home > ict > winter 2015 > probable cause
International Clinical Trials

Probable Cause

The last 25 years have seen some major advances in pharmaceutical R&D. Competitive enrolment, risk-based monitoring and adaptive study designs have been introduced into business processes. There has been startling technological progress, from the rise of interactive voice response, electronic data capture and electronic patient-reported outcomes, to the recent advent of virtual trials and websites to broker the enrolment of patients. And pharma continues to expand into emerging markets to secure more cost-effective studies with targeted patient populations.

Clearly, the speed and effectiveness of clinical trials have improved significantly over that time. But we are still left with some serious issues to overcome – the most important of which are generally agreed to revolve around patient recruitment. Feasibility assessments, in particular, stand out as there appears to be wide consensus that they are uniformly inaccurate, and yet they remain stubbornly immune to improvement efforts.

Flawed Model

The process generally used to assess the feasibility of trials – the paper-based feasibility questionnaire – is both fundamentally flawed and hopelessly outdated. We all know how it works: a summary of the protocol inclusion and exclusion criteria is sent to physicians, and they are asked to determine the number of patients they might be able to recruit per month. Sponsoring pharma companies provide limited information in the first place, ask for a reply in two weeks and do not pay physicians for their efforts (more popular investigative sites apparently receive eight to ten questionnaires a week). The results are returned and then halved for good measure as they are not believed.

But this model simply does not work. Moreover, it masks some much more serious issues. We are focused on the number of patients who fulfil the protocol inclusion and exclusion criteria at each site. While these have an effect, other much higher impact criteria must also be considered to determine the speed at which a trial will run.

These criteria are found further down the recruitment funnel and include the lack of patients ever presented with the option of a trial, the resources available at the site to run the study, and the impact of the research on the patient’s life.

Recruitment Funnel

The recruitment funnel tracks where a trial tends to lose patients. A recent article by Beth Harper and Gen Li, using their own data and that from several other trusted sources, shows that the loss of patients is predictable and has been fairly constant across multiple sources for the last 15 years (1).

There are two general categories: the parts of the funnel that are specific to the science in the protocol – screening criteria and such like; and the parts of the funnel which are patient-centric – namely, consent and drop-out postrandomisation. In summary, across all studies we can expect to lose 50% to 60% of patients at the point of consent after pre-screening, and 36% of the remainder entering screening.

Research on how many patients have the option of the trial presented to them is hard to find, but there is research indicating that no more than 30% of potential patients have the study discussed with them. The other 70% never hear about the trial (2).

Recent evidence from an analysis of electronic medical records at the Mayo Clinic in the US – also by Harper – found that, based on its case studies, the adjustment of screening criteria could improve recruitment by around 20% (3). The last piece of relevant data, from Tufts Center for the Study of Drug Development, showed only 14% of all trials use any other recruitment method over traditional methods – defined as physician referrals and mass media (4).

So, the impact of addressing resources and consent rates is much more powerful than addressing inclusion and exclusion criteria. Of course, if we could do all of this, recruitment would be enhanced significantly.

Addressing Resources

A feasibility assessment needs to consider primarily what we are going to do to help the sites and what they might accept. For example, changing the inclusion criteria is difficult, expensive and runs the risk of invalidating the science to some degree; but, if we place resources within sites to treat patients, we can increase the likelihood of the patients being made aware of the trial and ensure sites have the resources to look after all those they are able to recruit.

Specialist services can provide sites with assistance, while there are other approches too – for example, nurse banks are available for the site to hire from, or they can take on permanent staff. Each approach has its pros and cons, but the standardisation you get from a team of specially trained nurses under control of the sponsor, ensuring they are well-aligned with the aims of the study, can be particularly beneficial.

Other companies can provide different types of resources, such as study site coordinators (SSCs), to ensure that patients are made aware of the trial – although they cannot actually treat the patients, and so nurses may be required as well. There are not many of these companies around, but there is typically at least one in most countries that take part regularly in trials.

Another model is where the site is an independent business focused solely on running trials commercially; however, not many of these exist in Europe. The advantage here is that the resources are again well-aligned with the sponsor. The disadvantage is that the access to patient information in the referring site is not present, so it is not possible to ensure that the maximum number of patients have been made aware of the trial.

Better Consent Rates

The next challenge is how to improve consent rates. Once more, the options are limited. One solution is home trial support, where the trial visit takes place in the patient’s own home, typically for around 50% of the visits or in conjunction with a virtual trial design.

This approach, favoured by the Medical Research Network, has been shown to increase recruitment by 50% or more by improving consent rates, and increasing the pool of patients to consider for the trial to include those who live much further away – previously too far for multiple trial visits.

An additional method is to reduce the number of tests the patient has to undergo, the overall duration of the trial, and the number of invasive procedures – although these measures are not always appropriate, and could once again compromise the science. However, evidence suggests that we probably measure as much as 25% more data than we actually need in trials, so reductions are desirable and important to consider (5).

When selecting actual sites in a feasibility assessment, there needs to be – as now – a simple check against the original protocol design to ensure the patients exist. Added to this must be a careful assessment of whether the sites will need extra resources, such as nurses or SSCs. A further consideration is whether the sponsor is willing to refer to an independent site, or if they are willing to use other tools, such as home trial support, to reduce the impact of the trial on the patient’s life.

Digital Feasibility

A site-driven model for trial recruitment continues to be used – but in today’s digital world, this need not be the case. The reason for a site-driven approach is access to patients; pharma companies go to a large number of sites because they can find patients, yet they only have so many. However, it is possible to go direct to the patients by accessing social media and internet-based data.

Even compared with the 100 to 200 million prescription data records available within the US, collected over several years, today’s traditionally sourced information is dwarfed by the size and currency of the data available through social media.

Access to this sort of data to analyse for detailed protocol planning – where upward of 170 billion conversations can be analysed for relevant information, all collected in the last few months – is available through analysis software. This can build a very accurate map of the locations of high-quality screening candidates for trials worldwide.

It is possible to not only determine exactly where patients can be found geographically, but also to analyse the interactions between people in the dataset. You can map their connections and therefore learn exactly where they go for data about their illness and where they search for data about clinical trials.

By using this data, a sponsor can design an outreach campaign focused on exactly these influential sites, and through other tools – for instance, Google AdWords – produce a laser focus for advertising spend. This targeted approach gets the information to possible patients so they can decide to opt into the trial recruitment process.


Where does this leave the feasibility assessment? Using these tools, a much more intensive search can be carried out on whether these patients exist than has been possible before. The inclusion and exclusion criteria can be checked to significant depth (these patients are looking for trials, so publish their previous medication, diagnostic criteria, stage or grade, previous studies they have been in, or how long they have had the disease in question). Furthermore, it is possible to analyse where the competitive trials are taking place using other databases and then to place sites near where the patients are, rather than the other way around.

This data turns the whole process on its head, focusing on the patient first and then the site. Once the sponsor knows exactly where patients are and where to avoid – because of competitive trials – it is then necessary to create a screening website or a call centre that patients can contact to register their interest and get a referral to an investigative site.

Again, if this approach is combined with the correct resource allocation at the site and a willingness by the site to use tools such as home trial support, then many more patients could be recruited from further afield. Sites could look after 20 to 50 patients each, instead of 5 to 10 – significantly reducing the number of sites needed for the trial. As a consequence, under-performing sites would be eliminated, driving down costs and increasing efficiency and quality.

Call for Change

Feasibility assessments as they currently stand are almost worthless. When predicting the outcome of recruitment to a site, the focus on the number of patients who fulfil the inclusion and exclusion criteria is hugely outweighed by the site’s resources, as well their willingness to use tools that will reduce the impact of the trial on the patient’s life. If paper questionnaires focused on these issues, this alone would be a major improvement. However, the digital data available today is immensely powerful as a fundamental source of feasibility data. Changing the approach to locate patients directly, rather than through sites, and putting the whole set of solutions together would enable studies to be managed with fewer sites, for less expense and to higher quality.


1. Harper B and Li G, The PAREXEL biopharmaceutical R&D statistical source book 2013/2014. Emerging clinical trial recruitment benchmarking metrics: The recruitment funnel analysis, pp203-207
2. Simons MS et al, Factors associated with breast cancer clinical trials participation and enrollment at a large academic medical center, Journal of Clinical Oncology 22(11): pp2,046-2,052, June 2004
3. Harper B, The PAREXEL biopharmaceutical R&D statistical source book 2013/2014. The predictive value of EMR enabled approach to enrolment validation and protocol optimisation: a 2011 analysis, pp246-250
4. Getz K and Lamberti M, A new benchmark created for patient recruitment and enrolment: a 2013 analysis, Tufts Center for the Study of Drug Development Impact Report, January/February 2013
5. Getz K and Stergiopoulos S, Improving clinical research performance and reducing costs through clinical protocol designs, Tufts Center for the Study of Drug Development Impact Report, November/December 2012

Read full article from PDF >>

Rate this article You must be a member of the site to make a vote.  
Average rating:

There are no comments in regards to this article.

Dr Graham Wylie is the Chief Executive Officer of the Medical Research Network(MRN), responsible for both company strategy and making it stand out from the competition. With a BSc in Pharmacology, an MBBS from St Bartholomew’s College London, and more than 21 years of experience in the pharmaceutical industry with names such as Pfizer, PAREXEL and Healthcare at Home, Graham has successfully grown the MRN from a two-person start-up to a business with an annual turnover of several million pounds. Graham was previously Managing Director of Healthcare at Home’s clinical trials division, where he led the management buyout to create the MRN in November 2006.
Graham Wylie
Print this page
Send to a friend
Privacy statement
News and Press Releases

Charles River Laboratories Announces Fourth-Quarter and Full-Year 2021 Results

WILMINGTON, MA, February 16, 2022 – Charles River Laboratories International, Inc. (NYSE: CRL) today reported its results for the fourth-quarter and full-year 2021 and reaffirmed guidance for 2022. For the quarter, revenue was $905.1 million, an increase of 14.4% from $791.0 million in the fourth quarter of 2020.
More info >>

White Papers

Environmental Factors in Data Centres


In addition to temperature and humidity, there are many more constraints in the environment that increase failure rates, explains Emilio Sapiña of Secure Techincal Rooms.
More info >>




©2000-2011 Samedan Ltd.
Add to favourites

Print this page

Send to a friend
Privacy statement