역학 연구에서의 스터디 디자인


역학 연구에서의 스터디 디자인을 한 눈에 볼 수 있는 테이블이다. 각각의 스터디 디자인에서 기술 연구, 분석 연구를 할 때 일반적으로 어떤 measure를 사용하는지, 그리고 스터디 디자인마다 장, 단점은 무엇인지 파악할 수 있다.


Table 1: 
Main study designs in classic epidemiology.


Case reportsCase seriesEcological studiesCross sectional studiesCase controlCohortIntervention trials

Unit of study(Single case) 
individual
(>1 cases) individualPopulationIndividualIndividualIndividualIndividual

DescriptionDescribes unusual characteristics of a caseA group of casesComparing populations in different places at the same time or, in a time seriesStudy of population at, “point-in-time,”Study of two groups of subjects: (case; disease of interest and control; disease-free)Study of two groups of subjects (exposure and non-exposure groups)Study and examine two groups of subjects (intervention and control groups)

DirectionPresentPresentPresentPresentReverseForwardForward

Type of measurementReporting, descriptionReporting, descriptionCorrelationPrevalence, associationOdds ratioPrevalence, incidence, relative risk, attributable riskPrevalence, incidence, relative risk, attributable risk

AdvantagesQuick, having clinical importance, opportunities for physicians to exchange of thoughtsQuick, having clinical importance, opportunities for physicians to exchange of thoughtsQuick, inexpensive, group-level studies may also be the only way to study the effects of group-level constructs, for example, lawsEasy, inexpensive, useful for investigating fixed exposures such as blood group, most convenient in outbreaks of diseaseRelatively inexpensive good for rare diseases Efficient in resources and timeBetter for rare exposures, ability to determine causality relationsThe strongest evidence for causality, control of unknown confounders, fulfils the basic assumption of statistical hypothesis tests

limitationInability to determine statistical relations and analysisInability to determine statistical relations and analysisEcological fallacySusceptible to selection bias and misclassification, difficult to establish a putative, “cause”, Not good for rare diseases or rare exposuresSusceptible to selection bias and misclassification bias, may be difficult to establish that, “cause,” preceded effectCostly and time consuming, susceptible to selection bias. Relatively statistically inefficient unless disease is common.Expensive, time consuming, sometimes ethically generalizability problem

https://www.hindawi.com/journals/isrn/2013/952518/tab1/