Dataset Information

Examining multimorbidity differences across racial groups: a network analysis of electronic medical records.

ABSTRACT: Health disparities across ethnic or racial groups are typically examined through single behavior at a time. The syndemics and multimorbidity health disparities have not been well examined by race. In this study, we study health disparities by identifying the networks of multimorbidities among individuals from seven population groups based on race, including White, African American, Asian, Hispanic, Native American, Bi- or Multi-racial and Pacific Islander. We examined a large electronic medical record (EMR) containing health records of more than 18.7 million patients and created multimorbidity networks considering their lifetime history from medical records in order to compare the network properties among seven population groups. In addition, the networks at organ system level depicting the relationship among disorders belonging to different organ systems are also compared. Our macro analysis at the organ-level indicates that African-Americans have a stronger multimorbidity network followed by Whites and Native Americans. The networks of Asians and Hispanics are sparse. Specifically, the relationship of infectious and parasitic disorders with respiratory, circulatory and genitourinary system disorders is stronger among African Americans than others. On the other hand, the relationship of mental disorders with respiratory, musculoskeletal system and connective tissue disorders is more prevalent in Whites. Similar other disparities are discussed. Recognition and explanation of such differences in multimorbidities inform the public health policies, and can inform clinical decisions as well. Our multimorbidity network analysis identifies specific differences in diagnoses among different population groups, and presents questions for biological, behavioral, clinical, social science, and policy research.

SUBMITTER: Kalgotra P

PROVIDER: S-EPMC7419498 | biostudies-literature | 2020 Aug

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Examining multimorbidity differences across racial groups: a network analysis of electronic medical records.

Kalgotra Pankush P Sharda Ramesh R Croff Julie M JM

Scientific reports 20200811 1

Health disparities across ethnic or racial groups are typically examined through single behavior at a time. The syndemics and multimorbidity health disparities have not been well examined by race. In this study, we study health disparities by identifying the networks of multimorbidities among individuals from seven population groups based on race, including White, African American, Asian, Hispanic, Native American, Bi- or Multi-racial and Pacific Islander. We examined a large electronic medical ...[more]

PMID: 32782346

Similar Datasets

Project description:Clinically recorded pain scores are abundant in patient health records but are rarely used in research. The use of this information could help improve clinical outcomes. For example, a recent report by the Institute of Medicine stated that ineffective use of clinical information contributes to undertreatment of patient subpopulations--especially women. This study used diagnosis-associated pain scores from a large hospital database to document sex differences in reported pain. We used de-identified electronic medical records from Stanford Hospital and Clinics for more than 72,000 patients. Each record contained at least 1 disease-associated pain score. We found over 160,000 pain scores in more than 250 primary diagnoses, and analyzed differences in disease-specific pain reported by men and women. After filtering for diagnoses with minimum encounter numbers, we found diagnosis-specific sex differences in reported pain. The most significant differences occurred in patients with disorders of the musculoskeletal, circulatory, respiratory and digestive systems, followed by infectious diseases, and injury and poisoning. We also discovered sex-specific differences in pain intensity in previously unreported diseases, including disorders of the cervical region, and acute sinusitis (P = .01, .017, respectively). Pain scores were collected during hospital encounters. No information about the use of pre-encounter over-the-counter medications was available. To our knowledge, this is the largest data-driven study documenting sex differences of disease-associated pain. It highlights the utility of electronic medical record data to corroborate and expand on results of smaller clinical studies. Our findings emphasize the need for future research examining the mechanisms underlying differences in pain.This article highlights the potential of electronic medical records to conduct large-scale pain studies. Our results are consistent with previous studies reporting pain differences between sexes and also suggest that clinicians should pay increased attention to this idea.

Project description:ObjectiveAttention networks learn an intelligent weighted averaging mechanism over a series of entities, providing increases to both performance and interpretability. In this article, we propose a novel time-aware transformer-based network and compare it to another leading model with similar characteristics. We also decompose model performance along several critical axes and examine which features contribute most to our model's performance.Materials and methodsUsing data sets representing patient records obtained between 2017 and 2019 by the Kaiser Permanente Mid-Atlantic States medical system, we construct four attentional models with varying levels of complexity on two targets (patient mortality and hospitalization). We examine how incorporating transfer learning and demographic features contribute to model success. We also test the performance of a model proposed in recent medical modeling literature. We compare these models with out-of-sample data using the area under the receiver-operator characteristic (AUROC) curve and average precision as measures of performance. We also analyze the attentional weights assigned by these models to patient diagnoses.ResultsWe found that our model significantly outperformed the alternative on a mortality prediction task (91.96% AUROC against 73.82% AUROC). Our model also outperformed on the hospitalization task, although the models were significantly more competitive in that space (82.41% AUROC against 80.33% AUROC). Furthermore, we found that demographic features and transfer learning features which are frequently omitted from new models proposed in the EMR modeling space contributed significantly to the success of our model.DiscussionWe proposed an original construction of deep learning electronic medical record models which achieved very strong performance. We found that our unique model construction outperformed on several tasks in comparison to a leading literature alternative, even when input data was held constant between them. We obtained further improvements by incorporating several methods that are frequently overlooked in new model proposals, suggesting that it will be useful to explore these options further in the future.

Project description:Post-market medical device surveillance is a challenge facing manufacturers, regulatory agencies, and health care providers. Electronic health records are valuable sources of real-world evidence for assessing device safety and tracking device-related patient outcomes over time. However, distilling this evidence remains challenging, as information is fractured across clinical notes and structured records. Modern machine learning methods for machine reading promise to unlock increasingly complex information from text, but face barriers due to their reliance on large and expensive hand-labeled training sets. To address these challenges, we developed and validated state-of-the-art deep learning methods that identify patient outcomes from clinical notes without requiring hand-labeled training data. Using hip replacements-one of the most common implantable devices-as a test case, our methods accurately extracted implant details and reports of complications and pain from electronic health records with up to 96.3% precision, 98.5% recall, and 97.4% F1, improved classification performance by 12.8-53.9% over rule-based methods, and detected over six times as many complication events compared to using structured data alone. Using these additional events to assess complication-free survivorship of different implant systems, we found significant variation between implants, including for risk of revision surgery, which could not be detected using coded data alone. Patients with revision surgeries had more hip pain mentions in the post-hip replacement, pre-revision period compared to patients with no evidence of revision surgery (mean hip pain mentions 4.97 vs. 3.23; t?=?5.14; p?<?0.001). Some implant models were associated with higher or lower rates of hip pain mentions. Our methods complement existing surveillance mechanisms by requiring orders of magnitude less hand-labeled training data, offering a scalable solution for national medical device surveillance using electronic health records.

Dataset Information

Examining multimorbidity differences across racial groups: a network analysis of electronic medical records.

Publications

Examining multimorbidity differences across racial groups: a network analysis of electronic medical records.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets