A CTQ-related problem can be detected reliably while the process that produces it remains insufficiently controlled. That is the operational tension at the center of Quality by Design. Detection shows where a quality-relevant condition becomes visible. For risks that require upstream prevention, intervention must occur before the resulting condition reaches the detection point. When activity settles at the detection point, the trial can generate convincing evidence that problems are being found without showing that the source of those problems has changed. [1]
The distinction matters because the two control positions answer different questions. CTQ detection establishes whether a quality-relevant problem is present at the point of observation. Upstream control addresses whether the earlier process condition that gives rise to that problem has been reduced sufficiently. Those answers can move independently. A detection mechanism can perform well while the upstream risk remains largely unchanged.
When detection becomes the control endpoint
The operational failure begins when the control loop closes around a detected CTQ-related problem at the point of detection. The issue is recognized as quality-relevant, while control remains concentrated where the consequence has already become visible. In that state, detection functions as the practical endpoint of control. The system proves that it can identify the problem, but it does not establish that the earlier source of risk has been changed sufficiently.
What repeated CTQ detection reveals
That separation changes the meaning of repeated detection. The first detected occurrence establishes that a CTQ-related problem exists. When the same condition continues to reach the detection point, the repeated signal carries additional operational meaning. It can indicate that the upstream risk has not been changed sufficiently to stop the problem from reaching detection again. At that stage, stronger detection alone does not resolve the control question. It may improve confidence that recurrence will be seen, while leaving recurrence itself untouched. [2]
The evidence problem for oversight
This creates a difficult evidence problem for oversight. Detection activity is visible because it produces observable evidence. Those records establish that the resulting condition was identified, while leaving unresolved whether the earlier source of risk has changed sufficiently. A review can see a functioning detection process and still lack evidence that Quality by Design is reducing the production of CTQ-related risk before it reaches the observation point.
An FDA inspection of Study EMR100070-003 shows how a visible monitoring signal can capture only part of the process risk behind it. Routine performance indicators had identified low source-data-verification production by one CRA. The CRA was removed from one site because of workload concerns but continued to monitor three other sites. FDA later identified protocol deviations at the affected site that the CRA had not reported, and the inspection record noted that the CRA’s prior monitoring activity at the other three sites had not been reviewed. [3]
This leaves the evidence of effective control ambiguous. If reliable detection and upstream risk control are treated as equivalent evidence, accountability can center on finding the problem rather than establishing whether the process condition that produces it has been changed. The distinction is operational rather than semantic: one position confirms that a quality-relevant condition became observable, while the other concerns whether the process was controlled early enough to keep that condition from reappearing at the same point.
This ambiguity becomes sharper when leadership evaluates apparent improvement. Better detection performance can look like stronger quality control because fewer CTQ-related problems escape observation. That evidence has value, but its meaning is limited. It says more about the ability to identify a problem than about the status of the risk that precedes it. If the upstream source is unchanged, improved detection can coexist with continued production of the same quality-relevant condition.
Where leadership decisions become unclear
Leadership faces the decision once detection has established that the problem is real. At that moment, the available evidence can suggest either that the detection mechanism needs to become more capable or that control needs to move earlier in the process. Detection evidence alone cannot determine which interpretation is correct because it does not establish whether the upstream risk has already been controlled adequately.
That creates a decision-traceability problem. Leadership can see what was detected and when it became visible, while still lacking a clear connection to the earlier process condition that requires control. Without that connection, an apparent improvement in CTQ detection can be mistaken for an improvement in Quality by Design. The evidentiary record becomes strongest at the point where the problem appears, while the unresolved control question sits earlier in the process.
The unresolved control question
The tension does not make CTQ detection less necessary. Detection supplies the information that exposes the quality-relevant condition and makes repeated occurrence visible. Its limits arise when that information does not change where intervention occurs. The practical issue is whether detection informs earlier action or becomes the point at which the control process effectively stops.
For clinical operations leadership, the most consequential unresolved question is evidentiary: does the system show that CTQ-related problems are being found, or does it show that the upstream risk producing them is being controlled before they reach detection? Those forms of evidence are related, but they are not interchangeable. Quality by Design is not established by detection alone when the effectiveness of the relevant risk controls cannot be demonstrated. The unresolved distinction affects how oversight interprets control performance and how leadership reads evidence of improvement.
Sources:
[1] International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH). ICH E8(R1): General Considerations for Clinical Studies. Step 4 Guideline. Adopted 6 October 2021. ICH E8(R1) guideline.
[2] Meeker-O’Connell A, Glessner C, Behm M, et al. “Enhancing clinical evidence by proactively building quality into clinical trials.” Clinical Trials. 2016;13:439–444. PMID: 27098014. DOI: 10.1177/1740774516643491. PubMed record.
[3] U.S. Food and Drug Administration. Clinical Inspection Summary, BLA 761049, Bavencio™ (Avelumab). Center for Drug Evaluation and Research, Office of Scientific Investigations. FDA clinical inspection summary.