If you only discover trouble at final release, you’re basically speed-running regret. In-Process Quality Control (IPQC) flips that script. It watches your batch while it’s being made—spotting drift, tightening variability, and saving you from the “Why is this OOS now?” spiral. Think of IPQC as Find My iPhone for your process: live location, instant alerts, fewer panicked searches.
Benefits? fewer fire drills, cleaner audits, happier patients. This guide keeps your original structure and gets practical—what IPQC is, how it maps to FDA/EMA/WHO expectations, what to check (and when), the tools that actually help on the floor, how IPQC looks at each manufacturing stage, where teams stumble, how digital QMS ties it all together, and what’s next (hi, continuous and AI-assisted control).
Who’s this for? QA/QC leads, production managers, validation folks, and anyone who’s tired of late surprises and paper chases. Keep it open while you work on your SOPs—we’re aiming for zero fluff and 100% usable.
IPQC is the set of checks you run during manufacturing—not after—to keep a batch inside the guardrails. It’s how you spot drift early, adjust on the fly, and document that every critical step stayed in control.
Why it matters (in real life):
When granulation runs wetter than expected, or a filler starts creeping off target, IPQC is the early-warning system. You catch it, tune it, and keep moving—without gambling the whole batch or finding out at final testing that something went sideways hours ago.
| Aspect |
IPQC |
End-Product QC |
| Timing |
Real-time or near real-time during each unit operation |
After the batch is complete |
| Purpose |
Control and adjust the process |
Verify the outcome |
| Impact |
Prevents defects; reduces scrap/rework |
Detects defects; if it fails, you’re already late |
- Before a step starts: line clearance, equipment status, pre-run checks
- During the step: in-process tests and patrol checks tied to critical process parameters (CPPs) and critical quality attributes (CQAs)
- After the step: document results, trend the data, and decide—proceed, adjust, or escalate
All regulators want the same thing: prove the process stayed in control while you made the product—and show your math. Here’s the quick tour you can cite in SOPs.
- What it says: 21 CFR §211.110 requires sampling and testing of in-process materials to assure batch uniformity and integrity, with clear expectations for when to sample, how to evaluate, and how to quarantine/reject materials that don’t meet criteria.
- What’s new-ish: FDA’s Jan 2025 Draft Guidance explains practical ways to comply, including use of advanced manufacturing and process models inside your control strategy. (Draft = nonbinding, but very useful.)
- How to apply: Map each unit operation to required IPQC checks, define sampling plans and acceptance criteria, and document disposition logic (accept/adjust/hold).
- What it says: EudraLex Vol. 4 (notably Chapter 5 – Production) expects defined in-process controls embedded in written procedures, with results recorded and trended to support ongoing control and release decisions.
- How to apply: For each step, list IPCs, methods, limits, and frequency; ensure results feed into batch evaluation and—where relevant—continuous process verification.
- What it says: WHO’s GMP “main principles” require controls across all manufacturing stages, emphasizing validated steps, documented IPCs, and data that demonstrate the process stayed capable.
- How to apply: Align site SOPs to WHO’s principles (risk-based sampling, validated methods, change control) so multi-market supply files tell a consistent story.
Across FDA, EMA, and WHO, the theme is identical—risk-based IPQC linked to a documented control strategy, with data you can trend, defend, and act on
IPQC exists so you don’t “hope for quality” at the end—you prove control while you make the product. Here’s what it’s actually trying to achieve:
- Assure batch uniformity and integrity in real time, not post-mortem. (This is the core expectation in 21 CFR §211.110.)
- Detect drift early and correct fast—link checks to CPPs/CQAs so operators can tweak set-points before defects stack up. (Aligned to a PQS “state of control” per ICH Q10.)
- Demonstrate a risk-based control strategy—define what to sample, how often, and what triggers action using quality risk management principles (ICH Q9).
- Strengthen data integrity by design—capture results with traceable methods, limits, and audit trails so you can trend, defend, and pass inspections. (Consistent with ICH Q10 elements: monitoring, CAPA, change management.)
- Harmonize with EU/WHO GMP—documented in-process controls inside SOPs and batch records that regulators can follow without guesswork.
IPQC’s purpose is control —make quality visible, measurable, and adjustable while the batch is still in play.
There’s no one-size-fits-all list. IPQC is tailored to your control strategy and the unit operation in front of you. That said, most shops watch a common core of process and product attributes—sampled at defined intervals with action limits that keep the batch inside guardrails. Regulators expect you to define these checks, set statistically sound limits, and act when data drifts.
For solid oral dose (tablets/capsules)
- Blend uniformity / assay (often via NIR or grab samples) — confirms content is even before compression/encapsulation.
- Moisture / LOD — protects flow, compressibility, and stability.
- Tablet weight & weight variation — early signal of feeder/die issues.
- Hardness & thickness — ties directly to mechanical integrity and downstream performance.
- Friability — screens for chipping/abrasion risk in handling.
- Disintegration / (when justified) in-process dissolution — spot checks that support release later.
FDA flags hardness, disintegration, dissolution, and friability as critical attributes for certain tablet types—use that thinking to justify your IPQC set.
For liquids & semisolids
- pH and osmolality (where relevant) — formulation stability and patient compatibility.
- Viscosity / rheology — fill performance, syringeability, spreadability.
- Fill volume / weight — dose accuracy.
- Content uniformity — especially for suspensions and multi-dose presentations.
For sterile/aseptic operations
- Differential pressures & airflow direction/velocity — keep the cleanroom cascade healthy.
- Non-viable particle counts & viable EM trending — verify the environment while you manufacture, not just after.
- Filter integrity tests (pre- and post-use, including PUPSIT where applicable) — demonstrate the sterilizing filter actually performed.
EU Annex 1 explicitly calls out pre- and post-use integrity testing and robust EM/control of classified areas—anchor your sterile IPQC here.
For biologics
- pH, conductivity, osmolality — track process state during buffers/steps.
- Protein concentration / aggregation (at-/on-line where possible) — maintain product quality during purification.
- Bioburden at key points (as applicable) — keep upstream and harvest steps inside limits that protect downstream sterilizing steps.
- Tie each check to a CPP/CQA and define what happens at alert vs. action.
- Base limits on process averages/variability and statistical procedures (per 21 CFR 211.110), then keep refining with CPV data.
- Document frequency, sample size/locations, and disposition logic in SOPs and the batch record (EU Chapter 5 expectation).
Pick parameters that give operators fast, actionable signals. If a check can’t trigger a tweak, a hold, or an escalation, it’s noise—tighten it or retire it.
PQC is a toolbox. You’ll mix classic QC instruments with PAT sensors, SPC trending, and digital systems that route exceptions in real time.
- The classics (HPLC/UV, KF/LOD, GC, dissolution, hardness/friability, viscosity)
- Process Analytical Technology (PAT)
- SPC & trending (a.k.a. seeing drift early)
- Sampling that supports decisions
- Sterile operations: environment & filtration specifics
- Digital systems that make IPQC actionable
- Risk as the organizing principle
Pick instruments and sensors that give operators fast, defensible signals, wire them into SPC + eBR, and let your Q8/Q9/Q10 stack prove the control strategy is real—not just a slide.
Consider IPQC as a chain of small, smart gates across the line. Here’s a stage-by-stage playbook you can drop into SOPs and eBRs.
Raw material prep & dispensing
- Checks: ID/label match, status labels, weigh accuracy, sieve/mesh verification, environmental conditions (temp/RH for hygroscopic APIs).
- Why: Prevent mix-ups and keep inputs inside spec before value-adding steps begin.
- Frequency & triggers: Every lot/dispense; hold if ID, weight, or status is off; escalate per deviation SOP.
Blending (and wet/dry granulation)
- Checks: Blend uniformity (grab or NIR), LOD/moisture, torque trends, screen integrity.
- Why: Ensure content uniformity and flow/compressibility before compression.
- Frequency & triggers: Early/mid/late samples or PAT windows; adjust time/speed if RSD or moisture drifts; investigate if trends breach action limits.
Drying, milling, sizing
- Checks: Moisture to target, inlet/outlet temps, particle size distribution, screen checks, metal detection.
- Why: Protect downstream weight control, hardness, and dissolution.
- Frequency & triggers: IPC at defined intervals or end-point; tweak set-points; hold if PSD out of range.
Compression / encapsulation
- Checks: Tablet weight & variation, hardness, thickness, appearance/defects; capsule net weight, lock status.
- Why: Direct signals of process stability and dose accuracy.
- Frequency & triggers: Start-up, then time-based (e.g., every 15–30 min) or count-based; adjust feeder/compression; hold if trending breaches action rules (SPC).
Coating
- Checks: Weight gain, spray rate, atomization pressure, inlet/outlet temp, bed temp, appearance (mottle, peel).
- Why: Uniformity and target release profile.
- Frequency & triggers: Patrol checks + end-point; adjust airflow/spray; hold if weight-gain band missed.
Liquids & semisolids (solution/suspension/gel/cream prep + filling)
- Checks: pH, viscosity, content uniformity (esp. suspensions), fill volume/weight, container/closure checks.
- Why: Dose accuracy, stability, and usability.
- Frequency & triggers: Batch stages plus periodic fills; adjust speed/nozzles or mix parameters; hold line if CU or fill volume drifts.
Sterile/aseptic compounding, filtration, and filling
- Checks: Pre- and post-use sterilizing filter integrity (PUPSIT where applicable), differential pressure cascading, airflow direction/velocity, non-viable and viable environmental monitoring during manufacture, line clearance and interventions log.
- Why: Real-time assurance the sterile barrier and environment stayed under control while filling.
- Frequency & triggers: Per Annex 1 program—integrity test before and after use; EM at defined frequencies; stop/segregate if integrity fails or EM exceeds action limits.
Lyophilization (if applicable)
- Checks: Shelf temp, chamber pressure, product temp (TPc/TPf), endpoint (Pirani vs Capacitance), stopper placement.
- Why: Protect CQAs like residual moisture and structure.
- Frequency & triggers: Phase-based checks; adjust shelf profile; investigate if product temp deviates from model.
Packaging & labeling
- Checks: Print/label verification (code dates, lot), seal integrity, count checks, artwork/version control.
- Why: Final identity and patient-facing accuracy.
- Frequency & triggers: Start-up and periodic; line hold on any mismatch (treat as potential mix-up).
Quick Tip: Treat each unit operation like a mini control room. Define what you’ll check, how often, and exactly what happens when data moves—then let PAT/SPC and your eBR/QMS keep the loop tight and audit-ready.
IPQC is where theory meets the production floor—and the floor always fights back. Here are the usual plot twists and how to handle them without derailing the batch (or your audit).
1) Sampling bias (aka “we only checked the easy spot”)
- Symptom: Blend looks perfect… because you sampled only the top.
- Fix: Stratify by time and location (start/mid/end; top/middle/bottom). Use NIR/PAT where possible to widen coverage without slowing the line.
2) Tiny samples, big decisions
- Symptom: Three tablets decide the fate of 3 million.
- Fix: Size samples with a statistical rationale tied to risk and variability. Document why that “n” is enough to act.
3) Using specs as controls
- Symptom: You treat final-release specs like in-process limits, so action happens late.
- Fix: Set tighter control limits (SPC run rules), keep alert vs action bands, and escalate before you hit spec.
4) Results arrive after the moment has passed
- Symptom: By the time the lab posts LOD, you’ve already coated the lot.
- Fix: Move critical checks at-line/on-line (PAT, rapid tests). Define pre-approved adjustment playbooks so operators can act immediately.
5) Human error & data integrity gremlins
- Symptom: Illegible entries, missing initials, mystery units.
- Fix: eBR + barcode capture, enforced fields, timestamped entries, and role-based reviews. Make “no data = no move” a system rule, not a reminder.
6) Vague triggers = inconsistent decisions
- Symptom: Two shifts, two different reactions to the same drift.
- Fix: Write crystal-clear If/Then rules: what to tweak, when to hold, who to call. Put them on the batch record and the line poster.
7) Noise checks that don’t change outcomes
- Symptom: You’re running tests that never trigger action.
- Fix: Retire or repurpose. Every IPQC must tie to a CPP/CQA and a decision. If it can’t change your next move, it’s clutter.
8) Method transfer drift
- Symptom: Site B’s “same” method trends higher—cue arguments.
- Fix: Do formal equivalency checks, align system suitability, and schedule periodic method performance reviews with cross-site trending.
9) Equipment realities
- Symptom: Worn feeders, clogged screens, sneaky air leaks.
- Fix: Fold condition-based checks into IPQC (screen integrity, torque profiles, leak tests) and sync with maintenance/calibration windows.
10) Aseptic specifics (high stakes, zero chill)
- Symptom: EM action hits during filling or a filter integrity test flunks post-use.
- Fix: Stop/segregate, investigate fast, and document interventions. Pre-/post-use integrity testing (and PUPSIT where required) plus real-time EM trending keep surprises contained.
11) Supplier variability shows up on your line
- Symptom: Same recipe, different lot behavior.
- Fix: Tighten incoming checks, trend lot-to-lot impacts, and push feedback through supplier scorecards/CAPA so RM variability doesn’t become your rework.
12) Siloed systems = slow IPQC
- Symptom: LIMS says one thing, QMS another, eBR is… a PDF.
- Fix: Connect LIMS ↔ eBR ↔ QMS (and MES/ERP) with shared IDs. Auto-trigger NC/CAPA from IPQC breaches and show live trends on ops dashboards.
13) Change control lag
- Symptom: Process changed; SOPs and batch records didn’t get the memo.
- Fix: Link change control → document updates → training in one workflow with effective dates and read-and-sign tracking.
14) People load & training fatigue
- Symptom: IPQC feels like a second job. Corners get cut.
- Fix: Simplify forms, standardize intervals, use visual work aids, and rotate tasks. Celebrate catches—operators are your first-line quality team.
IPQC fails quietly when triggers are fuzzy, data is late, or tests don’t drive action. Build risk-based sampling, real-time signals, and crisp playbooks into the flow, and the process starts protecting itself.
Paper can record quality. Digital can enforce it. A modern eBR/QMS stack turns IPQC from “nice checks” into a closed-loop control system that catches drift, triggers the right actions, and proves compliance.
For example, eBR enforces steps, holds, and conditional logic; QMS auto-opens NC/CAPA/Change with e-sigs; LIMS streams at/online results; PAT + SPC power real-time dashboards; MES/ERP sync recipes, status, and genealogy. Integrity is by design (ALCOA+, Part 11/Annex 11, single IDs). Breaches trigger auto holds and operator playbooks; role-based visuals and handover notes keep the action moving. Analytics run themselves (CPV, practical AI for anomalies, OOT triage, action hints). Governance links changes to SOPs, eBR, and training. Track TTD/TTA, in-process catches, FPY, and rework to prove impact.
IPQC works best when it’s baked into the way you run the line—not tacked on like a sticker. Steal these field-tested plays.
- Start with a control strategy, not a checklist
Map CPPs/CQAs → IPQC tests → limits → actions for each unit operation. If a test isn’t tied to a decision, it doesn’t belong.
- Design sampling to answer a decision
Stratify by time and location (start/mid/end; top/middle/bottom). Pre-define the minimum “n” and exactly what a fail means (adjust vs. hold vs. investigate).
- Use control limits, not just specs
Set alert and action limits tighter than release specs. Teach teams to respond at alert (tweak) and stop at action (hold + NCR).
- Trend in real time
Put SPC charts on the few signals that matter (e.g., tablet weight, hardness, fill volume). Auto-notify on run-rule violations—not just spec failures.
- Push critical checks at-line or on-line
If the decision can’t wait, move it out of the lab: PAT (NIR/Raman), rapid moisture, in-line torque/pressure/temperature. Shorter TTD (time-to-detect) = fewer surprises.
- Make triggers executable
Write If/Then rules right into the eBR:
“IF weight trend breaches action limit, THEN pause feed, reduce turret speed 5–10%, re-sample in 10 minutes; IF still out, open NC and notify QA.”
No guesswork, no inbox ping-pong.
- Harden data integrity by design
Barcode materials/equipment, enforce required fields, capture timestamps and e-signatures, and keep a clean audit trail. One master ID for lots, samples, and methods.
- Close the loop automatically
When IPQC breaches: auto-create NC/CAPA, link evidence, require effectiveness checks before removing temporary controls. Let the system drive discipline.
- Treat sterile controls as live controls
For aseptic lines, elevate EM trends, differential pressures, and filter integrity to first-class IPQC. Pre- and post-use integrity tests aren’t paperwork—they’re gates.
- Keep the list lean
Quarterly, retire checks that never trigger action and add ones that would have caught recent issues. Less noise = faster reactions.
- Train for first-line quality
Give operators visual work aids (what good/alert/action looks like) and do short “drift drills.” Celebrate early catches; that’s the culture you want.
- Wire CPV into daily life
Feed IPQC data to Stage 3 CPV dashboards. Review trends in tier meetings, not just at APR/PQR season.
- Audit your own playbook
Do quick, on-the-floor audits: Can we see today’s alerts? Are actions documented? Did the hold actually hold? Fix gaps the same day.
10-Minute IPQC Health Check (use on any line):
- Do we know which signals matter most?
- Are alert/action limits visible at the point of use?
- Can operators act without waiting for QA emails?
- Are today’s trends on one screen?
- Did any IPQC breach auto-create the right record (NC/CAPA)?
Build IPQC around fast signals + clear actions + clean data. Do that, and your process starts guarding itself.
The next wave of IPQC is less “sample and hope” and more sense, decide, and adjust—with models and sensors doing the heavy lifting.
- Continuous manufacturing: IPQC becomes a continuous, residence-time-aware layer with PAT at chokepoints and auto set-point tweaks (per ICH Q13).
- RTRT: Validated in-process signals replace select end tests, guided by EMA RTRT and aligned with ICH Q13/Q14.
- Models on the shopfloor: Process models trigger actions under FDA §211.110 (draft) and are lifecycle-managed per ICH Q14.
- PAT + AI: NIR/Raman + streaming SPC evolve to prediction, catching weak signals (e.g., feeder wear, humidity) before limits break.
- Sterile gates: Annex 1 makes CCS, PUPSIT, and live EM hard gates—breach = immediate line action.
- Living validation: CPV is always-on dashboards, not annual reports—anchored to FDA process validation.
- Data > documents: Tight LIMS–eBR–QMS–MES, single IDs, and audit trails make APR/PQR a click and pave the way for RTRT/model-informed release.
So what? Build model-ready data, put PAT where decisions happen, treat Annex 1 gates as live controls, and wire CPV into daily tier huddles.
IPQC keeps inspectors calm, patients safe, and margins healthy. It’s in the regs: FDA demands in-process sampling with statistically sound limits; EU GMP/Annex 1 expects defined controls, live EM, and filter integrity; WHO reinforces validated steps across stages. For safety, IPQC supplies real-time signals to a risk-based PQS (ICH Q9/Q10), turning high-risk steps—especially aseptic—into hard gates. For the business, IPQC powers Stage-3 CPV, cuts rework, speeds release, and enables continuous manufacturing and RTRT, so decisions are faster and inventories leaner.
In-process quality control acts as your early-warning system. When you tie IPQC to a real control strategy, trend signals in real time, and trigger clean workflows for NC/CAPA and change, you cut rework, speed release, and keep patients safe. That’s the whole game: fewer surprises, calmer audits, better margins.
If you’re ready to turn this guide into daily reality, see how Qualityze Intelligent EQMS Suite makes IPQC truly in-process—with eBR enforcement, live SPC dashboards, PAT integrations, automated NC/CAPA, and audit-ready records.
Request a Qualityze Intelligent EQMS Demo
