AtlasSession 3L8: Study designp.86–90
Design tools
Fig. 1
PBL 3 mouse TACE study (IPN hydrogel vs lipiodol, 5 mice per group): risk of bias in seven domains, before and after design fixes
Fig. 1, A checklist before the experiment (First look). A mouse study: a new gel vs the standard oil for blocking a liver tumour’s blood supply, 5 mice per group, tumour damage scored under the microscope. No changes made yet. Traffic lights: Groups alike at the start? Can’t tell; Who got into the study? Low risk; Who got what, recorded right? Low risk; Same care in both groups? Can’t tell; Every mouse counted? Can’t tell; Results scored fairly? Can’t tell; All results shown? Can’t tell.
- Low risk
- 2/7
- Can’t tell
- 5
- High risk
- 0
Bias Scenario Classifier · First look
Level 1 · First look · Lecture 8 · p.86–90
Use a checklist before an experiment to catch problems early.
Use the risk-of-bias domains to design a study before collecting data and revise it until each domain is low risk; know the NC3Rs EDA and clinical-relevance tools exist.
Pre-flight checklist (the step you are on)
Pilots run their checks before take-off, not after landing.
Pilots run their checks before take-off, not after landing. The mouse plan before any changes: 2/7 areas low risk, 5 we can’t tell.