Gibson D. Lewis Library Libguides
Evidence-Based Practice
Questions? Ask Us!
AskALibrarian@unthsc.edu
817-735-2465
Which study design is best for my question?
Evidence is not created equal and higher quality evidence leads to better decisions! Different studies are more valid for different types of clinical questions. Remember- evidence alone isn't enough and should be combined with clinical expertise and the patient's preferences/needs.
Clinicians generally have five types of clinical questions: therapy, prognosis, harm/etiology, diagnosis, and prevention. The type of question determines the best evidence to look for and the most appropriate study design to answer it.
Use the table below to identify which study types offer the strongest evidence. Please note that the evidence should always be integrated with clinical expertise and the patients preferences and needs when making decisions.
Study Designs by Question
| Type of Question | Type of Study | |
|---|---|---|
| Therapy | Assessing how interventions impact patient outcomes, such as symptoms, function, morbidity, mortality, and costs. For a given disease or condition, does this treatment offer more benefits than risks? |
Systematic Review RCT Lab/Animal Study |
| Prognosis | Predicting the likely progression of a patient's condition based on factors other than interventions. What is the expected clinical trajectory over time, and what complications are likely to arise from the disease or condition? |
RCT Cohort Study Case-Control Study Cross-Sectional |
| Harm/Etiology | Determining the impact of potentially harmful agents on outcomes that matter to patients. Did this harmful exposure contribute to the development of the disease or condition? |
RCT* Case-Control Study Cohort Study Lab/Animal Study |
| Diagnosis | Assessing a test's ability to distinguish between individuals with and without a specific condition or disease. How accurate is this diagnostic test? |
Prospective, blinded assessment against a recognized benchmark; Cross-Sectional |
| Prevention | How to lower the risk of disease by identifying and addressing risk factors, and how to detect it early through screening. Will this preventive measure decrease or eliminate the likelihood of developing this disease or condition? |
Systematic Review RCT Cross-Sectional |
*Harm data can be part of the results in a randomized controlled trial (RCT) as adverse effects of a treatment. However, depending on the intervention, it might be unethical to intentionally expose patients to potential harm.
Table adapted from Evidence‐based Medicine; How to practice & teach EBM. David L. Sackett. Churchill Livingstone, 2000, p.19
Studies Defined
Case Series and Case Reports
A detailed account on the treatment of an individual patient. These are descriptive reports that have no statistical validity as they do not use control groups to compare outcomes.
Case Control Studies
These studies compare patients with a specific condition to those without it. The investigation starts with patients who have the outcome and examines potential exposures retrospectively. Researchers use medical records and patient recall for data to identify factors or exposures that may be linked to outcome. These studies are less reliable than randomized control trials and cohort studies, because finding a statistical association does not prove causation.
Cohort Studies
This study tracks a large group of people over time who are already using a certain treatment or have been exposed to something. Researchers compare this group to another similar group that has not gone through the treatment or have been exposed. These are observational studies and not as strong as randomized control trials (RCT) because the groups might differ in other ways then the exposure being studied. On the other hand, they can be more reliable than case-controlled studies because they collect data going forward.
Cross-Sectional Studies
These studies examine data at a single point in time, assessing the relationship between a specific exposure and an outcome in a defined population. They can identify associations but do not establish causation. Cross-sectional studies are useful for generating hypotheses and understanding the prevalence of conditions.
Lab/Animal Studies
These studies involve experiments conducted in controlled environments, often using animal models to investigate the effects of treatments or interventions before they are tested in humans. While valuable for initial research and understanding biological mechanisms, the results may not always translate to human applications due to differences in biology.
Meta-Analysis
This research looks at multiple well-conducted studies on a topic and combines their results using statistical methods to present them as if they were from one large study. The Cochrane Collaboration is highly regarded as the top authority for systematic reviews and meta- analyses.
Randomized Controlled Trials (RCTs)
These are well-designed experiments that test how a treatment or exposure affects real patients. Researchers use methods like random assignment and blinding to minimize bias then compare the results between the group receiving the treatment and the group not receiving Participants are randomly assigned to different groups and then tracked over time to see outcomes. RCTs can provide strong evidence of cause and effect.
Systematic Reviews
These studies focus on a specific clinical topic and question. They involve a thorough search for studies with reliable methods. The identified studies are reviewed for bias and their results are summarized based on the criteria set by the review. Systematic reviews focus on a specific clinical topic and research question. They involve a comprehensive search for multiple, well-conducted studies using reliable methods, typically narrowing down to a single study type (such as only Randomized Controlled Trials). The selected studies are rigorously evaluated for bias, and their findings are summarized based on predetermined criteria.
