A case study on barriers to the research implementation of a novel technology in an academic medical center

Natural Language Processing allows extracting unstructured text data from electronic health records (EHR), but historically required extensive coding and expertise. Amazon Comprehend Medical (ACM) offers a scalable solution for mining EHR data without extensive natural language processing expertise. This case study examined barriers and facilitators to implementing ACM in an academic medical center. We reviewed correspondence regarding ACM implementation between study investigators and respective experts within the medical center. We qualitatively coded the correspondence for barriers and facilitators using the Consolidated Framework for Implementation Research (CFIR) framework as a guide. Key findings included the involvement of non-traditional stakeholders in the approval process and unexpected limitations of anticipated facilitators. The study revealed that implementing novel technologies like ACM in academic medical settings requires careful consideration of safety protocols, which may slow adoption. Our findings can guide research teams in navigating the implementation of similar technologies, balancing innovation with necessary safeguards.