Thanks for visiting Scientific Computing World.

You're trying to access an editorial feature that is only available to logged in, registered users of Scientific Computing World. Registering is completely free, so why not sign up with us?

By registering, as well as being able to browse all content on the site without further interruption, you'll also have the option to receive our magazine (multiple times a year) and our email newsletters.

Joining up the data from clinical trials

Share this on social media:

Clinical trials, not finding new chemical entities, are the most costly aspect of drug development. Peter Rees looks at the software solutions that might make clinical trials cheaper and faster

Last year, drug safety scares sent shudders through the pharmaceutical industry. It wasn't just the immediate cost of drug withdrawals that worried executives, although Merck suffered from lost sales worth $2.5bn annually while its arthritis drug Vioxx was off the market (it was voluntarily withdrawn in September 2004 but an FDA panel voted to allow it back in February this year). The companies were more concerned that increased scrutiny by the US Food and Drug Administration could send the costs of safety testing to new highs. Clinical trials to test safety are at the heart of the drug approval process and they are driving up the cost of research as they become longer and more complex.

The average cost of developing a new drug reached $802m in 2000, according to the Tufts Center - 10 years previously, a similar study estimated the average cost at $231 million. (Note: This is not the amount of money actually spent to discover and develop new drugs - rather controversially, it includes the expense of using money for drug research rather than investing it elsewhere.)

The cost is so high partly because of the number of failures - just one out of 5,000 screened drug candidates makes it to the market as a new medicine. Many are weeded out before they reach the clinic. Nevertheless, only one in five of drugs that begin human trials is eventually approved for marketing.

And it is the rising cost of clinical trials that may really be hurting drug company profits. The process now accounts for more than half the total spending on developing a new drug. Clinical costs grew more than five times as fast as pre-clinical research and development, Tufts' researcher Dr Joseph DiMasi found. Factors driving up the cost were: a greater emphasis on developing treatments for chronic and degenerative diseases; rising subject recruitment costs; and increasing clinical trial sizes. The number of patients needed to complete a drug trial for the results to be acceptable to US drug regulators has almost doubled in the past 10 years.

It is a labour-intensive job to gather and integrate data from clinical research centres to prepare a licence application for drug regulators. For the past 15 years, companies have been trying to move away from paper by investing in technology to improve the management of trials. Many types of software - internally developed and commercially supplied - have been introduced: clinical data and clinical trial management systems (CDMS and CTMS); electronic patient reported outcomes; and patient diaries and electronic data capture (EDC) are just a few. But, as with many historical attempts to automate processes, the end-result has been standalone systems, despite the fact that these applications use overlapping data sets. Data is routinely exchanged manually or semi-manually through a series of exporting and importing steps. If this data is to be shared automatically between software systems, and used more widely through companies, agreed standards are needed. It is a time-consuming process agreeing and compiling these standards, but one that is now bearing fruit (see 'Setting standards' section below).

It has been a long road to reach even this level of automation. The most mature sector is electronic data capture (EDC), where patient data is recorded through an electronic process - directly in software or via a web browser on a network - rather than on paper. Introduced in the 1990s, it has taken some time for EDC to become firmly established. Bayer and Novartis have been early adopters of the technology, but their enthusiasm is not shared by all. Estimates of uptake vary but in 2002 a study, carried out by clinical trials listing service Center Watch and standards organisation CDISC, found that around 25 per cent of new clinical trials were using some sort of electronic data capture; perhaps half of these were wholly electronic.

A major problem is that, even with EDC, there is still a lot wasted effort. Human intervention is needed so that the data can be sent in different electronic formats to many different organisations taking part in a trial - the drug company, clinical investigators, an institutional review board, a contract research organisation, and others. At times, data may be printed out and re-keyed - leading to inconsistencies that need to be checked manually. Nevertheless, many experts expect most trials to use EDC by 2008.

Electronic data capture systems, and the clinical trial data management systems (CDMS) that store the data, are often home-grown systems, but the sector has two dominant players - Oracle and Phase Forward. They both compete strongly in the high end of the market, with Phase Forward's strength leaning towards EDC, while Oracle's is in CDMS. Nevertheless, their products overlap. Oracle has its own EDC - Clinical Remote Data Capture - while Phase Forward has a CDMS - Clintrial.

It is hard to gauge the relative positions of the two companies but competition for the big pharma customers is fierce. Oracle's most visible customer is Pfizer but other users include Bristol-Myers, Wyeth, Amgen, Merck, and Astrazeneca. Phase Forward's products are used by many of the top 15 pharmaceutical companies, although that doesn't always mean they all use the firm's complete trials solution. In 2003, Eli Lilly signed up for a package that added CDMS, and other software and services, to Phase Forward's InForm electronic data capture system. Then in 2004 GlaxoSmithKline did the reverse, settling on InForm to supplement the company's Clintrial clinical data management solution (CDMS).

Oracle isn't Phase Forward's only rival in the EDC sector. Others include Nextrials, eResearch Technologies, Medidata, DataTrak and e-trials. Historically the sector has been fragmented, but it is now consolidating as the market matures.

At the same time, pharmaceutical and biotech companies are starting to show an interest in wide-ranging applications, according to a report by US market research company Life Science Insights (LSI). Historically important technologies, such as EDC and CDMS, will continue to play an important role says the report, but it is systems designed to manage multiple aspects of clinical trials that are gaining ground. These CTMS systems will drive adoption of information technology for trials among drug and biotech companies, according to the report. And not just in the companies themselves, but in academic and government institutions and the contract research organisations that do much of the work.

As a result the clinical trial management systems sector should see a compound annual growth rate of nearly 13 per cent, claims the LSI study. The company's research analysts expect the market for software licenses to grow during the same forecast period, from $193 million to $360 million.

The current balance in supplier market is set to alter, according LSI research analyst Judy Hanover. 'The market for CTMS is still relatively immature. Current market leaders cannot afford to become complacent as this market evolves, for leadership and market-share positions remain in flux,' she said.

It is often hard to draw firm distinctions when sectors naturally overlap, but CTMS-based software companies that stand to benefit include ClinSource, Parexel, and Advanced Clinical Software. Nevertheless, the changing needs of the customers have already prompted a series of shifts among software suppliers as companies position themselves to take advantage of the increased need for integration. Sometimes it has been an internal process with predominantly EDC firms - such as ClickTrials and Medidata - developing additional trial management software or features. In other cases, it's been a case of consolidation and acquisition.

Last year, Oracle Clinical moved firmly into trial management when it acquired Siteworks and its SiteMinder and TrialMinder software (both of which use the Oracle Database). SiteMinder is used by clinical researchers to plan and manage the clinical trials they take part in, and TrialMinder does the same job for pharmaceutical companies and CROs. Since then, Oracle has been busy integrating the software into its clinical suite.

Another recent entrant to the CTMS sector is Nottingham-based ClinPhone with its Trialworks clinical trial management software, which is used by around 40 firms. The company's roots are in service-based technologies, but it has made significant steps to move beyond that approach. In early 2004, it acquired the Trialwork's developer, TrialTrac, and began to push sales of the software beyond its core market of small and medium-sized companies. Clinphone has now web-enabled Trailworks, to make it even more attractive to multinational companies. Trialworks covers a lot of ground and can track and report on IRB approvals, patient enrolment and status, contact information, study planning, investigator recruitment and other parameters. It also connects to ClinPhone's market-leading interactive voice response (IVR) tool for patient randomisation and experiential data.

It is not just CTMS that is set to boom. There is also a surge in interest in patient-reported outcomes systems. These focus on recording symptoms and quality-of-life measures from the clinical trial participants, rather than the physiological data and medical history recorded by doctors. Historically, patients were asked to fill in paper diaries with this information but, as with any diaries, entries were often missed completely or filled-in retrospectively. PDA-based electronic patient diaries (or e-diaries), such as those provided by Finland-based CRF, are proving increasingly popular because they are more accurate.

Drug companies are looking to software and service suppliers to provide evidence that their technology will ultimately cut costs. So the technologies that will win greatest favour will be those that truly increase efficiency as clinical trials reach new levels of complexity. Trial costs will remain high in the US because of recruitment and other problems, and while some trials are already international affairs, they are set to become increasingly so. More complexity will surely increase the rate of uptake of automated systems.

Setting standards

The US-based Clinical Data Interchange Standards Consortium (CDISC) is the leading standard-setting body for electronic data in clinical trials. It is an open organisation with a wide range of members, including pharmaceutical firms such as Merck, and software suppliers from Microsoft to Medidata.

The organisation is building data models to support the flow of information from the earliest part of the trial through to regulatory submission. It works closely with another standards development body - HL7 or Health Level 7 - which covers a much wider range of healthcare-related protocols.

CDISC's standards include: a study data-tabulation model (STDM), for submitting clinical trial data to regulatory bodies; an operational data model (ODM), for collecting and sending patient and other clinical data; a laboratory data model (LAB), for collecting and sending lab test data; and an analysis dataset model (ADaM), for processing and submitting statistical analysis results to regulatory bodies.

Last July the FDA roundly endorsed CDISC's work when it agreed to allow SDTM-compliant submissions. The agency said it expected the move to speed up FDA reviews of new drugs. The FDA may now adopt further CDISC standards and the organisation is working to persuade regulatory bodies in Europe and Japan to adopt its submission data standards.