This is a ready-to-use tracking log. Replace every <<FILL: ...>> placeholder with your own specifics and feed it into your standing quality metrics review. A worked filled specimen follows the template. The indicators here are starting points for discussion, not regulatory limits; the point is the trend, not any single period’s number.
Purpose
These four indicators, trended over time, surface data integrity culture and control problems before they surface as 483 observations or warning letter citations. A single period tells you little. The same number drifting in one direction across a year tells you something is changing in how work is actually done. This log makes that trend visible and reviewable.
Indicator definitions
| Indicator | Numerator | Denominator | What it signals |
|---|---|---|---|
| OOS invalidation rate | Count of OOS results closed as assignable laboratory error | Total OOS results in the period | Pressure to find a lab-error justification rather than accept a genuine result |
| Batch record amendment rate | Count of batch record amendments made after the recording shift/step | Total batch records in the period | Records not captured contemporaneously; late corrections clustered before release deserve scrutiny |
| After-hours access frequency | Count of unexplained system access events outside scheduled hours | Total access events in the period | Work happening outside oversight; investigate unexplained spikes, not the baseline |
| CDS re-injection rate | Count of re-injections | Total injections in the period | Testing into compliance; watch the trend, not the absolute number |
Tracking log
| Period | System(s) | OOS invalidation rate | Batch amendment rate | After-hours access (unexplained) | CDS re-injection rate | Reviewer | Review date | Escalated? |
|---|---|---|---|---|---|---|---|---|
<<FILL: e.g. Q1 2026>> | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> |
Review trigger
Escalate to a data integrity self-audit or a targeted investigation per <<FILL: SOP-ID>> when any indicator shows a sustained directional trend across three or more consecutive periods, not on a single period’s value. A single high-OOS-invalidation quarter driven by one known root cause (for example a documented instrument fault, resolved) is not automatically an escalation; the same pattern repeating after the fix is.
Instructions
- Calculate each indicator at a consistent cadence,
<<FILL: e.g. quarterly>>, from the same source systems each period so the trend is comparable. - Record the reviewer and review date even when no escalation is warranted; a log with gaps is itself a finding.
- Where a period’s number is elevated, record the known cause if one exists (instrument fault, staffing change, a specific investigation) so future reviewers can distinguish an explained blip from a genuine drift.
- Feed this log into the standing quality metrics and KPIs program so indicators are reviewed on a fixed cadence rather than only when someone remembers to look.
Retention
Retain this log for <<FILL: retention period>>, consistent with the retention period for the underlying quality records the indicators are drawn from.
References
FDA Data Integrity and Compliance With Drug CGMP guidance (December 2018). MHRA GxP Data Integrity Guidance and Definitions (March 2018).
Filled specimen
| Period | System(s) | OOS invalidation rate | Batch amendment rate | After-hours access (unexplained) | CDS re-injection rate | Reviewer | Review date | Escalated? |
|---|---|---|---|---|---|---|---|---|
| Q1 2026 | QC Analytical, CDS-HPLC-07 | 18% (4 of 22 OOS) | 6% | 2 events | 3.1% | R. Gomez | 2026-04-10 | No |
| Q2 2026 | QC Analytical, CDS-HPLC-07 | 31% (7 of 23 OOS) | 7% | 5 events | 5.8% | R. Gomez | 2026-07-08 | Yes, DI self-audit triggered |
| Q3 2026 | QC Analytical, CDS-HPLC-07 | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> |
The Q2 review shows the pattern this log exists to catch: OOS invalidation rate and re-injection rate both moved in the same direction in the same quarter, which is a stronger signal together than either number alone, and it was the trigger for the self-audit documented in the DI self-audit summary report for that period.
Common inspection findings this log prevents
- Leading indicators mentioned in a quality manual as a concept but never actually tracked as numbers over time.
- A rising re-injection or OOS invalidation rate that was visible in the raw data for a year before anyone connected it to a data integrity risk.
- Metrics reviewed once and then abandoned, with no reviewer or review date evidencing an ongoing cadence.
How to adapt this log
- Add indicators specific to your operation (for example an EBR correction rate for manufacturing-heavy sites) using the same numerator/denominator/signal structure.
- Set the review cadence and escalation threshold to match the criticality and volume of the systems tracked.
- Link an escalation directly to the DI self-audit checklist and its summary report template so a trend finding routes into the same structured process as any other self-audit finding.