Loading...
Pharma & Biotech
Clinical, lab, and quality data stays inside the study and the line a person is assigned to. Every read and every write lands in the audit trail on its own, so keeping the evidence never becomes one scientist's side job.
Bring case report forms and lab results from every site onto one standard, then list missing fields and out-of-range values as queries. The edit checks a data manager rewrites for each study stay in the app instead. CRAs see only the site they cover; the data manager sees the study.
PB-204 Phase 2 — data cleaning
GxPBlinding enforcedAudit trail onQA reviewerStudy PB-204| Subject ID | Site | Fields captured | Data status |
|---|---|---|---|
| 001-003 | BBoston 001 | 50/50 | Clean |
| 001-005 | BBoston 001 | 50/50 | Clean |
| 001-008 | BBoston 001 | 49/50 | Query raised |
| 001-011 | BBoston 001 | 48/50 | Needs review |
| 001-014 | BBoston 001 | 50/50 | Clean |
| 002-002 | DDenver 002 | 50/50 | Clean |
| 002-004 | DDenver 002 | 47/50 | Query raised |
| 002-007 | DDenver 002 | 50/50 | Awaiting SDV |
| 002-009 | DDenver 002 | 49/50 | Query raised |
| 002-012 | DDenver 002 | 50/50 | Clean |
| 003-001 | PPortland 003 | 50/50 | Query raised |
| 003-003 | PPortland 003 | 49/50 | Query raised |
| 003-006 | PPortland 003 | 50/50 | Clean |
| 003-008 | PPortland 003 | 44/50 | Withdrawn |
| 003-011 | PPortland 003 | 50/50 | Clean |
| 003-014 | PPortland 003 | 50/50 | Awaiting SDV |
Find past runs by compound, process condition, or date, and put batches made under the same conditions side by side. Researchers ask a question instead of digging through notebooks and shared folders. Every answer carries the notebook page and the signatures behind it, so it can be cited as it stands.
Experiment lookup — oral solid formulation
21 CFR Part 11Audit trail onQA reviewerFormulation records14 hits in 10,342 records
14 Jun 2023 · Study design — In the PB-31 10 mg controlled-release layer, the HPMC 2208 (K4M, 4,000 cP) level of 12% kept pushing tablet hardness against the upper action limit and produced repeated sticking, so a reduced-polymer formulation was screened. One granulation lot was split into 12%, 10%, and 8% arms and compressed at a fixed 12 kN, all else at commercial settings.
21 Jun 2023 · Pilot manufacture — The 8% arm was carried forward as the lead formulation and three pilot batches (B23-0912, B23-0913, B23-0914) were manufactured and placed at 40°C/75% RH for a six-month accelerated study. All three ran at one tenth of commercial scale on the same press and the same punch set, so scale was the only difference left.
21 Jun 2023 · Stability protocol — Samples are packed in HDPE bottles with a 1 g desiccant and held in chamber CH-02 at 40°C/75% RH, with 12 tablets pulled from each batch at 0, 1, 3, and 6 months. Dissolution runs in 900 mL of pH 6.8 phosphate buffer, paddles at 50 rpm, following method validation report v3. Chamber conditions log every 5 minutes and any excursion opens a deviation.
5 Jul 2023 · Release testing and compression follow-up — Initial hardness for the 8% arm averaged 94 N against 108 N for the 12% arm, and release dissolution at 1, 4, and 8 hours sat inside the profile window for both arms. Content uniformity gave an AV of 6.2 or lower across all three pilot batches against an acceptance value of 15. Light adhesion to the upper punch appeared over the last 20 minutes of the B23-0914 run; the lubricant level was raised from 0.5% to 0.75% for the next pilot, B23-0915, and the adhesion did not recur. That change is recorded in NB-2023-0502 p.35.
21 Sep 2023 · 3-month accelerated pull — Dissolution matched the original formulation at every time point with f2 = 61, and hardness had moved less than 3 N from release. No new related substance appeared above the 0.05% reporting threshold, and the chamber log for the interval showed no excursion.
26 Dec 2023 · 6-month accelerated pull and conclusion — Total related substances were 0.38% against a 1.0% specification and the largest individual impurity held at 0.11% against a 0.20% limit, so the 8% formulation is equivalent to the original through 6 months accelerated with the lubricant fixed at 0.75%. The 12-month long-term pull and a supplier viscosity comparison stay open. Because the change lowers the release-controlling polymer, it was assessed under SUPAC-MR as a Level 3 change and raised as CC-2023-118; QA approved the change control on 15 Jan 2024, it was filed as a Prior Approval Supplement, and commercial batches move to the 8% formulation only after FDA approval.
Check results against specification to flag OOS results, trend them against history to catch OOT signals, and keep every deviation's investigation status on one screen. The criteria live in the app, so a pass-down does not change the answer. Investigation and approval steps land in the GxP audit trail with the user, the time, and the reason.
Deviations & CAPA — drug product plant
GxP21 CFR Part 11Audit trail onQA reviewerDrug product lines5 open · 5 closed · 10 logged this quarter
On 11 Aug 2026 at 7:52 AM, batch B26-0417 on tablet press TP-03 ran above the 18.0 kN upper action limit for 4 minutes 12 seconds, peaking at 19.4 kN. The alarm sounded 1 minute 40 seconds late, so alarm delay entered the scope, and the operator stopped the press and segregated 12,400 tablets.
QA placed the full batch on hold the same day, and tablets from 30 minutes either side of the excursion were sampled at 10-minute intervals — 7 points, 20 tablets each, 140 in total. Hardness there averaged 112 N against a 110 N limit; thickness and weight variation held.
The load cell on TP-03's upper punch had drifted 0.9 kN from its last calibration on 19 May 2026; TP-01 and TP-02 sat at 0.1 and 0.2 kN, inside range. After recalibration a trial run agreed to within 0.1 kN, and a three-month interval was raised as CAPA-2026-0071.
The 12,400 quarantined tablets were classed for destruction and the rest of the batch retested: dissolution met specification on 12 units, content uniformity gave an AV of 5.8 against 15, related substances 0.21% against a 1.0% limit. Batch disposition stays with QA and follows this review.
In clinical and quality data, who read what and when matters as much as the value itself. Your databases stay where they are and inline AI owns only the gate in front of them, so control does not scatter across systems. When an inspector asks what changed, when, and by whom, that record is the answer.
Records follow ALCOA+. Reads, edits, and deletion attempts land in the audit trail with the user, the timestamp, and the stated reason.
Corrections never overwrite. When an app changes a batch record or a test result, the change lands as a new signed version with a documented reason for change, and the original value stays in the audit trail exactly as it was.
Clinical access follows your SSO groups. You do not rebuild blinding or per-study permissions in a second set of accounts.
Bring QA to the first call. Connection scope and validation get settled there.
Book a demo