Clinical Research: Research Design Comparison/Contrast

Types of Research Design  

Definition

Pros/Cons

Examples

Randomized controlled trial (RCT)

True experimental design which manipulates a therapeutic intervention; participants in the research are randomized to experimental or control groups; control may be placebo or standard treatment; answer the question: "Does the intervention make a difference?"

PRO: Randomization helps control for bias (inherent differences among groups); use of control groups provides better comparison, helps mitigate placebo effect; blinding (masking) when possible also helps; best for establishing efficacy; provide strong evidence of causality

CON: Not possible for some kinds of research that may present ethical dilemmas; take a long time; require sound methodology; expensive

George, J., Raskob, G., Vesely, S., Moore D Jr, ., Lyons, R., Cobos, E. et al. (2003). Initial management of immune thrombocytopenic purpura in adults: a randomized controlled trial comparing intermittent anti-D with routine care. American Journal of Hematology, 74(3), 161-9.

Cohort study

Data collected from a defined group of people (cohort); look forward in time, from an exposure, intervention, or risk factor to an outcome or disease; answer the question: What will happen?

PRO: Observe people in a natural setting; ethical; timing/time intervals of data collection provided possible associations of results

CON: No randomization; groups with possible inherent differences (selection bias);  attrition (participant dropout) may bias results; may require long follow-up; expensive

Glanz, J., France, E., Xu, S., Hayes, T.,  & Hambidge, S. (2008). A population-based, multisite cohort study of the predictors of chronic idiopathic thrombocytopenic purpura in children. Pediatrics, 121(3), e506-12.

Case control study

Look backward in time, from an outcome or disease to a possible exposure, intervention, or risk factor; answers the question: What happened?

PRO: Quick and cheap; good for rare disorders with a long time between exposure and outcome; efficient-data often collected from record reviews; convenient (patient already have disease)

CON: No randomization; groups with possible inherent differences (selection bias); difficult to choose appropriate control group

Berends, F., Schep, N., Cuesta, M., Bonjer, H., Kappers-Klunne, M., Huijgens, P. et al. (2004). Hematological long-term results of laparoscopic splenectomy for patients with idiopathic thrombocytopenic purpura: a case control study. Surgical Endoscopy, 18(5), 766-70.

Case series/case report

Describe observations that have occurred in a patient or a series of patients; call attention to unusual association; bring attention to a unique case

PRO: Preliminary observation of a problem; new or rare diagnosis; low cost; can lead to further studies

CON:  No control group; no statistical validity; not planned; no research hypothesis; limited scientific merit

Galbusera, M., Bresin, E., Noris, M., Gastoldi, S., Belotti, D., Capoferri, C. et al. (2005). Rituximab prevents recurrence of thrombotic thrombocytopenic purpura: a case report. Blood, 106(3), 925-8.