Risk for Clinically Relevant Adverse Cardiac Events in Patients With Chest Pain at Hospital Admission. (NCBI Article Link)
Is this a prospective or retrospective study? How can you tell? Give some strengths and weaknesses of this type of study.
The authors state that “all data were collected before our study began,” so this is a retrospective study. A study is considered “retrospective” if the events being measured occurred before the study was defined. Retrospective studies allow for less expensive ways to study larger populations than other study types. They can be useful for studying rare events. Unfortunately, because the investigator has no control over the conditions under which the data are collected, this type of study is prone to biases and confounders. Consequently, retrospective studies can be used to identify associations but cannot verify causation.
Suppose the authors were unable to randomize or blind treatment allocation in this study. Would their conclusions be as convincing? Why? (Consider how failure to randomize or blind might change their data in your answer.)
If EMS providers were allowed to choose which treatment to give to their patients, a selection bias would likely occur. First, it is unlikely that any EMTs would knowingly administer a placebo when prior practice dictated giving an active drug. Second, EMTs might believe one drug is superior to the other, and thereby match a drug with a certain patient profile. Thus, the populations that receive each drug would likely be different sizes and comprised of different types of patients. In this less robust version of this study, we wouldn’t be certain if differences in response to the drug were due to the drug administered or some baseline characteristic(s) of the patient groups themselves.
A focused investigation of expedited, stack of three shocks versus chest compressions first followed by single shocks for monitored ventricular fibrillation/ventricular tachycardia cardiopulmonary arrest in an in-hospital setting. (NCBI Article Link)
This is a great example of a natural or before-and-after experiment. How is a before-and-after experiment defined? This type of experiment typically cannot be used to establish causation due to the effect of confounders. Give an example of a confounder that might have influenced the findings in this study.
Before-and-after studies compare study groups with common characteristics prior to and after a policy or procedure change. For example, imagine a study of mortality in patients involved in motor vehicle accidents before and after seat belt laws went into effect. A confounder is some concurrent change in the study environment that systematically influences the outcome independent of the policy change. For example, curiously the last US recession started in December 2007. Perhaps that group of patients from 2008-2011 had less to live for?