This study examines the frequencies of behavioral health diagnoses and treatments provided by primary care versus psychiatry, offering a framework for primary care providers and trainees to focus on prevalent behavioral healthcare needs. A retrospective analysis of 57 million primary care visits and 18 million behavioral healthcare visits from a Military Health System (MHS) database evaluated frequencies and proportions of specific mental health diagnoses from 2017 to 2021. The analysis compared percentages between direct and purchased care (PC) and further compared these to behavioral healthcare visits. A subset of data from the final year was also analyzed for psychotropic prescription patterns. Antidepressants were the predominant psychotropic medication class prescribed, accounting for more than all other classes combined. Prevalence ratios for common mental health diagnoses in primary care indicated that anxiety and depression diagnoses were less frequent in direct care versus PC for both active duty and dependents. However, adjustment disorder diagnosis in primary care was 2.59 times (95% confidence interval: 2.57, 2.61) more frequent in direct care versus PC in dependents. This study continues the evaluation of primary care behavioral health care provision trends in the MHS, providing crucial data for determining the training needed for primary care providers to appropriately treat mental health conditions. Understanding these patterns can help primary care providers confidently address the mental health conditions most encountered in practice.
As interest grows in multimodal large language models (LLMs) for ophthalmic imaging, their reliability must be established before clinical use. A key prerequisite is determining whether image quality permits interpretation. However, their ability to assess retinal nerve fiber layer (RNFL) optical coherence tomography (OCT) scan quality remains uncertain. We compared a general-purpose AI tool (ChatGPT-5.1) with a clinical AI tool (DoxGPT) for RNFL OCT report-quality assessment. Both evaluated 145 bilateral reports from 11 patients in an enriched quality-control sample using global and stepwise prompts. One ophthalmologist’s responses to the same predefined quality questions served as the clinical reference for calculating percent agreement, linearly weighted Cohen’s κ with patient-level cluster-bootstrap 95% confidence intervals (CIs), and prevalence-adjusted bias-adjusted κ (PABAK). Global prompting showed higher observed agreement than stepwise prompting: ChatGPT, 56% versus 47% (κ = 0.36 [95% CI, 0.23–0.49] versus 0.12 [0.08–0.18]); DoxGPT, 59% versus 48% (κ = 0.41 [0.28–0.54] versus 0.13 [0.03–0.23]). Component agreement was higher for signal strength (ChatGPT/DoxGPT: 99%/99%) and optic nerve head centering (87%/96%), and lower for dropout (41%/43%) and segmentation error (64%/65%). Both tools showed limited agreement with the ophthalmologist for overall RNFL report reliability. Their accuracy requires independent validation, supporting continued clinician review.