This is a ready-to-use quality risk management (QRM) assessment for the share-versus-dedicate decision in a multi-product facility. It converts the HBEL and cleaning data into a documented, defensible conclusion: shared with validated cleaning, dedicated product-contact parts, dedicated equipment, dedicated area, or a self-contained facility. Replace every <<FILL: ...>> placeholder with your own specifics and route it through your QRM and change-control procedures. A worked specimen follows. Verify each cited regulation against the current source before you rely on it.
Document control header
| Field | Entry |
|---|---|
| Assessment title | Dedicated versus Shared Facility Decision, <<FILL: PRODUCT / EQUIPMENT TRAIN>> |
| Document number | <<FILL: RA-ID, e.g. QRM-CC-021>> |
| Version | <<FILL: version>> |
| Effective date | <<FILL: date>> |
| Facilitator / author | <<FILL: role>> |
| Cross-functional team | <<FILL: toxicology, cleaning validation, QC, engineering, QA>> |
| Trigger | <<FILL: new product / new equipment / new toxicology data / periodic review>> |
1. Objective and scope
Decide the level of equipment or facility dedication required for <<FILL: PRODUCT>> in <<FILL: EQUIPMENT TRAIN / AREA>>, under EU GMP Chapters 3 and 5 and ICH Q9(R1), so that cross-contamination of any co-produced product is prevented and residual risk is acceptable. Scope covers <<FILL: which trains, rooms, and shared utilities>>.
2. Methodology
The decision follows a gated QRM logic. A product is dedicated if it fails a mandatory class check or if any gate cannot bring residual risk below the health-based exposure limit (HBEL). The gates are applied in order; the first gate that cannot be satisfied determines the dedication level.
- Gate 0, mandatory class. Beta-lactams and live-organism biologicals are dedicated by rule; no HBEL argument applies.
- Gate 1, hazard. The PDE of the product as a contaminant. A low PDE (potent, narrow margin) raises the bar.
- Gate 2, detectability. Can a validated method quantify at or below the resulting limit with acceptable recovery?
- Gate 3, cleanability. Can the cleaning process reproducibly meet the limit across the worst case?
- Gate 4, airborne/utility. Can containment and HVAC keep dust, aerosol, and shared-air carryover below the HBEL?
3. Scoring scales
Use a simple, documented scale for each gate. Adapt the anchors to your QRM system; the point is that each score is defined, not subjective.
| Score | Hazard (Gate 1) | Detectability (Gate 2) | Cleanability (Gate 3) | Airborne (Gate 4) |
|---|---|---|---|---|
| 1 (low risk) | High PDE, wide margin | LOQ well below limit, recovery good | Reproducibly clean with margin | Closed process, contained |
| 2 (moderate) | Moderate PDE | LOQ near limit, recovery acceptable | Clean but limited margin | Partly open, controlled HVAC |
| 3 (high risk) | Low PDE, potent | LOQ at or above limit | Cannot reproducibly meet limit | Open process, shared air, dust |
4. Mandatory class check (Gate 0)
| Question | Answer | Consequence |
|---|---|---|
| Is the product a beta-lactam (penicillin, and certain cephalosporins/carbapenems)? | <<FILL: yes/no>> | If yes, dedicate self-contained facility with dedicated air. Stop. |
| Is the product a live-organism biological or otherwise covered by Annex 2 containment? | <<FILL: yes/no>> | If yes, dedicate per the biological containment requirement. Stop. |
If either is yes, the assessment ends here with a dedicated self-contained facility.
5. Gate assessment table
Complete one row per shared transition or per shared unit as appropriate.
| Gate | Input data | Score (1-3) | Evidence reference | Notes |
|---|---|---|---|---|
| 1. Hazard (PDE) | <<FILL: PDE mg/day, HBEL monograph ref>> | <<FILL>> | <<FILL: HBEL-ID>> | |
| 2. Detectability | <<FILL: method, LOQ, recovery %, vs limit>> | <<FILL>> | <<FILL: method validation ref>> | |
| 3. Cleanability | <<FILL: cleaning development / verification result vs limit>> | <<FILL>> | <<FILL: study ref>> | |
| 4. Airborne / utility | <<FILL: containment, HVAC, pressure cascade status>> | <<FILL>> | <<FILL: qualification ref>> |
6. Decision logic
- All gates score 1-2 with documented margin: shared with validated cleaning is justified.
- Gate 3 fails on specific components only: dedicate those product-contact parts, share the cleanable platform.
- Gate 2 or Gate 3 fails on the platform: dedicate the equipment.
- Gate 4 cannot be controlled: dedicate the area or the air system.
- Gate 0 triggered, or hazard combined with an uncontrollable pathway: self-contained facility.
Decision: <<FILL: shared / dedicated parts / dedicated equipment / dedicated area / self-contained>>
7. Mitigations and residual risk
| Pathway | Control applied | Residual risk | Acceptable? |
|---|---|---|---|
| Equipment surface carryover | <<FILL: validated cleaning to X µg/cm²>> | <<FILL>> | <<FILL: yes/no>> |
| Airborne / dust | <<FILL: containment, dedicated HVAC>> | <<FILL>> | <<FILL>> |
| Personnel / gowning | <<FILL: flow, gowning, training>> | <<FILL>> | <<FILL>> |
| Mechanical transfer / change parts | <<FILL: dedicated parts, line clearance>> | <<FILL>> | <<FILL>> |
Residual risk is acceptable when every credible pathway has at least one validated or verified control bringing exposure below the HBEL, and the QRM file shows each pathway was considered.
8. Acceptance criteria for the assessment
- The mandatory class check was performed and documented.
- A signed HBEL monograph exists and its PDE was used, not a historical default.
- Detectability and cleanability were tested against the real recovery-corrected limit, not asserted.
- Every credible pathway maps to a control and a residual-risk judgment.
- The decision and its evidence are recorded, and change control is set to re-trigger the assessment on a new product, formula, equipment, or toxicology change.
9. References
EU GMP Chapter 3 (Premises and Equipment) and Chapter 5 (Production), 2015 revision. EMA/CHMP/CVMP/SWP/169430/2012 (HBEL guideline) and EMA/CHMP/CVMP/SWP/246844/2018 (Q&A). ICH Q9(R1), Quality Risk Management. EU GMP Annex 1 (2022) for the contamination control strategy context; Annex 2 for biological containment. 21 CFR 211.42, 211.46, 211.67, 211.176 for the US equivalents (including the penicillin separation).
Confirm the current version and clause numbers of each reference before issue.
10. Approvals
| Role | Name | Signature | Date |
|---|---|---|---|
| QRM facilitator | <<FILL>> | ||
| Toxicology | <<FILL>> | ||
| Cleaning validation SME | <<FILL>> | ||
| QA | <<FILL>> |
Filled specimen
The following shows the assessment completed for an illustrative potent compound sharing a solid-dose train. Numbers are illustrative.
| Gate | Input | Score | Decision effect |
|---|---|---|---|
| 0. Mandatory class | Not a beta-lactam, not a live biological | pass | Proceed |
| 1. Hazard | PDE 50 µg/day (potent), HBEL-014 | 3 | Raises the bar |
| 2. Detectability | Specific HPLC, LOQ 2 µg/swab, recovery 82%, limit 6 µg/swab | 1 | Measurable with margin |
| 3. Cleanability | Platform cleans reproducibly; filling needles and one hopper fail at the worst case | 3 (parts), 1 (platform) | Parts fail only |
| 4. Airborne | Contained charging, dedicated dust extraction | 1 | Controlled |
Decision: dedicate the filling needles and the failing hopper; share the cleanable platform with validated cleaning. Residual risk after dedicating the two failing components is judged acceptable, because every remaining pathway has a validated control below the 50 µg/day HBEL. Change control is set so a new product on this train re-opens the worst-case and the gate scores.
This is the proportionate answer the article describes: dedicate the parts that cannot be cleaned to the limit, share the platform that can, and document why.
Common inspection findings this assessment prevents
- Sharing equipment for a potent compound with no documented risk assessment justifying it.
- A decision made on the platform’s average cleanability while ignoring specific change parts that fail.
- Detectability assumed rather than demonstrated against the real limit.
- No change-control trigger, so a new product enters a shared train with the old worst-case still assumed.
- Beta-lactam or biological separation decided on a PDE argument instead of the mandatory rule.
How to adapt this assessment
- Pull the PDE from the signed HBEL monograph; do not proceed on an unsigned number.
- Run the mandatory class check first; if it triggers, the assessment is over.
- Score each gate against real detectability and cleanability data, per equipment train.
- Apply the decision logic literally; dedicate at the lowest level that controls the real risk.
- Record residual risk per pathway and set the change-control trigger before closing.