Full text
| |
| Author(s): |
Carolina de Meneses Gaya
Total Authors: 1
|
| Document type: | Doctoral Thesis |
| Press: | Ribeirão Preto. |
| Institution: | Universidade de São Paulo (USP). Faculdade de Medicina de Ribeirão Preto (PCARP/BC) |
| Defense date: | 2011-08-19 |
| Examining board members: |
Antonio Waldo Zuardi;
Ronaldo Ramos Laranjeira;
Sonia Regina Loureiro;
Jose Antonio Baddini Martinez;
Sandra Cristina Pillon
|
| Advisor: | Antonio Waldo Zuardi |
| Abstract | |
Depressive disorders, tobacco dependence, and alcohol-related disorders (ARD) are highly prevalent and are associated with several comorbidities. Nonetheless, they are still under-diagnosed and under-treated in many health settings. Therefore, this study was aimed at (1) assessing reliability and validity parameters of screening instruments for depressive disorders (Patient Health Questionnaire 2 [PHQ-2] and Well-Being Index 5 [WHO-5]), tobacco dependence (Fagerström Test for Nicotine Dependence [FTND] and Heaviness of Smoking Index [HSI]), and alcohol dependence (Fast Alcohol Screening Test [FAST], CAGE, Alcohol Use Disorders Identification Test [AUDIT] and their brief versions); (2) performing the factorial analysis of the WHO-5, FTND, and AUDIT; and (3) assessing the comorbidity between the above-mentioned disorders. The test-retest reliability study involved 429 university students and the inter-rater reliability study included 41 patients of a Center for Psychosocial Attention Alcohol and Drugs (CAPS-AD, in the Portuguese acronym). The assessment of the predictive validity, internal consistency, comorbidity, and factorial analysis involved a sample of 530 patients of an emergency unit and the CAPS-AD. Reliability was estimated by means of kappa and interclass correlation (ICC) coefficients. In the predictive validity studies, the SCID-CV was used as the comparison diagnostic measure, with calculation of the areas under the ROC curve (AUC), sensitivity, specificity, accuracy, and positive and negative predictive values of the instruments. Concurrent validity was assessed using Spearmans coefficient. Cronbachs alpha coefficient was used to assess internal consistency. The exploratory factorial analysis was conducted according to Kaisers criteria. Comorbidities were analyzed by logistic regression raw and adjusted odds ratio. The level of statistical significance was set at p < 0.05, with a 95% confidence interval. The PHQ-2 and WHO-5 had high levels of reliability (> 0.98) and internal consistency (0.76 and 0.78, respectively). The WHO-5 had an AUC of 0.89; sensitivity of 0.85; and specificity of 0.81 for a cut-off of 9. The PHQ-2 had an AUC of 0.86; sensitivity of 0.74; and specificity of 0.87 for a cut-off of 3. The WHO-5 yielded one single factor, which explained 51% of the data variance. The reliability study of the FTND provided high ICC coefficients (0.92 for test-retest and 0.99 for inter-rater). In the factorial analysis, the FTND yielded two factors. The internal consistency of the FTND was high (0.83), but that of the HIS was below recommended levels (0.56). The FTND and the HIS had high indices of sensitivity, specificity, and accuracy. The AUDIT and its brief versions had internal consistency values between 0.83 and 0.94. In the study of predictive validity, these instruments had AUC between 0.92 and 0.96, with sensitivity levels between 0.84 and 0.93 and specificity levels ranging from 0.83 and 0.94 for the screening of ARD. The shortest version of the AUDIT (AUDIT-3) had excellent results. In the assessment of concurrent validity, expressive correlations were found between the AUDIT and its brief versions (0.91-0.99). In all analyses, the CAGE had satisfactory results, although inferior to those of the brief versions of the AUDIT. The results confirm the validity and reliability of the Brazilian versions of the WHO-5, PHQ-2, FTND, and HIS, as well as the efficacy of all the brief versions of the AUDIT, demonstrating that their psychometric qualities are as satisfactory as those of the full version of the instrument and superior to those of the CAGE. Therefore, the use of these instruments is supported by their viability and validity, which encourage their use in routine clinical practice and research settings. The results also showed that smoking and alcohol abuse and dependence are important predictors of depressive disorders and that alcoholics had a fourfold higher chance of being smokers. (AU) | |