Site selection is a crucial step in clinical research as it affects the quality, efficiency, and cost of the clinical trial. However, many challenges exist in the current site selection process, such as low investigator participation, poor recruitment performance, and lack of data-driven decision making. This article aims to highlight some of the key factors that need to be considered when selecting sites, and to provide suggestions on how to improve the process and leverage the potential of artificial intelligence (AI) and machine learning (ML) in the future.
According to a recent WCG survey, only 28% of investigators participate in more than one clinical trial, and 40% of trials fail to meet their recruitment targets. Some of the reasons for these low numbers are the lack of adequate training, resources, and incentives for investigators, as well as the mismatch between the trial protocol and site characteristics. Therefore, it is important to evaluate a site’s capabilities, experience, patient population, and infrastructure before selecting it for a trial. The decreasing number of investigators and increase in site burden, lead to the challenge of meeting recruitment targets and make site identification and selection an imperative part of clinical trial execution that should be continuously evaluated. It is essential to establish clear and transparent communication and collaboration between sponsors, sites, and patients at the start of the site identification/selection process and throughout the trial. This is to ensure that modifications, such as adding and/or removing sites or amending protocols, are included in all feedback loops.
Current Practises and Challenges in Site Selection The traditional site selection process consists of two main stages: site identification and site feasibility. Site identification involves using various data sources, such as historical performance, claims data, regulatory compliance, and quality indicators, to generate a list of potential sites that match the study criteria. Despite the number of data sources available, a recent poll conducted by WCG revealed that more than half of sponsors are still making their site identification and selection decisions exclusively based on relationships and historical experiences, as opposed to databased selection. The second step in the site selection process is site feasibility, which involves collecting data from the sites themselves such as, patient population, demographics, referral sources, and facilities to assess their suitability and interest in participating in the trial. However, in the same poll conducted by WCG, most respondents believe that sponsors have little trust in the feasibility responses they receive from sites. Sites agree that sponsors are not likely to believe their responses, particularly surrounding recruitment numbers.
