For decades, pharmaceutical developers have relied on clinical trials to provide evidence that a new medicine is safe and effective in its target indication. But now data collection is taking a turn away from typical gathering of information for clinical programmes – consisting of date from Phase I to III clinical trials.
While the clinical trial itself is likely to remain the mainstay of any marketing authorisation application its clear and well-recognised limitations, in particular the number of patients studied, is starting to be reassessed by companies concerned it is too small and narrowly defined compared to the population that will be exposed to a new medicine after launch.
For years, what has been missing from the picture is a focus on how a medicine functions in the real world – in other words, what happens when a product meets its primary market, the patient. Individuals taking a drug may have a co-morbid condition or belong to an ethnic group that has not been studied in sufficient numbers or was perhaps even deliberately excluded from pre-approval studies. Of course, the strictly-defined protocols in clinical trials cannot take into account variations in doctors' prescribing habits and the way the general public actually take medicines.
Data collected during the course of a subject's daily lifestyle can vary from data collected in a more structured environment, according to data capture technology and electronic diary company CRF Health's senior technical engineer John Hutchin.
"Psychological factors can influence how a subject feels and behaves," Hutchin says, pointing to the well-recognised phenomenon of "white coat hypertension", where blood pressure readings taken in a doctor's office are higher than those taken in the real world.
As a result, sponsors of new medicines have started to pay more attention to what happens to their products when they reach the market through greater use of observational, non-interventional studies.
Drug companies are now carrying out more post-approval (Phase IIIb and IV) studies, causing the post-approval study segment to grow at an annual rate of around 20%, compared with about 7% growth for Phase II-III studies, according to data from Tufts University in the US.
There are a number of drivers behind that trend, according to contract research organisation Parexel International's vice-president of per-approval operations Kate Trainer.
"Having real-world data regarding biopharmaceutical products is more important than ever for the industry due to expanding requirements for drug safety in larger populations, growing complexity driven by increasing regulatory demand and the need to demonstrate and maximise return on R&D investments," Trainer says.
High-profile product recalls in recent years, such as Merck & Co's withdrawal of painkiller Vioxx (rofecoxib) in 2004, have prompted greater scrutiny of safety signals after a new product's launch. All too often regulators are now making approval of a new medicine conditional on a commitment to carry out post-approval studies.
Another critical factor has been an increasing focus on health economics, and the growth of health technology assessment (HTA) agencies such as the UK's National Institute of Clinical Excellence (NICE). An increased need to satisfy regulators of the cost effectiveness of new medicines has introduced a requirement to generate economic evidence to achieve an optimal price and obtain reimbursement.
"Phase IV studies, registries and retrospective or prospective observational studies now almost invariably include economic endpoints," Trainer says.
Sharing the risk
Dominic Farmer, chief executive of Cisiv, a company that has developed a software platform specifically to handle the data coming from non-interventional trials, says he believes a clear indicator of this trend comes with recent risk-sharing agreements between drug makers and healthcare providers.
The first example of this came with Johnson & Johnson's cancer drug Velcade (bortezomib), which was turned down by NICE in early 2007. A deal forged between the company and the UK National Health Service means that drug will only be reimbursed if a blood test shows that it is effective. If not, the cost will be refunded by the drug manufacturer.
"Greater understanding of how a drug works in real-world patients will increase its overall cost-effectiveness," Farmer says. "A drug with a 90% success rate will clearly be more cost-effective than one which works in only half of patients."
Another driver is the risk of litigation when things go wrong and drugs must be removed from the market, according to Hutchin. "Liability issues due to unforeseen adverse effects that a drug may cause can lead companies to try and better characterise how their product will actually be used in a real-world setting," he says.
These drivers are also changing the way post-approval studies are being conducted. Not so long ago sponsors would design them like Phase III trials, at large medical institutions and under the oversight of experienced clinical investigators with a clearly-defined set of inclusion and exclusion criteria. Now, the focus is much more on primary care and doctors in the community.
These studies can, however, be challenging to implement, particularly on a global level since at present there is not a clearly defined conduct approach. Therefore it is important that pharmaceutical companies identify the best study design for the research objectives, apply the right regulatory approach and collect data in the most efficient manner adhering to quality standards.
Electronic data capture (EDC) is gradually replacing paper-based reporting in the clinical trials sector as companies realise the time- and cost-savings that can accrue. For example, the changeover can save a big pharma company up to $15m a year in mailing and protocol distribution costs alone, according to data published earlier this year by clinical industry analysis firm CenterWatch.
More recently, EDC has begun to find a role in observational studies as drug companies include them in development programmes but the requirements for these systems are somewhat different in this setting. Data capture systems for non-interventional studies need to be very easy to use and quick to deploy for non-expert users. While investigators in clinical trials may be familiar with EDC systems, observational studies rely on the input from primary care doctors.
"The adoption of EDC has grown dramatically in late phase studies since conducting these large studies without EDC can be cumbersome and expensive," Parexel's Trainer says. "For instance, a global registry with a large volume of geographically distributed sites and patients that would have been handled manually with paper-based documents is now much more efficient using electronic methods for data collection."
Using electronic rather than paper data capture systems also makes real-time monitoring of studies possible, which can make it possible to pick up problems, say if a study site stops completing patient data, and helps avoid "recall bias" that can skew findings. "The best decisions are made based on current information; troop movement information on the battlefield is not of much use when it is a month old," says Hutchin.