Study Design

Resources to assist you in developing a protocol process

The protocol development process should be well thought out and include expert consultations in the areas of statistical analysis, data management, clinical care, statistical analysis methods, basic research, and regulatory affairs.  Key areas of discussion should focus on defining study objectives, variables, control group, blinding, stratification, study populations, inclusion/exclusion criteria, primary and secondary analysis, data collection points, ethics, risk/benefit ratio, study calendar, and the feasibility of executing the study-required procedures.

Common Terms in Study Design


Un-blinded - refers to when the participant, investigator and evaluator [i.e. sponsor] know what treatment the participant is receiving; occurs in open label studies.

Single blind - refers to when one of the parties, the participant, investigator or evaluator do NOT know the treatment.

Double blind - refers to when two of the parties do not know what treatment the participant is receiving (Participant & Investigator or Evaluator).  This process reduces the possibility of study bias towards rating scales or clinical impressions of improvement.  (Participants on active drug might report more favorable responses, staff may be less likely to identify and report responses in no-treatment groups etc.)

Double dummy - is used to blind two agents that are not similar in appearance.  Participants take both dosage forms, one is active and one is placebo.


Parallel group - is when participants receive one treatment for a defined time period.

Crossover - occurs when participants receive more than one treatment during a study.  Participants take each treatment for a defined period of time.  A washout period, where participants take no treatment usually precedes a change in the treatment.

Factorial - occurs when participants receive one of multiple treatments for a defined time period and serve as their own control

Survival - participants receive one or more treatments until an event occurs.


The major purpose of control groups are to allow for identification of changes in symptoms, signs or even morbidity caused by the study medication to be discerned from natural disease progression, participant/research staff expectations, or other treatments.

Uncontrolled – study does not contain a control group

Placebo – used when study compares one or more active treatments to a dosage form that does not contain an active ingredient.

Active – when a study compares two or more treatments that contain different active ingredients.

Historical – when a study compares observations from a current study with similar studies that were previously conducted.

No treatment - when a study compares two groups and only one group receives treatment.


Randomization - is used to equally assign participants to treatment groups, eliminating the need for the investigator to choose which treatment the participant should receive. This process eliminates study bias. Many times randomization is handled through an automated telephone system, other times it may be handled through labels or some other mechanism.

Stratification - is used to control treatment group assignment of participants based on a variable thought to affect the study outcome (a study that is using 2 –10 years olds may be stratified to participants 2-4 years, 5-7 years and 8-10 years; can also be stratified in treatment groups/treatment arms)


Evaluates the actions of a drug in the human body throughout certain time points and provides information on the drugs' absorption, distribution, metabolism and excretion.  Usually occurs in phase I/II studies.


Evaluates physiologic effects of a drug in the human body through procedures such as heart rate, blood pressure or ECG.  Usually occurs as Phase I/II studies.


A retrospective chart review is a systematic investigation of pre-existing information in a medical record which is limited to the information existing in the medical record.  In a prospective chart review the investigator plans to collect and evaluate data that does not exist at the time of the study design.


A descriptive study of an individual person, small group or event reported as an interesting occurrence in which conclusions are only generalizable to the individual or group described.


Allowance by FDA to provide investigational drugs to individual patients for the treatment for serious or life threatening disease, such as cancer or AIDS, where no other comparable or satisfactory drug treatment is available.  The treatment must not pose unreasonable risks and may be effective.  This requires a protocol with an investigator's brochure, IRB approval and informed consent.


When there is not enough time for an IND submission to the FDA, the FDA may allow a one-time shipment of the drug prior to the application submission. The detail in the protocol is limited. Sponsor of the study must make an IND submission as soon as possible. IRB approval is required within 5 days of drug administration.


These studies are ONLY available to patients who are not eligible to participate in a trial being conducted under an IND and have no therapeutic alternatives. The detail of the protocol is limited.


Research involving human subjects who are in need of emergency medical intervention (e.g., comparison of methods for providing cardiopulmonary resuscitation), but who cannot give informed consent because of their life-threatening medical conditions and who do not have an available legally authorized representative to provide consent.