Bmc Med Inform Decis
-
Bmc Med Inform Decis · Apr 2020
Observational StudyPublicly available machine learning models for identifying opioid misuse from the clinical notes of hospitalized patients.
Automated de-identification methods for removing protected health information (PHI) from the source notes of the electronic health record (EHR) rely on building systems to recognize mentions of PHI in text, but they remain inadequate at ensuring perfect PHI removal. As an alternative to relying on de-identification systems, we propose the following solutions: (1) Mapping the corpus of documents to standardized medical vocabulary (concept unique identifier [CUI] codes mapped from the Unified Medical Language System) thus eliminating PHI as inputs to a machine learning model; and (2) training character-based machine learning models that obviate the need for a dictionary containing input words/n-grams. We aim to test the performance of models with and without PHI in a use-case for an opioid misuse classifier. ⋯ We demonstrate good test characteristics for an opioid misuse computable phenotype that is void of any PHI and performs similarly to models that use PHI. Herein we share a PHI-free, trained opioid misuse classifier for other researchers and health systems to use and benchmark to overcome privacy and security concerns.
-
Bmc Med Inform Decis · Apr 2020
Non-health outcomes affecting self-care behaviors and medical decision-making preference in patients with type 2 diabetes: a cross-sectional study.
The effects of patient sustained self-care behaviors on glycemic control are even greater than the effects of medical treatment, indicating the value of identifying the factors that influence self-care behaviors. To date, these factors have not been placed in a single model to clarify the critical path affecting self-care behaviors. The aims of this study were to explore the relationships of these factors and the differences in patient preference for medical decision-making. ⋯ Health literacy is a critical factor in improving self-care behaviors in patients with type 2 diabetes, and the effect of health literacy on self-efficacy was more significant in the shared decision-making than in the physician decision-making. Therefore, developing an effective health strategy to strengthen health literacy awareness and designing friendly, diverse health literacy materials, and application tools is the most important factor to facilitate self-care behaviors in this population.