Contributed Commentary by Zaneta Szkarlat, PhD and Dave Li MD, PhD., KCR Consulting 

August 18, 2023 | Selecting appropriate sites is one of the most crucial factors for determining successful completion of a clinical trial. It is important to ensure that potential trial sites have the capability to meet the protocol requirements and the capacity to deliver the study objectives in a safe and efficient manner (DOI: 10.1200/OP.20.00821). Sites that are unable to successfully recruit enough patients not only extend the enrollment period but also impose an unnecessary economic burden. Furthermore, sites lacking adequate experience could be more prone to protocol deviations and may lead to low-quality data, requiring additional training, on-site visits, and increased queries for clarification. All these factors contribute to increased costs and prolonged study duration (DOI: 10.1136/bmjopen-2016-014796). In this article we explore some practical strategies toward optimizing the site feasibility assessment. 

An important initial step to achieve successful site selection is to develop an adequate study-specific feasibility questionnaire. Questions should be carefully chosen to cover the most relevant aspects of study design: access to patient population, enrollment projections, investigator and site staff experience, available equipment, possibility to perform protocol required procedures, competing trials, and site interest. 

The American Society of Clinical Oncology recently issued recommendations for enhancing site feasibility assessments. They suggest minimizing and standardizing questions when evaluating a site's suitability for a trial to facilitate and accelerate the process. It encourages investigators to complete the questionnaire and eliminates delays resulting from their constrained availability. Questions should be standardized with common nomenclature and reduced to the minimum required for preliminary site selection. The primary focus is aimed on two key factors: the site's capacity to carry out clinical trials, and the feasibility of implementing the specific protocol. More detailed information is assumed to be collected during pre-study visit; therefore, it is important to avoid requesting redundant data. Implementing standardization decreases variability among assessments, streamlines the process for sites, and facilitates prompt responses with precise and consistent information (DOI: 10.1200/OP.20.00821). 

When developing the feasibility questionnaire, it is anticipated that the collected data will be synthesized and analyzed in a systemic manner, and therefore it can be conclusively reportable. It is advisable to avoid open questions whenever possible, as it can complicate the analysis process. 

To ensure an accurate assessment, it is necessary to provide sites with detailed protocol requirements and inclusion/exclusive criteria per protocol. This enables them to thoroughly evaluate their capabilities and provide precise and accurate responses to the questions. 

Under certain circumstances, waiving the completion of the questionnaire is worth considering: for instance, if the site is well-known to the sponsor and has similar exposure from other trials. In such cases, collecting additional data may be redundant and create unnecessary burden for the site. 

The most efficient method of collecting feasibility data is using an electronic platform where the site information is directly stored. This approach enables real-time analysis, accurate reporting, future retrieval, and supporting decision making. Furthermore, it facilitates the creation of a site-specific database and building working relationships for future collaborations.  

There are multiple options for data input: it can be done by the principal investigator, the study coordinator or collected during the interview by the feasibility process lead and subsequently entered the system. 

Collected data should be extracted from the system and presented clearly allowing for the review of site capabilities. Integral in this process is the critical analysis of obtained data to emphasize the site information deemed as the most important for the decision-making in the site selection process. Predictive analysis can be applied to streamline the assessment process. Scores are assigned or determined by site-specific factors or features from the questionnaire with relative importance according to study protocol and designs. This allows for the ranking of sites, enabling the selection of the most suitable ones.  

Overall ranking could be supported by artificial intelligence algorithms using machine learning or deep learning. These algorithms can be trained using data from past successful studies, enabling them to forecast the future performance of the sites for a new study objectively. Furthermore, the feasibility assessment algorithm could be reiterated for performance improvements with additional or substituted feature extractions for further optimization with expanded training datasets per specific study design in protocol.  

Multiple factors or features are associated with decision making in clinical study site selections which can easily exceed the human cognitive capability. It is expected that the comprehensive data analysis could ideally be fully automated with accumulative multiplexed data using an advanced deep learning approach in neural networks which the data representations could be more informative and meaningful because the deep leaning algorithms have multiple successive layers of data representation via transformations with a backpropagation workflow for adjusting the weights of each layer for a given task toward expected outputs.  

We perceive that the most important criteria for decision during site selection are usually enrollment projections, due to their significant impact on study completion. The other crucial aspects that require attention include the experience and capability of investigators/sites to perform study procedures.  

For example, gene therapy trials where oncolytic virus is delivered by intratumoral injection, require experience in handling GMO (genetically modified organisms) and collaboration of multidisciplinary team, including medical oncologist, radiologist, surgeons. All the above-mentioned aspects should be analyzed. In this example, ongoing competing trials, which are usually treated as negative factor for accrual, can be advantageous in the case of the highly complex protocols for intratumoral injection indicating the site has experience in the field from exposures to similar studies. 

Feasibility assessments for site selection are complex, consisting of multiple steps and affected by a variety of factors depending on the study design and objectives. Optimizing the process, using an automated platform for data collection and real-time analysis facilitated by AI tools may enhance the chances for successful study execution that is on time, on budget, and delivers desirable outcomes. 

Zaneta Szkarlat, PhD Dr. Szkarlat is a Consultant at KCR with over 5 years of experience within the clinical research area. Her professional expertise revolves around various aspects of clinical trials, with a strong focus on the feasibility process. She has a solid scientific background derived from obtaining PhD in Biotechnology complemented by experience in biotechnology industry. She can be reached at zaneta.szkarlat@kcrcro.com.  

Dave Li, MD, PhD Dr. Li is a principal consultant and Clinical Research Physician with the KCR Consulting. He is a medical oncologist and regulatory scientist, and an expert in molecular medicine, immuno-oncology, and clinical informatics. He was on the faculty of Johns Hopkins Medicine and served as a medical officer with the US/HHS FDA before joining KCR. He obtained his medical degree from the Sun Yat-sen University, and MSc/PhD at the University of Texas M.D. Anderson Cancer Center at Houston, Texas. He can be reached at Dave.Li@kcrcro.com.