A cross-sectional population-based study was conducted in the urban zone of the municipality of Pelotas between February and June 2012 to evaluate the health of adolescents, adults and elderly people. A sampling design of two-stage conglomerates with probability proportional to size was used. According to the 2010 Population Census there were 495 census tracts, the primary sampling units. The secondary sampling units were households. All private households with permanent resident as of December 2011 in the 130 census tracts randomly selected were listed. In each census tract drawn, around 12 households were randomly selected for the survey. All the people living in the households drawn who were 10 years of age or over were eligible. The participants were interviewed at home, by trained interviewers, through applying a structured questionnaire that included questions about their economic condition, schooling, marital status, skin color, occupation, health, and behavior. The adults (≥20 years of age) answered the PHQ-9 and EPDS questionnaires, and these were applied by general interviewers. Individuals who had cognitive or mental disabilities confirmed by the fieldwork supervisor, as well as those institutionalized (hospitals, elderly homes, among others), were excluded.
Validation studies on PHQ-9 [13] and EPDS [14] were conducted on a subsample of adults (≥20 years of age). The sampling process for the validation studies was conducted weekly, starting from the interviews that were conducted for the main study. Through simple random draws, one-third of the households included in the main study were selected for the validation studies. The person in charge of the draw was unaware of the results from the PHQ-9 or EPDS tests that were applied in the main study. In each household thus selected, all the people living there who were 20 years of age or over, independently of the PHQ-9 or EPDS scores, were invited to receive a second visit for a supplementary interview. This second interview was conducted by a mental health professional (psychiatrist, psychologist or medical resident in psychiatry), who had previously been trained to apply and interpret the gold-standard instrument and was blind to the scores achieved by the participant in the PHQ-9 and EPDS questionnaires. The participants were unaware of the professional training of these interviewers, so that this would not influence the responses.
The PHQ-9 consists of nine questions that assess the presence of each of the symptoms of MDE, as described in the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) [15] (depressed mood; loss of interest or pleasure in doing things; problems relating to sleep, tiredness or lack of energy; changes in appetite or weight; feelings of guilt or uselessness; problems of concentration; feelings of being slow or restless; and suicidal thoughts). The frequency of each symptom over the preceding 2 weeks was evaluated on a Likert scale from 0 to 3.
The EPDS was originally constructed to identify postpartum depression, but it can be applied to screen for depression in the community, including among men [16]. The EPDS consists of a scale of 10 items, each with four possible responses from 0 to 3, which express the intensity of depressive symptoms over the 7 days preceding the interview.
The questionnaire for the first interview (main study) was set up in sections. The PHQ-9 was applied after the participants had answered the questions in the sections relating to socio-demographic factors, behavioral factors, chronic diseases and use of medications. Following the PHQ-9 application, there were questions on the subjects’ use of and access to healthcare services and their dietary habits, and then the EPDS was applied. Further details on the methodology of the validation studies for the PHQ-9 and EPDS, along with the Portuguese-language versions used, can be obtained in other published papers [13, 14].
To calculate the sample size, the following parameters were used: sensitivity and specificity of 80%, acceptable error of 10% points upwards or downwards and significance level of 95%. Thus it was necessary to include around 200 subjects with MDE and 200 without MDE. Given that the point-prevalence of depressive symptoms among the adult population of Pelotas had been found to be around 30% [17], it was estimated that with a sample of around 600 individuals, it would be possible to locate around 200 with MDE.
The gold standard used was the Mini International Neuropsychiatric Interview (MINI) [18], which has been validated for use in Brazil [19]. This structured diagnostic questionnaire assesses the presence of mental disorders, in accordance with DSM-IV and ICD-10. For depressive disorders, it has sensitivity and specificity of 92% [19]. In the present study, the gold standard was used to diagnose the presence of MDE. All individuals who were considered to be positive for MDE gave responses to an additional group of questions that investigated other possible causes for the symptoms, such as direct effects of substances, organ disorders, medical illness or presence of psychotic symptoms, or whether the symptoms would be better explained as reactions to grief, for which the diagnosis of MDE would be rejected.
The data analysis included calculation of the sensitivity and specificity for each score on a continuous scale for each of the tests. For each PHQ-9 and EPDS cutoff point, the sensitivity (proportion of individuals with MDE according to MINI criteria that were correctly identified by the test), specificity (proportion of individuals without MDE according to the gold standard correctly identified as such by the test), positive predictive value (proportion of true positives among all positives identified by the test), accuracy (proportion of true positives and true negatives identified by the test), and positive likelihood ratio (the odds that the given cutoff point would be expected in an individual with in opposed to one without MDE according to the gold standard) with 95% confidence intervals were calculated.
To compare the accuracy of the tests for identifying individuals at risk of MDE, the sensitivity and 1-specificity values of each of the cutoff points for the PHQ-9 and EPDS were plotted on a single receiver operating characteristic (ROC) curve. The cutoff point with greatest sensitivity and specificity on the ROC curve was defined as the lowest value for the equation {(1 − sensitivity)2 + (1 − specificity)2}. The accuracies of the PHQ-9 and EPDS were compared by means of the areas under the respective ROC curves. The concordance between the two tests, i.e. beyond what would be expected by chance, was calculated by means of the kappa statistic.
The main study and the validation studies were approved by the Research Ethics Committee of the School of Medicine, Federal University of Pelotas, in accordance with protocols 77/2011 and 14/2012, respectively. A free and informed consent statement was signed by each participant in the main study before information was gathered. The individuals who were diagnosed as positive through the gold-standard assessment were attended at home and/or were referred to the healthcare services.