Guidance on minimising disruptions to the conduct and integrity of clinical trials of medicines during COVID-19
Actions that trial sponsors should consider to build resilience into clinical trial design
Introduction
This guidance follows on from that published on managing clinical trials during coronavirus (COVID-19) and has been developed to assist those involved in running clinical trials of medicines. This will avoid disruptions as a result of the continuing pandemic and will support appropriate incorporation of the available flexibilities into normal practice for the benefit of trial participants. It is relevant to those involved with ongoing studies, those wishing to restart trials that have been paused or temporarily halted, and those wishing to start new studies.
The HRA (Health Research Authority) have also issued guidance on conducting research in a public health emergency.
The first priority should be the safety of trial participants and this will remain our focus.
Preparation and training
We are aware that change management activities required to take full advantage of the available flexibilities were hampered during the peak of the pandemic. Preparation of Standard Operating Practices (SOPs) and training/guidance documents for staff not familiar with decentralised methods for trial conduct will help facilitate a smoother transition in the face of potential further disruption. Creating formalised, structured methods of conducting telemedicine will help in gathering consistent reliable data.
If direct-to-patient supply of the Investigational Medicinal Product (IMP) is appropriate, sponsors should consider any training or information the research staff and trial participant will need to allow self-administration of the IMP at home. If self-administration is not feasible the sponsor should consider the logistics of home visits and how to capture data from remote patient monitoring visits.
In preparation for moving to a decentralised model of trial conduct, sponsors should begin to set up vendors and conduct due diligence on providers, including home care providers.
We would also remind applicants that when preparing your submission to the MHRA you should consider our guidance on common errors identified during validation and assessment on clinical trial applications. Taking this into consideration can help increase the success of your application, avoid requests for further information and therefore reduce the time taken to obtain authorisation.
Investigational medicinal product dossiers (IMPDs)
Sponsors can avoid the need for substantial amendments to extend the shelf life of the investigational medicinal product (IMP) by including a shelf life extension proposal in the IMPD. This should ideally be done at the time of initial application but can be added as a subsequent amendment to ongoing studies.
Details of what would constitute an acceptable extension plan.
Protocols
One of the lessons learnt from COVID-19 is that protocols are often written too prescriptively, leading to unnecessary substantial amendments when a change needs to be made. The complexity of protocols has also increased to include, for example, multiple exploratory endpoints that may be impossible to measure in a pandemic situation. Sponsors should carefully consider what endpoints are needed to answer the research question and build flexibility into the protocol, without compromising on the safety of participants or impacting on the intended goal of the development program. It is entirely feasible and acceptable to prepare a protocol that incorporates appropriate descriptions of both the procedures for regulatory decision-making and flexibility in how clinical visits, monitoring of trial participants, follow-ups, etc are implemented. Use of ‘decentralised’ and digital/virtual elements in a study should be considered, as described in our guidance.
Hybrid designs may be more appropriate and we would recommend a careful review of the protocol schedule of assessments so that where clinical visits are unavoidable, they are consolidated. The aim is to have visits that are as efficient as possible, avoiding unnecessary travel and contact in preparation for social distancing requirements.
Recording protocol deviations
The purpose of recording protocol deviations during a trial is to ensure the safety of the participants and the reliability of the results. Protocol deviations can point towards non-compliances that need to be addressed and corrective and preventive actions being put in place. They can identify the need for a protocol amendment and improvements/updates to other key trial documents such as plans, Case Report Forms (CRFs), instructions or for further specific training. Significant protocol deviations, such as eligibility criteria and deviations that impact primary and secondary endpoints, need to be taken into account during analysis of the data, and if not documented and communicated can affect the reliability of the trial results.
Due to COVID-19, we understand that there will be an increase in protocol deviations due to trial management processes changing, participants not being able to attend visits, and protocol-defined timelines slipping.
It is expected that a risk-based approach is taken to the documentation of protocol deviations. Deviations that impact on participant safety and the reliability of results should be identified, documented and taken into account when the data are analysed.
Minor protocol deviations should be identified, and these can be documented at a trial level, rather than an individual level. For example, trial-level file notes can be used to document participant visits that were performed remotely, and IMP that was delivered directly to homes where this was not originally intended.
Out-of-window patient visits can also be documented at the trial level, if the window does not have an impact on safety or data integrity. Sponsors can set thresholds, with a rationale, and only those deviations outside the threshold documented at the participant level. These thresholds should be outlined in the protocol.
Examples of minor protocol deviations
Patient visit windows are generally minor deviations, but if critical parameters are not collected this could be major, especially if these parameters can change over time and therefore would be missed by extending the time window too much.
Minor breaches in patient confidentiality can also be considered minor deviations - for example, a single incident of patient-identifiable information leaving the trial site. If this is systematic, for example all consent forms are sent to sponsor without consent being obtained, then this would be a major breach.
Risk-based approach
Risk-based approaches are supported by regulators globally and are actively encouraged by the MHRA.
A comprehensive risk assessment is vital to the approach. The main goal of the risk assessment exercise should be to determine if and which trial modifications are necessary to protect trial participant safety and minimise impact on study integrity. The risk assessment should be performed in a blinded manner. Where available, standardised metrics of trial operational status should be employed to reinforce a consistent approach to risk monitoring and assessment.
Risk assessment should include assessment of the impact of the pandemic on:
- participant enrolment over time
- adherence to study treatment and use of concomitant medication
- ability to monitor and report safety data
- ability to monitor/verify source data for critical parameters
- validity of the methods for measuring primary (or surrogate for the primary endpoint) and secondary endpoints, if measured differently to what was originally intended
The sponsor should implement a risk-based monitoring program to help ensure that key data can be reviewed to meet the study completion deliverables, and to identify and prioritise at the onset the most important information to be collected in order to assess the primary endpoint(s).
Please refer to the MHRA Inspectorate blog, the HRA blog and the MHRA GCP Forum FAQs on monitoring/remote monitoring for further guidance.
There are a number of resources available to help with the risk-adapted approach, including FAQs and a blog. We have also published examples of risk assessments along with a researchers experience of using the approach, which is in two parts - part 1 and part 2. FAQs and examples of remote monitoring are also available.
Maintaining study and data integrity
Because of the impact of COVID-19 on clinical studies, the trial’s ability to meet its objectives could be compromised. Modifications to trial conduct, increase in protocol deviations, and protocol amendments are likely to occur, and these may raise important issues relating to study integrity and increase the potential for bias (operational and statistical). Moreover, planned statistical analyses (main and sensitivity analyses) may need to be revised or supplemented to provide a thorough and appropriate interpretation of trial results as a consequence of these modifications.
Assuring data integrity requires appropriate quality and risk management systems, including adherence to sound scientific principles. Systems and processes should be designed in a way that encourages compliance with the principles of data integrity, as set out in the MHRA data integrity guidance. In particular, data integrity and data flow need to be considered when using digital data collection tools, telemedicine, home visits, and local rather than central laboratories. The sponsor should consider how to get traceable/auditable data from non-traditional sources into electronic data capture systems.
A risk assessment of potential disruptions and their consequences for trial conduct, data integrity and interpretability should be carried out by the sponsor in order to determine if changes are required in the protocol. Consideration should be given to aspects such as impact on participants’ enrollment over time, adherence to study treatment, use of concomitant medication, and validity of the methods for measuring primary (or surrogate) and secondary endpoints, if measured in a different way to what was originally intended.
Changes to planned analysis methods may also be necessary, for example in order to handle pandemic-related missing data and/or intercurrent events. Below are some examples of aspects to take into consideration in the risk assessment that could be used to address the issues relating to study objectives, inference, and statistical analyses. A flexible monitoring strategy is key to mitigating study-level impact of the pandemic and ensuring data integrity.
The below bullets are adapted from Meyer RD et al, Statistical Issues and Recommendations for Clinical Trials Conducted During the COVID-19 Pandemic, Statistics in Biopharmaceutical Research, Statistics in Biopharmaceutical Research, June 2020.
1. Assess the impact of COVID-19 on trial integrity
- Changes to treatment course and adherence, protocol compliance, recruitment and retention
- Data quality, completeness or heterogeneity
- Trial power and possibility of success, and trial integrity and interpretability
2. Define mitigation measures
- Implement operational mitigation strategies and protocol amendments
- Revise estimands
- Modify planned analysis and/or add new analysis
- Change sample size, follow-up, interim analyses or other design elements
3. Document and communicate
- Implement changes in protocol, statistical analysis plan, DSMB charter, clinical study report
- Request the MHRA’s input
- Communicate with sites and other stakeholders
Sponsors should make every effort to minimise the amount of missing data and document the reasons for missing data (pandemic-related and non-pandemic related). The sponsor should assess potential impact of missing data on several aspects (e.g. feasibility of the planned estimation method, and study power).
Sponsors should identify, prioritise, and collect data that are essential for the interpretation of the trial (e.g. primary endpoint(s) and important safety endpoints followed by key secondary endpoints) despite missed visits. The data collection should be appropriately labelled in relation to the pandemic periods: “before”, “during,” and “after” the pandemic. The impact of alternative data collection modalities on data integrity should also be considered.
Trial modifications should only be introduced when there is a convincing scientific reason that they minimise the impact on integrity and interpretability of results. Such decisions should be based on blinded review of the accumulating data. The use of unblinded data is not recommended, as it can seriously compromise the integrity of the trial. The risk assessment should not become an unplanned interim analysis for efficacy.
The estimand framework (see ICH E9(R1) addendum) can be a very useful tool in the risk assessment exercise, even if the principles outlined in the framework were not originally employed in the design of the trial. The impacts of the pandemic on all estimand attributes (treatment, population, endpoint, intercurrent event (ICE), population level summary) should be considered.
The pandemic will likely lead to unforeseen ICEs, and these can affect the interpretation and/or the existence of the measurements associated with the clinical question of interest, as well as increase the occurrence of planned ICEs. It is recommended that ICEs are classified as being unequivocally the result of the pandemic or non-pandemic related (the latter should be handled as had been pre-specified in the protocol). It is important to capture the information associated with the pandemic-related ICEs, such as adherence to study treatment (e.g. timings, doses), study treatment accessibility (e.g. drug supply interruption), COVID-19 infection status (e.g. participant positive for COVID-19), and concomitant COVID-19 medication (e.g. participant admitted to intensive care). Once all potential ICEs are identified, suitable strategies for handling them should be chosen, considering their compatibility with pre-existing strategies, where applicable.
Population-level summaries describing outcomes for each treatment and comparisons between treatments should remain unchanged if possible. Alignment of the analysis methods, handling of missing data and relevant sensitivity and supplementary analyses with the primary estimand(s) are important for the validity of the trial results.
The consequences of the pandemic for the amount and/or distinct patterns of missing data will need to be considered. Sponsors should generally maintain the same approaches for handling of non-pandemic-related missing data as originally planned. If updates to pre-specified analyses (e.g. data handling strategies) are required, these should be made prior to unblinding.
Supplementary analyses may be useful in order to gain greater understanding of the general pandemic effect on trial outcomes. Subgroup analyses based on subgroups defined by pandemic impact, e.g. primary endpoint visits before or after pandemic onset, are recommended.
All the above described processes, from risk assessments to actual changes implemented in the protocol and statistical assessment, should be thoroughly justified and documented. Advice from the MHRA is recommended before important changes are made to the trial protocol.
Adoption of data-enabled approaches to patient recruitment
Use of ‘real world data’ sources, such as electronic patient health records (EHR) to conduct trial feasibility is a well-established practice. New approaches, such as the Clinical Practice Research Datalink (CPRD) data-enabled services, are now available for targeting and inviting patients who are potentially eligible for clinical trials, based on EHR searches against the protocol criteria. In addition to increased speed and broadening the scope of finding patients based on their medical history, this remote recruitment method augments other decentralised and digital adaptations that minimise the need for patients to physically attend trial sites.
Help from the MHRA
If you have any questions, contact the Clinical Trials Unit Helpline at Covid.clinicaltrials@mhra.gov.uk or 020 3080 6456. The Helpline will give inquiries relating to COVID-19 the highest priority, including the possibility for fast-tracking submissions.
Updates to this page
Published 11 November 2020Last updated 13 November 2020 + show all updates
-
Minor edits made to text and new tags added
-
First published.