Personalized treatment options for chronic diseases using precision cohort analytics

Abstract

To support point-of-care decision making by presenting outcomes of past treatment choices for cohorts of similar patients based on observational data from electronic health records (EHRs), a machine-learning precision cohort treatment option (PCTO) workflow consisting of (1) data extraction, (2) similarity model training, (3) precision cohort identification, and (4) treatment options analysis was developed. The similarity model is used to dynamically create a cohort of similar patients, to inform clinical decisions about an individual patient. The workflow was implemented using EHR data from a large health care provider for three different highly prevalent chronic diseases: hypertension (HTN), type 2 diabetes mellitus (T2DM), and hyperlipidemia (HL). A retrospective analysis demonstrated that treatment options with better outcomes were available for a majority of cases (75%, 74%, 85% for HTN, T2DM, HL, respectively). The models for HTN and T2DM were deployed in a pilot study with primary care physicians using it during clinic visits. A novel data-analytic workflow was developed to create patient-similarity models that dynamically generate personalized treatment insights at the point-of-care. By leveraging both knowledge-driven treatment guidelines and data-driven EHR data, physicians can incorporate real-world evidence in their medical decision-making process when considering treatment options for individual patients.

Prognostication for prelabor rupture of membranes and the time of delivery in nationwide insured women: development, validation, and deployment

“…A clinician can use population-level data case-by-case, assisted by machine. We did not only rely on the DI-VNN and PC-RF models but also datasets to estimate the model performances at individual level based on subpopulation similar to that individual, 56 collaborating machine learning and evidence-based medicine. 57…”

“…24 Our models also used a cohort paradigm to prevent temporal bias in the delivery prediction. 45 We did not only rely on the DI-VNN and PC-RF models but also datasets to estimate the model performances at the individual level based on subpopulations similar to that individual, 46 in a collaboration of machine learning and evidence-based medicine. 47 This framework also answers key challenges to evaluate model weaknesses, 48 verify if the predicted outcome is reasonable, 49 and make sensible clinical predictions by utilizing electronic medical records.…”

Diabetes medication recommendation system using patient similarity analytics

“…This study extends existing works by considering patients’ HbA 1c trajectories in addition to clinical profile when identifying similar patients. In contrast to studies 20 , 21 that provide recommendations for medications type only, our study provides both medication type and dosage recommendations. Integrating dosage recommendation increases the complexity of medication recommendation system as titration of dosages need to take into consideration the age of a patient, their clinical profile, and/or other interacting medications 31 .…”

Construction of Cohorts of Similar Patients From Automatic Extraction of Medical Concepts: Phenotype Extraction Study

“…The similarity of diseases is based on their respective ICD-10 similarity (using the ICD-10 coding tree structure). Ng et al [34] presented an insightful method based on a precision cohort (ie, patient-similarity cohorts) to help clinicians make treatment decisions for chronic diseases. They trained a global similarity model on a set of thousands of predefined variables (disease variables were constructed using their ICD-9 and ICD-10 codes, laboratory variables with their Logical Observation Identifiers Names and Codes, etc) that learns a disease-specific distance (for the 3 chronic diseases presented: hypertension, type 2 diabetes mellitus, and hyperlipidemia), with significant manual work to build the training data set.…”