A large hospital network wants to compare two pain management protocols after surgery. They decide to implement Protocol A at their 5 urban hospitals and Protocol B at their 5 rural hospitals, then compare patient pain scores. A statistician recommends against this design. What is the most serious concern?
- AFive hospitals per group is not enough for statistical power
- BHospital setting is confounded with the protocolCorrect
- CThe study should use patient satisfaction instead of pain scores
- DThe patients were not blinded to which protocol they received
Why B is right
Hospital setting (urban vs. rural) is completely confounded with treatment. Urban and rural hospitals may differ in patient demographics, surgical teams, resources, and recovery environments. Any observed difference in pain scores cannot be attributed to the protocol because it is impossible to separate the effect of the protocol from the effect of the hospital setting.
Why the other 3 choices miss
- A
- Incorrect because while 5 hospitals per group is small, the primary concern is confounding, not sample size.
- C
- Incorrect because the choice of outcome measure is not the fundamental design flaw; pain scores are a valid measure.
- D
- Incorrect because while blinding is important, the critical flaw is that treatment and location are completely confounded.