Evaluating ChatGPT responses to frequently asked patient questions regarding periprosthetic joint infection after total hip and knee arthroplasty

被引:5
作者
Hu, Xiaojun [1 ,2 ,3 ,4 ,5 ]
Niemann, Marcel [1 ,2 ,3 ,4 ]
Kienzle, Arne [1 ,2 ,3 ,4 ]
Braun, Karl [6 ]
Back, David Alexander [1 ,2 ,3 ,4 ]
Gwinner, Clemens [1 ,2 ,3 ,4 ]
Renz, Nora [1 ,2 ,3 ,4 ]
Stoeckle, Ulrich [1 ,2 ,3 ,4 ]
Trampuz, Andrej [1 ,2 ,3 ,4 ]
Meller, Sebastian [1 ,2 ,3 ,4 ]
机构
[1] Charite Univ Med Berlin, Ctr Musculoskeletal Surg, Augustenburger Pl 1, D-13353 Berlin, Germany
[2] Free Univ Berlin, Berlin, Germany
[3] Humboldt Univ, Berlin, Germany
[4] Berlin Inst Hlth, Berlin, Germany
[5] Seventh Peoples Hosp Chongqing, Dept Orthoped, Chongqing, Peoples R China
[6] Tech Univ Munich, Univ Hosp Rechts Isar, Dept Trauma Surg, Munich, Germany
关键词
Artificial intelligence; Chatbot; ChatGPT; OpenAI; patient education; periprosthetic joint infection; RISK-FACTORS; ANTIBIOTICS; REVISION; SURGERY;
D O I
10.1177/20552076241272620
中图分类号
R19 [保健组织与事业(卫生事业管理)];
学科分类号
摘要
Background Patients access relevant information concerning their orthopaedic surgery resources through multiple information channels before presenting for clinical treatment. Recently, artificial intelligence (AI)-powered chatbots have become another source of information for patients. The currently developed AI chat technology ChatGPT (OpenAI LP) is an application for such purposes and it has been rapidly gaining popularity, including for patient education. This study sought to evaluate whether ChatGPT can correctly answer frequently asked questions (FAQ) regarding periprosthetic joint infection (PJI).Methods Twelve FAQs about PJI after hip and knee arthroplasty were identified from the websites of fifteen international clinical expert centres. ChatGPT was confronted with these questions and its responses were analysed for their accuracy using an evidence-based approach by a multidisciplinary team. Responses were categorised in four groups: (1) Excellent response that did not require additional improvement; (2) Satisfactory responses that required a small amount of improvement; (3) Satisfactory responses that required moderate improvement; and (4) Unsatisfactory responses that required a large amount of improvement.Results From the analysis of the responses given by the chatbot, no reply received an 'unsatisfactory' rating; one did not require any correction; and the majority of the responses required low (7 out of 12) or moderate (4 out of 12) clarification. Although a few responses required minimal clarification, the chatbot responses were generally unbiased and evidence-based, even when asked controversial questions.Conclusions The AI-chatbot ChatGPT was able to effectively answer the FAQs of patients seeking information around PJI diagnosis and treatment. The given information was also written in a manner that can be assumed to be understandable by patients. The chatbot could be a valuable clinical tool for patient education and understanding around PJI treatment in the future. Further studies should evaluate its use and acceptance by patients with PJI.
引用
收藏
页数:14
相关论文
共 77 条
[1]   Risk factors, diagnosis and management of prosthetic joint infection after total hip arthroplasty [J].
Ahmed, Syed S. ;
Begum, Fahima ;
Kayani, Babar ;
Haddad, Fares S. .
EXPERT REVIEW OF MEDICAL DEVICES, 2019, 16 (12) :1063-1070
[2]   Perioperative and Modifiable Risk Factors for Periprosthetic Joint Infections (PJI) and Recommended Guidelines [J].
Alamanda V.K. ;
Springer B.D. .
Current Reviews in Musculoskeletal Medicine, 2018, 11 (3) :325-331
[3]   Microbiota-Pain Association; Recent Discoveries and Research Progress [J].
Alizadeh, Naser ;
Naderi, Ghazal ;
Kahrizi, Mohammad Saeed ;
Haghgouei, Tannaz ;
Mobed, Ahmad ;
Shah-abadi, Mehran Ebrahimi .
CURRENT MICROBIOLOGY, 2023, 80 (01)
[4]   Patients at a high risk of PJI: Can we reduce the incidence of infection using dual antibiotic-loaded bone cement? [J].
Berberich, Christof ;
Josse, Jerome ;
Ruiz, Pablo Sanz .
ARTHROPLASTY, 2022, 4 (01)
[5]   Intravenous antibiotic duration in the treatment of prosthetic joint infection: systematic review and meta-analysis [J].
Bouji, Nour ;
Wen, Sijin ;
Dietz, Matthew J. .
JOURNAL OF BONE AND JOINT INFECTION, 2022, 7 (05) :191-202
[6]  
Breznicky Jozef, 2020, Med Glas (Zenica), V17, P92, DOI 10.17392/1046-20
[7]   Social and Demographic Patterns of Health-Related Internet Use Among Adults in the United States: A Secondary Data Analysis of the Health Information National Trends Survey [J].
Calixte, Rose ;
Rivera, Argelis ;
Oridota, Olutobi ;
Beauchamp, William ;
Camacho-Rivera, Marlene .
INTERNATIONAL JOURNAL OF ENVIRONMENTAL RESEARCH AND PUBLIC HEALTH, 2020, 17 (18) :1-16
[8]   Risk Factors for Periprosthetic Joint Infection after Primary Total Knee Arthroplasty [J].
Carlos Rodriguez-Merchan, Emerito ;
Delgado-Martinez, Alberto D. .
JOURNAL OF CLINICAL MEDICINE, 2022, 11 (20)
[9]   State of the art of 18F-FDG PET/CT application in inflammation and infection: a guide for image acquisition and interpretation [J].
Casali, Massimiliano ;
Lauri, Chiara ;
Altini, Corinna ;
Bertagna, Francesco ;
Cassarino, Gianluca ;
Cistaro, Angelina ;
Erba, Anna Paola ;
Ferrari, Cristina ;
Mainolfi, Ciro Gabriele ;
Palucci, Andrea ;
Prandini, Napoleone ;
Baldari, Sergio ;
Bartoli, Francesco ;
Bartolomei, Mirco ;
D'Antonio, Adriana ;
Dondi, Francesco ;
Gandolfo, Patrizia ;
Giordano, Alessia ;
Laudicella, Riccardo ;
Massollo, Michela ;
Nieri, Alberto ;
Piccardo, Arnoldo ;
Vendramin, Laura ;
Muratore, Francesco ;
Lavelli, Valentina ;
Albano, Domenico ;
Burroni, Luca ;
Cuocolo, Alberto ;
Evangelista, Laura ;
Lazzeri, Elena ;
Quartuccio, Natale ;
Rossi, Brunella ;
Rubini, Giuseppe ;
Sollini, Martina ;
Versari, Annibale ;
Signore, Alberto .
CLINICAL AND TRANSLATIONAL IMAGING, 2021, 9 (04) :299-339
[10]  
Chang Wen-Pei, 2023, Surg Infect (Larchmt), V24, P671, DOI 10.1089/sur.2023.139