In the second of a series of podcasts Outsourcing-pharma.com spoke with Michael Howley, associate Clinical Professor at Drexel University’s Lebow College of Business, who said the various methods commonly used to measure performance are not always sufficiently detailed.
“The first is a fairly general approach were they ask senior pharmaceutical managers something like ‘How satisfied were you with CRO performance?’ and ‘Would you recommend them?’ which are fine questions, but are not detailed enough to give the contract manager that is thinking of hiring that CRO again any insight.”
The second common approach, Howley continued, is often too detailed.
“You get question like ‘how many days did it take to recruit subjects for this trial?’ with answers like ‘their average patient recruiting time is 55 days’. Does this mean that the CRO is good at recruiting patients?”
Instead Howley, in partnership with CRO Analytics’ Peter Malamis , is developing an approach that is somewhere between the established methods by asking pharma functional managers how the CRO performed in individual tasks that are important to a successful trial.
“We found that the best way to ask these questions is to break them into four stages and in that way we can reach the functional areas managers within the pharmaceutical firms that are actually overseeing [the CROs} and in a legal sense are witnesses to what the CRO actually did.”
“So in the patient recruitment phase we would ask ‘how well did the CRO keep you informed of its progress; in the sales and contracting phase ‘how well did they perform in the bid defence?”; later on towards the end of study conduct ‘we ask about the CROs data management practices?’”
This, he continued, enables the gathering of information on discrete tasks which, when bundled together, lead to a more accurate and useful measure of CRO performance.
Howley also urged any Outsourcing-pharma.com readers who are interested in participating in the project during the validation stage to contact him directly at email@example.com .