(Click the link to comment and to vote – voting not working through email, sorry!)
An investigator wishes to know whether PED patients seen with bloody diarrhea are prescribed antibiotics in the PED are more likely to develop hemolytic uremic syndrome (HUS). Over a 10-year period, several cases of patients with HUS diagnosed after a PED visit for diarrhea are identified, and for each of these patients, 3 other patients seen in the PED for diarrhea that did not develop HUS matched for age, gender, duration of illness, and stool culture findings are identified. The two groups of subjects are compared regarding whether they received PED antibiotics or not.
Loading ...
October 23, 2018 at 2:55 pm
E) Case-control study. In a case-control study similar subjects are divided into groups based on the presence or absence of the outcome of interest, and comparisons are made regarding the risk factors or predictors. Associations can be made but never proven in these types of studies. Case-control studies are particularly useful for studying rare outcomes.
In a cross- sectional study, the presence of a risk factor and the presence of an outcome are reviewed simultaneously in a population. This is the weakest type of study for causal inference.
In prospective cohort studies, subjects are divided into groups based on the presence or absence of a presumed risk factor and are followed prospectively. These are powerful studies but are often costly and take a long time to complete. Retrospective cohort studies (often chart reviews) identify a cohort and look back in medical records or other sources of information to assess risk factors. These are less rigorous than prospective cohorts since there can be missing information.
In randomized clinical trials, subjects are matched in as many ways as possible. One group receives an intervention and both groups are blinded. Randomized clinical trials, if conducted well, remove many sources of bias and enable improved causal inference.
In a meta-analysis, many studies that address the same topic are statistically analyzed, providing a consensus of multiple investigations. Often methodology and definitions vary from study to study, potentially posing some limitations to performing meta-analysis.