85% of clinical trials fail to retain enough patients, and approximately 86% of all trials are delayed by over 6 months, in part due to recruitment issues, indicating a significant and urgent need for improvement (1, 2).
The manual recruitment and processing of patients is one of the major obstacles to running financially and clinically viable trials, and currently, traditional methods rely on doctors to recruit patients, or via word of mouth which directly contributes to clinician and clinician bias. These biases affect representation of certain races, ages and genders – there is an over-representation of individuals from Western, Educated, Industrialised, Rich and Democratic societies, and an under-representation of individuals from societies that do not fall within this categorisation (3).
This recruitment method is time-consuming and error-prone – little awareness of what a clinical trial is, what the benefits and adverse effects is linked to a reduce interest in participating. On the other hand up to 80% of individuals with a certain condition may not be aware that they are eligible and suitable for impactful trials, highlighting the need to expedite and streamline participant recruitment and concomitantly increase diversity and inclusion to tackle misrepresentation in research practice.
A modern solution to the issues of (1) recruitment and (2) representation could be found in the recent technological developments of cognitive technologies that include artificial intelligence (AI; the mimicking of human cognition in data analysis, and dissemination) and natural language processing (NLP; a subfield of linguistics, computer science and AI used to process and analyse communication data).
AI relies on machines and computers to carry out tasks that were once deemed only feasible for humans. It enables the gathering and processing of large amounts of data from very disparate sources with extreme accuracy and organisation, eliminating the effects of cognitive and emotional bias that occurs when humans undertake a task. The field of AI has several subfields, including machine learning. With machine learning. The machine or computer is programmed to continuously learn and adapt through experience, constantly improving the algorithms of interest.
Prevention and the development of treatment techniques and improved patient outcomes are at the heart of the AI health-related aims. A crucial application of AI would be in analysing the diagnostic pathways, and practices such as treatment protocol development, drug development, and other patient-centred goals such as monitoring and care. Next, AI and its subfields could also be used to scan patient forums, public search results and websites to provide a rapid search of potential patients, and use the data to signpost interested and eligible participants to relevant trials.
Another possibility would be to filter data based on specific geographical locations, for site-specific recruitment. Taken together, these applications would have some key implications: improved patient satisfaction, reduced backlog in the healthcare and associated research fields, increased cost saving, and accurate and expedited research recruitment and trial phase completion.
There are already some preliminary findings suggesting that patient centric recruitment via web-based tools such as patient communities better engages patients, in comparison to traditional recruitment methods. Novel and targeted AI strategies focussing on participant outreach have directly contributed to higher engagement in younger age groups, highlighting the benefits of using AI for patient recruitment (4).
Next, NLP could be used to gather insights on (1) patient needs related to clinical trials (e.g., information about safety, role as a participant, implications), (2) specific wording used in search engines. NLP data could then we used for targeted messages to individuals in order to increase the likelihood of engagement and participation in trials of interest. At the same time, machine learning applications would be able to adapt to the individual’s engagement levels, and tailor the timing they would see relevant campaigns to ensure maximum engagement, and improve retention rates. Ultimately this would lead to improved clinical trial outcomes, conferring benefits to the patients, clinical research organisations, and the pharmaceutical companies that sponsor the trials (5).
Neucruit’s application programming interface (API) identifies all channels related to the inclusion criteria of trials, identifies eligible patients and launches targeted campaigns for specific demographics. Then, Neucruit will pre-screen interested participants and exclude those who do not meet the clinical trial’s inclusion criteria.
Additionally, Neucruit’s NLP will alter phrasing and clinical language for each user to provide personalised engagement throughout the recruitment process, which will allow us to immediately identify and address any concerns in order to maintain user engagement and retention.
Our aim at Neucruit is to guide future clinical trial recruitment using cognitive technologies and improve the design of clinical trial for the targeted demographics. Our goal is to make recruitment and participation accessible to all in order to expedite research with broad and poignant implications.
Gul RB, Ali PA.
Clinical trials: the challenge of recruitment and retention of participants. 2010;19(1‐2):227-33.
Gebauer M, Krämer A.
A female-led team transforming clinical trial efficiency, transparency and diversity by improving accessibility to all patient communities. Neucruit is an intelligent software for clinical trial recruitment that redefines patient recruitment. Our technology aggregates real-time data from the over 25 million health-related conversations initiated online everyday to facilitate planning and recruitment in clinical trials. We support biopharmaceutical companies, site teams and investigators enhance site selection, optimise recruitment materials and reach groups that cannot be easily accessed through traditional methods.