. Data quality control was performed after the enrolment of the first 60 cases: data were checked by the coordinating institute, and a report with corrective recommendations was sent to each centre. Common errors were the absence of demographic information (e.g. age, sex, marital status, hours spend at the bedside). The coordinating centre recommended that steps be taken to complete the missing values, and in the final analysis, the percentage of missing values was very low. Informed consent was signed and confidentiality was preserved by recording the data in de-identified form. It was hoped this procedure would encourage participants to express honest opinions and feelings. The study was approved by the ethical committee of the coordinating centre and the approval was ratified by the participating institutions. Assessment Protocol Neurologists and neuropsychologists with experience in DOC administered to caregivers the following self-report tests: the beck depression inventory (BDI-II) (Ghisi, Flebus, Montano, Sanavio, & Sica, 2006), the Spielberger state trait anxiety inventory-Y (STAI-Y), the caregiver needs assessment (CNA) (Moroni, Sguazzin, & Filipponi, 2008), the family strain questionnaire (FSQ) (Rossi Ferrario, Bacchetta, Omarini & Zotti, 1998) and a socio-demographic questionnaire. The assessment lasted approximately 30–45 min. BDI-II is a 21-item self-report inventory that evaluates presence and severity of depressive symptoms. It measures cognitive manifestations such as pessimism, guilt, selfcriticism and self-esteem and somatic-affective symptoms, including loss of interest, loss of energy, changes in appetite and sleep, agitation and crying. An overall index, ranging from 0 to 63, is provided and higher scores indicate more severe symptoms. Two typical sample item of BDI-II assess ‘‘sadness’’ and ‘‘nervousness’’. Scores are also grouped in categories of minimal, mild, moderate and severe symptoms. STAI-Y is an inventory that measures feelings of apprehension, tension, nervousness and worry. Two main factors are independently calculated to differentiate between the temporary condition of anxiety or ‘‘State Anxiety’’, and the more general and longstanding property
123
J Clin Psychol Med Settings
of anxiety or ‘‘Trait Anxiety’’. STAI-Y is composed of 40 items ranging from 1 to 4 and both ‘‘State Anxiety’’ and ‘‘Trait Anxiety’’ scores range from 20 to 80. Higher scores suggest higher level of anxiety. Only ‘‘State Anxiety’’ was used in this study and two typical sample items are ‘‘I feel good’’ and ‘‘I feel calm’’. The CNA is a 17-item questionnaire that assesses needs expressed by caregivers of chronic and severely impaired patients, including social, relational, emotional, physical, cognitive, behavioural, and spiritual needs. A sample item for each type of caregiver need is as follows: for social needs, ‘‘I need to easily contact the medical staff of my relative;’’ for relational needs, ‘‘I need to exchange views with other caregivers of patients with the same clinical condition;’’ for emotional needs, ‘‘I need to be reassured;’’ for physical needs, ‘‘I need to know what is done and will be done to treat my relative;’’ for cognitive needs, ‘‘I need to collaborate with and be involved in the decisions related to my relative;’’ for behavioural needs, ‘‘I need to learn how to cope with behavioural problems of my relative;’’ and for spiritual needs, ‘‘I need spiritual support.’’. Scores range from 0 to 51 and higher values indicate stronger intensity of expressed needs. The FSQ is a semi-structured interview of 30 dichotomous items that investigates psychological problems of caregivers, including emotional burden, problems in social involvement, need for knowledge about the disease, satisfaction with family relationships and thoughts about death. Typical sample items for each content area are as follows: for emotional burden, ‘‘I suffer from insomnia, digestive problems, headache and tiredness’’ and ‘‘I am pessimistic about the future;’’ for problems in social involvement, ‘‘I gave up my personal interests’’ and ‘‘The clinical condition of my relative makes me tired;’’ for need for knowledge about the disease, ‘‘I feel I need suggestions on how to assist my relative’’ and ‘‘I need to talk with an expert (e.g., doctor, clinical psychologist);’’ for satisfaction with family relationships, ‘‘In case there is a need, I know that I can rely on somebody else in the family’’ and ‘‘In this moment, I feel very close to relatives and friends;’’ for thoughts about death, ‘‘I feel helpless about the disease’’ and ‘‘I sometimes think of the death of my relative.’’. An index of overall burden is provided, with higher score indicating higher presence of perceived burden. Statistical Analyses Prior to analyses, data were screened for univariate and multivariate outliers, non-normal distribution responses and missing values following the procedures outlined by Tabachnick and Fidell (2007). Frequencies and percentages were used to report descriptive statistics of categorical variables. Cronbach’s alpha for each test was calculated as
123
a measure of internal consistency and alpha values higher than 0.80 indicated a good consistency. Regression models were adopted to investigate two alternative models of the combined influence of depressive symptoms and perceived burden as predictors of DOC caregivers expression of needs. The models consider two different ways to chain the three variables together. As shown in Table 3, Model 1 examines a depression–burden– needs chain, i.e., the model treats perceived burden as a mediating variable that influences the relationship between caregivers’ depressive symptoms and their expression of needs. Model 2 reverses the order of the two predictors in the chain and examines a burden–depression–needs chain, i.e., the model treats depressive symptoms as a mediating variable that influences the relationship between caregivers’ perceived burden and their expression of needs. Age, gender, disease duration and STAI-Y Trait Anxiety scores were entered as control variables in both Model 1 and Model 2. Additional regression models were used to investigate two further models that include state anxiety as a variable that influences DOC caregivers expression of needs. As shown in Table 4, Model 3 examines an anxiety–burden– needs chain, i.e., the model treats perceived burden as a mediating variable that influences the relationship between caregivers’ anxiety and their expression of needs. Model 4 reverses the role of the two predictors in the chain and examines a burden–anxiety–needs chain, i.e., the model treats state anxiety as a mediating variable that influences the relationship between caregivers’ perceived burden and their expression of needs. Age, gender, disease duration and BDI-II scores were entered as control variables in both Model 3 and Model 4. For a variable to be considered a mediator four criteria need to be met, as proposed by Baron & Kenny (1986): (1) Independent variable should significantly predict mediator (path a); (2) mediator should significantly predict dependent variable (path b); (3) independent variable should predict dependent variable (path c); and (4) the relationship between dependent and independent variable should be reduced or no longer significant in the presence of mediator (path c’). In order to corroborate this effect, the significance of each mediation model was assessed by testing the joint significance of the independent variable to the mediator (path a) and of the mediator to the dependent variable (path b), following the suggestion of MacKinnon et al. (2002). The Sobel test, which yields a z-value, was used to test the significance of the joint effect of the a and b paths. All path coefficients represents standardised betas (regression coefficients). Caregivers’ gender and age as well as patient’s disease duration were entered as covariates in all regression models. Significance level a = 0.05 was adopted, and all statistical analyses were performed using SPSS v. 18.0.
J Clin Psychol Med Settings Table 1 Patient-related socio-demographic characteristics Patient diagnosis—n (%) SV
272 (68.5)
SMC
125 (31.5)
Patient’s disease duration—mean (SD), in years
3.4 (3.3)
Patient’s place of residence—n (%) With family
51 (12.8)
Post-acute rehabilitation centre
233 (58.7)
Long term care centre
113 (28.5)
Relationship to the patient—n (%) Partner
154 (38.8)
Relative
84 (21.2)
Son/Daughter
72 (18.1)
Grandparent, grandson/daughter, mother-in-law, sisterin-law and others
87 (21.9)
Table 2 Caregiver’s scores obtained at protocol assessment BDI-II score—mean (SD)
15.4 (10.1)
Minimal
165 (41.6)
Mild Moderate
59 (14.9) 32 (8.1)
Severe
141 (35.5)
STAI-Y trait anxiety score—mean (SD)
46.1 (10.4)
FSQ perceived burden score—mean (SD)
13.2 (6.1)
CAN needs expressed score—mean (SD)
33.8 (9.57)
Results A total of 487 caregivers of children and adult patients in VS or MCS were invited to participate, and 397 (81.5 %) questionnaires were completed and considered for analyses. Participants were mainly females (n = 266; 67 %), married (n = 301; 75.8 %), with a mean age of 51.5 years (SD = 12.9) and 10.6 years (SD = 4.5) of education. Patient-related socio-demographic characteristics and scores obtained by caregivers at protocol assessment are displayed in Tables 1 and 2, respectively. Procedures for screening raw data for completeness prior to statistical analysis identified few missing values, less than 5 % for all the variables. Missing values where replaced by the median value of the variable in order not to inflate correlation between variables. FSQ score was positively skewed and so a square-root transformation was applied. No cases were found to be a univariate (z-score = ± 3 or
beyond) or multivariate outlier (Mahalanobis distance at p \ 0.01). Regarding Cronbach’s alpha, internal consistency was good for STAI-Y Anxiety (a = 0.87), CNA (a = 0.87), BDI-II (a = 0.89), and FSQ (a = 0.92). In Model 1, greater depressive symptoms were significantly associated with greater perceived burden (path a) which, in turn, was positively associated to needs expressed (path b). Greater depressive symptoms were also significantly associated to greater needs expressed (path c) and this correlation was no longer significant when perceived burden was entered as mediator variable (path c’ in Fig. 1). Tests of the indirect or joint effect of the paths a and b revealed that perceived burden was a significant intervening variable between depressive symptoms and needs expressed (Sobel test, z = 4.639, p \ .001), as shown in Table 3. In Model 2, paths a, b and c were all significant and path c’ was still significant when depressive symptoms were entered as mediator (p \ .001), and the Sobel test was far from being significant (z = .214, p = .830). In Model 3, similar to Model 1, greater anxiety symptoms were significantly associated with greater perceived burden (path a) which, in turn, was positively associated to needs expressed (path b). Greater anxiety symptoms were also significantly associated to greater needs expressed (path c) and this correlation was no longer significant when perceived burden was entered as mediator variable (path c’ in Fig. 2). The Sobel test showed that perceived burden was a significant intervening variable between anxiety symptoms and needs expressed (z = 5.232, p \ .001). In Model 4, similar to Model 2, paths a, b and c were all significant and path c’ was still significant when anxiety symptoms were entered as mediator (p \ .001), and the Sobel test was far from being significant (z = .385, p = .700), as shown in Table 4. Although the relationship between variables in Tables 3 and 4 are referred to as ‘‘paths,’’ they are not the same as ‘‘paths’’ calculated in Path Analysis. In the current paper, the term ‘‘path’’ is used to refer to a regression coefficient that quantifies the association between two variables. The same holds true in Figs. 1 and 2, i.e. the arrow vectors shown in those Figures are referred to as ‘‘paths,’’ but they should not be confused with ‘‘paths’’ as calculated in Path Analysis. They are regression coefficients.
Discussion To the best of our knowledge, this is the first study that shows the important role of perceived psychosocial burden in determining the effects of depression and anxiety on DOC caregivers expression of needs. Our study shows that the relation between these symptoms and caregivers’ needs is mediated by burden, where higher burden accentuates and lower burden mitigates the needs expressed by
123
J Clin Psychol Med Settings Fig. 1 Depression-burdenneeds chain. This model treats perceived burden as a mediating variable that influences the relationship between caregivers’ depressive symptoms and their expression of needs
Perceived burden (FSQ) Path a .333 (p
