Abstracts

BRIEF REPORT

Family Burden interview scale for relatives of psychiatric patients (FBIS-BR): reliability study of the Brazilian version

Escala de sobrecarga de familiares de pacientes psiquiátricos (FBIS-BR): estudo de confiabilidade da versão brasileira

Marina Bandeira^I; Maria Gláucia Pires Calzavara^I,II; Lucas Cordeiro Freitas^III; Sabrina Martins Barroso^IV

^IResearch Laboratory in Mental Health (LAPSAM), Universidade Federal de São João del Rei (UFSJ), São João del Rei (MG), Brazil

^IIUniversidade Federal de São João del Rei (UFSJ), São João del Rei (MG), Brazil

^IIISpecial Education post graduation program, Universidade de São Carlos (UFSCar), São Carlos (SP), Brazil

^IVUniversidade Federal de Minas Gerais (UFMG), Belo Horizonte (MG), Brazil

^{Correspondence} Correspondence: Marina Bandeira Laboratório de Pesquisa em Saúde Mental Universidade Federal de São João del Rei Praça Dom Helvécio, 74 36301-160 São João del Rei, MG, Brazil E-mail: bandeira@ufsj.edu.br

ABSTRACT

OBJECTIVE:Psychiatric deinstitutionalization resulted in greater involvement of families in daily care of patients. Family caregivers burden has been described by international researches. In Brazil very few studies investigated familia burden with validated scales, perhaps due to the restrict availability of scales to measure this construct. The present study examined the reliability of the Brazilian version of the Family Burden Interview Schedule.

METHOD:Scale reliability was assessed through its internal consistency and temporal stability. A sample of 243 family caregivers of psychiatric patients attending public outpatient services were interviewed with the scale. Socio-demographic data were collected with a standardized form. A sub-sample of 42 relatives participated in a re-test, after 3 weeks, for assessment of the scale temporal stability.

RESULTS:Cronbach alpha coefficients ranged from 0.58 to 0.90 for global and domains scores. Pearson correlation coefficients and intra-class correlation coefficients for test and retest ranged from 0.54 to 0.90.

DISCUSSION:Results for internal consistency were similar to the original scale. Four subscales presented alpha coefficients superior to the original ones. The temporal stability cannot be compared because this analysis was not performed for the original scale.

CONCLUSION:The Brazilian version of the scale had good psychometric properties of temporal stability and internal consistency.

Descriptors: Reliability of tests; Family; Burden cost of illness; Mentaly ill persons; Deinstitutionalization

RESUMO

OBJETIVO:A desinstitucionalização psiquiátrica resultou em maior envolvimento das famílias nos cuidados diários dos pacientes. Pesquisas internacionais constataram a sobrecarga dos familiares, mas, no Brasil, poucas pesquisas a investigaram utilizando escalas validadas, talvez devido à carência de instrumentos de medida deste construto. Este estudo avaliou a confiabilidade da versão brasileira da escala Family Burden Interview Schedule.

MÉTODO: Avaliou-se a confiabilidade, em termos da consistência interna e estabilidade temporal. Foram entrevistados 243 familiares cuidadores de pacientes psiquiátricos de serviços públicos de saúde mental. Os dados sociodemográficos foram obtidos por questionário padronizado. Uma subamostra de 42 familiares foi entrevistada novamente, após três semanas, para avaliação da estabilidade temporal teste-reteste.

RESULTADOS:Os coeficientes alfa de Cronbach para os escores globais e domínios variaram de 0,58 a 0,90. Os coeficientes de correlação de Pearson e de correlação intraclasse para o teste-reteste foram positivos e significativos, variando de 0,54 a 0,90.

DISCUSSÃO:Os resultados da consistência interna foram semelhantes aos da escala original, mas quatro subescalas apresentaram valores de alfa superiores. A estabilidade não pode ser comparada, pois esta não foi avaliada para a escala original.

CONCLUSÃO:A versão brasileira da escala apresentou boas propriedades psicométricas de estabilidade temporal e consistência interna.

Descritores: Confiabilidade dos testes; Família; Efeitos psicossociais da doença; Pessoas mentalmente doentes; Desinstitucionalização

Introduction

As a result of the psychiatric deinstitutionalization policy and the implementation of community mental health services, family members have been increasingly involved in the daily care of the patients, therefore assuming new responsibilities.^1-2 Some researchers pointed out to the fact that family caregivers often take the role of "case managers" in order to fulfill the patient's needs, as a result of the services' difficulties in maintaining this role and as a consequence of a lack of integration between services.^2-3

The increasing demands to assist patients in daily activities, to maintain their medications and supervise their problematic behaviors contribute to a disruption in the caregiver's social, professional and family life, causing feelings of burden. Family caregivers also lack appropriate information about the patient's disease and the necessary skills to deal with the patients in daily life. Patients' caregiving represents an additional role to the family and requires that caregivers consider patients' needs as a priority, putting their own needs in second place.^4-6

The concept of family burden includes two aspects, objective burden related to the performance of daily assistance activities, financial impact, behavior supervision and disruption of family routine, and subjective burden, concerning worries about the patients and feelings of being disturbed by caregiving activities.⁴

International research results have indicated that family caregivers suffers from a high level of burden. Several studies have investigated the factors mainly associated with the relatives' burden of care, in order to get information for developing family orientation interventions.^4-6

In Brazil, although there is an increasing interest in studying family caregivers,^7-8 very few researches specifically assessed caregivers' burden using validated scales. The existing studies are limited to samples of relatives of old patients suffering from dementia or depression.^9-10 Moreover, the scale used in these studies was originally designed for assessment of burden in relatives of old patients suffering with dementia and it evaluates only the subjective aspect of burden and provides only a burden global score. The lack of Brazilian validated scales for assessing burden of care in relatives of a broader patients' age range and kinds of diagnostics may explain the limited number of researches developed up to now. In order to improve research in this area, the present study examined the reliability of the Family Burden Interview Schedule - FBIS, elaborated by Tessler & Gamache,¹¹ which evaluates both objective and subjective aspects of burden, provides independent scores for different burden domains and can be applied to a broad target population.

Method

1. Sample

A sample of 243 family caregivers of psychiatric patients attending five Brazilian public outpatients' mental health services participated in this research. Subjects were randomly selected from those attending the services. Patients' main caregivers above 18 years old were included in the sample. Family caregivers with psychiatric disorders or elders with dementia were excluded. Sample size was determined by the time schedule and availability to collect the data. A sub-sample of 42 relatives was randomly selected from the original sample to participate in a second application of the scale, in order to test the scale's temporal stability.

2. Instrument

The original FBIS scale was adapted to Brazilian population, with the designation of FBIS-BR, in a study following the international guidelines for cross-cultural scales adaptation, including translation and back translation by bilingual translators, panel of experts and pilot study^12-14 to assure semantic and cultural equivalence to the original scale ¹⁵). The scale has 52 items assessing five dimensions of family burden: 1) assistance to the patient in daily life; 2) supervision of patients' problematic behaviors; 3) financial burden; 4) impact on family routine; 5) worries about the patients' present and future life.

The scale evaluates both objective and subjective burden. Objective burden is assessed by items measuring the frequency of assistances and supervisions provided to the patient and by the disruptions produced in family life. Subjective burden is assessed by items measuring the degree of relatives' feeling of being disturbed when assisting the patient and when dealing with their problematic behaviors and by the caregivers' worries about the patients' life. Alternative responses to the scale items are distributed in a Likert-type scale, ranging from 1 (never) to 5 (every day) for frequency of occurrences and from 1 (not at all) to 4 (very much) for degree of disturbance. Assessment concerned the month prior to the interviews. Duration of scale application was approximately 60 min, the same as the original scale.

3. Procedure

After the research project was approved by the institutions' research and ethics evaluation committee, family caregivers were contacted and the interviews were individually scheduled. Interviews were usually conducted at the institutions' facilities, where patients received treatment and less often at their home, depending on the subjects' time schedules. Family caregivers were informed about research objectives and were asked to sign an informed consent form. Interviewers were undergraduate psychology students trained by the researchers to assure standardized application procedure even thought the scale application does not require a special clinical knowledge. During interviews, subjects could visualize an answer card containing the response alternatives to each question, in order to help them remember the items.

The sub-sample of 42 relatives was submitted to a second application of the scale, by the same interviewers. Interviews were performed after a 3 week-time interval from the first application.

4. Data analysis

The scale internal consistency was evaluated by the Cronbach alpha coefficient analysis. An item-total correlation above 0.20 was the criterion for retaining items in the scale. The scale temporal stability was evaluated comparing data from test and retest applications, using Pearson correlation coefficient analysis and intra-class correlation coefficients, with 95% confidence interval. Data analysis was implemented in a 13.0 SPSS program version.

Results

1. Sample characteristics

The study sample of 243 family caregivers included 188 female (77. 40%) and 55 males (22. 60%). The sample mean age was 54.51, with minimal age of 18 and maximum age of 87. Caregivers were mostly parents (47.30%), but also brothers or sisters (17.30%), partners (14.00%) or other relatives (9.50%) of the patients.

The patients were 40.14 years old in average, with minimum age of 14 and maximum of 71. Patients were 158 male (65%) and 85 female (35%), with the following diagnoses according to the services they attended: schizophrenia, schizo-affective disorder, acute psychotic disorder of schizophrenic type, non-organic psychotic disorder.

The re-test sub-sample of 42 family caregivers was 53.45 years old in average,ranging from 31 to 74 years old. This sample included 34 females (81%) and 8 males (19%). They were mostly brothers or sisters (35.7%), but also parents (28.6%), partners (21.4%), sons (4.8%) or others relatives (9.5%) of the patients. Patients were 44.50 years old in average, 22 male (52.4%) and 20 female (47.6%).

2. Scale reliability

1) Internal consistency

Table 1 shows the results concerning the FBIS-BR internal consistency, for the global objective and subjective burden scores and for the dimension scores. Alpha coefficient for global objective burden was 0.82 and item-total correlations ranged from 0.21 to 0.52. For global subjective burden, alpha coefficient was 0.92 and item-total correlations ranged from 0.31 to 0.97.

Regarding the burden dimensions, alpha coefficients ranged from 0.58 to 0.90 and item-total correlations ranged from 0.23 to 0.88. Three items from dimension B, both for subjective and objective evaluations, were eliminated because their item-total correlations were below 0.20.^16-17 Dimension C was not submitted to internal consistency analysis because open questions referring to the patient's expenses are not appropriate to this kind of statistical analysis. Two items from dimension D regarding complementary information not measured as continuous variables were not included in this analysis.

2) Temporal stability (test-retest)

Table 1 also shows the results concerning the scale temporal stability. Test-retest Pearson correlation scores for global objective burden was 0.83 (p < 0.05) and for global subjective burden was 0.74 (p < 0.05). For the scale dimensions, correlation coefficients were all significant and ranged from 0.54 to 0.92. The test-retest reliability estimated using intra-class correlation coefficients showed similar results, with scores ranging from 0.54 to 0.92.

Discussion

Results for the FBIS-BR internal consistency were similar to those obtained for the original scale, although alpha coefficients were even higher in 4 sub-scales of the Brazilian version: the one regarding the subjective burden in assisting the patient in daily life (0.90 compared to 0.74 in the original scale), the two sub-scales assessing burden in supervising patients problematic behaviors, the objective one (0.78 compared to 0.65 in the original scale) and the subjective one (0.78 compared to 0.64 from the original scale) and the sub-scale assessing the impact in caregivers' life (0.79 versus 0.57).

The objective burden sub-scale regarding assisting the patient in daily activities had a coefficient of 0.74, which is slightly inferior to the one obtained for the original scale (0.78). The sub-scale assessing the caregivers' worries about the patient also had an alpha coefficient (0.58) inferior to the one obtained for the original scale (0.89).

The results of the present study can be considered as good indications of the FBIS-BR's internal consistency, since it is more frequent to obtain Cronbach coefficients inferior to the ones observed in the original scales, in cross-cultural adaptation studies.¹² In the present study, scales and sub-scales alpha coefficients scores are adequate considering their number of items.^16-17 Also, most sub-scales had superior alpha scores compared to the original scale and only two had inferior values. The extensive and detailed process of adaptation of the scale during pilot study and the inclusion of examples in the items to assure better understanding by respondents¹⁵ can explain the high consistency of responses obtained for most sub-scales. Differences in caregivers' life conditions may also contribute to different results in both studies, for example the higher proportion of co-residence in the Brazilian sample. In particular, this condition can explain the low total variability of responses observed for the worries sub-scale resulting in a lower alpha coefficient.

Regarding the FBIS-BR's temporal stability, results showed significant positive test-retest Pearson correlations for all sub-scales as well as for the two global scales, varying from 0.54 to 0.92 (p < 0.01). Test-retest reliability estimated using intra-class correlation coefficient, which controls the effect of chance, showed results that were mostly identical to those obtained with Pearson correlations. This result may indicate that sample size was adequate and measurement errors were minimized. Results for the scale stability cannot be compared to the original scales, since this analysis was not implemented for the original scale.

Conclusion

The FBIS-BR scale is a reliable instrument for assessing the relatives' burden of care. The global and domains scores showed adequate psychometric properties of reliability, regarding internal consistency and temporal stability.

The Brazilian version of the scale can be applied to assess objective and subjective aspects of relatives' burden of care and also to assess different burden dimensions by means of independent scores. These independent measures constitute the innovative characteristics for this scale compared to other ones and they are useful to investigate factors affecting differentially each of these dimensions.

References

1. St-Onge M, Lavoie F, Cormier H. Les difficultés perçues para des mères de personnes atteintes de troubles psychotiques face au système de soins professionnels. Sante Ment Que. 1995;20(1):89-117.

2. Hanson JG, Rapp CA. Families' perceptions of community mental health programs for their relatives with a severe mental illness. Community Ment Health J. 1992;28(3):181-97.

3. Solomon P, Beck S, Gordon B. Family members' perspectives on psychiatric hospitalization and a discharge. Community Ment Health J. 1988;24(2):108-17.

4. Maurin JT, Boyd CB. Burden of mental illness on the family: a critical review. Arch Psychiatr Nurs. 1990;4(2):99-107.

5. Loukissa AD. Family burden in chronic mental illness: a review of research studies. J Adv Nursing. 1995;21(2):248-55.

6. Rose LE. Families of psychiatric patients: a critical review and future research directions. Arch Psychiatr Nurs. 1996;10(2):67-76.

7. Villares CC. Adaptação transcultural de intervenções psicossociais na esquizofrenia. Rev Bras Psiquiatr. 2000;22(Supl 1):53-5.

8. Villares CC, Mari JJ. Esquizofrenia e contexto familiar. In: Shirakawa I, Chaves AC, Mari JJ, eds. O desafio da esquizofrenia. São Paulo: Lemos Editorial; 1998. p. 243-5.

9. Garrido R, Menezes PR. Impacto em cuidadores de idosos com demência atendidos em um serviço psicogeriátrico. Rev Saude Publica. 2004;38(6):835-41.

10. Scazufca M, Menezes PR, Almeida OP. Caregiver burden in an elderly population with depression in São Paulo, Brazil. Soc Psychiatry Psychiatr Epidemiol. 2002;37(9):416-22.

11. Tessler RC, Gamache GM. The Family Burden Interview Schedule - Short Form (FBIS/SF). In: Sederer, L, Dickey B, eds. Outcome assessment in clinical practice. Baltimore: Williams & Williams; 1996. p. 110-2.

12. Vallerand RJ. Vers une méthodologie de validation trans-culturelle de questionnaires psychologiques: implications pour la recherche en langue française. Can Psychol. 1989;30(4):662-80.

13. Guillemin F, Bombardier C, Beaton D. Cross-cultural adaptation of health-related quality of life measures: literature review and proposed guidelines. J Clin Epidemiol. 1993;46(12):1417-32.

14. Jorge MR. Adaptação transcultural de instrumentos de pesquisa em saúde mental. In: Gorestein C, Andrade LHS, Zuardi AW, eds. Escalas de avaliação clínica em psiquiatria e psicofarmacologia. São Paulo: Lemos Editorial; 2000. p. 53-8.

15. Bandeira M, Calzavara MG, Varella AA. Escala de sobrecarga dos familiares de pacientes psiquiátricos: adaptação transcultural para o Brasil (FBIS-BR). J Bras Psiquiatr. 2005;54(3):206-14.

16. Gulliksen H. Theory of mental test. New York: John Wiley; 1950.

17. Martinez AR. Psicometria: Teoria de los tests psicoógicos y educativos. Madrid: Editorial Síntesis; 1995.

Financing:Conselho Nacional de Desenvolvimento Científico e Tecnológico(CNPq), n. 40166/2004-4-Ed 322004 and Fundação de Amparo à Pesquisa do Estado de Minas Gerais (FAPEMIG), nr. SHA-71/04

Conflict of interest: None

Submitted: March 8, 2006

Accepted: June 19, 2006

Publication Dates

History