Study Designs Study Documents

Study designs Aims of epidemiological research • Descriptive studies 🡪 Who? What? Where? When? o “Describe health status of a population” o Describes disease by person, place, and time • Analytic studies 🡪 why? o “Identify causes of disease” o Analyses associations between exposure and outcome Observational vs. experimental • Observational studies o Investigate and record exposures and outcomes as they naturally occur o Epidemiologist has an ‘observational’ role o May be descriptive or analytical • Experimental (intervention) studies o Involves an active attempt to change some factor to alter the progress of a disease or outcome o Epidemiologist ‘intervenes’ in determining type of exposure o Always analytical Why are some studies stronger? • Because of confounding (to confuse/muddle) and bias • In epidemiology, confounding is a mixing or muddling of the effects of 2 or more factors (“exposures”) 🡪 all or part of an apparent (observed) effect is due to some factor other than your primary exposure of interest • There is a lack of exchangeability between exposed and unexposed groups 37 In practice • Studies usually begin simply, as descriptive studies (case reports, case series, cross-sectional studies) • Progress to more complex analytic studies (cohort, case-control studies) • And then to experimental studies (RCTs , where appropriate) Analytical study types • Observational o Case-control studies o Cohort studies • Experimental studies o Randomised controlled trials o Quasi-experimental studies Randomised controlled trials • Best way to minimise confounding 🡪 makes comparison groups exchangeable • Involves an intervention • At least two groups o Experimental o Control or comparison • Individuals are followed to ascertain effects of intervention • Limitations: o Generalisability of findings may be limited if the criteria to get into the study are too narrow relative to the population to which you would like to apply the results ▪ E.g. A study that includes only men aged 20-40 years from a single ethnic background: How applicable are the results to women, to older people, or to people from other ethnic backgrounds? o Not all questions can be answered by an RCT 🡪 ethics, practicalities ▪ E.g. the feasibility of an RCT to determine the effectiveness of parachutes in averting death when jumping from a plane, or the ethics of using an RCT to study the effect of smoking during pregnancy on newborns Cohort studies • Compare the rate of disease in a group of people exposed to a factor, with a group that is not exposed • Main feature is the observation of large numbers of people over a long period of time • Aim is to identify associations between possible risk factors or prognostic factors and outcome of interest • Usually prospective (but may be retrospective) 38 • Advantages o Valuable when exposure is rare o Can examine multiple effects of a single exposure o Can often better determine temporal relationship between exposure and outcome o Allows measurement of incidence • Disadvantages o Validity affected by losses to follow-up (selection bias) o Biased distribution of confounders (compared to RCT) o Inefficient for evaluation of rare diseases o If prospective, expensive and time-consuming o If retrospective, requires good records • Interpretation o Who is risk? 🡪 All participants (exposed and unexposed) must be at risk of developing outcome o Who is exposed? 🡪Clear unambiguous definition of exposure at the outset o Who is an appropriate unexposed group? 🡪 Unexposed should be similar to the exposed in all respects except for the lack of exposure o Have outcomes been equally assessed? 🡪 Defined in advance. Clear, specific and measurable Case control studies • Subjects are selected on the basis of disease status o Cases have disease (or outcome) of interest o Controls: no disease (or outcome) of interest • Two groups compared with respect to past exposure history – determined retrospectively • Ideally should study incident (new) cases • Cases and controls should be representative of all cases • Study base should be well defined: time, people, place • Defining the source population i.e. eligibility criteria for cases to enter the study: o i.e. a case definition where people who develop the disease of interest in the source population are included as cases o If cases are a representative sample of all cases in an identified population, and controls are sampled directly from this source population 🡪 populationbased o Doesn’t have to be the general population, geographic boundary Cross sectional studies (analytical) • Measure prevalence of exposure and disease at the same time • Rate of exposure is correlated with rate of disease frequency (or other health outcome)