Linea Schmidt

Linea Schmidt

Auto-generated profile, based on publicly available information

Linea Schmidt

Hasso Plattner Institute, Digital Engineering Faculty, University of Potsdam, Potsdam, Germany

Research Focus & Metrics

Academic Career Length

4Years (2023-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
14 June 2024 in Journal of Medical Internet Research
Verified
Journal Article
Open access
LMLars Masanneck
Lars Masanneck
LSLinea Schmidt
Linea Schmidt
ASAntonia Seifert
Antonia Seifert
TKTristan Kölsche
Tristan Kölsche
NHNiklas Huntemann
Niklas Huntemann
MPMarc Pawlitzki
Marc Pawlitzki
Background: Large language models (LLMs) have demonstrated impressive performances in various medical domains, prompting an exploration of their potential utility within the high-demand setting of emergency department (ED) triage. This study evaluated the triage proficiency of different LLMs and ChatGPT, an LLM-based chatbot, compared to professionally trained ED staff and untrained personnel. We further explored whether LLM responses could guide untrained staff in effective triage. Objective: This study aimed to assess the efficacy of LLMs and the associated product ChatGPT in ED triage compared to personnel of varying training status and to investigate if the models’ responses can enhance the triage proficiency of untrained personnel. Methods: A total of 124 anonymized case vignettes were triaged by untrained doctors; different versions of currently available LLMs; ChatGPT; and professionally trained raters, who subsequently agreed on a consensus set according to the Manchester Triage System (MTS). The prototypical vignettes were adapted from cases at a tertiary ED in Germany. The main outcome was the level of agreement between raters’ MTS level assignments, measured via quadratic-weighted Cohen κ. The extent of over- and undertriage was also determined. Notably, instances of ChatGPT were prompted using zero-shot approaches without extensive background information on the MTS. The tested LLMs included raw GPT-4, Llama 3 70B, Gemini 1.5, and Mixtral 8x7b. Results: GPT-4–based ChatGPT and untrained doctors showed substantial agreement with the consensus triage of professional raters (κ=mean 0.67, SD 0.037 and κ=mean 0.68, SD 0.056, respectively), significantly exceeding the performance of GPT-3.5–based ChatGPT (κ=mean 0.54, SD 0.024; P<.001). When untrained doctors used this LLM for second-opinion triage, there was a slight but statistically insignificant performance increase (κ=mean 0.70, SD 0.047; P=.97). Other tested LLMs performed similar to or worse than GPT-4–based ChatGPT or showed odd triaging behavior with the used parameters. LLMs and ChatGPT models tended toward overtriage, whereas untrained doctors undertriaged. Conclusions: While LLMs and the LLM-based product ChatGPT do not yet match professionally trained raters, their best models’ triage proficiency equals that of untrained ED doctors. In its current form, LLMs or ChatGPT thus did not demonstrate gold-standard performance in ED triage and, in the setting of this study, failed to significantly improve untrained doctors’ triage when used as decision support. Notable performance enhancements in newer LLM versions over older ones hint at future improvements with further technological development and specific training.

24 May 2024 in npj Digital Medicine
Verified
Journal Article
Open access
PDF file
LSLinea Schmidt
Linea Schmidt
MPMarc Pawlitzki
Marc Pawlitzki
BRBernhard Y. Renard
Bernhard Y. Renard
SMSven G. Meuth
Sven G. Meuth
LMLars Masanneck
Lars Masanneck
The 2019 German Digital Healthcare Act introduced the Digital Health Application program, known in German as ‘Digitale Gesundheitsanwendungen’ (DiGA). The program has established a pioneering model for integrating Digital Therapeutics (DTx) into a healthcare system with scalable and effective reimbursement strategies. To date, the continuous upward trend enabled by this framework has resulted in more than 374,000 DiGA prescriptions, increasingly cementing its role in the German healthcare system. This perspective provides a synthesis of the DiGA program’s evolution since its inception three years ago, highlighting trends regarding prescriptions and pricing as well as criticisms and identified shortcomings. It further discusses forthcoming legislative amendments, including the anticipated integration of higher-risk medical devices, which have the potential to significantly transform the program. Despite encountering challenges related to effectiveness, evidence requirements, and integration within the healthcare system, the DiGA program continues to evolve and serves as a seminal example for the integration of DTx, offering valuable insights for healthcare systems globally.

3 November 2025 in JAMIA Open
Verified
Journal Article
Open access
PGPablo Guerrero
Pablo Guerrero
MEMorten Ernebjerg
Morten Ernebjerg
THThomas Holst
Thomas Holst
DWDavid Weese
David Weese
HDHerve DiBello
Herve DiBello
Girish N. NadkarniGirish N. Nadkarni
Girish N. Nadkarni
Objective: To present the Artificial Intelligence-Ready Mount Sinai (AIR·MS) platform-unified access to diverse clinical datasets from the Mount Sinai Health System (MSHS), along with computational infrastructure for AI-driven research and demonstrate its utility with 3 research projects. Materials and Methods: AIR·MS integrates structured and unstructured data from multiple MSHS sources via the OMOP Common Data Model on an in-memory columnar database. Unstructured pathology and radiology data are integrated through metadata extracted from and linking the raw source data. Data access and analytics are supported from the HIPAA-compliant Azure cloud and the on-premises Minerva High-Performance Computing (HPC) environment. Results: AIR·MS provides access to structured electronic health records, clinical notes, and metadata for pathology and radiology images, covering over 12M patients. The platform enables interactive cohort building and AI model training. Experimentation with complex cohort queries confirm a high system performance. Three use cases demonstrate, risk-factor discovery, and federated cardiovascular risk modeling. Discussion: AIR·MS demonstrates how clinical data and infrastructure can be integrated to support large-scale AI-based research. The platform’s performance, scale, and cross-institutional design position it as a model for similar initiatives. Conclusion: AIR·MS provides a scalable, secure, and collaborative platform for AI-enabled healthcare research on multimodal clinical data.

14 January 2026 in Communications Medicine
Verified
Journal Article
Open access
PDF file
LSLinea Schmidt
Linea Schmidt
SISusanne Ibing
Susanne Ibing
FBFlorian Borchert
Florian Borchert
JHJulian Hugo
Julian Hugo
AMAllison A. Marshall
Allison A. Marshall
RURyan C. Ungaro
Ryan C. Ungaro
Background: Real-world studies based on electronic health records often require manual chart review to derive patients’ clinical phenotypes, a labor-intensive task with limited scalability. Here, we developed and compared computable phenotyping based on rules using the spaCy framework and a Large Language Model (LLM), GPT-4, for sub-phenotyping of patients with Crohn’s disease, considering age at diagnosis and disease behavior. Methods: For our rule-based approach, we leveraged the spaCy framework and for the LLM-based approach, we used the GPT-4 model. The underlying data included 49,572 clinical notes and 2204 radiology reports from 584 Crohn’s disease patients. A test set of 280 clinical texts was labeled at sentence-level, in addition to patient-level ground truth data. The algorithms were evaluated based on their recall, precision, specificity values, and F1 scores. Results: Overall, we observe similar or better performance using GPT-4 compared to the rules. On a note-level, the F1 score is at least 0.90 for disease behavior and 0.82 for age at diagnosis, and on patient level at least 0.66 for disease behavior and 0.71 for age at diagnosis. Conclusions: To our knowledge, this is the first study to explore computable phenotyping algorithms based on clinical narrative text for these complex tasks, where prior inter-annotator agreements ranged from 0.54 to 0.98. There is no statistical evidence for a difference to the performance of human experts on this task. Our findings underline the potential of LLMs for computable phenotyping and may support large-scale cohort analyses from electronic health records and streamline chart review processes in the future.

2 October 2023 in JMIR Preprints
Verified
Preprint
Open access
LMLars Masanneck
Lars Masanneck
LSLinea Schmidt
Linea Schmidt
ASAntonia Seifert
Antonia Seifert
TKTristan Kölsche
Tristan Kölsche
NHNiklas Huntemann
Niklas Huntemann
MPMarc Pawlitzki
Marc Pawlitzki
BACKGROUND Large language models (LLMs) have demonstrated impressive performances in various medical domains, prompting an exploration of their potential utility within the high-demand setting of emergency department (ED) triage. This study evaluated the triage proficiency of different LLMs and ChatGPT, an LLM-based chatbot, compared to professionally trained ED staff and untrained personnel. We further explored whether LLM responses could guide untrained staff in effective triage. OBJECTIVE This study aimed to assess the efficacy of LLMs and the associated product ChatGPT in ED triage compared to personnel of varying training status and to investigate if the models’ responses can enhance the triage proficiency of untrained personnel. METHODS A total of 124 anonymized case vignettes were triaged by untrained doctors; different versions of currently available LLMs; ChatGPT; and professionally trained raters, who subsequently agreed on a consensus set according to the Manchester Triage System (MTS). The prototypical vignettes were adapted from cases at a tertiary ED in Germany. The main outcome was the level of agreement between raters’ MTS level assignments, measured via quadratic-weighted Cohen κ. The extent of over- and undertriage was also determined. Notably, instances of ChatGPT were prompted using zero-shot approaches without extensive background information on the MTS. The tested LLMs included raw GPT-4, Llama 3 70B, Gemini 1.5, and Mixtral 8x7b. RESULTS GPT-4–based ChatGPT and untrained doctors showed substantial agreement with the consensus triage of professional raters (κ=mean 0.67, SD 0.037 and κ=mean 0.68, SD 0.056, respectively), significantly exceeding the performance of GPT-3.5–based ChatGPT (κ=mean 0.54, SD 0.024; P<.001). When untrained doctors used this LLM for second-opinion triage, there was a slight but statistically insignificant performance increase (κ=mean 0.70, SD 0.047; P=.97). Other tested LLMs performed similar to or worse than GPT-4–based ChatGPT or showed odd triaging behavior with the used parameters. LLMs and ChatGPT models tended toward overtriage, whereas untrained doctors undertriaged. CONCLUSIONS While LLMs and the LLM-based product ChatGPT do not yet match professionally trained raters, their best models’ triage proficiency equals that of untrained ED doctors. In its current form, LLMs or ChatGPT thus did not demonstrate gold-standard performance in ED triage and, in the setting of this study, failed to significantly improve untrained doctors’ triage when used as decision support. Notable performance enhancements in newer LLM versions over older ones hint at future improvements with further technological development and specific training.

28 July 2025 in medRxiv
Verified
Preprint
Open access
PDF file
ANAnatol-Fiete Näher
Anatol-Fiete Näher
LSLinea Schmidt
Linea Schmidt
MKMarvin Kopka
Marvin Kopka
MSMatthias Schulte-Althoff
Matthias Schulte-Althoff
AAAkhyar Ahmed
Akhyar Ahmed
The rapid adoption of digital health technologies (DHT) in clinical trials promises to enhance trial representativeness by eliminating geographical barriers and reducing participant burden. However, their actual impact on participant representation remains poorly understood. Our analysis of 68,206 US-based trials from ClinicalTrials.gov uncovers gaps in participant representation in trials utilizing DHT for data collection (DHTcd). While DHTcd trials achieved overrepresentation of female participants (+1.7% versus US Census data and +5.5% vs. non-DHT trials) and reduced underrepresentation of Black participants by 4.2% (vs. non-DHT trials), they showed underrepresentation of adults aged 65 years and older (−8.8% vs. non-DHT trials). These effects vary by therapeutic area. We furthermore find that DHTcd trials report group-specific sample compositions less frequently than non-DHT trials, suggesting that growing DHTcd adoption has not been systematically used to improve trial representativeness. We conclude that evidence-based and indication-specific implementation strategies are needed to enhance standards for DHT-based data collection in clinical trials and to prevent a widening digital divide in trial participation.