A pharmaceutical company researcher observes that a new antidiabetic drug reduces HbA1c by a statistically significant 0.1% (p < 0.001, n = 50,000). A clinician states this result is not clinically significant. This discrepancy illustrates:
- A Type II error — the study failed to detect a true clinical effect
- B The distinction between statistical significance and clinical significance; large sample sizes detect tiny effects that may be clinically irrelevant ✓
- C Publication bias favouring positive results
- D Regression to the mean artificially inflating the drug's apparent effect
Explanation
With very large sample sizes (n = 50,000), even trivially small differences become statistically significant (p < 0.001) because statistical significance depends on both effect size and sample size. A 0.1% HbA1c reduction is clinically irrelevant (minimum clinically important difference is ~0.4–0.5%). Statistical significance (p-value) indicates the probability of the observed result under the null hypothesis but says nothing about the magnitude or clinical importance of the effect. This is why effect sizes, confidence intervals, and clinical significance thresholds (like MCID) are essential alongside p-values.
Reference: Park's Textbook of Preventive and Social Medicine, 27th ed.
High-yield for: NEET PGINI-CETNExTFMGEUSMLEPLABMRCP
Written and medically reviewed by the StethoPrep medical team.