ABSTRACT
Nonalcoholic fatty liver disease （NAFLD） is an abnormal lipid metabolic disorder of the liver characterized by accumulation of a large amount of lipids in the liver， and it is currently the most common liver disease around the world. Mendelian randomization （MR） incorporates genomic data into traditional epidemiological study designs to infer the causal relationship between exposure factors and disease risk. In recent years， MR has been widely used in studies on inference of the etiology of NAFLD. This article systematically summarizes the advances in the application of MR in NAFLD research， so as to provide new ideas for understanding the nature of the disease and scientific interventions.
ABSTRACT
The purpose of this paper was to introduce the methods of identifying causal effects based on instrumental variables， distinguishing different models with data， and using SAS software to realize calculation. Firstly， the four main contents of causal graph theory were introduced， including sources of association， statistical properties of causal models， identification and adjustment， and instrumental variables. Secondly， for two examples and with the help of the CAUSALGRAPH procedure in SAS/STAT， the following two tasks were completed： the first task was to identify causal effects using instrumental variables； the second task was to use data to distinguish different models.
ABSTRACT
Mendelian randomization is an approach using the genetic variants as instrumental variable to estimate and assess the casual relationship between exposure of interest and outcomes. As a valid instrument, genetic variants have to meet the assumptions of strong correlation with exposure but without pleiotropic effect with the outcomes. However, pleiotropy of the variants is usually inevitable, owing to the existence of complex biological effects. Thus, correction methods related to pleiotropic bias are introduced in this paper regarding the selection of instrumental variables, testing of invalid instrumental variables, construction of pleiotropic effect correction models and sensitivity analysis of the robust results. For practical application, investigators should take consideration on the following areas including the types of data, sample size and other relative aspects, thereby selecting the suitable method for the inference of consistent and robust casual estimation.
ABSTRACT
Los conceptos de causalidad y sesgo están en la base de la investigación biomédica moderna, desde el análisis de cientos de factores de exposición, hasta los megaestudios para evaluar intervenciones. Los consumidores de estos productos de la investigación, vemos con desconcierto, que una conclusión que se formula hoy, se pone en duda mañana, y se desecha poco tiempo después, para eventualmente ser retomada en el futuro bajo otras ópticas u otros presupuestos. Aunque no es el único responsable, el sesgo metodológico juega un papel importante como determinante de esta realidad. Este artículo tiene el propósito de destacar el concepto de sesgo, relevante, entre otras posibles acepciones, para la investigación biomédica contemporánea, y su asociación con la definición técnica de confusión, exponer la visión moderna sobre el significado práctico de una causa y examinar críticamente dos modernos recursos analíticos para afrontar el problema del sesgo y la causalidad: los puntajes de susceptibilidad y las variables instrumentales
The concepts of causation and bias are crucial to modern biomedical research, ranging from the analysis of hundreds of exposure factors to megatrials, in order to assess the impact of interventions. As consumers of these research products, we are amazed that a statement made today is put into question tomorrow, discarded afterwards, and eventually retaken in the future from different perspectives or under different assumptions. Although the methodological bias is not the only culprit, it plays an important role as determinant of this reality. This paper intended to clarify the concept of bias, which is relevant, among other possible meanings, to contemporary biomedical research, and its association with the technical meaning of confounding. Other objectives were to present the current vision on the practical meaning of cause in epidemiological causal inference, and to critically review two modern analytical tools to deal with bias and causation such as propensity scores and instrumental variables