ABC | Volume 112, Nº2, February 2019

Original Article d’Almeida et al Cut-point for Dietary Sodium Restriction Questionnaire Arq Bras Cardiol. 2019; 112(2):165-170 Adult patients, with a diagnosis of HF – reduced or preserved left ventricular ejection fraction (LVEF) 9 – were included. Patients in outpatient treatment (compensated) and those admitted to emergency rooms due to acute HF (decompensated) participated in this study. It was used a convenience sample, with a total of 431 HF patients (206 compensated and 225 decompensated). Patients with cognitive impairment or barrier (e.g., decreased hearing acuity, neurological sequelae) were excluded since these impairments could make it difficult for patients to fill out the questionnaire. Data collection Clinical and sociodemographic data were collected from medical records. TheQRSDswere administeredby the researchers in a private room, with a mean duration of 40 minutes. The Brazilian version of the DSRQ comprises 27 items, 11 descriptive questions and 16 questions divided into three subscales, which are scored using the 5-point Likert scale: 10 a) Attitude and subjective norm (nine items, with scores ranging from nine to 45) – assesses the patient’s beliefs regarding the results of performing a diet with reduced sodium and the importance of other people’s approval or disapproval of this practice; b) Perceived behavioral control (four items with scores ranging from four to 20) – assesses the patient’s ability to identify facilitators and barriers related to the reduction of sodium in their diet; C) Dependent behavior (three items with scores ranging from three to 15) – assesses the presence or absence of resources and constraints for a patient to follow a sodium-reduced diet. In the first subscale – attitude and subjective norm – the lowest score indicates a “strong disagreement” and the highest, a “strong agreement”. In the second and third subscales – perceived behavioral control and dependent behavior – the minimum score indicates “not at all”, while the maximum indicates “a lot”. 8 This study was approved by the Ethics Committee of the institutions involved and all participants signed a written informed consent form before taking part in this study. Data analysis Data were analyzed using the Statistical Package for Social Sciences version 18.0. Continuous variables with normal distribution were expressed as mean and standard deviation and without normal distribution, as median and interquartile range. Categorical variables were expressed as absolute numbers and relative frequency. To compare continuous variables, unpaired Student’s t-test or Mann-Whitney test were used, according to data distribution. Associations between categorical variables were analyzed using the chi-square test or Fisher's exact test. A 5% significance level was adopted. To define the cut-points, the QRSD scores were compared between compensated and decompensated patients. A ROC curve was constructed for each subscale, and an additional comparison of patients by functional class (I - II) and (III - IV) was performed to determine the best point of sensitivity and specificity regarding adherence to the diet. determine the best point of sensitivity and specificity regarding adherence to the diet. Results A total of 431 HF patients participate in the study. Of the total, 206 were in outpatient treatment (compensated) and 225 patients sought emergency care (decompensated). Sociodemographic and clinical characteristics of the studied population are shown in Table 1. Mean age was 63 ± 13 years, and 59.2% of the participants were male; mean LVEF was 36.8 ± 14.0%. Regarding theQRSD scores, comparedwith decompensated patients, compensated patients had better scores, showing greater adherence in all subscales. Mean scores for compensated and decompensated groups, and for categories of functional classes are shown in Table 2. According to the ROC curve analysis, the area under the curve was 0.725 (95%CI; 0.677 to 0.772) for the attitude and subjective norm subscales; 0.670 (95%CI; 0.620 to 0.721) for the perceived behavioral control subscale; and 0.544 (95%CI; 0.489 to 0.598) for the dependent behavior subscale (Figure 1). The results of the functional class analysis were 0.631 (95%CI; 0.578 to 0.685) for the attitude and subjective norm subscales; 0.628 (95%CI; 0.574 to 0.682) for the perceived behavioral control subscale; and 0.561 (95%CI; 0.506 to 0.617) for the dependent behavior subscale. Sensitivity and specificity were, respectively, 53.8 and 83.5 for the attitude and subjective norm subscales; 68.0 and 58.3 for perceived behavioral control subscale; and 60.9 and 51.0 for dependent behavioral subscale. Cut-off points for adherence were scores greater than or equal to 40 points in the attitude and subjective norm subscale; lower than or equal to eight points for perceived behavioral control; and lower than or equal to three points for Dependent Behavior (Table 3). Discussion This is the first study conducted in a clinical scenario that tried to establish cut-points for the DSRQ/QRSD regarding adherence. This instrument considers the knowledge, barriers and attitudes of patients with HF regarding sodium restriction in the diet. Adherence can be defined as the degree to which individuals comply with recommendations (related to pharmacological treatment of changes in lifestyle) from the health team. 14 In the context of HF, treatment adherence is considered an essential component to the success of self-care and prevention of complications, including hospitalizations. 15 The sample was predominantly male patients older than 60 years, poorly educated, and with predominantly reduced LVEF, similar to other studies that addressed adherence in patients with HF. 8,16,17 Compared with compensated patients, in decompensated patients’ group, there were fewer men, fewer people with white ethnicity and a greater number of people living alone. These characteristics have already been related to lower adherence in previous studies. Lennie et al. 11 investigated 166