Full text loading...
Chapter 6 : Methodologic Issues in Perinatal Research
This chapter presents an overview of the methods and study designs used to assess the possible association between sexually transmitted diseases (STDs) and adverse pregnancy outcomes. The chapter defines what is meant by random and systematic error, explains how they relate to specific types of epidemiologic study designs, and suggests how errors can be minimized at both the design and analysis phases of investigation. The types of epidemiologic research designs and the quantitative estimates of effect that derive from the study designs is reviewed. The various sources of systematic error (bias), with particular emphasis on how they relate to specific types of epidemiologic studies and how they can be reduced by appropriate design and analytic strategies, is delineated. The two broad types of error can be characterized as random error and systematic error. Perinatal epidemiologic studies should focus on outcomes of major importance to the fetus, infant, or mother, including stillbirth, neonatal death, neonatal infant morbidity, major congenital malformations, and maternal mortality and morbidity. Reverse causality bias is particularly likely to occur in cross-sectional studies. Reverse causality is eliminated by ensuring that exposure occurs prior to the development of the outcome. Finally, differential information bias can be minimized by ensuring adequate blinding of subjects, investigators, and care givers. It is also often useful to blind care givers to avoid the occurrence of co-interventions, i.e., additional interventions in one of the treatment arms capable of influencing the outcome and thereby confounding the effect of the study treatment.