IJCS | Volume 31, Nº4, July / August 2018

415 Guerra et al. Screening: depression in heart failure Int J Cardiovasc Sci. 2018;31(4)414-421 Original Article Geriatric Depression Scale of Yesavage concluded that both instruments are useful in identifying depression in elderly individuals, with a rho correlation of 0.387 (p < 0.000), kappa reliability 0.41 (p < 0.001), sensitivity 80%, specificity 44%, and moderate agreement. A randomized study conducted by Freedland et al. 14 evaluated the efficacy of an integrative cognitive behavior therapy intervention for depression (measured by the scales BDI-II [46%] and HAM-D [51%]) and self-care in outpatients with HF (n = 158); the results of the study showed that the interventionwas effective for depression, but not for self-care. Another study, 15 conducted to determine the best sensitivity and specificity values of the BDI-II and HAM-D in patients (n = 73) seen at a reference center in neuropsychiatric showed that the BDI-II had higher sensitivity and specificity (94.4% and 90.6%, respectively) than the HAM-D (95% and 75.5%, respectively). The BDI-II has been described as a gold-standard instrument to screen for depression in HF. 3,16 However, as a self-rating scale, the BDI-II has limited application in patients with cognitive impairment or low education level. In these cases, interviewer-rated scales may be applied by professionals experienced in interviewing depressed patients. 15,17 The objective of this study was to determine the prevalence of depression and the agreement among methods of screening for depression in patients with HF. Methods Type of study Cross-sectional study with a consecutive sample, carried out between March 2015 and January 2017 in a clinic specialized in HF in Universidade Federal Fluminense (UFF), Niterói, RJ, Brazil. Participants The participants were enrolled in a multidisciplinary program of a clinic specialized in HF (UFF). The participants included in the study were patients enrolled in the clinic’s program, diagnosed with HF according to the criteria by McMurray et al. 18 , of both genders, and with the ability to answer the questionnaires. We excluded patients with cognitive impairment, as identified in their medical records, difficulty or inability of understanding the instruments, and with a prior history of cognitive therapy or use of antidepressants. In all, 76 patients comprised the final study sample and were assessedwith three questionnaires for the screening of depressive symptoms: HAM-D, BDI-II, and PHQ-9. The interviewwas conducted by a single interviewer, who followed the guiding protocol of the interviewer- rated questionnaire (HAM-D). 17 The questionnaires BDI-II and PHQ-9 were applied as recommended in the literature. The questionnaires were applied during the same interview, conducted by a single examiner. In the case of the self-rating questionnaires (BDI-II and PHQ-9), the patients were informed that these questionnaires were meant to evaluate their mental health status. They were then instructed to read the questionnaires attentively and mark the answers according to the intensity of their symptoms, with all following the same direction (i.e., the greater the severity of the symptom, the higher the score to be checked). There was no time constraint for the participants to complete the questionnaires and the examiner did not interfere with the reading of the questions to avoid a biased interpretation by the patient. In regards to the HAM-D questionnaire, whose interview follows a protocol, the examiner conducted structured interviews and assigned a score to each response according to the intensity of the patient’s signs and symptoms. Instruments Hamilton Depression Rating Scale The HAM-Dwas the first interviewer-rated scale, i.e., applied by an interviewer. This scale was designed and developed by Hamilton at the end of the 1950s decade. 19 In 1994, the HAM-D was adapted to the Brazilian population as a valid instrument for an early diagnosis of a depressive episode. The HAM-D scale initially had 21 items but was subsequently reduced to a 17-item version after some items were removed (paranoid symptoms, obsessional symptoms, derealization, and diurnal variation), due to the low incidence or reliability of these items in relation to the measure of depression. 11,19 In 1988, a structured manual was prepared for the HAM-D scale interview 20 to standardize the questions by the interviewer. Hamilton did not establish a cutoff value to discriminate normality frommorbidity. Currently, it is accepted that scoreswithmore than 25 points characterize