ABC | Volume 110, Nº1, January 2018

Original Article Santos et al Development and psychometric validation of HIPER-Q Arq Bras Cardiol. 2018; 110(1):60-67 Methods Conception and procedures This study was approved by the research ethics committee of Santa Catarina State University (UDESC) (approval number 159.213/2012). The study had a cross-sectional, observational design. In the first stage of the study, a literature review was performed to identify the pieces of knowledge about SAH considered relevant to hypertensive individuals. 3,24 The bibliographic search was performed in Pubmed database from January 2010 to September 2016. The questionnaire was constructed and revised by a commission composed of 17 health specialists, with experience in CR. These specialists carried out an analysis of content and clarity of the instruments, to verify its adequacy to hypertensive patients participating in CR programs. The second stage was a pilot study to evaluate both applicability and reproducibility of the instrument, as well as patients’ understanding of the items (clarity). A convenience sample of hypertensive patients, who participated in CR programs, was studied, and the results were used for refinement of the HIPER-Q instrument. Patients of the pilot study did not participate in the psychometric validation. The third step was the psychometric validation. The refined tool was used in a larger convenience sample, composed of hypertensive patients participating in CR programs at the Clinic of Cardiology and Cardiopulmonary and Metabolic Rehabilitation (Cardiosport), the Center of Cardiology and Sports Medicine ( Núcleo de Cardiologia e Medicina do Esporte, NCME) of the clinic, and the Santa Catarina Institute of Cardiology (ICSC). Data were collected between November 2015 and May 2016. Participants Patients of the pilot study and patients of the psychometric validation group were recruited from the CR programs mentioned above if they met the following inclusion criteria: clinical diagnosis of SAH, age ≥ 18 years, participation in a CR program for a period longer than one month, and agreement to participate in the study by signature of the informed consent form, according to the CNS 466/12 resolution. Patients with cognitive dysfunctions that could make the completion of the questionnaire difficult, i.e ., who did not demonstrate a minimal understanding of socio-demographic questions were excluded, at the investigator’s discretion. Measurements To assess the clarity of the instrument, participants of the study pilot were asked to classify each item of the questionnaire in a 1 (not clear) to 10 (very clear) scale. 25 Also, these patients answered the HIPER-Q at two different occasions with a 14‑day interval for analysis of the reproducibility of the instrument. Patients who participated in the psychometric validation were characterized by sex, age, educational attainment, comorbidities, time in CR, cardiac risk factors and clinical history. These characteristics were self-reported. Statistical analysis Sample calculation for the psychometric analysis was performed according to Hair & Anderson’s 26 who recommend a minimal sample size of 10 subjects per item and/or a minimum of 100 participants. Since the questionnaire was composed of 17 items, a sample of 170 hypertensive subjects was considered sufficient. Test-retest reproducibility of the instrument was validated in the pilot study group using the intraclass correlation coefficient (ICC). The items should meet the minimal recommended standard – ICC > 0.7. 27,28 Psychometric properties of the new tool were assessed by analysis of internal consistency, criterion validity and factorial structure. First, internal consistency was analyzed in the psychometric validation group by Cronbach's alpha, reflecting the internal correlation between items and factors. 26 Values greater than 0.60 are generally considered acceptable. 29 Second, criterion validity was analyzed by relating the HIPER-Q scores to patients’ educational attainment and family income, using the Spearman correlation. Third, the dimensional structure (as well as the construct validity) was evaluated by exploratory factor analysis. A component method for factor extraction was performed, considering only those factors with characteristic values > 1.0. When necessary, items with low factor loading (< 0.35) were excluded. Once the factors were selected, a correlation matrix was generated, in which the associations between items and factors were identified by factorial loadings greater than 0.30 in only one factor. The promax method was used for matrix interpretation, 30 and the Spearman correlation was used for analysis of criterion validity. Finally, a descriptive analysis of HIPER-Qwas performed using mean values and standard deviations of normally distributed variables, and median and interquartile ranges for variables with non-normal distribution. Data normality was evaluated by the Kolmogorov-Smirnov test. Due to non‑normality of data, we used the chi-square test to evaluate the association between the HIPER-Q scores based on patients’ sociodemographic and clinical characteristics. Patients’ total knowledge was represented by the median of total score. Statistical analyses were performed using the Statistical Package for Social Sciences ( SPSS) version 20 (IBM Inc. 2011, NYC), and the level of significance was set at 5% for all tests. Results Participants For content validation, 17 health professionals with experience in CR were consulted: 6 (35.5%) physicians, 6 physiotherapists (35.3%), 2 nurses (11.8%), 2 physical educators (11.8%) and 1 dietitian (5.9%). For the pilot test, 30 hypertensive patients participating in CR programs were recruited by convenience to answer the questionnaire; 11 (22%) of them were women, with mean age of 62 ± 8 years. For psychometric validation, 184 hypertensive patients with mean age of 60.5 ±10 years and median time of diagnosis of 8 years (interquartile range 18 years) completed the HIPER-Q. 61