Aspect-Oriented Summarization for Psychiatric Short-Term Readmission Prediction

Recent progress in large language models (LLMs) has enabled the automated processing of lengthy documents even without supervised training on a task-specific dataset. Yet, their zero-shot performance in complex tasks as opposed to straightforward information extraction tasks remains suboptimal. One feasible approach for tasks with lengthy, complex input is to first summarize the document and then apply supervised fine-tuning to the summary. However, the summarization process inevitably results in some loss of information. In this study we present a method for processing the summaries of long documents aimed to capture different important aspects of the original document. We hypothesize that LLM summaries generated with different aspect-oriented prompts contain different information signals, and we propose methods to measure these differences. We introduce approaches to effectively integrate signals from these different summaries for supervised training of transformer models. We validate our hypotheses on a high-impact task -30-day readmission prediction from a psychiatric discharge -using real-world data from four hospitals, and show that our proposed method increases the prediction performance for the complex task of predicting patient outcome. * 12/31/19**: Born * Unknown date: Diagnosed with Schizoaffective disorder, bipolar type * Unknown date: Diagnosed with borderline personality disorder * Unknown date: Diagnosed with polysubstance disorder, including alcohol and cannabis use * Unknown date: Multiple prior hospitalizations * Unknown date: History of self-injury by cutting, no known suicide attempts * 9/12/2030: Admitted to psychiatric hospital for evaluation and treatment of confusion/disorganization in the context of worsening depression and alcohol/cannabis use * 9/28/2030: Discharged from psychiatric hospital, diagnosed with Schizoaffective disorder, bipolar type, and prescribed lithium carbonate *** mg capsule nightly at bedtime.