I’m seeing Clinical QA bundled with vendor oversight, CSV, and study start-up under one mid-level posting at two companies this week… For those negotiating scope, are you using hard metrics — CAPA cycle time, audit closure within 30 days, deviation recurrence rate — to justify a dedicated QA headcount and a clean RACI, or accepting the consolidation?
I’ve gotten leadership to unbundle by converting CAPA cycle time, audit backlog, and recurrence into a capacity model (QA hours/quarter) plus a simple ‘cost of poor quality’ estimate, then framing the inspection risk via ICH Q9 (https://database.ich.org/sites/default/files/ICH_Q9(R1)_Guideline_Step4_2023_0216.pdf) — ‘a Swiss Army knife isn’t a scalpel.’ If consolidation is non-negotiable, I push for a signed RACI with WIP limits on QA sign-offs so timelines don’t quietly slip. Do you have a defensible FTE-per-study or per-audit ratio you’re using, @OP?
Quick step: run a 90‑day pilot keeping QA independent for audits/CAPA while ops holds vendor oversight/CSV, and track two leading indicators plus your lagging ones — time to risk classification and ‘audit closure within 30 days’ — to show the delta in inspection readiness. Would you get air cover for a pilot from your study lead or CMO? Referee ≠ coach, and I’d cite the independence expectation in ICH E6(R2) with a one‑page RACI and this link: ICH Official web site : ICH.
I’d stop debating headcount and quantify the conflict‑of‑interest risk. Take 12 months of ‘audit closure within 30 days’ and deviation recurrence and compare periods where QA also owned vendor oversight/CSV; if closures slip and recurrence ticks up, you’ve got evidence. Pair that with MHRA GCP inspection metrics showing CAPA effectiveness is a top finding (https://www.gov.uk/government/publications/gcp-inspection-metrics) and ask whether they want the owner checking their own work or to fund a clean RACI.
One lever that’s worked: track “finding density” (major/minor per audit hour) before and after QA starts carrying ops work; it usually drops, which makes a clean QA lane easier to sell — like asking your goalie to coach mid‑game. If you’re sub‑50 staff, I’d tolerate a temporary blend with a clear trigger (≥3 active protocols or first Phase 2) to split; @lucy_taylor47 have you looked at detection rate or first‑pass audit‑ready as leading indicators?
Add one metric leadership understands: ‘inspection concordance’ — the share of external observations QA had already flagged internally; if folding QA into ops drives that down, it’s your canary to keep QA independent. Building on @sthomp48, can you pull the last two inspections and show concordance drift alongside any vendor escalations?
Bundle if you must, but write it into the RACI that anyone running study start‑up or CSV can’t audit or approve those same areas, then report the recusal rate and its impact on ‘CAPA cycle time’ each month. When the excluded audit universe or CAPA age breaches your SLA, that’s your data‑backed trigger for a dedicated QA FTE, and it aligns with ICH E6(R2) independence for audits (5.19: https://database.ich.org/sites/default/files/E6_R2_Addendum.pdf). Consolidation under a mid‑level posting is only workable if that firewall and trigger are explicit.
What’s worked for me is a short pilot where we add a “QA gate before DB lock” (and SIV) and ring‑fence QA to review/release only, not execute the workstreams. We track right‑first‑time on validation/CSV packages and TMF QC plus change‑control cycle time; when RFT climbs and CC time drops, it’s a clean data point for a dedicated QA lane. If you can’t get a pilot, set a hard stop tied to risk staleness (e.g., no go if top risks sit >14 days), which keeps QA authority visible even in a bundled model.
I’ve had luck modeling the dollar hit: tag ‘CRO change orders due to compliance’ and ‘reconsent events per amendment’ for a quarter, then project savings if QA stays separate from execution — nothing kills the bundling pitch faster than a $75k “oops” line item. Building on @emily_ward83’s pilot angle, run a 60–90 day trial with finance coding those two buckets and a go/no‑go tied to fewer change orders. Can you get finance to stand up those cost codes this month?
But , when QA is also running CSV or start‑up, the fox ends up auditing the henhouse… @jordan_miller77 do you track inspector prompts like ‘who approved their own work?’ as a proxy for independence risk. I carve QA into its own cost center and report “audit closure within 30 days” plus eTMF QC pass rate and conflict‑of‑interest flags in approvals to make the case.
And > 30 days” plus eTMF QC pass rate and conflict‑of‑interest flags in approvals to make the case. Agree on the timing piece — what moved the needle for me was a one‑page eQMS “self‑approval” heatmap per study signed by the COO, showing dual‑role signoffs jump whenever QA reports into Ops; that visual got us a separate reviewer FTE — @jordan_miller77, do you track it monthly?
And i’ve had success showing an “independence score” — count self‑review events and segregation‑of‑duties breaches, plus rework hours linked to them — and mapping the cost delta when QA is standalone vs bundled, which resonates because it’s like asking the goalie to take the penalty kicks too. If leadership insists on consolidation, I insist on eQMS role walls and a dotted line to Quality with a quarterly independence audit; @andrew_p86 would a short pilot to baseline that score fly at your shop?
Quick win: run a 6‑week pilot where QA is firewalled from CSV/start‑up and track “defect escape rate” to production plus rework hours versus the blended model; convert the delta to $/study‑month as the headcount case — finance even smiles at a clean before/after chart. Pair it with a timestamped independence‑exception log to show trend, and @jordan_miller77 have you seen inspectors respond well to that level of evidence? If leadership still balks, propose a hybrid RACI where QA owns release gates while ops keeps vendor comms, then revisit after one quarter.