ICR/EMWA conference on writing clinical trial protocols: Audio wrap-up

Following on from the ICR/EMWA joint symposium that took place earlier this week (and was partially reported on this blog), I have interviews with Wendy Kingdom and Adam Jacobs, who chaired the morning and afternoon sessions respectively.

These will be posted on the CRfocus website (www.crfocus.org) but as I’m currently having technical difficulties there, I’ll post them here as well…

Update: These files are now available from the main CRfocus website, and have been removed from the temporary host:

Interview: Wendy Kingdom on ICR/EMWA joint symposium on Writing Clinical Trial Protocols

Interview: Adam Jacobs on ICR/EMWA joint symposium on Writing Clinical Trial Protocols

Sue Mackay on the study nurse view of trial protocols

Sue Mackay of CDSS leads a team of field-based study nurses, who deal with protocols on a practical level, often interacting with protocols when they are complete, and sometimes feels that she is trying to fit the square peg of a protocol into a round hole of practicality, and she thinks it a pity that more practical issues are not considered earlier in the process.

Sue described the typical role of the research nurse, listing the many activities that they do, from patient identification to AE reporting and data queries. She highlighted site study feasibility and protocol review as key areas for this conference.

A well-written and consistent trial protocol is very welcome, and makes the implementation far easier. As topics like subject selection and study procedures flow from protocols, advice on how they are done in the field could be flowed back into protocol development.

For example, she suggested that we could be missing patients because of slight differences in inclusion criteria; surely it would be better to consider this at a draft stage rather than asking for patient exemptions. She also discussed timelines of first visit to site vs time to consent: too short can be impractical but too long can cause identified patients to lose interest.

Similarly, timelines for study procedures and frequency of patient visits might not be feasible or acceptable for patients. Other hospital departments might not be flexible enough to meet the timelines of the protocol, given their primary role in general patient care.

Sue also suggested that we think about aspects of the protocol from the patient’s point of view: timing of procedures might not fit easiy with patients’ family and work commitments, in terms of duration and of time of day.

Sue again stressed the importance of clarity and consistency within the protocol, as it can cause confusion and waste time when the protocol is implemented. She gave an example of a ‘requested’ screening procedure in an asthma study that would not be routine treatment in that therapeutic area.

Next, Sue discussed subject selection and compliance. We need to be careful in where we place studies, that the site has access to the right type of patient. Similarly, is the duration of the study or the frequency of visits appropriate to the therapeutic area, eg, oncology patients perhaps deteriorating during the course of a study. She also discussed the implications of the economic conditions on willingness to take part in a trial: routine commitment to site visits during a trial might make a patient less employable, so they might decide not to take part.

Offering some solutions, Sue suggested more, broader communication is vital, although the route for communication between nurses and medical writers is less clear. Similarly involvement of other site staff (eg, labs, radiography, physiotherapy etc.) in protocol development would be beneficial. She reiterated the call for consistency.

In summary, she stressed that these ’square pegs’ are people, and that they need to be considered more in order to retain their goodwill and willingness to take part in studies.

One delegate asked how site staff use a protocol; Sue answered that she requires her team to read the protocol and investigator’s brochure in detail and refer to them regularly, even though it is used in conjunction with other notes.

A project manager in the audience mentioned that she sends draft protocols to investigators, and was a little surprised that it isn’t circulated to the site study team at that stage. Perhaps research nurses could help to educate their investigators on consulting more widely at the draft stage. Sue agreed that it would be helpful to be more pro-active in this area. Another project manager in the audience said that it is difficult to contact study nurses directly, with contact going through investigators.

Adam Jacobs on statisticians in protocol development

Adam Jacobs spoke next. Adam has a background in medical writing, but is also a statistician and sits on an ethics committee. His main point was that it’s never too early to ask a statistician.

Adam discussed military precision in writing protocols. The most important of these is to Keep It Simple. A protocol should be as simple as it should be and still achieve its objectives, for ease of implementation. Once an objective is selected, they must be maintained, with anything else being omitted. However, morale is also important, so its vital to keep key stakeholders informed, so that any problems can be identified early. Flexibility and contingency planning are also important, as is robust communication.

Adam went into communications in more detail, particularly for dealing with statisticians, who can seem to speak a different language to the rest of us! However, effective communication is vital. His top tip is that there is no substitute to face to face meetings. He said that statisticians are used to not being understood, and to explaining again if you let them know what you haven’t understood.

Adam described the statistician’s role: specification of objectives, trial design, analysis method, timing and choice of outcome methods and sample size. These are performed by the whole group, but led by the statistics.

Objectives are important to agree in detail as early as possible, particularly in later phase studies, and must be in a form that can be tested statistically.

Study design may be dictated by these objectives, but there are other elements that need to be considered (eg, blinded, parallel or crossover etc.) Methods must be carefully chosen to avoid bias; however these are not always possible in the real world, due to practical or ethical constraints. Compromises may need to be made to make the study achieveable, and this may require careful negotiation.

Selection of endpoints is largely a clinical decision, but statisticians can advise on freedom from bias and ease of analysis. This can also impact on sample size. Again, compromise might be required.

Analysis methods are mostly left to the statistician, but clinical input is still essential, as different analyses might be possible, and they might provide subtly different clinical information. Examples include repeated measures vs endpoint analysis, number of adverse events vs number of patients with events etc.

Again, even in sample size, many other non-statistical inputs are required, some of which might be subjective! This includes consideration of a ‘clinically relevant difference’, which has a significant impact on the sample size and needs to be discussed more widely. Sample sizes can be significantly more than imagined by non-statistians, which can raise budgetary issues.

Adam moved on to consider protocols from his perspective as an ethics committee statistician. He highlighted the importance of completing the ethics application form correctly, so it gives the committee the information it needs to review the study. The application form is the primary document, rather than the protocol, so don’t assume that the protocol will be read. Also, the committee contains lay members, so the application needs to explain some concepts in more detail than might be expected for medics alone.

When reviewing a protocol, Adam made particular mention of the sample size: too large is unethical (but rare, because of cost), and too small is also unethical as the study won’t answer the question! The committee needs to decide that the balance between risks and benefits involved with the study is acceptable. This depends on the scientific validity of the study.

Other common problems his ethics committee sees include: poorly written patient information sheets (in as much as 50% of cases!) including careful description of risky or unpleasant procedures, poorly explained methods, and inconsistencies in the application and between the application and the protocol.

In response to a question, Adam agreed that patient information sheets are getting longer. Certain elements are required by ICH GCP and templates are available from the NRES website. Adam suggested that a study summary would enable the patient to decide whether to read a more detailed subsequent description. However, some studies fall outside ICH GCP (eg, of a registered drug within its indication) a simpler approach would be beneficial.

Another question mentioned adaptive trials. Adam said that expert advice is a pre-requisite, but also suggested that checking assumptions of sample size midway through the study can be beneficial. This can be tested without breaking the blind, and Adam suggested we do this on any ongoing study; increasing the sample size mid-study could avoid a study failing to show a conclusive result.

Pharma view on trial protocols

Sandra Waechter, Senior Project Manager with Janssen-Cilag, gave the pharma view of protocols. She outlined the overall process for developing a protocol, moving from global and regional product strategies, through research concepts and development plans, down to the level of individual study protocols. The development plan contains input from various stakeholders (eg, medical, regulatory, health economics etc.) on registration requirements, how to develop an indication, line extension etc. From this, it can be decided what studies are necessary to close any knowledge gaps.

A research concept is developed to plug a knowledge gap. This is sometimes developed with or without a specialist medical writer, but always includes primary and secondary objectives, scientific rationale, dosage, population, statistical plan etc. This concept is then reviewed and approved within the company to ensure that any potential safety and efficacy concerns are assessed. This is a global review, which can be complex due to the range of stakeholders and the need to align with global straetgy.

From this stage, the protocol itself is developed. Medical writers are always involved from this stage. A physician is responsible for the study, to discuss key features with key stakeholders (within and outside the organisation), prepare the synopsis and then forward it to the medical writer for further development. This is done using a standard template, based on the concept, timeline and event schedule. Sections are assigned to other specialists to manage delivery of, eg, the statistics section. It’s important to meet appropriate pharmacovigilance requirements, and that consistent terminology, structure and content are used, as inconsistency can trigger queries or imperfect implementation by site investigators. The medical writer then distributes the draft protocol to team members (including local operations teams) for review, specifying timelines for response.

Comments should be consolidated and reviewed, with the medical writer arbitrating changes if necessary. Moving on, the medical writer prepares and circulates the final draft. This needs to go for broader approval by global company stakeholders, including statisticians, medics, lawyers etc. After any changes, the final protocol is finalised and signed off.

Sandra shared her company’s definition of protocol revisions vs amendments: basically, whether it has been distributed to competent authorities, ethics etc prior to the change request arising. Any revisions or amendments need to be reviewed in a similar way as before. Again, Sandra highlighted the importance of version control throughout this process.

Sandra then spoke about her role as a project manager. She assembles the protocol team, covering all relevant areas of knowledge, defining tasks and setting timelines and lines of communication. She manages the development process, developing a budget, creating a realistic milestone timetable and ensuring appropriate quality processes are followed and ultimately tracking progress and driving execution to time and budget.

She outlined her expectations of medical writers: to develop a deliberated and well-written protocol that clearly addresses the describes the research question and study objective. The introduction is vital, which must be up to date with current literature, positioning the research question in this context with correct citations. The protocol must contain enough detail to enable investigators to conduct the study, including consistency in wording. The medical writer should be pro-active in approaching stakeholders to collect information, organise copies of all study related materials to be included in the appendices before submission, and ensure that the document complies with any relevant regulations and guidelines.

In response to a question, Sandra said that the whole process should take around 3 months.

ICR/EMWA conference on trial protocols

This morning, I’m live-blogging from the ICR/EMWA joint symposium on collaboration in developing clinical trial protocols. This meeting will be reported in a future issue of CRfocus and The Write Stuff, with the afternoon session reported by Alexandra Dedman.

Wendy Kingdom opened the meeting and then handed over to Debbie Reynolds, Senior Medical Writer at Dianthus Medical. She discussed the details and diplomacy involved in writing a study protocol from the perspective of a medical writer.

She described the people involved in the process: medical writer, statistician, investigator (internal and/or external), sponsor representative and hopefully a monitor to advise on the practicality of the process.

Speaking as a medical writer, Debbie highlighted that they are specialists in putting together complex documents, getting input from other experts, and devising the protocol with the eventual study report in mind, making it easy to understand and to implement. She put the medical writer’s role in the centre in terms of diplomacy to resolve disagreements between other contributors.

Debbie then went on to discuss some of the common problems. These included agreeing a detailed synopsis, coordinating the team, resolving disagreements (eg, power vs cost), meeting all the requirements of sponsor, investigator etc., coordinating comments and version control between the multiple other contributors.

She then discussed examples of how it can all go wrong, such as a team member missing a deadline, with knock-on time issues, team members changing their minds, confusion of differing versions etc.

Suggesting how the process could be improved, Debbie suggested using a specialist medical writer, agree on a detailed synopsis, put somebody in charge of coordination, be strict on version control and deadlines, and finally to trust the medical writer’s skills.

Bringing her presentation to a close, Debbie discussed the CDISC protocol standard to enable machine-readable protocols, using XML to tag data fields in a structure. This makes it easier to generate CRFs and study databases (and, of course, a human readable version). Other advantages include that each data field is used only once, so a single change can be flowed through the whole of the document, rather than needing to make the same change multiple times.