This is a ready-to-use worksheet for scoring and classifying the individual findings a data integrity gap assessment produces. It operationalizes the numeric severity times likelihood times detectability method described in the parent article as a fillable tool, so two assessors working from the same evidence land on the same classification without arguing about the label itself. It is not the per-system control assessment, see the Data Integrity Risk and Gap Assessment (Per System) for that layer, and it is not the roll-up report, see the Report: Data Integrity Program Gap Assessment for that document. This worksheet sits between the two: it takes findings already identified during evidence gathering and turns each into a defensible, ranked classification before either of those documents is written. Replace every <<FILL: ...>> placeholder with your own specifics. A worked filled specimen follows. This is educational structure to adapt, not legal or regulatory advice.
Document control header
| Field | Entry |
|---|---|
| Document title | Data Integrity Gap Assessment Finding Scoring and Classification Worksheet |
| Document number | <<FILL: reference, e.g. DI-SCORE-2026-01>> |
| Assessment date | <<FILL>> |
| Scored by | <<FILL: lead assessor name>> |
| Parent charter / report reference | <<FILL>> |
1. Purpose
This worksheet scores each finding from a data integrity gap assessment on three factors, severity, likelihood, and detectability, multiplies them into a risk priority number, and bands the result into the same four-level classification used in the assessment report, so the classification is reproducible from the evidence rather than a judgment call taken on trust.
2. Scoring scales
| Factor | 1 | 3 | 5 |
|---|---|---|---|
| Severity: consequence if the gap were exploited or the condition ran uncorrected | Negligible effect on record trustworthiness | Affects a non-critical decision | Directly compromises a release, safety, or reportable-result decision |
| Likelihood: how plausible the failure is under normal conditions | Requires an unusual, deliberate action to occur | Plausible under normal operating pressure | Already occurring, or the normal way the system or process gets used |
| Detectability: how likely an existing control would catch it before harm | Would be caught immediately by an independent control | Might be caught at periodic review | Leaves no trace; nothing in the current control set would ever surface it |
Score each factor 1, 2, 3, 4, or 5; intermediate values (2 and 4) are for a factor that sits between two anchor descriptions.
3. Risk priority number bands
| RPN (severity x likelihood x detectability) | Classification | Typical target timeframe |
|---|---|---|
| 75 to 125 | Critical | Immediate, in parallel with formal CAPA |
| 30 to 74 | Major | 30 to 90 days |
| 10 to 29 | Minor | 90 to 180 days |
| 1 to 9 | Observation | Tracked as continuous improvement |
4. Finding scoring table
One row per finding. Score independently of the system owner; the lead assessor owns this table.
| Ref | Finding | Severity | Likelihood | Detectability | RPN | Classification | Basis for scores |
|---|---|---|---|---|---|---|---|
<<FILL: F01>> | <<FILL>> | <<FILL: 1-5>> | <<FILL: 1-5>> | <<FILL: 1-5>> | <<FILL>> | <<FILL>> | <<FILL: one line justifying each score against the evidence>> |
5. Reconciliation, where two assessors disagree
If two assessors score the same finding differently, do not average the classifications; reconcile the three underlying factors instead.
| Finding ref | Assessor 1 scores (S/L/D) | Assessor 2 scores (S/L/D) | Factor with the disagreement | Resolution and final score |
|---|---|---|---|---|
<<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> |
6. Acceptance criteria
- Every finding carried forward from evidence gathering has a row in section 4 with all three factors scored and a stated basis for each score.
- The RPN is calculated correctly (severity multiplied by likelihood multiplied by detectability) and banded per section 3 without manual override of the band.
- Any classification that a reader would expect to be higher or lower than the RPN band suggests is either rescored with corrected factors or the override is explicitly justified in writing, an unexplained override defeats the purpose of a reproducible method.
- Any scoring disagreement between assessors is resolved through section 5, not through an unrecorded compromise.
- The completed worksheet is retained as a working paper supporting the assessment report.
7. References
ICH Q9(R1), Quality Risk Management, for the severity, likelihood, and detectability scoring logic this worksheet adapts. FDA, Data Integrity and Compliance With Drug CGMP: Questions and Answers (final, December 2018). MHRA, GXP Data Integrity Guidance and Definitions (Revision 1, March 2018).
Confirm the current version of each reference before issue.
8. Revision history
| Version | Date | Author | Summary of change |
|---|---|---|---|
<<FILL: 1.0>> | <<FILL: date>> | <<FILL: author>> | Initial issue. |
9. Approvals
| Role | Name | Signature | Date |
|---|---|---|---|
| Author (Lead Assessor) | <<FILL>> | ||
| Reviewer (QA) | <<FILL>> |
Filled specimen
The following shows the worksheet completed for five findings from the same site assessment, so you can see how the numeric method drives a ranked list rather than four buckets of unequal size. The company and findings are illustrative; replace them with your own.
Finding scoring table:
| Ref | Finding | Severity | Likelihood | Detectability | RPN | Classification | Basis for scores |
|---|---|---|---|---|---|---|---|
| F01 | Audit trail off for method changes on a release-testing CDS | 5 | 4 | 5 | 100 | Critical | Severity 5: feeds batch release directly. Likelihood 4: method edits are a routine analyst action. Detectability 5: no other control captures this event type. |
| F02 | Three users hold unneeded admin rights on the LIMS | 4 | 3 | 4 | 48 | Major | Severity 4: LIMS results feed disposition but with a second-person check downstream. Likelihood 3: misuse would require a deliberate act. Detectability 4: no periodic access review exists to catch it. |
| F03 | No audit trail review procedure exists for a Tier 2 environmental monitoring system | 3 | 4 | 3 | 36 | Major | Severity 3: supports trending, not direct release. Likelihood 4: absence of a procedure means the gap is already the normal state. Detectability 3: an annual system audit would eventually surface it. |
| F04 | Password expiration disabled on a label-printing workstation | 2 | 3 | 2 | 12 | Minor | Severity 2: label printing is a low-consequence function. Likelihood 3: plausible under normal IT workload pressure. Detectability 2: caught at the next periodic IT review. |
| F05 | Audit trail review procedure exists but was not updated after a recent configuration change | 2 | 2 | 2 | 8 | Observation | Severity 2: procedural currency gap, not an active control failure. Likelihood 2: requires the specific timing of a missed procedure update. Detectability 2: would be caught at the procedure’s own next periodic review. |
Reconciliation extract: F02 was initially scored Severity 3 by one assessor and Severity 4 by a second; the disagreement traced to whether the downstream second-person check should discount severity. Resolved at Severity 4, because the worksheet scores the gap’s inherent consequence, not the residual risk after a compensating control, which is scored separately in the per-system risk assessment.
The finding worth noting in this specimen: F01 and F02 both read as serious findings in narrative, but the RPN gap, 100 against 48, made the sequencing decision for the remediation roadmap immediate rather than a matter of the report writer’s judgment call.
Common inspection findings this worksheet prevents
- Two findings of similar apparent severity classified differently with no visible method, so the classification looks arbitrary when challenged.
- A Critical downgraded to Major with no recorded basis for the change.
- A large finding set with no way to produce a single ranked list, so remediation sequencing defaults to whoever wrote the report first.
- Disagreement between assessors resolved by an unrecorded compromise rather than a traceable reconciliation of the underlying factors.
How to adapt this worksheet
- Set your document number and scorer in the header, and link it to the assessment charter or report it supports.
- If your organization already uses a different severity, likelihood, or detectability scale for FMEA or another risk tool, align the anchor descriptions in section 2 to that scale so classification language stays consistent across your quality system.
- Do not let a classification override the RPN band silently; if you disagree with what the math produced, rescore the factors and show your work.
- File the completed worksheet as a working paper behind the assessment report, not as a replacement for it.
- Confirm every regulation in section 7 against the current published version before issue.