It is important to avoid bias (various distortions that can lead to ambiguous or erroneous conclusions). This can happen when there is confounding (i.e., when a variable other than the experimental intervention of interest is varies along with it). In such cases the results may be a consequence of the confounding variable, rather than the experimental intervention.

		This chart describes several types of confounding and how they can be avoided. Making sure to avoid all types of potential bias is a major challenge of designing excellent clinical trials.

 

PICO Outcome Issues (Endpoints)

Clinical trials assess the effect of different treatments on the course of disease by measuring specific outcomes (or endpoints) that indicate an intervention's effectiveness. Endpoints differ depending on the phase and type of trial.

The choice of outcomes or endpoints typically involves trade-offs due to conflicting priorities such as speed, completeness, and clinical value.

 

Primary endpoints (those of highest interest) are selected and the trial is designed to ensure that these endpoints can be adequately assessed.

• Tumor response rate (TRR) - the proportion of trial participants whose tumor was reduced in size by a specific amount, usually described as a percentage. If 7 of 10 patients responded, the response rate is 70 percent.
• Disease-free survival (DFS) - the amount of time a participant with localized disease survives without cancer occurring or recurring, usually measured in months.
• Progression-free survival (PFS) - the amount of time a participant with advanced disease survives without cancer occurring or recurring, usually measured in months.
• Response duration (RD) - the response duration is occasionally used to analyze the results of the treatment for the advanced disease. The event measured is progression of the disease (relapse). This endpoint involves selecting a subgroup of the patients and measures the length of time of the response in those patients who responded. The patients who don't respond to treatment aren't included in the endpoint analysis.
• Overall survival (OS) - the amount of time a participant lives, typically measured from the beginning of the clinical trial or the date of diagnosis until the time of death.

 

Secondary endpoints are of lesser interest. They are also specified in the protocol but the trial may not be powered (have enough numbers) to adequately assess them.

• Side-effect profile
• Quality of life (QOL)

The examples above can be used as primary or secondary endpoints depending on the study, although that is rarely done. Additionally, many current clinical trials include the collection of a host of demographic and biomarker measures (sometimes referred to as secondary endpoints) that are analyzed to identify interesting questions for future study. Analysis of these variables is call correlative science.