Machine-learning based prediction of hypo- and hyperglycemia from electronic health records (Preprint)

2022 ◽  
Author(s):  
Harald Witte ◽  
Christos Theodoros Nakas ◽  
Lia Bally ◽  
Alexander Benedikt Leichtle
Keyword(s):  
Electronic Health Records ◽  
Prediction Model ◽  
Hospital Admissions ◽  
Laboratory Data ◽  
Significant Risk ◽  
Routine Data ◽  
Therapeutic Interventions ◽  
Classification Model ◽  
Health Records ◽  
Electronic Health

BACKGROUND The increasing need for blood glucose (BG) management in hospitalized patients poses high demands on clinical staff and health care systems alike. Acute decompensations of BG levels (hypo- and hyperglycemia) adversely affect patient outcomes and safety. OBJECTIVE Acute BG decompensations pose a frequent and significant risk for inpatients. Ideally, proactive measures are taken before BG levels derail. We have generated a broadly applicable multiclass classification model for predicting decompensation events from patients’ electronic health records to indicate where adjustments of patient monitoring and/or therapeutic interventions are required. METHODS A retrospective cohort study was conducted of patients hospitalized at a tertiary hospital in Bern, Switzerland. Using patient details and routine data from electronic health records (EHRs), a multiclass prediction model for BG decompensation events (< 3.9 mmol/L (hypoglycemia), or > 10, > 13.9, or > 16.7 mmol/L (representing different degrees of hyperglycemia)) was generated, based on a second-level ensemble of gradient-boosted binary trees. RESULTS 63’579 hospital admissions of 33’212 patients were included in this study. The multiclass prediction model reached a specificity of 93.0%, 98.5%, and 93.6% and a sensitivity of 69.6%, 63.0%, and 65.5%, for the main categories of interest. i.e., non-decompensated cases, hypo- or hyperglycemia, respectively. The median prediction horizon was seven and four hours for hypo- and hyperglycemia, respectively. CONCLUSIONS EHRs hold the potential to reliably predict all kinds of BG decompensations. Readily available patient details and routine laboratory data can support the decisions for proactive interventions and thus help to reduce the detrimental health effects of hypo- and hyperglycemia.

scholarly journals Development and validation of techniques for phenotyping ST-elevation myocardial infarction encounters from electronic health records

JAMIA Open ◽  
2021 ◽  
Vol 4 (3) ◽  
Author(s):  
Sulaiman Somani ◽  
Stephen Yoffie ◽  
Shelly Teng ◽  
Shreyas Havaldar ◽  
Girish N Nadkarni ◽  
...  
Keyword(s):  
Myocardial Infarction ◽  
Electronic Health Records ◽  
Hospital Admissions ◽  
Coronary Intervention ◽  
St Elevation Myocardial Infarction ◽  
Health Records ◽  
St Elevation ◽  
Electronic Health ◽  
Cardiac Catheterizations ◽  
Discharge Summaries

Abstract Objectives Classifying hospital admissions into various acute myocardial infarction phenotypes in electronic health records (EHRs) is a challenging task with strong research implications that remains unsolved. To our knowledge, this study is the first study to design and validate phenotyping algorithms using cardiac catheterizations to identify not only patients with a ST-elevation myocardial infarction (STEMI), but the specific encounter when it occurred. Materials and Methods We design and validate multi-modal algorithms to phenotype STEMI on a multicenter EHR containing 5.1 million patients and 115 million patient encounters by using discharge summaries, diagnosis codes, electrocardiography readings, and the presence of cardiac catheterizations on the encounter. Results We demonstrate that robustly phenotyping STEMIs by selecting discharge summaries containing “STEM” has the potential to capture the most number of STEMIs (positive predictive value [PPV] = 0.36, N = 2110), but that addition of a STEMI-related International Classification of Disease (ICD) code and cardiac catheterizations to these summaries yields the highest precision (PPV = 0.94, N = 952). Discussion and Conclusion In this study, we demonstrate that the incorporation of percutaneous coronary intervention increases the PPV for detecting STEMI-related patient encounters from the EHR.

scholarly journals Electronic health records in ambulances: the ERA multiple-methods study

2020 ◽  
Vol 8 (10) ◽  
pp. 1-140
Author(s):  
Alison Porter ◽  
Anisha Badshah ◽  
Sarah Black ◽  
David Fitzpatrick ◽  
Robert Harris-Mayes ◽  
...  
Keyword(s):  
Emergency Department ◽  
Health Services ◽  
Electronic Health Records ◽  
Case Studies ◽  
Routine Data ◽  
Health Economy ◽  
Multiple Methods ◽  
Health Records ◽  
Electronic Health ◽  
Future Work

Background Ambulance services have a vital role in the shift towards the delivery of health care outside hospitals, when this is better for patients, by offering alternatives to transfer to the emergency department. The introduction of information technology in ambulance services to electronically capture, interpret, store and transfer patient data can support out-of-hospital care. Objective We aimed to understand how electronic health records can be most effectively implemented in a pre-hospital context in order to support a safe and effective shift from acute to community-based care, and how their potential benefits can be maximised. Design and setting We carried out a study using multiple methods and with four work packages: (1) a rapid literature review; (2) a telephone survey of all 13 freestanding UK ambulance services; (3) detailed case studies examining electronic health record use through qualitative methods and analysis of routine data in four selected sites consisting of UK ambulance services and their associated health economies; and (4) a knowledge-sharing workshop. Results We found limited literature on electronic health records. Only half of the UK ambulance services had electronic health records in use at the time of data collection, with considerable variation in hardware and software and some reversion to use of paper records as services transitioned between systems. The case studies found that the ambulance services’ electronic health records were in a state of change. Not all patient contacts resulted in the generation of electronic health records. Ambulance clinicians were dealing with partial or unclear information, which may not fit comfortably with the electronic health records. Ambulance clinicians continued to use indirect data input approaches (such as first writing on a glove) even when using electronic health records. The primary function of electronic health records in all services seemed to be as a store for patient data. There was, as yet, limited evidence of electronic health records’ full potential being realised to transfer information, support decision-making or change patient care. Limitations Limitations included the difficulty of obtaining sets of matching routine data for analysis, difficulties of attributing any change in practice to electronic health records within a complex system and the rapidly changing environment, which means that some of our observations may no longer reflect reality. Conclusions Realising all the benefits of electronic health records requires engagement with other parts of the local health economy and dealing with variations between providers and the challenges of interoperability. Clinicians and data managers, and those working in different parts of the health economy, are likely to want very different things from a data set and need to be presented with only the information that they need. Future work There is scope for future work analysing ambulance service routine data sets, qualitative work to examine transfer of information at the emergency department and patients’ perspectives on record-keeping, and to develop and evaluate feedback to clinicians based on patient records. Study registration This study is registered as Health and Care Research Wales Clinical Research Portfolio 34166. Funding This project was funded by the National Institute for Health Research (NIHR) Health Services and Delivery Research programme and will be published in full in Health Services and Delivery Research; Vol. 8, No. 10. See the NIHR Journals Library website for further project information.

scholarly journals Prediction of Postoperative Infection for Patients Undergoing Gastrointestinal Surgery: Findings from Electronic Health Records

2020 ◽  
Vol 11 (2) ◽  
pp. 36-46
Author(s):  
Kyoko Nakazawa ◽  
Takashi Ishikawa ◽  
Akira Toyama ◽  
Toshifumi Wakai ◽  
Kohei Akazawa
Keyword(s):  
Risk Factors ◽  
Electronic Health Records ◽  
Surgical Procedures ◽  
Postoperative Infection ◽  
Gastrointestinal Surgery ◽  
Significant Risk ◽  
Estimating Equation ◽  
Endoscopic Surgical Procedures ◽  
Health Records ◽  
Electronic Health

Introduction: Postoperative infection is a major cause of morbidity and prolonged hospitalization in patients undergoing gastrointestinal surgery. This observational study aimed to investigate the risk factors associated with postoperative infection and to develop a prediction model for postoperative infections that occur after gastrointestinal surgery. Methods: The study population comprised 1637 patients who underwent gastrointestinal surgery at Niigata University Medical and Dental Hospital between June 2013 and May 2017. Observational data from 1883 surgical procedures were used in the statistical analyses (including 198 patients who underwent several operations). Results: The generalized estimating equation (GEE) was used to detect significant risk factors, including older age, history of smoking, body temperature greater than 38 °C, non-endoscopic surgical procedures, surgery in the thoracic or lower gastrointestinal tract, and use of medical nutritional products during surgery. The sensitivity and specificity of the GEE model were 88.2% and 55.1%, respectively. Conclusion: This study established a predictable GEE model, incorporating the data of patients who were hospitalized several times into a prediction analysis, even though the sensitivity was not sufficiently high. The GEE model, which is considered clinically useful, can be constructed using a variety of variables, including those obtained from electronic health records.

scholarly journals Comparison of physical examination and laboratory data between a clinical study and electronic health records

PLoS ONE ◽  
2020 ◽  
Vol 15 (7) ◽  
pp. e0236189
Author(s):  
Yi-An Ko ◽  
Yingtian Hu ◽  
Arshed A. Quyyumi ◽  
Lance A. Waller ◽  
Eberhard O. Voit ◽  
...  
Keyword(s):  
Clinical Study ◽  
Electronic Health Records ◽  
Physical Examination ◽  
Laboratory Data ◽  
Health Records ◽  
Electronic Health

Concept Attribute Labeling and Context-Aware Named Entity Recognition in Electronic Health Records

2018 ◽  
Vol 7 (1) ◽  
pp. 1-15 ◽  
Author(s):  
Alexandra Pomares-Quimbaya ◽  
Rafael A. Gonzalez ◽  
Oscar Mauricio Muñoz Velandia ◽  
Angel Alberto Garcia Peña ◽  
Julián Camilo Daza Rodríguez ◽  
...  
Keyword(s):  
Electronic Health Records ◽  
Ad Hoc ◽  
Named Entity Recognition ◽  
Ensemble Classification ◽  
Entity Recognition ◽  
Classification Model ◽  
Health Records ◽  
Named Entity ◽  
Electronic Health ◽  
Concept Attribute

Extracting valuable knowledge from Electronic Health Records (EHR) represents a challenging task due to the presence of both structured and unstructured data, including codified fields, images and test results. Narrative text in particular contains a variety of notes which are diverse in language and detail, as well as being full of ad hoc terminology, including acronyms and jargon, which is especially challenging in non-English EHR, where there is a dearth of annotated corpora or trained case sets. This paper proposes an approach for NER and concept attribute labeling for EHR that takes into consideration the contextual words around the entity of interest to determine its sense. The approach proposes a composition method of three different NER methods, together with the analysis of the context (neighboring words) using an ensemble classification model. This contributes to disambiguate NER, as well as labeling the concept as confirmed, negated, speculative, pending or antecedent. Results show an improvement of the recall and a limited impact on precision for the NER process.

The factors influencing community nurses' usage of electronic health records: findings from focus groups

2021 ◽  
Vol 26 (12) ◽  
pp. 604-610
Author(s):  
Ruth Lezard ◽  
Toity Deave
Keyword(s):  
Electronic Health Records ◽  
Focus Groups ◽  
Hospital Admissions ◽  
Specific Training ◽  
Health Records ◽  
Mobile Access ◽  
Factors Influencing ◽  
The Social ◽  
Electronic Health ◽  
Mobile Working

Electronic health records (EHRs) are integral to community nursing, and mobile access aids seamless, responsive care, prevents repetition and reduces hospital admissions. This saves time and money, enabling smoother workflows and increased productivity. Common practice among community nurses is to return to workbases to access EHRs. This research was conducted to explore what leads to inconsistency in EHR use. Focus groups were held with community nurses, and reflexive thematic analysis of the data was undertaken. Nurses who used EHRs during consultations described the practice as integrative and informed, promoting collaborative care. Those who did not described EHRs as time-consuming, template-laden and a barrier to nurse-patient communication. One barrier to mobile working is the threat to collegiate teamworking and the social and clinical supports it provides. This study suggests specific strategies could increase mobile EHR engagement: role-specific training for effective EHR use; clear organisational directives; innovative team communication; and peer-to-peer coaching.

scholarly journals Graphical display of diagnostic test results in electronic health Records: a comparison of 8 systems

2015 ◽  
Vol 22 (4) ◽  
pp. 900-904 ◽  
Author(s):  
Dean F Sittig ◽  
Daniel R Murphy ◽  
Michael W Smith ◽  
Elise Russo ◽  
Adam Wright ◽  
...  
Keyword(s):  
Electronic Health Records ◽  
Laboratory Test ◽  
Clinical Laboratory ◽  
Negative Impact ◽  
Laboratory Data ◽  
Test Results ◽  
Health Records ◽  
Laboratory Test Results ◽  
Data Points ◽  
Electronic Health

Abstract Accurate display and interpretation of clinical laboratory test results is essential for safe and effective diagnosis and treatment. In an attempt to ascertain how well current electronic health records (EHRs) facilitated these processes, we evaluated the graphical displays of laboratory test results in eight EHRs using objective criteria for optimal graphs based on literature and expert opinion. None of the EHRs met all 11 criteria; the magnitude of deficiency ranged from one EHR meeting 10 of 11 criteria to three EHRs meeting only 5 of 11 criteria. One criterion (i.e., the EHR has a graph with y-axis labels that display both the name of the measured variable and the units of measure) was absent from all EHRs. One EHR system graphed results in reverse chronological order. One EHR system plotted data collected at unequally-spaced points in time using equally-spaced data points, which had the effect of erroneously depicting the visual slope perception between data points. This deficiency could have a significant, negative impact on patient safety. Only two EHR systems allowed users to see, hover-over, or click on a data point to see the precise values of the x–y coordinates. Our study suggests that many current EHR-generated graphs do not meet evidence-based criteria aimed at improving laboratory data comprehension.

Sign in / Sign up

Export Citation Format

Share Document