RESUMEN
INTRODUCTION: Government and insurance sponsored exercise programs have demonstrated decreased hospitalizations, but it is unclear if this is the case for self-referred programs. METHODS: In this retrospective cohort study from 2013 to 2020, older adults who participated for at least three months at a community-based exercise center (participants) were compared with those who did not (nonparticipants). Each completed a baseline physical assessment and periodic reassessments thereafter. These data were paired with regional hospital data and a national mortality database. Statistical analysis and modeling were performed from 2020 to 2023. Survival to all-cause hospitalization was assessed with a priori subgroup comparison by gender and cox proportional hazard modeling by age, gender, and comorbidities. RESULTS: The cohort included 718 adults, mean age 69.5 years (SD 8.4), with 411 (57.2%) participants and 307 nonparticipants. Mean follow-up was 26.7 months. Participants had similar baseline measures of fitness (p>0.05) but were more likely to be retired and less likely to have diabetes or prior stroke than nonparticipants. Sustained participation was associated with a reduced rate of all-cause hospitalization (9.0% vs. 12.7%, p=0.02), even when adjusted (HR 0.54; 95% CI 0.34, 0.87, p=0.01). This decrease was noted only in women (p=0.03) but not in men (p=0.49), gender was nonsignificant after adjustment for comorbidities (p=0.15). CONCLUSIONS: Exercise program participation was independently associated with decreased risk of all-cause hospitalization, with possible differential effects by gender. Further randomized trials of the benefits of personalized exercise programs are warranted to assess sex- and gender-specific effects.
RESUMEN
Objectives: This study presents a design framework to enhance the accuracy by which large language models (LLMs), like ChatGPT can extract insights from clinical notes. We highlight this framework via prompt refinement for the automated determination of HEART (History, ECG, Age, Risk factors, Troponin risk algorithm) scores in chest pain evaluation. Methods: We developed a pipeline for LLM prompt testing, employing stochastic repeat testing and quantifying response errors relative to physician assessment. We evaluated the pipeline for automated HEART score determination across a limited set of 24 synthetic clinical notes representing four simulated patients. To assess whether iterative prompt design could improve the LLMs' ability to extract complex clinical concepts and apply rule-based logic to translate them to HEART subscores, we monitored diagnostic performance during prompt iteration. Results: Validation included three iterative rounds of prompt improvement for three HEART subscores with 25 repeat trials totaling 1200 queries each for GPT-3.5 and GPT-4. For both LLM models, from initial to final prompt design, there was a decrease in the rate of responses with erroneous, non-numerical subscore answers. Accuracy of numerical responses for HEART subscores (discrete 0-2 point scale) improved for GPT-4 from the initial to final prompt iteration, decreasing from a mean error of 0.16-0.10 (95% confidence interval: 0.07-0.14) points. Conclusion: We established a framework for iterative prompt design in the clinical space. Although the results indicate potential for integrating LLMs in structured clinical note analysis, translation to real, large-scale clinical data with appropriate data privacy safeguards is needed.