IJCS | Volume 31, Nº4, July / August 2018

418 Guerra et al. Screening: depression in heart failure Int J Cardiovasc Sci. 2018;31(4)414-421 Original Article The prevalence rates of depression diagnosed by the instruments were: 72.4% (n = 55) by the HAM-D, 67.1% (n = 51) by the BDI-II, and 40.8% (n = 31) by the PHQ-9, as shown in Table 2. Considering a simultaneous diagnosis by all three instruments, the prevalence of depressionwas 28.9% (n = 22). The three instruments showed a diagnostic agreement (presence or absence of depression) in only 47.4% of the sample (n = 36). Comparing the three instruments in terms of diagnosing depression in the HF population, there was superficial agreement (k F =C k = 0,27, as evaluated by Fleiss’ kappa (k F ) and Krippendorff’s alpha coefficients (C k ), and moderate consistency (significantly not null, p = 0.000), as assessed by Cronbach’s alpha coefficient (C C ). Table 3 shows the quality measures of the HAM-D and PHQ-9 instruments as diagnostic tests for depression in outpatients with HF, using the BDI-II as the gold- standard instrument. The HAM-D scale proved to be the best instrument to diagnose depression, as it showed higher accuracy and sensitivity and a lower percentage of false-negative results. The PHQ-9 instrument was Table 2 - Prevalence (P) of depression in outpatients with heart failure Instrument Results HAM-D 72.4% (n = 55) BDI-II 67.1% (n = 51) PHQ-9 40.8% (n = 31) Agreement among all three instruments k ↓ F = C ↓ k = 0.27 Consistency of the three instruments ( ↓ C 0.602; p < 0.000 n: number of patients; HAM-D: Hamilton Depression Rating Scale; BDI-II: Beck Depression Inventory-II; PHQ-9: Patient Health Questionnaire-9. Table 3 - Quality measures of the HAM-D and PHQ-9 instruments as diagnostic tests for depression in outpatients with heart failure Instrument Accuracy Sensitivity Specificity % of false- positives % of false- negatives HAM-D 76.3 86.3 56.0 14.5 9.2 PHQ-9 55.3 47.1 72.0 9.2 35.5 HAM-D: Hamilton Depression Rating Scale; PHQ-9: Patient Health Questionnaire-9. conservative in diagnosing depression, with a high percentage of false-negative results and a low sensitivity in identifying patients who in fact had depression. Considering the BDI-II as the gold-standard instrument in diagnosing depression, we investigated the association of depression with the patients’ characteristics. There was no significant association between depression and the following factors: gender (p = 0.291), self-declared color (p = 0.976), education level (p = 0.918), obesity (p = 0.324), diabetes (p = 0.316), dyslipidemia (p = 0.056), COPD (p = 0.250), stroke (p = 0.296), and CRF (p = 0.536). The age and income of the patients with and without depression were also not associated with depression (p = 0.862 [unpaired Student’s t test] and p = 0.776 [Mann- Whitney test], respectively). Therefore, none of the variables included in this study was associated with depression or was a risk factor for depression in outpatients with HF. Discussion This is the first study comparing screening methods for depression in outpatients in a multidisciplinary clinic specialized inHF. Depression has not been systematically analyzed in patients with HF, but when specifically researched, has been observed to be frequent in this population. 28-36 This condition affects between 14.0% and 26.0% of the patients without HF, but the incidence increases to 24.0% to 85.0% in patients with HF. 33,34 The main findings of this study indicate a relevant prevalence of depression in HF patients when screened by the HAM-D, BDI-II, and PHQ-9 instruments. BDI- II has been considered the gold-standard instrument for depression screening in patients with HF, but in individuals with cognitive impairment or illiterate, this instrument is not recommended. 10,12,21 Therefore, the results of our investigation in relation to the internal