PVminer: A Domain-Specific Tool to Detect the Patient Voice in Patient Generated Data

TL;DR

PVminer is a domain-specific NLP framework that effectively detects and structures the patient voice in patient-generated data, leveraging BERT models and topic modeling to improve performance over existing methods.

Contribution

The paper introduces PVminer, a novel NLP framework that combines patient-specific BERT encoders and topic modeling for multi-label detection of patient voice in healthcare communication.

Findings

PVminer outperforms baseline models with F1 scores over 80%.

Incorporating author identity and topic augmentation improves detection accuracy.

Pre-trained models and datasets will be publicly released.

Abstract

Patient-generated text such as secure messages, surveys, and interviews contains rich expressions of the patient voice (PV), reflecting communicative behaviors and social determinants of health (SDoH). Traditional qualitative coding frameworks are labor intensive and do not scale to large volumes of patient-authored messages across health systems. Existing machine learning (ML) and natural language processing (NLP) approaches provide partial solutions but often treat patient-centered communication (PCC) and SDoH as separate tasks or rely on models not well suited to patient-facing language. We introduce PVminer, a domain-adapted NLP framework for structuring patient voice in secure patient-provider communication. PVminer formulates PV detection as a multi-label, multi-class prediction task integrating patient-specific BERT encoders (PV-BERT-base and PV-BERT-large), unsupervised topic modeling for thematic augmentation (PV-Topic-BERT), and fine-tuned classifiers for Code, Subcode, and Combo-level labels. Topic representations are incorporated during fine-tuning and inference to enrich semantic inputs. PVminer achieves strong performance across hierarchical tasks and outperforms biomedical and clinical pre-trained baselines, achieving F1 scores of 82.25% (Code), 80.14% (Subcode), and up to 77.87% (Combo). An ablation study further shows that author identity and topic-based augmentation each contribute meaningful gains. Pre-trained models, source code, and documentation will be publicly released, with annotated datasets available upon request for research use.