Site-Level Quality by Design: Putting Risk-Proportionate GCP into Practice
Published November 2026 — A practical guide for sites and site networks on applying the quality-by-design and proportionality principles of ICH E6(R3) to everyday trial conduct.
Quality Is Designed In, Then Checked
ICH E6(R3) reframed Good Clinical Practice around a small set of principles. Two of them shape daily work at sites more than any others. Quality by design asks that quality be built into a trial from the start by identifying the factors critical to participant protection and reliable results. Proportionality asks that processes and controls match the risks involved, with effort concentrated where it matters.
The Principles and Annex 1 were finalized in January 2025, became operative in the EU in July 2025, and were published as final FDA guidance in September 2025. With Annex 2 adopted in June 2026, the revision is complete. Sponsors have rewritten monitoring plans and quality management systems in response. The corresponding change at the site level has received less attention.
This paper describes what quality by design looks like at a site: six elements of a site-level approach, how to apply them through the life of a study, and the mistakes that most often undermine them. The aim is a method a coordinator and investigator can use on the next protocol.
Six Elements of a Site-Level Approach
These elements translate the E6(R3) principles into things a site can do and show.
For each protocol, identify the small number of things that must go right to protect participants and produce reliable results: eligibility decisions, consent, the primary endpoint assessment, dosing, and safety reporting are the usual candidates. Everything else is supporting activity.
Ask what could cause each critical factor to fail at this site, with these staff and this patient population. A complex endpoint assessment performed by one trained rater is a different risk from the same assessment performed by five. Record the risks and what the site will do about each.
Put the strongest controls on the critical factors: second-person checks of eligibility, rater certification, dosing verification. Apply lighter controls elsewhere. Checking everything to the same depth spreads attention thin and hides the errors that matter.
The investigator's oversight should be visible in the record. Regular documented review of eligibility, safety events, and delegated work shows more than a signature on a log. Oversight of service providers who perform trial activities belongs here too.
E6(R3) gives data governance its own section. Sites should know, for each system they use, who has access, how changes are tracked, how data are protected, and how records are retained. This applies to site-owned systems and to those the sponsor provides.
Treat each important deviation as information about the process. Find the cause, decide whether it could recur, and change the control if it could. A deviation log that is only a list does not improve quality.
Applying It Through the Study Lifecycle
Quality by design is not a separate project. It fits into steps a site already takes for every study.
At Feasibility: Read the Protocol for Risk
When a protocol arrives for feasibility, look beyond patient numbers. Note which procedures are unusual for the site, which assessments depend on a single person or device, and where the visit schedule is tight. Raising these points with the sponsor early is part of quality by design, and sponsors increasingly expect it.
At Startup: Write a Short Site Quality Plan
Before activation, record the critical-to-quality factors for the study, the main site-level risks, and the controls the site will apply. Two pages is enough. The plan gives new staff a clear picture of what matters most and gives monitors and inspectors the reasoning behind the site's approach.
During Conduct: Check the Critical Factors First
Build routine checks around the plan. Review eligibility before randomization, verify primary endpoint data soon after each visit, and reconcile product at defined points. A brief standing meeting between the investigator and coordinators to review these items keeps oversight continuous.
When Things Change: Reassess
A protocol amendment, staff turnover, a new vendor, or an unexpected pattern of deviations is a reason to revisit the risk assessment. Update the plan and note what changed and why. Proportionality means the level of control follows the current risk.
At Close-Out: Capture What Was Learned
Before the study file is archived, record which risks materialized, which controls worked, and what the site would do differently. Those notes are the starting point for the next study's plan and, in a network, for other sites running similar protocols.
Common Pitfalls
Sites that struggle with the new approach usually run into one of these.
Uniform checking feels thorough but takes time away from eligibility, endpoints, and safety, where an error has real consequences.
A risk assessment completed once at startup and never opened again does not guide anyone. It should be short enough to use and revisited when circumstances change.
Sponsors design the trial-level quality approach, but only the site can see site-level risks such as staffing gaps, local workflow, and patient population.
Investigators often oversee closely and record little of it. If the review is not documented, it cannot be shown at inspection.
Imaging centers, local labs, and home health providers performing trial activities fall under the investigator's oversight. Their qualifications and performance should be reviewed.
Correcting a single deviation without asking why it happened leaves the same failure waiting for the next participant.
The Role of Network Governance
A single site builds its quality approach from its own experience. A governed network can pool that experience. Risk assessments for similar protocols can be shared, controls that worked at one site can be adopted by others, and deviation patterns can be seen across sites before they become a finding at any one of them.
This is the reasoning behind ClinFirst™, the governance framework that Clinitiative network sites operate under, and behind the continuous evaluation provided by SiteScore™. Common standards give sponsors a consistent baseline across sites, and site-level performance data show where attention is needed.
Conclusions
E6(R3) asks sites to think before they check: to decide what is critical in each study, put their effort there, and be able to explain the reasoning. For most sites this is a change of emphasis more than a change of workload, and it tends to reduce time spent on low-value verification.
The sites that adapt fastest will be those that make the approach routine: a short quality plan for every study, regular documented investigator review, and a habit of asking why when something goes wrong. This paper is a summary for planning purposes and should be read alongside the guideline text and applicable regional guidance.
Want Sites That Work This Way?
Learn how ClinFirst™ governance standardizes quality across the Clinitiative network.