Definition
Percentage of dismissed diversion alerts where no subsequent evidence of diversion or integrity compromise emerged within six months. Measures triage quality, not alert volume.
The Measurement Gap
The Gap: Supply chain delivery metrics are clean. On-time, in-full, within temperature range. Serialization compliance metrics are green. But the quality of diversion alert investigation (whether dismissed alerts later proved genuine) is measured nowhere. The investigation team cannot investigate every alert with limited staff. They triage by apparent severity and dismiss the rest. The question is not just how many alerts are false but of the alerts dismissed, how many turned out to be real.
The Decision:
Any confirmed miss (a dismissed alert later linked to a verified diversion event) triggers full review of dismissal criteria. DAIQ below 90% triggers immediate investigation capacity expansion. The metric transforms diversion alert management from a resource-constrained triage operation into a measured quality process where dismissal accuracy is tracked and accountability exists.
Formula
Notation:
DAIQ = (Σ Correctly_Dismissed / Σ Total_Dismissed) × 100
Variables:
- Correctly_Dismissed = dismissed alerts where no subsequent evidence of diversion or integrity compromise was identified within 6 months
- Total_Dismissed = all diversion alerts dismissed without full investigation during the measurement period
Worked Example: Over 12 months, the model generated 780 alerts. 624 dismissed (80% rate). Rolling 6-month retrospective of the first 312 dismissed alerts: 304 showed no subsequent evidence (correctly dismissed), 5 associated with channels where regulatory actions or product recalls subsequently occurred (concerning), 3 directly linked to confirmed diversion events. DAIQ = 304/312 = 97.4%. Looks high, but the 3 confirmed misses represent diversion events the organization's monitoring failed to catch.
Edge Cases:
- 6-month lookback is proposed; some diversion events take longer to surface. Track rolling annual as well.
- "Subsequent evidence" must be defined: confirmed counterfeit detection, regulatory enforcement action, verified diversion event in the affected channel
- Low dismissed alert count per period makes the metric unstable. This is inherently a low-frequency, high-consequence domain. Track cumulative rolling.
Sign convention:
Higher DAIQ is better. 100% means every dismissal was correct.
Proposed formula, pending validation pilots.