Methods This narrative review drew on targeted searches of PubMed/MEDLINE and arXiv (January 2020–October 2025), verification of peer-reviewed versions via ACL Anthology for selected preprints, and citation tracking of seminal literature. In this review, we trace the methodological evolution from rules to encoder-based models and LLMs, propose a multidimensional evaluation framework for real-world deployment—which includes accuracy, structural quality, human-in-the-loop effort, stability, and compliance—and develop an operational governance checklist to support auditable and reproducible implementations. Using representative tasks—diagnosis extraction, medication records, clinical trial data, and phenotype integration—we summarize the improvements and failure modes of LLM-based extraction and analyze key challenges, including domain shift, factual “hallucinations,” privacy and regulatory constraints, and cost/latency trade-offs. Finally, we outline future directions through which multimodal and cross-lingual extensions, human–machine collaborative annotation, and standardized reporting practices can advance precision medicine and sustainable, high-quality clinical research.
Background: Notifiable infectious diseases remain a public health priority in China. Understanding long-term epidemiological patterns and developing low-cost, automated forecasting tools may support timely prevention efforts at the county level. Methods: We analyzed 26,756 valid cases of 29 statutory infectious diseases reported in Ziyang County (2013–2023). Diseases with ≥ 500 cumulative cases were selected for primary analysis. Seasonal autoregressive integrated moving average (SARIMA) models were fitted for tuberculosis (TB), influenza, and hand–foot–mouth disease (HFMD), with two training windows (2013–2021 and 2018–2021) used for influenza to account for surveillance updates. Model selection was based on the Akaike information criterion, and forecasts were validated using data from 2022–2023. Results: TB, influenza, and HFMD accounted for more than 55% of the cases. The TB incidence steadily decreased, with spring–summer peaks, whereas the incidence of influenza increased after 2019 and peaked in the winter. Short-horizon forecasts were robust across training windows but less accurate in 2022–2023, likely reflecting post-COVID-19 behavioral and policy shifts. HFMD forecasts were unstable because of zero-inflated data. Conclusions: The incidence of tuberculosis decreased consistently, whereas the incidence of influenza rebounded sharply after nonpharmaceutical interventions were lifted. The SARIMA models reproduced seasonal patterns but showed limited accuracy for influenza and HFMD, reflecting the influence of surveillance changes, behavioral shifts, and sparse data. These findings highlight the need for complementary approaches—such as models explicitly accounting for structural breaks—to improve the reliability of local infectious disease forecasts.
To develop and validate a risk prediction model for incident chronic kidney disease (CKD) in individuals with prediabetes by integrating clinical, biochemi