Clinical Trial Registration Audit Report

The requirement to register 

The UK policy framework for health and social care research expects information about all research studies to be in the public domain before the research starts:

Whilst this is good practice for all research, there are stronger requirements for clinical trials. Clinical trials are:

• Clinical trials of investigational medicinal products (CTIMPs) 
• Clinical investigations or other studies of medical devices 
• Combined trials of investigational medicinal products investigational medical devices 
• Other clinical trials to study novel interventions or randomised clinical trials to compare interventions in clinical practice 

It is a condition of the Research Ethics Committee (REC) favourable opinion that a clinical trial is entered onto a publicly accessible register. Registration should occur before, or no longer than six weeks after, the first participant is recruited. 

All CTIMPs must be entered onto the EU Clinical Trials Register, with the exception of phase one clinical trials in adults which do not form part of an agreed paediatric investigation plan. This requirement is fulfilled by the Medicines and Healthcare products Regulatory Agency (MHRA), which enters these studies onto the EU Clinical Trials Register.

All other clinical trials should be entered onto a register on the World Health Organisation list of primary registers or the list of registers accepted by the International Committee of Medical Journal Editors.

For some research there may be reasonable grounds, such as commercial sensitivity around early phase trials, to defer registration. Researchers may apply to the HRA for permission to defer registration provided there is no legal requirement to register. 

It is also a condition of the REC favourable opinion that research tissue banks are registered on a publicly accessible database. 

Applicants for research ethics review can register their study before making their application and provide registration details in the application. If they register the study after approval, they are not required to submit registration information to the HRA, but may do so when updating study information, such as requesting approval for an amendment to the study. 

About the audit 

The purpose of this audit was to determine the number of clinical trials, given a favourable opinion between 1 January 2017 and 30 June 2017, that met the condition to register on a publicly accessibly database. Additionally, it was hoped that this audit would provide reasons for non-compliance, and therefore highlight any areas that could be improved. It was not designed to increase compliance although it did prompt some researchers to register their studies. Previous audit reports are available on the HRA website.

The audit followed the same methodology as the previous year’s audit of 2016 data. There were three main steps:

• Step 1: Data collection 
• Step 2: Registration search 
• Step 3: Follow-up emails to researchers 

Step 1: Data collection 

To identify studies for the audit, data was extracted from the HRA Assessment Review Portal (HARP). HARP is a database which records and tracks the progress of research applications submitted to the HRA. The following data was extracted from HARP on 11 December 2018:

• studies that received a favourable REC opinion between 01 January 2017 and 30 June 2017 (UK wide) and fell into one of the following categories: 
  - clinical trial of investigational medical product - limited to phase one trials involving healthy volunteers 
  - clinical investigation or other study of a medical device
  - combined trial of an investigational medicinal product and an investigational medical device
  - other clinical trial to study a novel intervention or randomised clinical trial to compare interventions in clinical practice.

The data did not include: 

• studies that received an unfavourable or provisional opinion
• all other clinical trials of investigational medicinal products 

The extract contained basic details about each study as well as any available registration details. Where a registration number was present HARP, this was recorded on the audit log. 

Step 2: Registration search 

For studies without a registration number in HARP, registration details were searched for online using the search engine ‘Google’. Searches took place between 7 January and 11 January 2019 using full title, short title and REC reference.

Where registration details were located, the registration number and registry name were recorded on the audit log as well as updated in HARP.

Where no registration details were located online, the study entry in HARP was further reviewed to look for evidence of a valid registration deferral or evidence that the study was closed early prior to recruitment. The audit log was updated accordingly.

Step 3: Follow-up emails to researchers 

Where no registration details could be found for a study and no valid reason for this could be found in HARP records, the Chief Investigator (for phase one CTIMPs) or Sponsor and Chief Investigator (for all other studies) were contacted via email.

All emails were sent out on 15 or 16 January 2019 with response deadlines of 29 or 30 January 2019. A reminder was sent one week after the first email was sent. All responses were recorded on the audit log. Responses received after the deadline were not included in the audit findings.

The email requested that they either provided registration details or a reason why registration had not taken place. Responses were categorised by:

a) registration status: 

• registered (despite not being found in steps 1 and 2) 
• registration in progress (registered but pending registration number)
• registered during the audit (audit email acting as a prompt)
• registered on a database outside of the HRA's requirements
• not registered but intend to register (audit acting as a prompt) 
• not registered and no confirmation whether registration is intended 
• no response 

b) reason for non-compliance (if provided):

• stated in A50 of IRAS that they would not register 
• did not know about the requirement to register 
• forgot to register/oversight 
• did not know how to register 
• did not consider study to be a clinical trial 
• thought that publication of research summary on HRA website was sufficient
• technical issue with registry website 

Responses were categorised based on the last correspondence received from the Chief Investigator (CI) or Sponsor.

Results 

Overall compliance 

The results show that the overall rate of compliance for the 581 studies involved in this audit was 81%.

In this report, ‘compliant’ refers to studies that had:

• registered
• registration in progress (registered but pending registration number)
• a valid registration deferral in place 
• not registered because the study had not started recruitment 

The date of registration (i.e. whether a study was registered within six weeks of recruitment of first participant as per the requirements) was not taken into consideration as this date is not recorded in HARP. It therefore means that some studies that had registered are categorised as ‘compliant’ but may have registered later than six weeks after recruitment of the first participant.

In this report, 'non-compliant' refers to studies that had: 

• not started the registration process at the time of the audit (though some of these did go on to register after the audit)
• registered on a database outside of the HRA's requirements (e.g. on a university website)

Figure 1: Pie chart showing the percentage of studies that were compliant with the registration requirements 

The pie chart shows that of all 581 studies included in the audit, 81% were compliant with the registration requirements, 14% were non-compliant with the requirements and 5% did not respond to the audit emails. These findings are similar to the 2017 (2016 data) audit results.

Download a csv file for data from figure 1. 

Compliance according to study type 

Compliance was also analysed per study type. For simplicity, this report uses abbreviations for the four clinical trial study types as follows:

• 'Phase 1': CTIMPs limited to Phase 1 trials involving healthy volunteers 
• 'Devices': Clinical investigation or other study of a medical device 
• 'Combined': Combined trial of an investigational medicinal product and an investigational medical device
• 'Other': Other clinical trial to study a novel intervention or randomised clinical trial to compare interventions in clinical practice

Looking at compliance rate according to clinical trial study category, the results show that almost all the phase 1 studies 1 (97%) were compliant with registration.

Figure 2: Bar chart showing the percentage of studies that were compliant with the registration requirements, per study type

The bar chart shows that for phase 1 studies, 97% were compliant, 3% non-compliant. For device studies, 82% compliant, 15% non-compliant and 3% did not respond to audit emails. For other studies, 77% were compliant, 15% non-compliant and 7% did not respond to audit emails. Of the 581 studies included in the audit, none were combined trials.

Download a csv file for data from figure 2.

Responses to follow up emails 

Following steps 1 and 2 of the audit, there were 183 studies where registration details could not be located either in HARP records or following an online search. The following results relate to the responses received to follow-up emails that were sent to these 183 studies (email sent to CI and/or Sponsor). The results show that 23% of respondents confirmed that the study was not registered but that they intended to register the study after the audit, an additional 14% of respondents did not confirm if they would register their study and 11% reported that they had already registered the study before the audit although the registration details had not been found during the online search, this is most likely because a different study title was used in the registry compared to the REC application/HARP record.

Although the purpose of the audit was not to enforce compliance, at least 70 studies were registered or going to be registered because of receiving the audit follow-up email, which seemed to act as a prompt to register studies.

Figure 3: Bar chart to show responses received to the follow-up emails regarding registration status 

The bar chart shows the number of responses to follow-up emails that fall into each registration status category. The graph also shows an additional breakdown for each study type. Of the 183 emails sent, the most common responses was that 23% stated that their study was not registered but they intended to register it after the audit, this was followed by 15% of respondents stated that the study was not registered because the study never started or recruitment had not started yet, and 14% confirming that the study was not registered and no confirmation whether registration was intended. Of the 183 emails sent, 16% did not receive a response. Most of the non-responses were within the ‘other’ study category and all of those contacted regarding phase 1 studies responded.

Download a csv file for data from figure 3. 

Reasons given for non-compliance 

Those respondents who reported that the clinical trial had not been registered gave several different reasons for non-compliance. Of the 74 who gave a reason for not registering, 38% reported that they had forgotten to register, some explained that there had been changes of staff and roles within the research team and that registration had been overlooked during the handover process. Most of these went on to confirm that the email had prompted them to begin the registration process.

Around a third of respondents said that they did not consider the study to be a clinical trial. Most of these responses were regarding ‘other’ types of clinical trials. Common reasons for not considering the studies as clinical trials were because the they were student, feasibility or non-interventional studies. Some respondents stated that they did not know about the registration requirements, one of these referred to a Phase 1 study where they were unaware that phase 1 trials, conducted solely in adults and which are not part of an agreed paediatric investigation plan, are not public in the European Union Clinical Trial Registry.

Figure 4: Bar chart to show responses received regarding the reasons for non-compliance 

The bar chart shows that the most common reason for non-compliance (38%) was that registration had been an oversight or forgotten. This is closely followed by researchers not considering their study to be a clinical trial (31%) and then 14% stating that the did not know about the registration requirements. Other reasons included technical issues with the registry, assuming that the research summary on the HRA website was sufficient and not knowing how to register.

Download a csv file for data from figure 4.

Types of registries 

The results show that clinicaltrials.gov and ISRCTN were the two main registries chosen to register clinical trials. All phase 1 studies were registered on clinicaltrials.gov, whilst 42% of ‘other’ clinical trials chose ISRCTN. 

Figure 5: Bar chart to show where studies were registered

The bar chart shows that all Phase 1 studies were registered on the clinicaltrials.gov registry. For device studies, 72% were registered on clinicaltrials.gov, 22% on ISRCTN and 4% on other registries. For other clinical trials, 53% were registered on clinicaltrials.gov, 42% on ISRCTN, 3% on clinicaltrials.gov and ISRCTN and 4% on other registries. Some ‘other’ clinical trials confirmed that they were registered on multiple registries.

‘Other registry’ includes: Australian New Zealand Clinical Trials Registry (ANZCTR), Brazilian Clinical Trials Registry (REBEC), German Clinical Trials Registry (DRKS) and Research Registry.

Download a csv file for data from figure 5.

Observations

Overall, the results of this audit show that there is not full compliance with the requirement to register a clinical trial on a publicly accessible database.  Most of those who had not registered their studies at the time of the audit expressed a willingness to register. Although 81% of the studies in this audit were compliant, by the end of the audit, 91% of all the studies were either registered or researchers had stated that they intended to register. The reasons given for non-compliance were lack of awareness of registration requirements, both in relation to the study category type or where it should be registered. 

Comparisons to 2016 data 

The results are very similar to the previous year’s audit, showing that the compliance rate has neither improved nor declined. The two main differences between this audit and the 2017 audit were:

• In the 2017 audit, 90% of device studies were registered on clinicaltrials.gov whereas this audit showed a wider range of registries being used, 72% on clinicaltrials.gov.

his audit had a better response rate of 83%, an increase of 21%. This may have been due to the time of year that the audit was conducted, as the previous audit was conducted in August when it is more likely that individuals would be on annual leave.

Willingness to register 

Of the 581 studies involved in the audit, 91% were compliant with the requirement to register or had made plans to become compliant. Of the remaining 9%: 4% did not confirm whether they would register their study and 5% did not respond to the audit emails.

Most responses to the audit follow-up emails were apologetic and researchers showed eagerness to register as soon as possible or had already registered the study before the audit deadline. The majority of those that responded indicated that they would be happy to pay a registration fee. This audit has been the catalyst for registration and many clinical trials would not have been registered if contact had not been made.

Registry type 

The audit showed that some Sponsors and CIs were aware that they needed to register but were not aware of the accepted registries. Some thought that the publication of the research summary on the HRA website was sufficient, and many assumed that publication on an organisation website was enough to fulfil the registration requirement. 

Definition of a clinical trial 

One of the most common reasons for non-compliance was because the CI or Sponsor did not consider the study to be a clinical trial, even though a clinical trial category had been selected in IRAS and the REC final opinion letter included wording that stated that the study was a clinical trial requiring registration. 

Responsibility for registration

The most common reason for non-compliance was oversight or forgetting to register. Most responses went on to explain that there had been changeover in staff and that the task of registration had not been handed over. This may highlight that registration is not prioritised well within research teams and there appears to be a lack of certainty regarding which individual or team is responsible for registration. 

Identifying registration entries 

Searching for studies in Google has proved challenging as study titles are often different in the registry from their IRAS application and HARP. It was noticed that the REC reference number was usually included in the ISRCTN registry, which helped with the search, but was often not included in other registries. 

Actions

There is a clear need to increase awareness amongst Sponsors and CIs of the requirement to register clinical trials. We will clarify our guidance and introduce new learning as part of our work to implement the Make it Public research transparency strategy.