This is a ready-to-use protocol for a formal process capability study. Replace every <<FILL: ...>> placeholder with your own specifics, set your document numbers and dates, and route it through your normal document control, review, and approval before execution. A worked filled specimen follows the template. Verify each cited regulation against the current source before you rely on it. This protocol pairs with the Worksheet: Process Capability Index Calculation and, for study parameters already under continued monitoring, the Worksheet: Control Limit Establishment and Signal Rule Configuration.
Approval page (execute only after all signatures)
| Field | Entry |
|---|---|
| Protocol title | Process Capability Study for <<FILL: PRODUCT, PROCESS STEP, CHARACTERISTIC(S)>> |
| Protocol number | <<FILL: PROT-ID, e.g. VAL-CAP-011>> |
| Version | <<FILL: version, e.g. 1.0>> |
| Effective date | <<FILL: date>> |
| Supersedes | <<FILL: prior version or "New">> |
| Manufacturing site | <<FILL: site>> |
| Reason for this study | <<FILL: e.g. PPQ Stage 2 capability confirmation, annual re-baseline, post-change verification, new low-volume product first capability assessment>> |
| Role | Name | Signature | Date |
|---|---|---|---|
| Author (Quality Engineering / Statistician) | <<FILL>> | ||
| Process / Manufacturing SME | <<FILL>> | ||
| Analytical / QC | <<FILL>> | ||
| Quality Assurance (approver) | <<FILL>> |
Approval before execution is mandatory. Data collected against an unapproved protocol carries no capability claim.
1. Objective
To determine, with a defensible statistical basis, whether <<FILL: PROCESS / CHARACTERISTIC>> reliably meets its specification limits, and to report the potential (Cp/Pp) and actual (Cpk/Ppk) capability with the assumptions behind each number stated and checked. This protocol does not itself establish process control; it assumes control has been demonstrated separately and confirms that assumption before calculating anything (section 5).
2. Scope
This protocol applies to <<FILL: the specific process step(s), product(s), and characteristic(s) in scope>>. It does not cover characteristics evaluated elsewhere under a separate capability or monitoring protocol, and it does not replace the ongoing continued process verification (CPV) program; a capability study is a point-in-time confirmation, CPV is the ongoing monitoring. See continued process verification (CPV).
3. System / process description
| Field | Entry |
|---|---|
| Process step(s) under study | <<FILL>> |
| Characteristic(s) and units | <<FILL>> |
| Specification limits (LSL / USL) and source (registration, internal spec) | <<FILL>> |
| Target value, if defined | <<FILL>> |
| Measurement method and instrument(s) | <<FILL>> |
| CQA / CPP / key parameter designation | <<FILL>> |
4. Prerequisites
| Prerequisite | Status | Reference |
|---|---|---|
| Measurement system analysis (Gauge R&R) completed and acceptable | <<FILL: Yes/No, date>> | <<FILL: report reference>> |
| Instrument(s) qualified and in current calibration | <<FILL: Yes/No>> | <<FILL>> |
| Process demonstrated in a state of statistical control for the study period (control chart reviewed) | <<FILL: Yes/No>> | <<FILL: chart / worksheet reference>> |
| Sampling plan for data collection defined and approved | <<FILL: Yes/No>> | <<FILL>> |
| Personnel executing the study trained on the sampling and measurement procedure | <<FILL: Yes/No>> | <<FILL>> |
Do not proceed past this section unqualified. A capability study executed before its prerequisites are satisfied produces a number that cannot be defended, regardless of how the arithmetic in section 7 turns out.
5. Data collection plan
| Field | Entry |
|---|---|
| Number of subgroups (k) planned | <<FILL: commonly 25 or more for an established estimate>> |
| Subgroup size (n) and rationale | <<FILL: see the rational subgrouping guidance in the source article; state what source of variation each subgroup is designed to expose>> |
| Sampling frequency | <<FILL>> |
| Sources of variation the data set is designed to span | <<FILL: shifts, operators, raw material lots, equipment trains>> |
| Data collection window | <<FILL: start / end dates or batch range>> |
| Data recording method | <<FILL: system name, form reference>> |
| If the data set will be provisional (fewer subgroups than the established minimum) | <<FILL: state the count planned and that the result will be labeled provisional>> |
6. Acceptance criteria (set before data collection begins)
| Criterion | Threshold |
|---|---|
| Statistical control confirmed before capability is calculated | Required; no exceptions |
| Normality checked; transformation or non-normal method applied and documented if the characteristic is not normal | Required |
| Minimum Cpk / Ppk for this characteristic | <<FILL: e.g. Cpk >= 1.33, Ppk >= 1.00, or your internal standard>> |
| Measurement system contribution to observed variation | <<FILL: e.g. below 10 percent of tolerance>> |
| Gap between Cp and Cpk flagged for review if it exceeds | <<FILL: e.g. 0.20>> |
| Gap between Cpk and Ppk flagged for review if it exceeds | <<FILL: e.g. 0.20>> |
Setting these after the data arrives, so the threshold conveniently matches the result, defeats the purpose of a protocol; the criteria in this section are fixed before section 8 is executed.
7. Roles
| Role | Responsibility |
|---|---|
| Author / Statistician | Designs the sampling plan, performs the calculation, applies non-normal or short-run methods where required |
| Process owner / Manufacturing SME | Confirms the process was run under normal, representative conditions during data collection |
| QC / Analytical | Generates or verifies the measurement data and confirms instrument status |
| Quality Assurance | Approves the protocol before execution, reviews the report, dispositions any deviation |
8. Test cases (execution)
| ID | Step | Expected | Actual | Pass/Fail | Executed by | Date |
|---|---|---|---|---|---|---|
| TC-1 | Confirm all section 4 prerequisites satisfied before data collection | All Yes | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> |
| TC-2 | Collect data per the section 5 plan | Data set complete, matches planned k and n | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> |
| TC-3 | Confirm the process was in statistical control during collection | Control chart shows no unresolved special cause | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> |
| TC-4 | Test normality; apply transformation or non-normal method if needed | Method selected and documented | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> |
| TC-5 | Calculate Cp, Cpk, Pp, Ppk (or the one-sided equivalents) | Calculation complete per the capability worksheet | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> |
| TC-6 | Compare results against the section 6 acceptance criteria | Meets, or documented as not meeting with a stated path forward | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> |
9. Deviation handling
Any departure from the approved protocol, an unplanned gap in data collection, a special-cause event discovered mid-study, a prerequisite found unsatisfied after execution began, is recorded here and assessed for impact on the study’s validity before the conclusion is written.
| # | Deviation | Impact assessment | Disposition | Approved by |
|---|---|---|---|---|
<<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> |
10. Summary and conclusion
| Field | Entry |
|---|---|
| Cp / Cpk result | <<FILL>> |
| Pp / Ppk result | <<FILL>> |
| Normality method applied | <<FILL>> |
| Acceptance criteria met? | <<FILL: Yes/No, with detail on any miss>> |
| Provisional or established result? | <<FILL>> |
| Interpretation (centering vs spread vs between-subgroup drift) | <<FILL>> |
| Recommendation | <<FILL: accept, recenter, investigate, increase monitoring, re-run at N additional batches>> |
| Feeds into | <<FILL: PPQ report / CPV plan / annual product review>> |
11. Attachments
- Raw data set and its source extract.
- Control chart covering the study period.
- Measurement system analysis (Gauge R&R) report.
- Completed Worksheet: Process Capability Index Calculation.
12. References
FDA, Process Validation: General Principles and Practices (2011), Stage 2/3. ICH Q8(R2) (Pharmaceutical Development), ICH Q9(R1) (Quality Risk Management). ISO 22514-2, Statistical methods in process management, Capability and performance, Part 2: Process capability and performance of time-dependent process models (replaced the withdrawn ISO 21747:2006). 21 CFR 211.110 (Sampling and testing of in-process materials and drug products). See statistics in quality: Cpk and control charts.
Confirm the current version of each reference before issue.
13. Revision history
| Version | Date | Author | Summary of change |
|---|---|---|---|
<<FILL: 1.0>> | <<FILL: date>> | <<FILL: author>> | Initial issue. |
14. Approvals (post-execution)
| Role | Name | Signature | Date |
|---|---|---|---|
| Author | <<FILL>> | ||
| Reviewer (Process Owner) | <<FILL>> | ||
| Approver (Quality Head) | <<FILL>> |
Filled specimen
A drug substance chromatography step, yield percent, LSL 85.0 percent, no upper limit, target 92.0 percent, run as an annual re-baseline capability study.
Section 4, prerequisites: Gauge R&R on the yield calculation inputs (UV concentration assay), 4 percent of tolerance, acceptable, completed 3 January 2026. In-line UV analyzer calibrated, current. Control chart (I-MR) reviewed for the prior 30 batches, no unresolved special cause, confirmed in control. Sampling plan approved 10 January 2026. Analysts trained on the yield calculation procedure.
Section 5, data collection plan: 30 batches (established), one result per batch, spanning three shifts and two resin lots, collection window 1 February 2026 to 30 April 2026, recorded in the batch record and pulled via validated report RPT-YLD-004.
Section 6, acceptance criteria set in advance: Ppk minimum 1.00 (one-sided, lower limit only), measurement system below 10 percent of tolerance, Cpk-to-Ppk gap flagged for review above 0.20.
Section 8, execution: TC-1 through TC-4 pass as planned. TC-5: mean 91.6 percent, within-subgroup sigma (MR-bar/1.128) 1.8 percent, overall sigma 2.4 percent. Cpl = (91.6 - 85.0)/(3 x 1.8) = 6.6/5.4 = 1.22. Ppl = (91.6 - 85.0)/(3 x 2.4) = 6.6/7.2 = 0.92. TC-6: Ppl of 0.92 falls short of the 1.00 minimum; acceptance criterion not met.
Section 9, deviation: None during execution; the shortfall in TC-6 is a study finding, not a protocol deviation, and is carried into the conclusion rather than dispositioned as a deviation against this protocol.
Section 10, conclusion: Cpl (1.22) comfortably exceeds Ppl (0.92), a gap of 0.30 against a flag threshold of 0.20, indicating the process wanders between batches more than the within-batch spread alone would suggest, consistent with the two-resin-lot span of the data. Recommendation: investigate resin-lot-to-resin-lot yield variation as the likely driver, hold the current alert limit in the CPV program, and re-run this study after the resin qualification range is narrowed or a root cause is implemented, targeting Ppl at or above 1.00 at the next annual re-baseline.
Common inspection findings this protocol prevents
- A capability study executed and reported with no evidence the prerequisites (measurement system, statistical control) were checked first.
- Acceptance criteria that were not defined until after the result was known.
- A capability study report with no raw data attachment, so the calculation cannot be reconstructed.
- A study that reports Cpk alone, omitting Ppk, on a process later found to drift between subgroups.
- A missed acceptance criterion carried forward with no stated recommendation or re-study trigger.
How to adapt this protocol
- Set the protocol number, characteristic, and reason for the study in the approval page.
- Confirm every section 4 prerequisite with a real reference, not a checkbox with nothing behind it.
- Set the section 6 acceptance criteria from your internal standard or the registration commitment, before data collection starts.
- Attach the completed capability worksheet as the calculation record.
- Where the characteristic is monitored on an ongoing basis under CPV, cross-reference the control limit establishment worksheet so the two documents do not disagree about the baseline data set.
- Confirm every regulation in section 12 against the current published version before issue.