From design to data: achieving quality in clinical trials

Emma Law, Head of Clinical Quality Assurance, writes for The Research Quality Association (RQA).

As a quality assurance professional, I have always been interested in the evolution of quality and how it can improve the conduct of clinical trials.

Interestingly, the origins of quality can be traced back to the guilds of medieval Europe. The guilds developed strict rules for product quality, which were enforced by inspection and confirmed by applying a special mark to flawless goods.

Over time, the approach to achieving quality has shifted away from focusing purely on inspecting the end product (quality control) to prospectively building quality into the design of products from the outset (quality assurance).

As W Edwards Deming, the 20^th century quality expert, put it:

“Inspection does not improve the quality, nor guarantee quality. Inspection is too late. The quality, good or bad, is already in the product.”

Instead, everything must be designed in such a way that the end result could only be a high-quality product or service.

“Quality by design” in clinical trials

The same shift towards quality assurance has occurred in the design and delivery of clinical trials – including the use of “quality by design” principles focused on building quality into the design of trials from the outset.

In recent years, guidance has been published to encourage quality by design, for example by the Good Clinical Trials Collaborative and the Clinical Trials Transformation Initiative.

My role as Head of Clinical Quality Assurance at Protas is to understand what we mean by quality in a clinical trial setting, and to ensure we’re achieving it in trial design and operationally, and even in our culture, right across our organisation.

Our approach to quality is to focus on the issues that materially matter to the credibility of trial results and the care of future patients, such as:

• Using quality by design principles to design streamlined and informative trials, avoiding complex protocols which demand too much data and long participant visits.
• Building quality into trial operations including defining and mitigating against any important risks upfront.
• Utilising Cantata, Protas’ purpose-built clinical trials management platform, to make our trials easier to conduct at research sites by guiding site staff through the requirements of each participant visit.

Implementing risk-based monitoring strategies focused on identifying any important issues quickly, rather than implementing a predefined schedule of on-site visits.

Management of trial data

High-quality data ensures trial results are credible. It is vitally important to achieving high-quality clinical trials.

However, it is important to recognise that high-quality data in clinical trials does not mean “perfect” data. Any errors in the data are likely to be unbiased with respect to the randomised treatment allocation (in other words, be of about the same size and frequency in each arm) so they do not materially affect the randomised analyses.

For example, a moderate number of errors in the measurement of blood pressure might affect the average blood pressure in an individual, but if similar errors affect another individual in the other arm of the trial then these errors “cancel each other out” in the randomised analyses, and the effect of treatment on blood pressure can still be estimated accurately.

In this way, not all data errors matter to the same extent when assessing the quality of trial data. Efforts to identify and correct data errors should focus on those errors which, if left uncorrected, would have a material impact on decision-making in relation to the reliability of the results or the rights, safety and wellbeing of the participants.

The best way to ensure high-quality data is by designing the trial and its data collection tools in a way that maximises the chances of important data being correct at the point of data entry. This mirrors the NHS “Getting It Right First Time” programme, which was implemented to tackle variations in clinical care and to drive up quality standards.

However, despite the more widespread use of quality by design methodology, the approach to ensuring the quality of data is still more focused on quality control ie checking that data is correct once it has been collected.

There are many reasons that could delay or prevent the correction of data after it’s collected, such as overly complex protocols which demand too much data, overstretched resources at research sites and participant dropout. Therefore, this is not an optimal way of guaranteeing data quality.

Steps to ensure data is correct at the point of entry could include:

• Streamlining data collection so that only the data required to answer the research question and to ensure the safety of participants is collected.
• Determining the operational feasibility of data collection tools with research sites during the trial design stage, including ensuring the case report form is structured around a logical flow which could be easily implemented at sites.
• Utilising electronic data capture systems to include validations at the point of data entry to prevent impossible or improbable data being entered, and ensuring that all necessary steps have been completed during the visit before the form can be submitted.

There is still an important role for quality control activities in clinical trials to support an overall quality assurance approach, particularly as part of the trial monitoring process. However, a proportionate and risk-based approach should be implemented focusing on the areas that are critical to trial quality ie factors that will affect the credibility of the trial results and the safety of participants.

Disproportionate approaches to quality checking data not only add cost and complexity to the conduct of trials, but also risk detracting away from core trial activities and affecting overall trial quality.

Ultimately, the goal is to help patients to get the treatment they need by ensuring that clinical trials are far more likely to provide useful answers. It’s important for trial quality activities to be proportionate and remain focused on this end goal – in which quality plays a critical role.

This article was first published in Quasar, the membership magazine of the RQA.