2.1.9 Inclusion and exclusion criteria
A. Background & Definitions
Inclusion criteria are criteria that qualify subjects or specific observations for inclusion in the study or in the study analysis.
Exclusion criteria are criteria that disqualify subjects or specific observations or data points from inclusion in the study or in the study analysis.
Inclusion / exclusion criteria may apply not only to individual subjects or observations but may also be pre-specified as "test acceptance criteria” or "pass / fail criteria on experimental level". For example, one may include a positive control in the study and would therefore need to pre-specify whether the entire study will be considered as “failed” if a positive control fails. Another example is given by pre-specification of performance of vehicle controls (e.g. against historical values).
B. Guidance & Expectations
Inclusion and exclusion criteria should be pre-specified when the experiment is being designed (i.e. should be established prospectively) and it should be explicitly documented (e.g. listed and saved in the study protocol).
The best way to incorporate pre-specification into your experiment is by storing pre-specified information such as inclusion / exclusion criteria in a laboratory notebook (e.g. electronic) as part of a study protocol. Pre-specification may also be conducted by means of 2.1.11 Preregistration of the study protocol.
If exclusion criteria were not pre-specified, certain values may still be excluded, but only if the decision is taken by a person blinded to the treatment conditions or other study design aspects that could bias the decision (i.e. ideally, the decison is made during the primary data review and verification before the study is unblinded).
Existence of any inclusion and exclusion criteria must be explicitly stated in the study reports (including scientific publications).
Inclusion criteria (examples):
- Species, strain, sex, weight and age of animals included in the study
- Transgene copy number
- Conditions/ performance criteria that must be fulfilled (e.g., after amphetamine administration, rats showing ≥5 full turns/min in the direction ipsilateral to the unilateral nigrostriatal lesion are selected for further experiments; Dekundy et al. 2006)
- Area of the calibration curve where the measured values should be in order for the study to be declared valid
Exclusion criteria (examples):
- Statistical outliers
- Pre-specify methods to be used to identify outliers and decision-making
- Exclusion of subjects, observations or data points if the study or data collection was or could be affected by unforeseen environmental or technical circumstances:
- Loud noise or vibration (e.g. due to ongoing construction in the facility)
- Malfunctioning of the research equipment
- Technical errors during the study conduct (e.g. failed injection procedure)
- Exclusion of subjects, observations or data points due to animals' health (e.g. exclusion of SOD1 mice that died from non-ALS causes; Scott et al., 2008)
PLEASE DO NOT FORGET
- To check whether there are any pass/fail criteria on experimental level that may be applicable
* To consider adding this subject to a training program for new employees or refresher training
C. Resources
- Online Grubb’s test for outlier detection - [1]
Guidelines on reporting of inclusion and exclusion criteria (in vivo research):
back to Toolbox
Next item: 2.1.10 Plausibility check
