DecisionView’s Consortium Approach to Optimizing Enrollment

By Deborah Borfitz   

April 17, 2012 | eCliniqua | For industry leader Merck, country-specific and site-level performance metrics have been among the early wins from the DecisionView Enrollment Benchmarks data set, says Brendan O’Neill, Merck’s director and head of global trial optimization. The pharmaceutical giant last June joined peers GlaxoSmithKline and Roche in working with DecisionView to turn their combined operational patient enrollment data into a cross-industry set of clinical trial enrollment benchmarks.   

Linda Drumright

Enrollment Benchmarks is available to DecisionView StudyOptimizer customers who contribute their completed trial data, which is anonymized to ensure the sponsor’s confidentiality, then aggregated into the larger data set, explains CEO Linda Drumright. Data contributors alone have access to the aggregated data, which they can then slice and dice internally to suit their trial planning needs as well as use to invigorate dialogue with their outsourcing partners. The benchmark data is useful primarily to “test and vet assumptions” and help companies build a reasonable risk mitigation plan, she adds, which StudyOptimizer customers already routinely do with their own historical data.    

Although Merck has a lot of capabilities in the respiratory arena, looking more broadly at enrollment patterns for respiratory trials ensures future trials are planned using the most robust and timely benchmarks, O’Neill says. The data set is especially useful in terms of trial placement and setting realistic budgets and timelines. Country performance, in particular, is hugely variable. Site information is used to steer modeling and planning rather than to pick individual investigators. Where the figures provoke more questions than answers, he adds, the benchmarks serve as a “nice…complement to other data sets out there.”   

The data set has been especially valuable in helping Merck more intelligently plan its work in oncology, where it has but five years of experience from which to draw. “We struggled to find a data set to help us plan in this [therapeutic] area.” But even those with vast experience in oncology have much to gain from the enrollment benchmarks, he adds, since there are many types of cancers and therapies targeting one type may prove more beneficial in treating another.   

Although the inaugural issue of the Enrollment Optimization Report wasn’t designed to elicit a “big aha,” says Drumright, findings from the more specific analyses were compelling. For example, phase II cycle times for oncology patient enrollment have steadily increased while phase III enrollment times increased from 2006 to 2009 before dropping in 2010. But is that because trials were placed in more or fewer countries or because of changes in the size of trials or the nature of the patient population being targeted? Organizational knowledge is required to interpret the results—or perhaps, in the case of Enrollment Benchmark customers, further manipulation of the raw data, says Drumright.   

Another interesting trend highlighted in the Enrollment Optimization Report is a reduction in the number of sites utilized for all oncology trial phases from 2006 to 2010, which Drumright presumes is partly a reflection of industry efforts to choose better sites. The oncology analysis also found a dip and rebound in the median number of patients per study in phase III and IV trials, which perhaps mirrors changes in protocol design.   

The low number of cancer patients per country in phase III trials relative to other therapeutic areas is an unsurprising finding, given the ongoing difficulty in finding treatment-naïve participants who also meet all other qualifying criteria. “That’s why companies are looking at BRIC countries,” says Drumright. Among the raw comparative facts about BRIC in the Enrollment Optimization Report are that China screens and randomizes significantly more patients per study; Brazil has the lowest screening and randomization rates but also a significantly lower dropout rate; India and Brazil have the shortest enrollment cycle times (26 weeks) and China the longest (41 weeks); and Russia has the shortest start up time (31 weeks) while Brazil has the longest (55 weeks). That suggests “a lot of tradeoffs and choices” for trial planners, notes Drumright.   

Data analysis for Enrollment Benchmark customers will be especially easy later this year, when they start receiving the comparative industry benchmarks as a direct upload into StudyOptimizer, says Drumright.   

At the moment, DecisionView Enrollment Benchmarks is built on 700 studies from the three initial sponsors, says Drumright. But others will be getting on board, adding statistical validity to the dataset as well as the quarterly teaser reports that will be available for complimentary download.