Timothy Keeley
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Timothy Keeley has a total of 14 co-authors. Here are the top co-authors by number of shared publications:
- Alexandra So1 shared publication
- Ashley E. Levack
Department of Orthopaedic Surgery & Rehabilitation, Loyola University Medical Center, Maywood, IL
1 shared publication - Derek Ward
Department of Orthopaedic Surgery, University of California, San Francisco, 1500 Owens St, San Francisco, CA 94158, USA
1 shared publication - Douglas W. Lundy
St. Luke's University Health Network, Bethlehem, PA
1 shared publication - Emma Grellinger
School of Medicine, University of California San Francisco, San Francisco, CA 94143, USA
1 shared publication - Eric Meinberg
HealthPartners Medical Group, Bloomington, MN
1 shared publication
Latest Publications
- Open Access
- 1 November 2026
- Published by Wolters Kluwer Health in JAAOS: Global Research and Reviews
- Vol. 10 (11)
American Indian and Alaska Native Patients Undergoing Hip Fracture Surgery: A National Retrospective Cohort Study Between 2016 and 2022
Introduction: Racial and ethnic disparities have been reported in the treatment of hip fractures; however, American Indian and Alaska Native (AI/AN) patients are often excluded. No existing racial disparity studies have evaluated the management and outcomes of hip fractures among AI/AN patients. We aim to assess whether there are disparities in risk factors, management, and complications in the surgical treatment of hip fractures among AI/AN patients. Methods: This retrospective cohort study used the American College of Surgeons National Surgical Quality Improvement Program (NSQIP) database to analyze AI/AN and White patients undergoing surgical treatment of hip fractures between 2016 and 2022. Descriptive statistics, t -tests, chi-squared tests, and multiple regression models were used to assess differences in patient demographics, preoperative comorbidities, time to surgery (TTS), and postoperative complications. Results: AI/AN patients were significantly younger than White patients yet had higher rates of preoperative comorbidities (diabetes [ P < 0.001], smoking [ P < 0.001], dialysis [ P < 0.001], anemia [ P = 0.007], and delirium [ P < 0.001]). When controlling for patient characteristics, TTS and mortality were similar between AI/AN and White patients ( P = 0.31, P = 0.38, respectively). However, AI/AN patients had significantly more postoperative complications including stroke ( P = 0.04) and pneumonia ( P < 0.01), had a shorter length of stay (LOS) ( P < 0.001), and were less likely to be discharged to a facility ( P < 0.001). Discussion: AI/AN patients with hip fractures had a higher burden of preoperative comorbidities and postoperative complications when compared with White patients. Reassuringly, AI/AN and White patients had similar TTS and 30-day mortality. When controlling for patient characteristics, AI/AN patients had a shorter LOS and were more likely to be discharged home. This may be inferred as a lack of access to rehabilitation facilities on discharge or more support from patient families and communities at home. Overall, AI/AN patients should be included in current and future research to better assess disparities in orthopaedic care.
- 1 November 2026
- Published by Wolters Kluwer Health in JAAOS: Global Research and Reviews
- Vol. 10 (11)
Open AccessAmerican Indian and Alaska Native Patients Undergoing Hip Fracture Surgery: A National Retrospective Cohort Study Between 2016 and 2022
Introduction: Racial and ethnic disparities have been reported in the treatment of hip fractures; however, American Indian and Alaska Native (AI/AN) patients are often excluded. No existing racial disparity studies have evaluated the management and outcomes of hip fractures among AI/AN patients. We aim to assess whether there are disparities in risk factors, management, and complications in the surgical treatment of hip fractures among AI/AN patients. Methods: This retrospective cohort study used the American College of Surgeons National Surgical Quality Improvement Program (NSQIP) database to analyze AI/AN and White patients undergoing surgical treatment of hip fractures between 2016 and 2022. Descriptive statistics, t -tests, chi-squared tests, and multiple regression models were used to assess differences in patient demographics, preoperative comorbidities, time to surgery (TTS), and postoperative complications. Results: AI/AN patients were significantly younger than White patients yet had higher rates of preoperative comorbidities (diabetes [ P < 0.001], smoking [ P < 0.001], dialysis [ P < 0.001], anemia [ P = 0.007], and delirium [ P < 0.001]). When controlling for patient characteristics, TTS and mortality were similar between AI/AN and White patients ( P = 0.31, P = 0.38, respectively). However, AI/AN patients had significantly more postoperative complications including stroke ( P = 0.04) and pneumonia ( P < 0.01), had a shorter length of stay (LOS) ( P < 0.001), and were less likely to be discharged to a facility ( P < 0.001). Discussion: AI/AN patients with hip fractures had a higher burden of preoperative comorbidities and postoperative complications when compared with White patients. Reassuringly, AI/AN and White patients had similar TTS and 30-day mortality. When controlling for patient characteristics, AI/AN patients had a shorter LOS and were more likely to be discharged home. This may be inferred as a lack of access to rehabilitation facilities on discharge or more support from patient families and communities at home. Overall, AI/AN patients should be included in current and future research to better assess disparities in orthopaedic care.
- Open Access
- 1 May 2026
- Published by Wolters Kluwer Health in OTA International
- Vol. 9 (2)
Deriving the OTA/AO fracture classification from routinely collected radiology reports using a large language model
Objectives: Fracture classification plays a pivotal role in research and quality assurance; despite its wide acceptance, the OTA/AO classification is seldom documented in patients' electronic medical records, which impedes fracture registry creation and effective interdisciplinary communication. In this study, we investigate “off-the-shelf” large language models (LLMs) in translating free text in radiology reports into OTA/AO classification labels. Methods: We employed a Health Insurance Portability and Accountability Act-compliant LLM to classify 109 fracture descriptions from randomly selected radiology reports in a deidentified electronic medical record database. Ground-truth classifications were assigned by expert orthopaedic traumatologists based on corresponding radiographs. Multiple prompting strategies were tested, including zero-shot prompting, zero-shot chain-of-thought prompting, and retrieval-augmented generation. We additionally asked the LLM to assign classification labels to “ideal” fracture descriptions written according to the 2018 OTA/AO Fracture and Dislocation Classification Compendium. Model performance was assessed using Cohen kappa and accuracy against ground-truth labels. Results: The 3 prompting strategies tested yielded similar classification performance on radiology report fracture descriptions, with almost perfect agreement at the bone and bone and location levels. Performance declined to slight agreement at the subgroup level. The best performance was observed using ideal fracture descriptions with retrieval-augmented generation, in which the agreement between the full LLM-generated and ground-truth labels remained moderate. Classification errors were largely due to imprecise descriptions, hallucinated information, or incorrect application of factually correct information. Conclusions: Our study demonstrates some potential for LLMs to translate free-text fracture descriptions into OTA/AO classifications, allowing for efficient labeling of large datasets of radiology reports. Future work should focus on refining model classification capabilities using more sophisticated prompting methods. Level of Evidence: Level III.
- 1 May 2026
- Published by Wolters Kluwer Health in OTA International
- Vol. 9 (2)
Open AccessDeriving the OTA/AO fracture classification from routinely collected radiology reports using a large language model
Objectives: Fracture classification plays a pivotal role in research and quality assurance; despite its wide acceptance, the OTA/AO classification is seldom documented in patients' electronic medical records, which impedes fracture registry creation and effective interdisciplinary communication. In this study, we investigate “off-the-shelf” large language models (LLMs) in translating free text in radiology reports into OTA/AO classification labels. Methods: We employed a Health Insurance Portability and Accountability Act-compliant LLM to classify 109 fracture descriptions from randomly selected radiology reports in a deidentified electronic medical record database. Ground-truth classifications were assigned by expert orthopaedic traumatologists based on corresponding radiographs. Multiple prompting strategies were tested, including zero-shot prompting, zero-shot chain-of-thought prompting, and retrieval-augmented generation. We additionally asked the LLM to assign classification labels to “ideal” fracture descriptions written according to the 2018 OTA/AO Fracture and Dislocation Classification Compendium. Model performance was assessed using Cohen kappa and accuracy against ground-truth labels. Results: The 3 prompting strategies tested yielded similar classification performance on radiology report fracture descriptions, with almost perfect agreement at the bone and bone and location levels. Performance declined to slight agreement at the subgroup level. The best performance was observed using ideal fracture descriptions with retrieval-augmented generation, in which the agreement between the full LLM-generated and ground-truth labels remained moderate. Classification errors were largely due to imprecise descriptions, hallucinated information, or incorrect application of factually correct information. Conclusions: Our study demonstrates some potential for LLMs to translate free-text fracture descriptions into OTA/AO classifications, allowing for efficient labeling of large datasets of radiology reports. Future work should focus on refining model classification capabilities using more sophisticated prompting methods. Level of Evidence: Level III.