Ava Homiar

Excited to share my first-author paper in @BMJMentalHealth! We developed GPT-4o prompts for screening in a living systematic review, achieving 100% sensitivity for included studies & simulating a 65-85% reduction in workload. mentalhealth.bmj.com/content/28/1/e…

AI models developed for use in general populations are not uniformly transportable into oncology. Few rigorous studies of large language models have been done in cancer care. nature.com/articles/s4174…

We are excited to share our latest pre-print enabling effective human oversight for AMIE, our research diagnostic dialogue AI. We introduce a new asynchronous oversight paradigm, decoupling history-taking by AMIE from sharing a human-approved diagnosis – a thread 🧵:

Congratulations, Ava, on your first-author publication in @BMJMentalHealth 🎊 A significant contribution demonstrating the potential of GPT-4o to enhance screening efficiency in living systematic reviews 👏🏼 Impressive results and exciting implications for the future of evidence…

Healthcare leaders face an impossible choice: deploy AI quickly and risk patient safety, or move slowly and risk losing a competitive advantage. The DCI Network has developed an actionable framework. Read the paper: sciencedirect.com/science/articl… Highlights: linkedin.com/pulse/actionab…

Excited to share our new work on knowledge tracing approach to personalized education, with a real-world, prospective evaluation in @harvardmed AI in Medicine 2 course zitniklab.hms.harvard.edu/AIM2, led by Grey Kuling 🧠📚 arxiv.org/abs/2507.00032 What's new? Knowledge tracing blends…

Can machine learning help #psychiatrists predict outcomes in first-episode #psychosis? A new study @BMJMentalHealth finds #MLMs perform similarly to #clinicians but don’t consistently improve accuracy yet may help in tough cases. Integration challenges remain. Link:…

🕰️ What is Temporal Drift? Data that clinical prediction models face when implemented can differ from the data they were trained on, even in the same setting. This can be due to changing patients, treatments or practices The result? The model’s predictions start to go wrong.

NEW Umbrella Review: Digital health interventions for mental health disorders: an umbrella review of meta-analyses of randomised controlled trials. @CristinaCrocamo @Unimib_Psychiat @giuscarra @JohnTorousMD @fbartoli_unimib Read it here: thelancet.com/journals/landi…

Check out my latest reflection on our DCI Network conference: Reflecting on Patient-Powered Progress and Looking Toward Collaborative AI Research linkedin.com/pulse/reflecti…

Nice new review of generative AI in mental health in @JMIR_JMH with an impressive number of studies reviewed (n=79) and breakdown of use cases and quality: free @ mental.jmir.org/2025/1/e70610

Why are so many people with psychosis missed before their first episode? We studied over 1,500 patients to find out whether those missed by early intervention services show different early signs. Here’s what we found. 🧵👇

This study developed and evaluated an autonomous clinical artificial intelligence agent leveraging GPT-4 and multimodal precision oncology tools to support personalized cancer decision-making nature.com/articles/s4301…

Remarkable experiments by @mlpowered @wesg52 et al. @anthropic suggesting that LLMs have quasi-symbolic internal representations and multiply redundant reasoning abstraactions/paths I packaged their manuscript “On the Biology of a Large Language Model” into a 40 minute podcast…

🧠📱 Check out our new open-access paper: “A Survey of Preferences for Sensing Technologies in People with Severe Mental Illness in Acute Psychiatric Care” published in #AnnalsofBiomedicalEngineering Read it at link.springer.com/article/10.100…

How do people recover from persecutory delusions? Hear from patients on this topic. Delighted to share this latest paper, published in the journal Psychosis.tandfonline.com/doi/full/10.10… With @shioma_lei @ThomasKabir @ProfDFreeman @FelicityWaite #psychosis #IPA #paranoia #psychology

Urgent Need for Standards and Safeguards for Health-Related Generative Artificial Intelligence doi.org/10.7326/ANNALS…

📢 New AI for rare disease diagnosis: SHEPHERD shows how simulation + knowledge-grounded AI = deep learning for ultra‑low label domains nature.com/articles/s4174… This may shorten the journey from phenotype to diagnosis. Excited to help clinicians leverage fewer labels + more…

🧠📊 Survey of 700+ mental health researchers: ~70% use LLMs (mostly ChatGPT), mainly for proofreading & coding. Top concerns? Accuracy, ethics & bias. Most want clearer guidelines & disclosure norms. #MentalHealthResearch #AI #LLMs #ChatGPT @SGoldbergPhD @JohnTorousMD To…

International Guidelines for the Algorithmic Treatment of Schizophrenia (INTEGRATE) - The Lancet Psychiatry Robert McCutcheon and others led a fantastic effort in creating guidelines for schizophrenia treatment based on an algorithmic logic of decision-making. The guidelines…