Regresión logística / Logistic regression

La regresión logística es un grupo de técnicas estadísticas que tienen como objetivo comprobar hipótesis o relaciones causales entre una variable dependiente categórica y otras variables independientes que pueden ser categóricas y cuantitativas. Mediante este modelo pretendemos estudiar la probabilidad de que ocurra el evento estudiado en función de unas variables que suponemos que son relevantes o influyentes. En este método es necesario detectar las variables modificadoras de efecto y las de confusión. Sus parámetros se estiman con el método de la máxima verosimilitud a través de un proceso con sucesivas iteraciones (AU)

Logistic regression is a group of statistical techniques that aim to test hypotheses or causal relationships between a categorical dependent variable and other independent variables that can be categorical and quantitative. Through this model we intend to study the probability that the event studied will occur based on some variables that we assume are relevant or influential. In this method it is necessary to detect effect modifier and confounding variables. Its parameters are estimated with the maximum likelihood method through a process with successive iterations (AU)

[Logistic regression]. / Regresión logística.

Logistic regression is a group of statistical techniques that aim to test hypotheses or causal relationships between a categorical dependent variable and other independent variables that can be categorical and quantitative. Through this model we intend to study the probability that the event studied will occur based on some variables that we assume are relevant or influential. In this method it is necessary to detect effect modifier and confounding variables. Its parameters are estimated with the maximum likelihood method through a process with successive iterations.

La curva ROC / ROC curve

La curva ROC es una herramienta estadística que se utiliza para evaluar la capacidad discriminativa de una prueba diagnóstica dicotómica. Se trata de curvas en las que se presenta la sensibilidad en función de los falsos positivos (complementario de la especificidad) para distintos puntos de corte. Son útiles para elegir el punto de corte más adecuado de una prueba, conocer el rendimiento global de esta y comparar la capacidad discriminativa de 2 o más pruebas diagnósticas. (AU)

The ROC curve is a statistical tool used to evaluate the discriminative capacity of a dichotomous diagnostic test. These are curves in which sensitivity is presented as a function of false positives (complementary to specificity) for different cut-off points. They are useful for choosing the most appropriate cut-off point for a test, knowing its overall performance and comparing the discriminative capacity of 2 or more diagnostic tests. (AU)

[ROC curve]. / La curva ROC.

The ROC curve is a statistical tool used to evaluate the discriminative capacity of a dichotomous diagnostic test. These are curves in which sensitivity is presented as a function of false positives (complementary to specificity) for different cut-off points. They are useful for choosing the most appropriate cut-off point for a test, knowing its overall performance and comparing the discriminative capacity of 2 or more diagnostic tests.

[Intraclass correlation coefficient]. / Coeficiente de correlación intraclase.

The intraclass correlation coefficient makes it possible to measure the general concordance between two or more measurements that involve variables of a quantitative nature, obtained with different measuring instruments or evaluators. It is based on an analysis of variance model with repeated measures. Its use is only possible if there is normality of the distributions of the variables, equality of variances and independence between the errors produced by the observers.

[Early detection and prevalence of risk of eating disorders in Primary Care in Guadalajara city]. / Detección precoz y prevalencia de riesgo de trastorno de la conducta alimentaria en Atención Primaria en Guadalajara capital.

OBJECTIVES: To know the percentage of adolescent patients at risk of developing an eating disorder (ED); to identify the prevalence of risk for Anorexia Nervosa, Bulimia Nervosa and Unspecified Eating Disorders (NERD) according to age, sex, health centers, nationality and family data; determine the percentage of young people with subclinical forms and study the feasibility of using tools for the early detection of this pathology in Primary Care. MATERIAL AND METHODS: Observational, cross-sectional and multicentre study developed in the five health centers of the city of Guadalajara. Sample of 291 adolescents (124 males and 167 females) obtained through stratified, multi-stage and proportional random sampling according to age, sex and health centers. Sociodemographic, anthropometric and family variables were studied and the EAT-26, SCOFF and EDE-Q6.0 tests were applied in the first round and the EDE-12 interview in the second round. RESULTS: The global prevalence of risk of eating disorders was 17.5% (95% CI: 13.7%-19.4%). The variables, having siblings, being single-parent and having obesity were statistically significant. Among the various types of EDs, the NERD were statistically significant with respect to Anorexia and Bulimia. The subscale concern for the silhouette was statistically significant with respect to the others. The SCOFF questionnaire showed a sensitivity of 91.6% and a specificity of 91.0% and the EDE-Q, 99.3% and 91.7% respectively. CONCLUSIONS: High percentage of adolescents at risk of developing an eating disorder. The unspecified type of eating disorders and the concern for the silhouette subscale predominate. The SCOFF test proves to be a good screening instrument in Primary Care.