Designing an Outcomes Research Study

A health outcomes research study design is a plan for executing the study. At a minimum, the design depicts the groups studied; for example, treatment and control group, instances of the treatment and the timing, and frequency of health outcomes measures. The design provides a high-level overview of the health outcomes study and insights into the plan for analysis. Finally, the design should specify whether the individuals studied are randomly assigned to either receive the treatment of interest or no treatment , also referred to as a control group. Control over treatment assignment through randomization is the basis for distinguishing two types of outcomes studies: experiments and quasi-experiments. Random assignment of subjects is central to controlling for extraneous differences between groups, but it does not guarantee comparability ; it simply asserts that any differences are due to chance. Without randomization of study participants, the outcomes researcher runs the risk of individuals with particular characteristics having a higher probability of being included in the study or one of the study groups. These differences can arise from patient self-selection or from clinician decisions about who should get treatment. Selection bias or self-selection has the potential to confound the treatment-outcome relationship, thereby biasing results. Some of these differences can be measured and controlled for in the analysis , but others may remain unmeasured and uncorrected. Overall, selection bias may be the greatest threat to the validity of health outcomes research studies.