A Data Driven Approach

In the last decade the ever increasing number of drugs in development has resulted in a significant demand for patients to participate in clinical trials. The 2011 R&D pipeline in the United States alone included more than 3,000 medicines in clinical trials or awaiting FDA review. As the competitive landscape for clinical trials increased, the ability to recruit patients in the United States decreased, and biotech/pharma companies opened clinical trial sites in emerging markets in order to achieve patient recruitment timelines. However, there remains a significant pool of patients in the United States that are not currently entering clinical trials.

This shift to emerging markets impacted oncology clinical trials disproportionately based on pipelines. For example, biotechnology companies in 2011 had 901 medicines in development with 39% of those in oncology indications. As more oncology trials were opened in the United States, average accrual rates decreased from an average of 2.0-plus patients/site/month (psm) in 2000 to an average rate of 0.1-0.2 psm in 2011. Consequently, many companies opened clinical trial sites in emerging markets such as Eastern Europe, Latin America, and Asia. In 2001, 75% of the global clinical trial sites were located in the United States while Eastern Europe, Latin America, and Asia accounted for only 2.9%, 2.3%, and 1% respectively. By 2011 those numbers had shifted with the United States only having 60% of trial sites and Eastern Europe, Latin America, and Asia having 8.5%, 5.6%, and 4.8% respectively. These regions were extremely favorable for patient accrual due to a socialized medicine structure with high numbers of potential clinical trial subjects.

While the ability to access oncology patients in emerging markets has seemed to be the answer over the last decade, it is an expensive solution. In most of these countries, the sponsor company must provide experimental medications for the protocol as well as any applicable standard of care medications. This requires not only supplying the drugs but incurring the costs of packaging, labeling, blinding, and shipping. Additionally, in most emerging markets sponsor companies are also required to cover all costs including items that are standard of care such as imaging, labs, hospitalization, con meds, etc. Given the increasing costs associated with the utilization of emerging markets, our industry must seek methods to increase patient accrual in the United States.

However, the solution to increasing patient participation in oncology trials cannot be focused on patient recruitment alone. Industry data suggest that fewer than 4% of US physicians participate in clinical trials. Worse yet, a 2008 Women's Health Research survey indicated that 94% of Americans have never been informed by their doctors of medical research studies. The physicians in the United States are seeing patients; they just are not recruiting them into clinical studies. Many of these sites are not even aware they have target patients in their practices. Industry data indicates that approximately 25% of US trial sites will never accrue a single patient. This waste of resources on non-productive sites is a key issue for clinical trial recruitment.

The ability in 2011 to access private and publicly available databases (claims, payor, CTG, BMIS, etc) in a HIPPA compliant, de-identified manner provides transparency and insight into physician patient volumes at the site level that in the previous decade was not available. Mining this data using ICD-9, CPT, and drug codes and leveraging this data to better select sites where known patients exist are key components in the strategy to drive recruitment back to the United States. The basic strategy is to decrease the overall number of non-productive US sites, thereby increasing overall recruitment rates per site. If we can increase the number of US sites, and correlatively the rate of US patient accrual, we should see a decreased overall accrual timeline and to some extent a decrease in overall costs compared to extensive emerging market involvement.

A recent retrospective analysis of a sponsor company's global pancreatic cancer trial demonstrated that US site recruitment rates were an average of 0.1 psm while Eastern European sites were accruing at three to 10 times that rate. Initially the sponsor company had indicated the desire to increase sites in the Eastern Europe and close US sites. However, via the data mining of a number of claims and payor databases, it was discovered that 30% of the US sites selected by the sponsor company had seen one or fewer patients with an ICD-9 code for pancreatic cancer in the last 12 months. This same analysis identified 50 US sites that had seen in excess of 25 patients with an ICD-9 code for pancreatic cancer in the same timeframe. A cross-analysis indicated that of those 50 sites, a large proportion had previously filed a 1572 indicating that the sites had clinical trial experience. Had this analysis been done prior to site selection it is highly likely the sponsor company could have increased US sites and decreased the need for the EEU sites, or kept both and decreased their overall clinical trial timeline.

The ability to mine this type of data can drive US recruitment at higher rates than historically seen by giving the ability to proactively select sites where patients that meet protocol criteria exist. In addition to this type of data mining, in the oncology recruitment arena where patients in the United States tend to be heavily focused on support groups, the use of tools such as social media for recruitment cannot be overlooked. These types of tools coupled with processes such as data mining are an integral solution to driving US patient accrual.

Increasing participation

NCI data indicates that less than 5% of oncology patients participate in clinical trials. Projections indicate that if doubled to 10%, the average study could be completed in one year instead of the current three to five years required. A starting point to increase participation is to select oncology sites with pools of eligible patients. The ability to mine available data sets creates an evidence-based approach to ensure productive site selection. Decreasing non-productive US sites and increasing US accrual using an evidence-based approach provides an alternative to utilizing emerging markets for increased patient recruitment.

Kent R. Thoelke is Executive Vice President, Scientific and Medical Affairs at PRA, 4130 ParkLake Avenue, Suite 400, Raleigh, NC, e-mail: thoelkekent@praintl.com.