Large language models and the future of rheumatology: assessing impact and emerging opportunities.

PURPOSE OF REVIEW: Large language models (LLMs) have grown rapidly in size and capabilities as more training data and compute power has become available. Since the release of ChatGPT in late 2022, there has been growing interest and exploration around potential applications of LLM technology. Numerous examples and pilot studies demonstrating the capabilities of these tools have emerged across several domains. For rheumatology professionals and patients, LLMs have the potential to transform current practices in medicine. RECENT FINDINGS: Recent studies have begun exploring capabilities of LLMs that can assist rheumatologists in clinical practice, research, and medical education, though applications are still emerging. In clinical settings, LLMs have shown promise in assist healthcare professionals enabling more personalized medicine or generating routine documentation like notes and letters. Challenges remain around integrating LLMs into clinical workflows, accuracy of the LLMs and ensuring patient data confidentiality. In research, early experiments demonstrate LLMs can offer analysis of datasets, with quality control as a critical piece. Lastly, LLMs could supplement medical education by providing personalized learning experiences and integration into established curriculums. SUMMARY: As these powerful tools continue evolving at a rapid pace, rheumatology professionals should stay informed on how they may impact the field.

Healthcare disparities: patients' perspectives on barriers to joint replacement.

OBJECTIVE: Racial and ethnic disparities in arthroplasty utilization are evident, but the reasons are not known. We aimed to identify concerns that may contribute to barriers to arthroplasty from the patient's perspective. METHODS: We identified patients' concerns about arthroplasty by performing a mixed methods study. Themes identified during semi-structured interviews with Black and Hispanic patients with advanced symptomatic hip or knee arthritis were used to develop a questionnaire to quantify and prioritize their concerns. Multiple linear and logistic regression analyses were conducted to determine the association between race/ethnicity and the importance of each theme. Models were adjusted for sex, insurance, education, HOOS, JR/KOOS, JR, and discussion of joint replacement with a doctor. RESULTS: Interviews with eight participants reached saturation and provided five themes used to develop a survey answered by 738 (24%) participants; 75.5% White, 10.3% Black, 8.7% Hispanic, 3.9% Asian/Other. Responses were significantly different between groups (p < 0.05). Themes identified were "Trust in the surgeon" "Recovery", "Cost/Insurance", "Surgical outcome", and "Personal suitability/timing". Compared to Whites, Blacks were two-fold, Hispanics four-fold more likely to rate "Trust in the surgeon" as very/extremely important. Blacks were almost three times and Hispanics over six times more likely to rate "Recovery" as very/extremely important. CONCLUSION: We identified factors of importance to patients that may contribute to barriers to arthroplasty, with marked differences between Blacks, Hispanics, and Whites.

A Novel Approach for Mixed-Methods Research Using Large Language Models: A Report Using Patients' Perspectives on Barriers to Arthroplasty.

OBJECTIVE: Mixed-methods research is valuable in health care to gain insights into patient perceptions. However, analyzing textual data from interviews can be time-consuming and require multiple analysts for investigator triangulation. This study aims to explore a novel approach to investigator triangulation in mixed-methods research by employing a large language model (LLM) for analyzing data from patient interviews. METHODS: This study compared the thematic analysis and survey generation performed by human investigators and ChatGPT-4, which uses GPT-4 as its backbone model, using data from an existing study that explored patient perceptions of barriers to arthroplasty. The human- and ChatGPT-4-generated themes and surveys were compared and evaluated based on their representation of salient themes from a predetermined topic guide. RESULTS: ChatGPT-4 generated analogous dominant themes and a comprehensive corresponding survey as the human investigators but in significantly less time. The survey questions generated by ChatGPT-4 were less precise than those developed by human investigators. The mixed-methods flowchart proposes integrating LLMs and human investigators as a supplementary tool for the preliminary thematic analysis of qualitative data and survey generation. CONCLUSION: By utilizing a combination of LLMs and human investigators through investigator triangulation, researchers may be able to conduct more efficient mixed-methods research to better understand patient perspectives. Ethical and qualitative implications of using LLMs should be considered.

Risk factors and clinical outcomes associated with multiple as opposed to single pathogens detected on the gastrointestinal disease polymerase chain reaction assay.

BACKGROUND: The use of gastrointestinal disease multiplex polymerase chain reaction (GI PCR) testing has become common for suspected gastrointestinal infection. Patients often test positive for multiple pathogens simultaneously through GI PCR, although the clinical significance of this is uncertain. METHODS: This retrospective cohort study investigated risk factors and clinical outcomes associated with detection of multiple (as opposed to single) pathogens on GI PCR. We included adult patients who underwent GI PCR testing from 2020 to 2023 and had one or more pathogens detected. We compared patients with multiple versus those with single pathogens and hypothesized that immunosuppression would be a risk factor for detection of multiple pathogens. We further hypothesized that, during the 90 days after GI PCR testing, patients with multiple pathogens would have worse clinical outcomes such as increased rates of emergency department (ED) visits, death, hospitalization, or ambulatory care visits. RESULTS: GI PCR was positive in 1341 (29%) of tested patients; 356 patients had multiple pathogens and 985 had one pathogen. The most common pathogens included Enteropathogenic Escherichia coli (EPEC, 27%), norovirus (17%), and Enteroaggregative E. coli (EAEC, 14%) in both multi- and singly positive patients. Immunosuppression was not associated with multiple pathogens (adjusted odds ratio [aOR] 1.35, 95% CI 0.96, 1.86). The factors most associated with multiple pathogens were Hispanic ethnicity (OR 1.86, 95% CI 1.42, 2.45) and chronic kidney disease (OR 1.69, 95% CI 1.13, 2.49). Patients with multiple pathogens were more likely to have ED visits during the 90 days after GI PCR testing (40% vs. 32%, p < 0.01), but they were not more likely to die, be hospitalized, or to have ambulatory medical visits. CONCLUSIONS: Immunosuppression was not associated with detection of multiple as opposed to single pathogens on GI PCR testing. There were worse clinical outcomes associated with detection of multiple pathogens, although these effects were modest.

Clinical and Histological Features of Prosthetic Joint Infections May Differ in Patients With Inflammatory Arthritis and Osteoarthritis.

Background: Patients with inflammatory arthritis are at increased risk of prosthetic joint infections (PJIs), but diagnosis in these patients can be challenging because active inflammatory arthritis produces elevated inflammatory markers that may mimic those seen in PJI. Purpose: In this pilot study, we sought to identify the clinical, microbiologic, and histopathologic features of culture-positive and culture-negative PJI in patients with inflammatory arthritis who underwent total hip arthroplasty (THA) or total knee arthroplasty (TKA). We also sought to obtain preliminary data to support a definitive study of optimal methods for PJI diagnosis in patients with inflammatory arthritis. Methods: We performed a retrospective analysis of TKA and THA patients treated for PJI from 2009 to 2018 at a single tertiary care orthopedic institution. Data were extracted from a longitudinally maintained hospital infection database. We reviewed hematoxylin and eosin slides of osteoarthritis and inflammatory arthritis PJI cases matched 3:1, respectively, by age, sex, and culture status. Clinical characteristics were evaluated using the Fisher exact test, χ2 test, Student t test, and Mann-Whitney U test where appropriate. Results: A total of 807 PJI cases were identified (36 inflammatory arthritis and 771 osteoarthritis cases). Patients with inflammatory arthritis presented younger, had a higher Charlson Comorbidity Index, more frequently used glucocorticoids, were more likely women, and had a higher proportion of culture-negative PJI compared with osteoarthritis patients. Of the 88 inflammatory arthritis cases reviewed for histopathology, a higher proportion of culture-positive than culture-negative PJI cases had >10 polymorphonuclear leucocytes per high-power field and met Musculoskeletal Infection Society criteria but presented with less chronic inflammation. Conclusions: This retrospective prognostic study suggests that culture-negative PJI may be more frequent in patients with inflammatory arthritis than in those with osteoarthritis. Chronic infections, antibiotic use, or misdiagnosis may be contributing factors to unclear PJI diagnoses among culture-negative cases. This preliminary work supports the need for further studies to assess the differences in clinical features between culture-negative and culture-positive PJI in patients with inflammatory arthritis and the ability of biological diagnostic markers to discriminate between them in this population.