In August 2023, Evidence Prime conducted a survey in order to understand the perception of health professionals towards use of Artificial Intelligence in literature review software. Respondents from all continents participated, representing a spectrum from individual contributors to organizations exceeding 10,000 employees. The diversity of perspectives was also ensured by the variety of organization types which included government institutions, non-profits, life science companies, consultancies and the academic sector.
The survey indicates that the majority (77%) of respondents are unhappy or at best neutral towards the literature review tools they use today. Users would want the entire literature review process to be more automated, especially screening and data extraction. Use of AI raises some serious concerns including risk of making erroneous decisions, difficulties validating the output and risk of bias when training the AI engines. Despite the concerns of AI in literature review software, 45% of respondents expect AI to replace human involvement in literature reviews to a significant extent and another 38% - to some extent.
Majority of the professionals raised lack of time and shortage of resources as their most significant pain points. Consequently, the most significant benefit they would expect from AI technology is efficiency - as measured by reduction of manual efforts, faster processes and increased productivity overall. Insights from the survey that we will look into are:
1. Unhappiness with Literature Review Tools
2. Strong Desire for Automation
3. Unaddressed Concerns Raised About AI in Healthcare
4. AI Viewed as Inevitable, Despite Concerns
5. Biggest pain point for literature reviews: Time & Resources
6. Biggest benefit AI could bring? Efficiency
How satisfied are you with the existing tools your organization is using for literature reviews?
UNHAPPINESS WITH LITERATURE REVIEW TOOLS 77% of respondents expressed feeling neutral or dissatisfaction with their existing literature review tools. These findings highlight a substantial opportunity for improvement in the quality and effectiveness of literature review tools. It's evident that there is a strong desire for better solutions among users.
Which areas of literature review would you personally want to become automated as soon as possible?
Respondents are eager to see automation applied across virtually the entire literature review process. Screening and Data Extraction emerge as the top priorities for automation, highlighting the desire for AI to take over these time consuming and labor-intensive tasks.
What concerns, if any, do you personally have about using AI in literature reviews?
Respondents have voiced significant concerns about the integration of AI in literature reviews, indicating the need for addressing these apprehensions. Interestingly, the fear of AI threatening roles and jobs was not a significant concern among respondents.
To what extent do you expect artificial intelligence to replace human involvement in literature reviews over the next 5 years?
Despite their concerns, respondents view AI as an inevitable force that will gradually enhance human involvement in literature reviews. A substantial 45% of respondents anticipate that AI will replace human involvement "to a significant extent" within the next 5 years.
WHAT IS YOUR SINGLE MOST SIGNIFICANT PAIN POINT TODAY RELATED TO LITERATURE REVIEWS?
TIME & RESOURCES
The most significant pain point related to literature reviews today is the time constraints and lack of resources. The substantial time required to complete systematic reviews takes away from tasks that prove to be more valuable.
WHAT ONE, SIGNIFICANT BENEFIT COULD AI BRING TO EVIDENCE-BASED MEDICINE?
1. Reduce time required for various tasks.
2. Reduce manual efforts.
3. Speed up the identification and extraction of data.
4. Accelerate the overall pace of tasks, increasing productivity.
The survey results imply that the majority of literature review software and tools in use today fall short of users' expectations. Simply put, healthcare professionals are in dire need of reliable automation. They are well aware of Artificial Intelligence's potential to transform literature reviews in the quest for efficiency and smart allocation of resources. But the evidence-based healthcare sector will only adopt AI-based technologies that convincingly and exhaustively address stakeholders' concerns including reliability of data, ability to validate decisions and data privacy
A business developer at Evidence Prime. She combines her passion for medicine with her extensive experience in software sales to advance Evidence Prime and its products in the Evidence-Based Medicine industry.
He makes sure that the Evidence Prime team understands what customers need and that Evidence Prime’s products meet and exceed these expectations. Meet Piotr at ISPOR Europe 2023
Related blog posts: