Following an interaction with a patient, physicians are responsible for the submission of clinical documentation, often organized as a SOAP note. A clinical note is not simply a summary of the conversation but requires the use of appropriate medical terminology. The relevant information can then be extracted and organized according to the structure of the SOAP note. In this paper we analyze two different approaches to generate the different sections of a SOAP note based on the audio recording of the conversation, and specifically examine them in terms of note consistency. The first approach generates the sections independently, while the second method generates them all together. In this work we make use of PEGASUS-X Transformer models and observe that both methods lead to similar ROUGE values (less than 1% difference) and have no difference in terms of the Factuality metric. We perform a human evaluation to measure aspects of consistency and demonstrate that LLMs like Llama2 can be used to perform the same tasks with roughly the same agreement as the human annotators. Between the Llama2 analysis and the human reviewers we observe a Cohen Kappa inter-rater reliability of 0.79, 1.00, and 0.32 for consistency of age, gender, and body part injury, respectively. With this we demonstrate the usefulness of leveraging an LLM to measure quality indicators that can be identified by humans but are not currently captured by automatic metrics. This allows scaling evaluation to larger data sets, and we find that clinical note consistency improves by generating each new section conditioned on the output of all previously generated sections.

The scarcity of public datasets for the summarization of medical conversations has been a limiting factor for advancing NLP research in the healthcare domain, and the structure of the existing data is largely limited to the simple format of conversation-summary pairs. We therefore propose a novel Incremental Note Generation (ING) annotation framework capable of greatly enriching summarization datasets in the healthcare domain and beyond. Our framework is designed to capture the human summarization process via an annotation task by instructing the annotators to first incrementally create a draft note as they accumulate information through a conversation transcript (Generation) and then polish the draft note into a reference note (Rewriting). The annotation results include both the reference note and a comprehensive editing history of the draft note in tabular format. Our pilot study on the task of SOAP note generation showed reasonable consistency between four expert annotators, established a solid baseline for quantitative targets of inter-rater agreement, and demonstrated the ING framework as an improvement over the traditional annotation process for future modeling of summarization.

Summarization of doctor-patient conversations into clinical notes by medical scribes is an essential process for effective clinical care. Pre-trained transformer models have shown a great amount of success in this area, but the domain shift from standard NLP tasks to the medical domain continues to present challenges. We build upon several recent works to show that additional pre-training with in-domain medical conversations leads to performance gains for clinical summarization. In addition to conventional evaluation metrics, we also explore a clinical named entity recognition model for concept-based evaluation. Finally, we contrast long-sequence transformers with a common transformer model, BART. Overall, our findings corroborate research in non-medical domains and suggest that in-domain pre-training combined with transformers for long sequences are effective strategies for summarizing clinical encounters.

Scoring the factuality of a generated summary involves measuring the degree to which a target text contains factual information using the input document as support. Given the similarities in the problem formulation, previous work has shown that Natural Language Inference models can be effectively repurposed to perform this task. As these models are trained to score entailment at a sentence level, several recent studies have shown that decomposing either the input document or the summary into sentences helps with factuality scoring. But is fine-grained decomposition always a winning strategy? In this paper we systematically compare different granularities of decomposition - from document to sub-sentence level, and we show that the answer is no. Our results show that incorporating additional context can yield improvement, but that this does not necessarily apply to all datasets. We also show that small changes to previously proposed entailment-based scoring methods can result in better performance, highlighting the need for caution in model and methodology selection for downstream tasks.

Fine-tuning pretrained models for automatically summarizing doctor-patient conversation transcripts presents many challenges: limited training data, significant domain shift, long and noisy transcripts, and high target summary variability. In this paper, we explore the feasibility of using pretrained transformer models for automatically summarizing doctor-patient conversations directly from transcripts. We show that fluent and adequate summaries can be generated with limited training data by fine-tuning BART on a specially constructed dataset. The resulting models greatly surpass the performance of an average human annotator and the quality of previous published work for the task. We evaluate multiple methods for handling long conversations, comparing them to the obvious baseline of truncating the conversation to fit the pretrained model length limit. We introduce a multistage approach that tackles the task by learning two fine-tuned models: one for summarizing conversation chunks into partial summaries, followed by one for rewriting the collection of partial summaries into a complete summary. Using a carefully chosen fine-tuning dataset, this method is shown to be effective at handling longer conversations, improving the quality of generated summaries. We conduct both an automatic evaluation (through ROUGE and two concept-based metrics focusing on medical findings) and a human evaluation (through qualitative examples from literature, assessing hallucination, generalization, fluency, and general quality of the generated summaries).

Multi-label document classification (MLDC) problems can be challenging, especially for long documents with a large label set and a long-tail distribution over labels. In this paper, we present an effective convolutional attention network for the MLDC problem with a focus on medical code prediction from clinical documents. Our innovations are three-fold: (1) we utilize a deep convolution-based encoder with the squeeze-and-excitation networks and residual networks to aggregate the information across the document and learn meaningful document representations that cover different ranges of texts; (2) we explore multi-layer and sum-pooling attention to extract the most informative features from these multi-scale representations; (3) we combine binary cross entropy loss and focal loss to improve performance for rare labels. We focus our evaluation study on MIMIC-III, a widely used dataset in the medical domain. Our models outperform prior work on medical coding and achieve new state-of-the-art results on multiple metrics. We also demonstrate the language independent nature of our approach by applying it to two non-English datasets. Our model outperforms prior best model and a multilingual Transformer model by a substantial margin.

Most classification models work by first predicting a posterior probability distribution over all classes and then selecting that class with the largest estimated probability. In many settings however, the quality of posterior probability itself (e.g., 65% chance having diabetes), gives more reliable information than the final predicted class alone. When these methods are shown to be poorly calibrated, most fixes to date have relied on posterior calibration, which rescales the predicted probabilities but often has little impact on final classifications. Here we propose an end-to-end training procedure called posterior calibrated (PosCal) training that directly optimizes the objective while minimizing the difference between the predicted and empirical posterior probabilities. We show that PosCal not only helps reduce the calibration error but also improve task performance by penalizing drops in performance of both objectives. Our PosCal achieves about 2.5% of task performance gain and 16.1% of calibration error reduction on GLUE (Wang et al., 2018) compared to the baseline. We achieved the comparable task performance with 13.2% calibration error reduction on xSLUE (Kang and Hovy, 2019), but not outperforming the two-stage calibration baseline. PosCal training can be easily extendable to any types of classification tasks as a form of regularization term. Also, PosCal has the advantage that it incrementally tracks needed statistics for the calibration objective during the training process, making efficient use of large training sets.