Types of Scientific Studies: The Basics

by Thomas Lynch, MD

The purpose of this article is to provide a basic overview of the major types of studies used to answer important clinical questions. When designing a study the first decision is whether to answer a clinical question prospectively or retrospectively. Prospective studies start in the present and collect data moving forward in time. Examples include observational cohort studies and experimental clinical trials. Retrospective studies are designed to look back on previously collected data to answer clinical questions. Retrospective studies are all observational in that they analyze data or present descriptive information that has already been obtained.

Case Series/Report: Retrospective studies that focus on a group of patients with a similar diagnosis. These studies often report on new or rare conditions where the prevalence of the pathology is not sufficient enough for a higher evidence level study. Case series are often limited to three patients and studies including more than three patients require IRB approval. Case reports are descriptions of rare disease or injury outcomes in a single patient. It is important to note these studies are purely descriptive in nature and lack the ability to provide analytical information for answering a research question. 

Case-control studies: Retrospectively look at an outcome or disease within a particular group and compare exposure histories of this group with a control group that does not have the particular outcome being investigated. For example, if you are trying to figure out why a group of tibia fractures went on to nonunion a case control study would compare this group with a group of tibia fractures that healed. The investigator would look at characteristics of patient history such as smoking, open versus closed fracture, and infection as possible variables that would predispose a tibia fracture to nonunion. These studies allow investigators to determine odds ratios associated with increased or decreased risk of developing the outcome of interest. The main limitation of case-control studies is obtaining a comparatively appropriate control group. In the aforementioned example, similarly based age and gender distribution of the outcome and control groups would strengthen the study.

Cohort Studies: Most commonly prospective studies that look at the relative risk of developing an outcome associated with a particular exposure. For example, comparing a group of smokers and nonsmokers to develop a relative risk of getting lung cancer. This is the only type of observational study that can directly establish risk of disease. One advantage of this type of study is that multiple different outcomes can be studied between the exposure group and the unexposed group. It is due to this fact that cohort studies can be conducted retrospectively in certain circumstances. Lastly, cohort studies are particularly good for studying rare exposures, but limited in their ability to look at rare outcomes.

Randomized Control Trial: These studies designed to prospectively allocate interventions and track clinical outcomes with the goal of answering a specific hypothesis. Interventions are randomized amongst the subjects. When the study permits both the patients and physicians are blinded to reduce bias and improve the quality of the experimental process. When effectively executed these studies provide the most powerful level of evidence in dictating if a treatment intervention is superior, inferior, or equivalent to the current industry standard. Even these studies may not produce reliable results when applied to general practice as they can be conducted in special homogenous populations, utilize varying statistical approaches, and have industry driven funding to name a few potential problems experienced with randomized control trials.

Systematic Review: Secondary study focusing on a specific answerable question. The goal of these studies is to critically evaluate available evidence to provide a concluding answer for a particular clinical question. Systematic reviews are transparent and conducted in such a way they can be replicated by anyone. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines provide a checklist for standards of the design, execution, and presentation of a well conducted systematic reviews. These reviews can be quantitative, as described in meta-analysis section, or qualitative in providing a conclusion based on review of available research meeting predefined inclusion criteria. Systematic reviews are limited by the quality of the studies included. This means a review that is five level one studies and one level four study it automatically is a level four study.

Meta-Analysis: A quantitative secondary research study that combines data from multiple individual studies. The requirement to be a meta-analysis is the combination of quantitative data from two or more similar studies. The inclusion of all relevant studies reporting on the topic on hand is not required, however a meta-analysis with absolute inclusion of all studies deemed relevant would also be considered a systematic review. These studies combine multiple randomized control trials that fit within defined criteria. The more stringent the criteria reduces the number of studies included but strengthens the internal validity of the study. For further reading on the nuances of conducting systematic reviews and meta-analyses, Research Pearls: The Significance of Statistics and Perils of Pooling. Part 3: Pearls and Pitfalls of Meta-anlyses and Systematic Reviews by Harris et al. from Arthroscopy 2017 is a great reference.