Page 9 - Read Online

Page 234 Giakas et al. Art Int Surg 2024;4:233-46 https://dx.doi.org/10.20517/ais.2024.24

information to patients and in helping them become more informed about their pathology in general. However, it
often lacks the specific patient context necessary to provide patients with personalized, accurate insights into their
prognosis and medical options.

Keywords: Artificial intelligence, ChatGPT, lumbar decompression, spine surgery

INTRODUCTION
In recent decades, internet search engines have transformed the way patients seek health information and
[1,2]
interact with the health care system . Accordingly, almost 75% of adult internet users have searched for
health information online. Within the orthopedic patient population, up to two-thirds of patients have used
[3,4]
the Internet to search for information about their conditions . Recently, open-access artificial intelligence
(AI) has become a popular tool for information-gathering in healthcare . ChatGPT, launched in
[5-7]
November 2022, is an AI-powered language processing tool that has the ability to integrate billions of
[8,9]
datapoints and fine-tune successive responses to mimic human conversation . Recently, the AI chatbot
garnered attention within the medical community for earning a passing grade on a United States Medical
Licensing Examination practice test, an exam that involves the application of knowledge and critical
thinking .
[10]
As patients become more familiar with AI platforms, they may increasingly look to these programs to
provide instantaneous answers to questions about their health. Previous studies have examined ChatGPT’s
utility in clinical decision making and patient education within orthopedics [11-16] . In a recent paper, Mika
et al. attempted to determine if ChatGPT could answer frequently asked questions related to total hip
[17]
arthroplasty . In their conclusion, the authors noted ChatGPT’s utility in the clinical space by providing
evidence-based responses, but acknowledged that responses often required at least some clarification.
Similarly, in another paper, AI-generated answers to common patient questions regarding minimally
invasive spine surgery (MISS) were explored. While authors acknowledged ChatGPT’s ability to accurately
describe procedures, indications, outcomes, and complications of MISS, they also noted its tendency to
make unsubstantiated claims and potentially confuse patients as they make decisions regarding their
treatment . Given the concerning discrepancies that these prior studies have identified, additional scrutiny
[18]
into AI’s ability to accurately aid patients and physicians in the shared decision-making process is
necessary .
[19]

Degenerative spine conditions are a prominent health concern, particularly within the aging population,
that can cause patients significant physical, mental, and emotional distress. Posterior lumbar decompression
(PLD), a procedure performed to treat degenerative spine conditions, is one of the most commonly
performed procedures by spine surgeons [20,21] . The utility and limitations of ChatGPT in answering common
preoperative questions for patients considering PLD have yet to be explored. Furthermore, while prior
studies in the adult reconstruction literature have attempted to score and analyze ChatGPT responses, no
previous study has methodologically scored AI’s responses to common questions asked by patients
considering spine surgery. As such, the purpose of the present study is to evaluate the clinical accuracy and
applicability of ChatGPT answers to commonly asked questions from patients considering PLD.