RESUMEN
BACKGROUND: Despite the substantial impact of depression on individuals and healthcare utilization, little is known about the specific relationship between depression severity and total cost of care (TCC). This study evaluates the association between depression symptom severity and TCC and how changes in severity affect TCC. METHODS: The analysis was conducted using insurance claims data and data from electronic health records between January 1, 2019 and December 31, 2020. Inclusion criteria comprised insured individuals with coverage during 2019 or 2020, aged one year or older, and identified as having depression in at least one year of the study. Depression symptom severity was assessed using the screening Identification and Stratification (IDS) framework and data available to the research team. The main outcome was TCC per member per month (PMPM) evaluated across the two-year period. RESULTS: Across 2019 and 2020, 744,854 members met inclusion criteria. A total of 369,460 members were studied across both years. Greater depression symptom severity was associated with higher TCC across both years. Unchanged severity was associated with limited change in TCC from 2019 to 2020. Decrease in depression symptoms was associated with an average $41 reduction in PMPM spend, whereas increase in depression symptom severity was associated with an average $608 increase. LIMITATIONS: Limitations include fragmented data, retrospective design that limits causality, and the IDS framework design. CONCLUSION: Changes in depression symptom severity were significantly associated with changes in TCC. Findings reveal financial and clinical opportunities associated with early identification and targeted management of depression.
RESUMEN
Artificial Intelligence (AI) language models continue to expand in both access and capability. As these models have evolved, the number of academic journals in medicine and healthcare which have explored policies regarding AI-generated text has increased. The implementation of such policies requires accurate AI detection tools. Inaccurate detectors risk unnecessary penalties for human authors and/or may compromise the effective enforcement of guidelines against AI-generated content. Yet, the accuracy of AI text detection tools in identifying human-written versus AI-generated content has been found to vary across published studies. This experimental study used a sample of behavioral health publications and found problematic false positive and false negative rates from both free and paid AI detection tools. The study assessed 100 research articles from 2016-2018 in behavioral health and psychiatry journals and 200 texts produced by AI chatbots (100 by "ChatGPT" and 100 by "Claude"). The free AI detector showed a median of 27.2% for the proportion of academic text identified as AI-generated, while commercial software Originality.AI demonstrated better performance but still had limitations, especially in detecting texts generated by Claude. These error rates raise doubts about relying on AI detectors to enforce strict policies around AI text generation in behavioral health publications.