INFORMACIONES PSIQUIÁTRICAS 247

14 Informaciones Psiquiátricas 2022 - n.º 247 Measures Sociodemographic, clinical and living- arrangement variables were collected from participants, both patients and caregivers. Objective and subjective caregiver burden was assessed using the FBIS-II questionnai- re (Family Burden Interview Schedule-Short Form) in its validated Spanish-language ver- sion (ECFOS II) (24). This structured inter- view was heteroadministered by expert staff (60 minutes) and provides different quan- tifiable scales on objective and subjective aspects of the caregiver of a patient with a serious mental disorder over the prece- ding month. It takes into account the time spent, the discomfort felt and the need for help in each assessed aspect (care for daily activities, containment of behavioral dis- turbances, financial burden, routines and life of the carer, concerns, help available, effects on his/her own health and overall impact on caregiver and family patient’s in- teraction). Most of the items are answered on a four or five points Likert scale from none/never to too much (value to subjec- tive mesures) or all days/always (value to objective mesures). In this work, in order to facilitate the analysis, the items of each module have been grouped into summary variables. Thus, the "Caregiver activity required" variable refers to the number of aspects of each mo- dule where the presence of the caregiver is necessary. The "Frequency" variable refers to the average number of times per week in the preceding month that the caregiver needs to intervene or that there is an im- pact on him/her. The "Caregiver discomfort" variable refers to the average amount of dis- comfort felt by the carer when carrying out the tasks. The approximate total hours of care per week are also collected. Perceived social support was assessed using the version of the Duke-UNC-11 ques- tionnaire validated for the Spanish popula- tion (25). This instrument gives measures of confidential social support (the degree to which one can count on people close to him/her), affective social support (the de- gree to which a person receives demonstra- tions of love, affection and empathy) and overall social support; responses are expres- sed on a Likert scale from 1 to 5. Analysis A descriptive study of the variables was carried out, checking normality and ho- moscedasticity (Kolmogorov-Smirnov and Levene test). The relationship between the variables was based on contingency tables for the qualitative variables (Chi-square or Fisher's exact test), on the Student's t-test or Kruskal-Wallis test for the quantitative/ qualitative relationship, and on Pearson and Spearman's correlation test, represented graphically by means of scatter plots, for the quantitative variables. A regression analysis was performed with the variables exhibiting a significant asso- ciation (less than 0.05) and no collineari- ty in order to propose a model indicating which socio-demographic, housing, clinical or caregiving variables explained the va- riance for the variables of burden of care (FBIS II) and perceived social support (Duke UNC 11). Multiple linear regression was used for the "caregiver activity required", "Fre- quency " and "Caregiver discomfort" varia- bles of modules “care for daily activities”, “containment of behavioral disturbances”, “routines and life of the carer”, “concerns” and “overall impact on the life of the ca- rer and their family” (FBIS II) and the to- Emilio Rojo / María Jesús San Pío / Gemma Marcet / Francisco José Eiroa-Orosa / Josep Rojo