This is a ready-to-use risk assessment for setting up Continued Process Verification when the batch count is too low to support a fully powered control chart, the normal condition for a clinical-stage program and for many low-volume commercial cell, gene, and RNA-based products. It documents the seeding basis for provisional limits, or the basis for a qualitative trend review where even provisional limits cannot be supported, and states the plan for reconciling against real commercial data as it accumulates. This is a different document from a control limit recalculation justification, which assumes established limits already exist; this one covers the period before they do. Replace every <<FILL: ...>> placeholder with your own specifics. A filled specimen follows.
Document control header
| Field | Entry |
|---|---|
| Document title | Provisional CPV Limits Justification for <<FILL: product / process / parameter set>> |
| Document number | <<FILL: e.g. CPV-PROV-2026-01>> |
| Version | <<FILL>> |
| Effective date | <<FILL>> |
| Linked CPV plan | <<FILL: document number>> |
| Document owner | <<FILL: e.g. Statistician / Validation Lead>> |
1. Purpose
This assessment documents the basis for using provisional control limits, or a documented qualitative trend review in place of a formal control chart, for <<FILL: parameter or parameter set>> given the current commercial batch count, and states the plan for moving to a fully powered chart as data accumulates.
2. Current batch count and working minimum
| Field | Entry |
|---|---|
| Commercial batches available for this parameter | <<FILL>> |
| Working minimum for the intended chart type (e.g. approximately 20 to 25 points for I-MR) | <<FILL>> |
| Gap between available data and the working minimum | <<FILL>> |
| Batches per year at current or planned manufacturing cadence | <<FILL>> |
| Estimated time to reach the working minimum at that cadence | <<FILL>> |
3. Method selected for the interim period
| Option | Selected | Basis |
|---|---|---|
| Provisional control limits seeded from PPQ data | <<FILL: yes/no>> | <<FILL: state which PPQ batches and why they represent the commercial process>> |
| Provisional control limits seeded from late-stage development runs on equivalent equipment/material | <<FILL: yes/no>> | <<FILL: state the equivalence basis; equipment, scale, material lineage>> |
| Specification used as an interim proxy pending sufficient data for a true control limit | <<FILL: yes/no>> | <<FILL: state explicitly that this is a weaker interim substitute, not a control limit>> |
| Documented qualitative trend review, no numeric chart | <<FILL: yes/no>> | <<FILL: state why even provisional numeric limits cannot be supported yet>> |
Select the strongest option the available data actually supports; do not default to the qualitative review if a provisional numeric chart is genuinely supportable from PPQ or development data.
4. Qualitative trend review method, if selected
| Field | Entry |
|---|---|
| Each result plotted against | <<FILL: specification and full available history>> |
| Reviewer | <<FILL: role>> |
| Frequency | <<FILL: every batch>> |
| Required reviewer output | <<FILL: an explicit written conclusion each cycle, not a passive log entry>> |
5. Patient- or donor-driven variability, if applicable
For an autologous or heavily donor-variable process, complete this section; otherwise mark not applicable.
| Question | Answer |
|---|---|
| Does this parameter measure process performance directly, or does it conflate process performance with starting-material variability | <<FILL>> |
| If conflated, is a normalized metric (e.g. fold expansion, recovery percentage) available or planned instead | <<FILL>> |
| Is input-material stratification used or planned | <<FILL: yes/no, and basis>> |
6. Reconciliation plan
| Field | Entry |
|---|---|
| Trigger for moving from provisional to established limits | <<FILL: e.g. accumulation of 20 commercial batches>> |
| Who performs the reconciliation | <<FILL>> |
| How a material difference between provisional and reconciled limits will be handled | <<FILL: e.g. change control, communication to any affected batch disposition>> |
7. Approvals
| Role | Name | Signature | Date |
|---|---|---|---|
| Author (Statistician / Validation lead) | <<FILL>> | ||
| Reviewer (Process owner / MSAT) | <<FILL>> | ||
| Approver (Quality Assurance) | <<FILL>> |
Filled specimen
Extract for an example autologous cell therapy with five commercial batches to date. Numbers are illustrative.
| Field | Entry |
|---|---|
| Commercial batches available | 5 |
| Working minimum for I-MR | Approximately 20 to 25 |
| Gap | 15 to 20 batches |
| Cadence | Approximately 30 batches per year at current commercial demand |
| Estimated time to working minimum | Approximately 6 to 8 months |
Method selected: provisional control limits seeded from the 4 PPQ batches plus the 5 commercial batches to date (n=9), combined because both sets used the same commercial-scale equipment, the same raw material specifications, and the same operators trained to the same procedure; the borrowing basis and small n are stated openly in the CPV plan rather than presented as an established limit.
Section 5, patient-driven variability: fold expansion ratio is used instead of absolute output cell count, specifically because absolute count conflates process performance with input leukapheresis quality; no formal input stratification is used yet at this batch count, planned for reassessment once 20 commercial batches accumulate and a stratification basis can be statistically supported.
Reconciliation plan: limits will be recalculated once 20 commercial batches are available, through the standard control limit recalculation form, referencing this document as the originating basis; if the reconciled limits differ materially from the provisional limits, the difference and its cause will be documented in that recalculation record, not silently absorbed.
Common inspection findings this justification prevents
- A control chart presented as mature and established when it is built from four or five points, with no disclosure of the small sample size.
- Absolute output metrics charted for an autologous or heavily donor-variable process with no acknowledgment that patient variability, not process variability, may be driving the pattern.
- No stated plan for when provisional limits will be reconciled against real commercial data, so the “provisional” label never actually resolves.
- A specification used as a control limit with no label distinguishing the two, contradicting the basic control-limit-versus-specification distinction the CPV program otherwise relies on.
How to adapt this justification
- State the real batch count and the real working minimum side by side; do not round either number in a way that makes the gap look smaller.
- Choose the strongest defensible method in Section 3; a documented qualitative review is a fallback for when provisional numeric limits genuinely cannot be supported, not a default for convenience.
- Complete Section 5 honestly for any autologous or high-donor-variability process, even if the answer is “not yet resolved, planned for the next revision.”
- Set a concrete reconciliation trigger in Section 6, tied to a batch count or a date, not an open-ended “when there is enough data.”