Research Identifies Clinician Support for AI Use in Musculoskeletal Care
A new study found that clinicians are in favor of using AI in healthcare generally and, more specifically, in musculoskeletal care.
Source: Thinkstock
By - A study published this week in BMC Medical Education identified substantial favorability toward artificial intelligence (AI) in healthcare and AI applied to musculoskeletal (MSK) care, such as skeletal radiography, among clinicians and medical students.
According to the study, traumatic MSK injuries are a common presentation in emergency departments. These are often initially assessed using standard radiography. However, the interpretation of the imaging usually falls to less experienced clinicians or non-radiologists, which can lead to assessment disparities or misdiagnosis.
To combat this, some have championed the use of AI to assist with image interpretation, as AI in medical imaging has piqued substantial interest in recent years. However, various challenges come with the application of AI in radiology, including integrating AI into legacy systems and assigning medical and legal liability to AI-based clinical decision support tools, the researchers explained.
Despite this, they noted that there is growing utilization of AI in skeletal radiology and indications of favorable patient attitudes toward the development of AI to assist clinicians with MSK radiograph interpretation. However, the research team stated that there is limited research evaluating clinicians’ opinions on the use of AI to address the increased burden of trauma imaging.
To bridge this research gap, the study aimed to present “novel data of clinicians’ confidence in interpreting trauma radiographs, their perception of AI in healthcare, and their support for the development of systems applied to skeletal radiography.”
To achieve this, the researchers distributed questionnaires to clinicians and medical students throughout southeast England, compiling responses into a database over three months. Questionnaires were designed using validated survey methodology and leveraged plain English to reduce ambiguity and misinterpretation of survey questions.
The first three questions were concerned with participant demographics. The fourth and fifth were divided into two parts, establishing participants’ self-assessed comprehension of AI technology and exposure to it in a healthcare setting using a 10-point Likert scale.
The sixth question was divided into three parts and sought to identify the participants’ confidence in interpreting plain skeletal radiographs for trauma using commonly cited factors known to negatively impact interpretation.
The seventh question identified key capabilities required to perform the initial interpretation of plain skeletal radiographs and asked respondents to indicate their confidence in their ability to do so.
The remaining four questions assessed participants’ favorability toward the general application of AI in healthcare and its specific use in trauma radiography.
Overall, responses from 297 participants were included. The mean score for self-assessed knowledge of AI in healthcare was 3.68 out of 10, with significantly higher knowledge reported by the most senior doctors, scoring an average of 4.88 out of 10.
Participants also indicated substantial favorability toward AI in healthcare and AI applied to skeletal radiography, with average ratings of 7.87 and 7.75 out of 10, respectively. However, there was a preference among participants for a hypothetical system indicating positive findings versus negative ones, with ratings of 7.26 and 6.20.
These findings suggest clear support for the general use of healthcare AI and its specific application in MSK care among clinicians and medical students, with the researchers concluding that the development of AI systems to address the burden of trauma imaging appears well-founded and popular. They also stated that the engagement of the reticent minority should be sought, and clinician education on AI should be improved.
