2025 – PAGE 391 – STATISTICS

SYSTEMATIC REVIEW AND META-ANALYSIS

Systematic reviews of randomized controlled trials are considered to have the greatest validity. Limitations come from the choice of relevant studies, “heterogeneity” (meaning the studies were not performed similarly), and the generalization of results.

RANDOMIZED CONTROLLED TRIALS

These are the gold standard of research designs and are most likely to yield valid information about the benefits and/or harms of an intervention. Strength comes from the ability to control many aspects of the study. This limits confounding variables. Weaknesses include the expense and time associated with them. Studying rare diseases is also difficult.

COHORT STUDIES

A cohort is a group, or sample of people, that can be followed over time and evaluated for exposures (or lack of exposures) that may be associated with the development of disease. The major strength of cohort studies is that a researcher can look at the risk of a particular exposure causing a particular outcome. The major weaknesses include the large sample size needed, and the difficulty in controlling exposure to other variables that could affect the outcome being studied (confounding variables).

PEARL: Because a large sample size is needed, cohort studies are NOT very useful in evaluating rare diseases.

PROSPECTIVE COHORT STUDIES

In a prospective cohort study, a cohort can be followed and evaluated for risks of developing a future outcome. Most commonly, this design is used to follow a cohort of people with and without an exposure in order to see if a particular disease develops in the future. The same reasoning can be used in patients with a disease and without a disease in order to follow the cohort for research on prognosis of a disease.

RETROSPECTIVE COHORT STUDIES

In a retrospective cohort study, also called a historical cohort, the conceptual design is identical to the prospective cohort study, but the events have already happened. The researcher selects cohorts with and without exposure at a starting point in the past, then follows them forward in time from that starting point, but still in the past with respect to the time of the research. For example, a researcher might use the records of a large health care system to select 100 five-year-olds who were TB skin test positive in 1990 and 100 who were negative, then follow their outcomes over 20 years. The key features of a retrospective cohort study are (1) the cohorts are identified by risk or exposure at the beginning of the period, not by eventual outcome and (2) as in a prospective study, the results are only valid to the extent that the entire cohort can be followed over time for the events of interest.

PEARL: Retrospective studies are always weaker than the best prospective studies, due to various complications and biases, but are cheaper and, of course, much faster than long-term prospective studies.