Manual and automatic analysis of patients values and preferences using Seton HCAHPS surveys

Understanding the values and preferences of patients is a key to patient-centered care. This study aims to identify these values and preferences by analyzing the data provided by The Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS). The data is presented as a free-form survey of patients of a variety of hospitals in the United States. The survey was conducted via telephone and concluded with free-form comments that provide a unique way to understand positive and negative sides of health care services. We use topic modeling to automatically identify the crucial aspects of patient-centered care as well as obtain the most high- and low-rated traits of the healthcare system. In particular, we use Latent Dirichlet Allocation (LDA) and Latent Semantic Analysis (LSA) to identify the extent to which the topics correspond to values and preferences. In this study we present (1) an approach to manual and automatic evaluation of patients values and preferences and (2) a list of the aspects of the modern healthcare system that are crucial to its patients. This list can be used by hospitals in order to improve the quality of services they provide. Our findings have important implications for patient satisfaction, patient-provider relationships, and, ultimately, health outcomes.