Accepted for/Published in: Journal of Medical Internet Research
Date Submitted: Apr 3, 2024
Date Accepted: Aug 16, 2024
Warning: This is an author submission that is not peer-reviewed or edited. Preprints - unless they show as "accepted" - should not be relied on to guide clinical practice or health-related behavior and should not be reported in news media as established information.
Trends in Studies of Artificial Intelligence/Machine Learning Registered on ClinicalTrials.gov 2010 to 2023: Cross-sectional study
ABSTRACT
Background:
The rapid growth of research in artificial intelligence (AI) and machine learning (ML) continues, but the main driving factors behind this growth are unclear.
Objective:
To assess temporal trends in AI/ML studies registered in ClinicalTrials.gov and identify variations that are not apparent from aggregated totals alone.
Methods:
We conducted a cross-sectional analysis of AI/ML studies started from 1 January 2010 to 31 December 2023. Studies were selected if any of the following terms were in the official title, detailed descriptions, brief summary, interventions, primary outcome, and sponsors’ keywords: AI-based, artificial intelligence, machine learning, deep learning, artificial neural network, convolutional neural network, deep neural network, multilayer perceptron, Bayesian network, naïve Bayes, classification tree, gradient boosting, k nearest neighbor, random forest, regression tree, vector machine, or natural language processing.
Results:
Of 3106 studies included, 235 (7.6%) were FDA-regulated. The most common study characteristics with material growth over time were as follows: randomized trials (56.2%; 670/1193 interventional) and prospective designs (58.9%; 1126/1913); studies purposed for diagnostic (28.2%; 335/1190) and treatment (24.4%; 290/1190); studies sponsored by hospital/clinic (44.2%; 1373/3106), academia (28.0%; 869/3106) and in the areas of neoplasms (12.9%; 420/3245); nervous system (12.2%; 395/3245), cardiovascular (11.1%; 356/3245), pathological conditions (10.0%; 325/3245) (multiple counts per study possible). Enrollment data was skewed to the right (maximum 13,977,257; mean 16,962; median 255); the ≤100 and 101-1000 categories increased the fastest over time. Compared with conventional drug or device studies, higher proportions of large studies (>100; >1000) were evident even among randomized trials. Study locations were predominantly in high-income countries (75.3%; 2340/3106), but upper-middle-income countries (21.7%; 675/3106) also gradually increased. The fastest-growing characteristics over time were as follows: Europe, Asia, and North America (study location); diagnostic and treatment (primary purpose); hospital/clinic and academia (lead sponsor); and randomized and prospective designs, which showed growth rates of 22.6% and 42.1%, respectively, from 2017 to 2023. Of completed studies, the results availability on ClinicalTrials.gov was limited (5.6%; 47/842) even after one year of study completion (6.5%; 45/691) with a median of 505 days to results posting; this trend remained constant over time.
Conclusions:
AI/ML studies continue to proliferate predominantly in high-income countries; this may have implications for health disparities for data-poor regions in tasks assisted or augmented by AI/ML. The overall research trend shows promise; the increase in randomized or prospective designs and relatively larger sample size may be a reflection of enough studies gradually progressing from the in-silico evaluation stage towards the comparative prospective evaluation stage. However, given the diverse nature of AI/ML use cases and the high volume of non-regulated AI/ML studies, timely posting of results is crucial in monitoring where the evidence pipeline is heading and informing prospective users of its applicability to their local context. Clinical Trial: Not applicable
Citation
