- Before
- Signs sections whose figures she cannot trace without asking someone to re-run the numbers.
- Now
- Clicks any number to see its query; every query is re-run on the frozen data at her signature.
One PBRER, from data lock point to QPPV signature
Monday, 3 August. The Veltrimab PBRER for 1 Jul 2025 – 30 Jun 2026 is due on 8 September, and Pravonix is due in six days. Here is the QPPV’s morning, screen by screen, in the working solution.
- 01Morning
Every report that needs her, on one screen
Dr. Maya Chen · QPPVFive reports in progress, nine sections and final reports waiting for her signature, six open review comments — each with an edit or reply already proposed — and 1,286 numbers in open reports, all traced to a query. The greeting says what matters: Pravonix is due in 6 days, and Veltrimab Section 5 is ready for her review.
Agents at work: “Section writer — Drafted 11 Halvexa sections — 162 numbers, each linked to its query.”
- 02Opened
The Veltrimab PBRER, section by section
Dr. Maya Chen · QPPVData lock point 30 Jun 2026, ICH E2C(R2), submitted in the EU as a PSUR under the EURD list. Fifteen of 20 sections are drafted and nine are signed. The activity trail shows how it got here: on 1 July the data agent froze 18,940 cases at the data lock point and ran 38 queries; the consistency checker then checked 214 numbers across sections and the prior report — 0 differences.
- 03One click
“Draft 5 remaining sections”
The agentsSections 14, 16, 17, 18 and 19 are still empty. The data agent re-runs 32 queries on the data frozen at 30 Jun 2026, the signal and risk writer drafts the sections from the query results, the prior report and the risk management plan, and the consistency checker checks every number against Sections 5 and 6 and the prior report. The drafts go to Dana Okafor, Dr. Maya Chen and Omar Haddad for review, and the report reaches 20 of 20 sections drafted.
The benefit-risk conclusion is a proposal for the QPPV, never final.
- 04Section 5
Click a number, see the query behind it
Dr. Maya Chen · QPPVSection 5, Estimated exposure and use patterns, carries 18 numbers, all linked to a query. Maya clicks the cumulative exposure, 224,177 patient-years. The panel says it in plain words — pens shipped from 30 Dec 2021 to 30 Jun 2026, returns removed, divided by 13 pens per patient-year — then shows the query, the result by region and the checks it passed.
“Sum of 4 regions = 224,177” · “Matches the source” · Sales and exposure extract, 6 Jul 2026.
- 05Review comments
A reviewer’s question, already answered
Sam Patel · EpidemiologySam, who reviews the exposure section, asked: “Japan also sells the 75 mg pen for patients under 60 kg. Are those counted as half a dose?” The review assistant checked the query: it already counts two 75 mg pens as one 150 mg dose, so no number changes. It proposes one sentence for the method in 5.2 instead.
Proposals only — a person accepts or rejects every one.
- 06Accepted
The edit lands, and the reconciliation is shown
Dr. Maya Chen · QPPVMaya accepts the edit; the sentence appears in 5.2, highlighted, and Sam is notified. Lena Ortiz from regulatory affairs had asked for the cumulative exposure to reconcile with last year’s report. The answer is on the comment, with its source: 133,054 patient-years in the PBRER for DLP 30 Jun 2025 (Section 5.2, Table 5.3) plus 91,123 in this interval equals 224,177. Maya resolves it.
- 07Signed
Every query re-run before her signature
Dr. Maya Chen · QPPVWith no comments open, Maya signs the section. Before the signature goes through, the 18 queries are re-run on the frozen data — every number matches the draft — and the cumulative is checked against prior plus interval. Her name, the date and time and the meaning “I approve this section for the final report” are recorded, linked to this version. The history shows each step, and 10 of 20 sections are now signed.
“Section 5 signed by Dr. Maya Chen · 10 of 20 sections signed”
- 08Check numbers
One difference, found before anyone signs it
Consistency checkerMaya runs the full check: every query re-run on the frozen data (214 numbers), every total and sub-total, cumulative equals prior plus interval, the same figure the same in every section. One difference: Section 11 says 1,912 literature records were screened, but the literature log returned 1,914 after its refresh on 31 Jul 2026, when two records wrongly marked as duplicates were restored. The fix waits for her: “Use 1,914”.
The checker never changes a draft without approval — every finding shows both values.
- 09Due in 6 days
Pravonix, ready for the final signature
Dr. Maya Chen · QPPVThe Pravonix PBRER has all 20 sections signed. The consistency checker re-checked 238 numbers against the frozen data before the final signature — 0 differences — and the submission package is built. Maya signs the final report with the meaning “I approve this report for submission as QPPV”, and it goes to regulatory operations for submission.
- 10Last 12 months
Every periodic report, on time
Dr. Maya Chen · QPPVThirteen of 13 reports submitted on time, on average 4 days before the deadline. A median of 13 days from data lock point to full draft — it was 46 before the agents. Every one of 8,214 numbers since go-live traced to a query, and 91% of 604 review comments closed within two days. Number questions fell from 31% to 22% of comments once every number opened its own query.