A rapid-cycle assessment strategy for understanding the opioid overdose epidemic in local communities.

Background: Certain communities in the United States experience greater opioid-involved overdose mortality than others. Interventions to stem overdose benefit from contextual understandings of communities' needs and strengths in addressing the opioid crisis. This project aims to understand multiple stakeholder perspectives on the opioid epidemic in communities disproportionately affected by opioid-involved overdose mortality. Methods: We performed a rapid-cycle qualitative assessment study utilizing in-depth interviews with community stakeholders and observations of community meetings in eight communities in Allegheny County, Pennsylvania, USA, disproportionately impacted by opioid-involved overdose mortality. Stakeholder categories included: current and past illicit users of opioids; medical and social service providers; emergency medical services; law enforcement; spouse or other family members of illicit users/former users of opioids; government officials; school officials; community members. Content analysis was utilized to identify themes and answer study questions. Regular feedback to stakeholders was provided to support targeted interventions. Results: We performed semi-structured, in-depth interviews with 130 community stakeholders and 29 community meeting observations in 2018-2019. Participants perceived similar economic and social determinants as origins of the opioid overdose epidemic including lack of economic resources, loss of jobs, transient populations and dilapidated housing. However, they differed in their awareness of and attention to the epidemic. Awareness was dependent on the visibility of opioid use (presence of paraphernalia litter, location of drug users, media coverage, and relationship to users). Overall, there was good knowledge of naloxone for opioid overdose reversal but less knowledge about local syringe services programs. Perceptions of harm reduction efforts were ambivalent. Conclusions: Members of communities impacted by the opioid epidemic perceived that economic downturn was a major factor in the opioid overdose epidemic. However, the varied beliefs within and between communities suggest that interventions need to be tailored according to the cultural norms of place.

AI-Assisted Summarization of Radiologic Reports: Evaluating GPT3davinci, BARTcnn, LongT5booksum, LEDbooksum, LEDlegal, and LEDclinical.

BACKGROUND AND PURPOSE: The review of clinical reports is an essential part of monitoring disease progression. Synthesizing multiple imaging reports is also important for clinical decisions. It is critical to aggregate information quickly and accurately. Machine learning natural language processing (NLP) models hold promise to address an unmet need for report summarization. MATERIALS AND METHODS: We evaluated NLP methods to summarize longitudinal aneurysm reports. A total of 137 clinical reports and 100 PubMed case reports were used in this study. Models were 1) compared against expert-generated summary using longitudinal imaging notes collected in our institute and 2) compared using publicly accessible PubMed case reports. Five AI models were used to summarize the clinical reports, and a sixth model, the online GPT3davinci NLP large language model (LLM), was added for the summarization of PubMed case reports. We assessed the summary quality through comparison with expert summaries using quantitative metrics and quality reviews by experts. RESULTS: In clinical summarization, BARTcnn had the best performance (BERTscore = 0.8371), followed by LongT5Booksum and LEDlegal. In the analysis using PubMed case reports, GPT3davinci demonstrated the best performance, followed by models BARTcnn and then LEDbooksum (BERTscore = 0.894, 0.872, and 0.867, respectively). CONCLUSIONS: AI NLP summarization models demonstrated great potential in summarizing longitudinal aneurysm reports, though none yet reached the level of quality for clinical usage. We found the online GPT LLM outperformed the others; however, the BARTcnn model is potentially more useful because it can be implemented on-site. Future work to improve summarization, address other types of neuroimaging reports, and develop structured reports may allow NLP models to ease clinical workflow.

Genome Sequences of Cluster G Mycobacteriophages Cambiare, FlagStaff, and MOOREtheMARYer.

Mycobacteriophages Cambiare, FlagStaff, and MOOREtheMARYer are newly isolated phages of Mycobacterium smegmatis mc(2) 155 recovered from soil samples in Pittsburgh, PA. All three genomes are closely related to cluster G mycobacteriophages but differ sufficiently in nucleotide sequence and gene content to warrant division of cluster G into several subclusters.