top of page
Search

Critical-to-Quality factors and Quality Tolerance Limits: what are they and how do we choose?

Updated: Jun 26


To start, here's a definition of Critical-to-Quality (CtQ) factors from ICH E8 (R1) [1]: “A basic set of factors relevant to ensuring study quality should be identified for each study. … These critical to quality factors are attributes of a study whose integrity is fundamental to the protection of study participants, the reliability and interpretability of the study results, and the decisions made based on the study results.”


According to this Pubmed article a Quality Tolerance Limit (QTL) is "a level, threshold, or value associated with a parameter which is critical to quality."


So essentially CtQ factors represent those aspects of the study that are vital to ensuring successful analysis of drug efficacy and safety along with assurance of patient safety. The Quality Tolerance Limits (QTLs) are thresholds that indicate the CtQ factor is at risk (of not being achieved or being achieved, depending on how it's defined). For example, if including a certain number of genetic subtype participants is a CtQ factor, then the QTL might be a specific dropout rate of those patients.

Quick note. I do want to clarify my stance on CtQ factors vs. "other" factors or data points:


Defining CtQ factors places added emphasis and focus on more important data points; however this should not result in ignoring other data -sometimes this is data required to meet CDISC submission standards, important for secondary analysis or otherwise may require narratives and justifications if missing, illogical or erroneous. Essentially, all data should be subject to minimally standard checks, monitoring and cleaning -doing so at the point of capture through automated logic checks is a great way to handle certain non-critical data.



For a given clinical trial there are some general areas to help define CtQ factors:


  1. Sample size and population: What is the minimum number of participants, disease traits, demographic representation and number of sites needed to provide statistical power so that the results are representative and analyzable?

    1. Examples:

      1. A minimum of 250 participants are needed.

      2. Genetic trait XYZ must be present in order for the treatment to be evaluated.


  2. Efficacy: what are the parameters (assessments, biomarkers, lab values...) that must be present in order to properly understand the efficacy profile of the treatment?

    1. Examples:

      1. HbA1c must be collected at baseline, at least twice during the study and at end of treatment in order for a study participant to evaluable.

      2. The MADRS score for suicidal thoughts will be required for primary efficacy assessment, even if the other portion of the MADRS is incomplete.


  3. Safety: what are those parameters (relevant medical history, Adverse Events, vital signs...) that must be captured to ensure the safety of participants and to properly evaluate the safety profile of the treatment?

    1. Examples:

      1. Liver function enzymes must be checked weekly based on the potential for treatment-related toxicity.

      2. Serious Adverse Events must be fully evaluated for all criteria as well as relatedness. All SAE data must reconcile with medical history, concomitant medications, vital signs and physical exams.


  4. Protocol adherence: what protocol requirements if not met would undermine the study results?

    1. Examples:

      1. ICF was not signed by the participant.

      2. SAEs were not reported due to inappropriate PI judgement.


  5. Data integrity: what within the protocol operations would undermine the results?

    1. Examples:

      1. Certain lab results were not masked/blinded and potentially biased the site staff.

      2. Participants were incorrectly assigned treatment.


There are resources on the web to help with thinking through some of the CtQ factors, including:



Here's the expanded definition of Quality Tolerance Limits (QTLs):


"A QTL is a level, threshold, or value associated with a parameter which is critical to quality. QTLs are set for risks identified at the trial level. A deviation from a threshold during the conduct of the trial may indicate a systematic issue that could impact participants’ safety or reliability of trial results." -see this article which includes this definition and provides a nice summary of defining QTLs for clinical trials.


So important aspects of QTLs:

  1. They are related to CtQ factors.

  2. They are data points or thresholds that allow specific tracking and reporting.

  3. They are study level (impacting overall study success), as opposed to Key Risk Indicators (KRIs) which can be study, country, site and participant level.


So in thinking about that definition of QTLs, here are some considerations:


  1. Define your QTLs based on your CtQ factors: if you have honed in on a primary efficacy endpoint as a CtQ factor, the QTL could include a study-level trend of missing primary endpoint values across a certain % of participants.


  2. The QTLs should be easily reportable, and should be considered "critical alert" data. Critical alerts should have reporting mechanisms that are triggered, e.g. a notification to key study team members. If that is not possible then the QTL should be part of a regular review, for instance as part of a regular study team call.


  3. QTLs represent the "limit", a level of escalation. Since this is the case it can be helpful to have KRIs that provide earlier signals for proactive action. Since KRIs can be defined at the participant, site and country levels they can be used to take action at those levels, and before the trend at a study level becomes "critical" to the study success.



Critical-to-Quality factors and Quality Tolerance Limits help focus the study design. The goal of defining CtQ factors and QTLs is to center the study team around what matters. Then managing the risks associated with these quality parameters can be delivered through a robust RBQM process.


Comments

bottom of page