Uses of electronic health records for public health surveillance to advance public health

Mdpg: a novel multi-disease diagnosis prediction method based on patient knowledge graphs

Diagnosis prediction, a key factor in enhancing healthcare efficiency, remains a focal point in clinical decision support research. However, the time-series, sparse and multi-noise characteristics of electronic health record (EHR) data make it a great challenge. Existing methods commonly address these issues using RNNs and incorporating medical prior knowledge from medical knowledge bases, but they neglect the local spatial characteristics and spatial-temporal correlation of the data. Consequently, we propose MDPG, a diagnosis prediction model based on patient knowledge graphs. Initially, we represent the electronic visit records of patients as a patient-centered temporal knowledge graph, capturing the local spatial structure and temporal characteristics of the visit information. Subsequently, we design the spatial graph convolution block, temporal self-attention block, and spatial-temporal synchronous graph convolution block to capture the spatial, temporal, and spatial-temporal correlations embedded in them, respectively. Ultimately, we accomplish the prediction of patients' future states through multi-label classification. We conduct comprehensive experiments on two real-world datasets independently and evaluate the results using visit-level precision@k and code-level accuracy@k metrics. The experimental results demonstrate that MDPG outperforms all baseline models, yielding the best performance.

Enhancing severe hypoglycemia prediction in type 2 diabetes mellitus through multi-view co-training machine learning model for imbalanced dataset

Patients with type 2 diabetes mellitus (T2DM) who have severe hypoglycemia (SH) poses a considerable risk of long-term death, especially among the elderly, demanding urgent medical attention. Accurate prediction of SH remains challenging due to its multifaced nature, contributed from factors such as medications, lifestyle choices, and metabolic measurements. In this study, we propose a systematic approach to improve the robustness and accuracy of SH predictions using machine learning models, guided by clinical feature selection. Our focus is on developing long-term SH prediction models using both semi-supervised learning and supervised learning algorithms. Using the action to control cardiovascular risk in diabetes trial, which includes electronic health records for over 10,000 individuals, we focus on studying adults with T2DM. Our results indicate that the application of a multi-view co-training method, incorporating the random forest algorithm, improves the specificity of SH prediction, while the same setup with Naive Bayes replacing random forest demonstrates better sensitivity. Our framework also provides interpretability of machine learning models by identifying key predictors for hypoglycemia, including fasting plasma glucose, hemoglobin A1c, general diabetes education, and NPH or L insulins. The integration of data routinely available in electronic health records significantly enhances our model's capability to predict SH events, showcasing its potential to transform clinical practice by facilitating early interventions and optimizing patient management. By enhancing prediction accuracy and identifying crucial predictive features, our study contributes to advancing the understanding and management of hypoglycemia in this population.

CliqueFluxNet: Unveiling EHR Insights with Stochastic Edge Fluxing and Maximal Clique Utilisation Using Graph Neural Networks

Electronic Health Records (EHRs) play a crucial role in shaping predictive are models, yet they encounter challenges such as significant data gaps and class imbalances. Traditional Graph Neural Network (GNN) approaches have limitations in fully leveraging neighbourhood data or demanding intensive computational requirements for regularisation. To address this challenge, we introduce CliqueFluxNet, a novel framework that innovatively constructs a patient similarity graph to maximise cliques, thereby highlighting strong inter-patient connections. At the heart of CliqueFluxNet lies its stochastic edge fluxing strategy - a dynamic process involving random edge addition and removal during training. This strategy aims to enhance the model's generalisability and mitigate overfitting. Our empirical analysis, conducted on MIMIC-III and eICU datasets, focuses on the tasks of mortality and readmission prediction. It demonstrates significant progress in representation learning, particularly in scenarios with limited data availability. Qualitative assessments further underscore CliqueFluxNet's effectiveness in extracting meaningful EHR representations, solidifying its potential for advancing GNN applications in healthcare analytics.

Lyme Disease Surveillance and Epidemiology in the United States: A Historical Perspective

In the 40 years since Steere and colleagues first described Lyme disease, the illness has increased in incidence and distribution to become the most common vector-borne disease in the United States. Public health officials have developed, implemented, and revised surveillance systems to describe and monitor the condition. Much has been learned about the epidemiology of the illness, despite practical and logistical constraints that have encumbered the collection and interpretation of surveillance data. Future development of automated data collection from electronic health records as a source of surveillance and clinical information will address practical challenges and help answer ongoing questions about complications and persistent symptoms. Robust surveillance will be essential to monitor the effectiveness and safety of future vaccines and other preventive measures.

Validation of an acute respiratory infection phenotyping algorithm to support robust computerised medical record-based respiratory sentinel surveillance, England, 2023

IntroductionRespiratory sentinel surveillance systems leveraging computerised medical records (CMR) use phenotyping algorithms to identify cases of interest, such as acute respiratory infection (ARI). The Oxford-Royal College of General Practitioners Research and Surveillance Centre (RSC) is the English primary care-based sentinel surveillance network.AimThis study describes and validates the RSC's new ARI phenotyping algorithm.MethodsWe developed the phenotyping algorithm using a framework aligned with international interoperability standards. We validated our algorithm by comparing ARI events identified during the 2022/23 influenza season in England through use of both old and new algorithms. We compared clinical codes commonly used for recording ARI.ResultsThe new algorithm identified an additional 860,039 cases and excluded 52,258, resulting in a net increase of 807,781 cases (33.84%) of ARI compared to the old algorithm, with totals of 3,194,224 cases versus 2,386,443 cases. Of the 860,039 newly identified cases, the majority (63.7%) were due to identification of symptom codes suggestive of an ARI diagnosis not detected by the old algorithm. The 52,258 cases incorrectly identified by the old algorithm were due to inadvertent identification of chronic, recurrent, non-infectious and other non-ARI disease.ConclusionWe developed a new ARI phenotyping algorithm that more accurately identifies cases of ARI from the CMR. This will benefit public health by providing more accurate surveillance reports to public health authorities. This new algorithm can serve as a blueprint for other CMR-based surveillance systems wishing to develop similar phenotyping algorithms.

Public Health Surveillance in Electronic Health Records: Lessons From PCORnet

Introduction

PCORnet, the National Patient-Centered Clinical Research Network, is a large research network of health systems that map clinical data to a standardized data model. In 2018, we expanded existing infrastructure to facilitate use for public health surveillance. We describe benefits and challenges of using PCORnet for surveillance and describe case studies.

Methods

In 2018, infrastructure enhancements included addition of a table to store patients' residential zip codes and expansion of a modular program to generate population health statistics across conditions. Chronic disease surveillance case studies conducted in 2019 assessed atrial fibrillation (AF) and cirrhosis. In April 2020, PCORnet established an infrastructure to support COVID-19 surveillance with institutions frequently updating their electronic health record data.

Results

By August 2023, 53 PCORnet sites (84%) had a 5-digit zip code available on at least 95% of their patient populations. Among 148,223 newly diagnosed AF patients eligible for oral anticoagulant (OAC) therapy, 43.3% were on any OAC (17.8% warfarin, 28.5% any novel oral anticoagulant) within a year of the AF diagnosis. Among 60,268 patients with cirrhosis (2015-2019), common documented etiologies included unknown (48%), hepatitis C infection (23%), and alcohol use (22%). During October 2022 through December 2023, across 34 institutions, the proportion of COVID-19 patients who were cared for in the inpatient setting was 9.1% among 887,051 adults aged 20 years or older and 6.0% among 139,148 children younger than 20 years.

Conclusions

PCORnet provides important data that may augment traditional public health surveillance programs across diverse conditions. PCORnet affords longitudinal population health assessments among large catchments of the population with clinical, treatment, and geographic information, with capabilities to deliver rapid information needed during public health emergencies.

Factors Influencing Data Quality in Electronic Health Record Systems in 50 Health Facilities in Rwanda and the Role of Clinical Alerts: Cross-Sectional Observational Study

BACKGROUND

Electronic health records (EHRs) play an increasingly important role in delivering HIV care in low- and middle-income countries. The data collected are used for direct clinical care, quality improvement, program monitoring, public health interventions, and research. Despite widespread EHR use for HIV care in African countries, challenges remain, especially in collecting high-quality data.

OBJECTIVE

We aimed to assess data completeness, accuracy, and timeliness compared to paper-based records, and factors influencing data quality in a large-scale EHR deployment in Rwanda.

METHODS

We randomly selected 50 health facilities (HFs) using OpenMRS, an EHR system that supports HIV care in Rwanda, and performed a data quality evaluation. All HFs were part of a larger randomized controlled trial, with 25 HFs receiving an enhanced EHR with clinical decision support systems. Trained data collectors visited the 50 HFs to collect 28 variables from the paper charts and the EHR system using the Open Data Kit app. We measured data completeness, timeliness, and the degree of matching of the data in paper and EHR records, and calculated concordance scores. Factors potentially affecting data quality were drawn from a previous survey of users in the 50 HFs.

RESULTS

We randomly selected 3467 patient records, reviewing both paper and EHR copies (194,152 total data items). Data completeness was >85% threshold for all data elements except viral load (VL) results, second-line, and third-line drug regimens. Matching scores for data values were close to or >85% threshold, except for dates, particularly for drug pickups and VL. The mean data concordance was 10.2 (SD 1.28) for 15 (68%) variables. HF and user factors (eg, years of EHR use, technology experience, EHR availability and uptime, and intervention status) were tested for correlation with data quality measures. EHR system availability and uptime was positively correlated with concordance, whereas users' experience with technology was negatively correlated with concordance. The alerts for missing VL results implemented at 11 intervention HFs showed clear evidence of improving timeliness and completeness of initially low matching of VL results in the EHRs and paper records (11.9%-26.7%; P<.001). Similar effects were seen on the completeness of the recording of medication pickups (18.7%-32.6%; P<.001).

CONCLUSIONS

The EHR records in the 50 HFs generally had high levels of completeness except for VL results. Matching results were close to or >85% threshold for nondate variables. Higher EHR stability and uptime, and alerts for entering VL both strongly improved data quality. Most data were considered fit for purpose, but more regular data quality assessments, training, and technical improvements in EHR forms, data reports, and alerts are recommended. The application of quality improvement techniques described in this study should benefit a wide range of HFs and data uses for clinical care, public health, and disease surveillance.

Preventive Service Usage and New Chronic Disease Diagnoses: Using PCORnet Data to Identify Emerging Trends, United States, 2018-2022

Background

Data modernization efforts to strengthen surveillance capacity could help assess trends in use of preventive services and diagnoses of new chronic disease during the COVID-19 pandemic, which broadly disrupted health care access.

Methods

This cross-sectional study examined electronic health record data from US adults aged 21 to 79 years in a large national research network (PCORnet), to describe use of 8 preventive health services (N = 30,783,825 patients) and new diagnoses of 9 chronic diseases (N = 31,588,222 patients) during 2018 through 2022. Joinpoint regression assessed significant trends, and health debt was calculated comparing 2020 through 2022 volume to prepandemic (2018 and 2019) levels.

Results

From 2018 to 2022, use of some preventive services increased (hemoglobin A1c and lung computed tomography, both P < .05), others remained consistent (lipid testing, wellness visits, mammograms, Papanicolaou tests or human papillomavirus tests, stool-based screening), and colonoscopies or sigmoidoscopies declined (P < .01). Annual new chronic disease diagnoses were mostly stable (6% hypertension; 4% to 5% cholesterol; 4% diabetes; 1% colonic adenoma; 0.1% colorectal cancer; among women, 0.5% breast cancer), although some declined (lung cancer, cervical intraepithelial neoplasia or carcinoma in situ, cervical cancer, all P < .05). The pandemic resulted in health debt, because use of most preventive services and new diagnoses of chronic disease were less than expected during 2020; these partially rebounded in subsequent years. Colorectal screening and colonic adenoma detection by age group aligned with screening recommendation age changes during this period.

Conclusion

Among over 30 million patients receiving care during 2018 through 2022, use of preventive services and new diagnoses of chronic disease declined in 2020 and then rebounded, with some remaining health debt. These data highlight opportunities to augment traditional surveillance with EHR-based data.

MENDS-on-FHIR: leveraging the OMOP common data model and FHIR standards for national chronic disease surveillance

Objectives

The Multi-State EHR-Based Network for Disease Surveillance (MENDS) is a population-based chronic disease surveillance distributed data network that uses institution-specific extraction-transformation-load (ETL) routines. MENDS-on-FHIR examined using Health Language Seven's Fast Healthcare Interoperability Resources (HL7® FHIR®) and US Core Implementation Guide (US Core IG) compliant resources derived from the Observational Medical Outcomes Partnership (OMOP) Common Data Model (CDM) to create a standards-based ETL pipeline.

Materials and Methods

The input data source was a research data warehouse containing clinical and administrative data in OMOP CDM Version 5.3 format. OMOP-to-FHIR transformations, using a unique JavaScript Object Notation (JSON)-to-JSON transformation language called Whistle, created FHIR R4 V4.0.1/US Core IG V4.0.0 conformant resources that were stored in a local FHIR server. A REST-based Bulk FHIR $export request extracted FHIR resources to populate a local MENDS database.

Results

Eleven OMOP tables were used to create 10 FHIR/US Core compliant resource types. A total of 1.13 trillion resources were extracted and inserted into the MENDS repository. A very low rate of non-compliant resources was observed.

Discussion

OMOP-to-FHIR transformation results passed validation with less than a 1% non-compliance rate. These standards-compliant FHIR resources provided standardized data elements required by the MENDS surveillance use case. The Bulk FHIR application programming interface (API) enabled population-level data exchange using interoperable FHIR resources. The OMOP-to-FHIR transformation pipeline creates a FHIR interface for accessing OMOP data.

Conclusion

MENDS-on-FHIR successfully replaced custom ETL with standards-based interoperable FHIR resources using Bulk FHIR. The OMOP-to-FHIR transformations provide an alternative mechanism for sharing OMOP data.

Electronic Health Records Versus Survey Small Area Estimates for Public Health Surveillance

INTRODUCTION

Electronic health records (EHRs) are increasingly being leveraged for public health surveillance. EHR-based small area estimates (SAEs) are often validated by comparison to survey data such as the Behavioral Risk Factor Surveillance System (BRFSS). However, survey and EHR-based SAEs are expected to differ. In this cross-sectional study, SAEs were generated using MDPHnet, a distributed EHR-based surveillance network, for all Massachusetts municipalities and zip code tabulation areas (ZCTAs), compared to BRFSS PLACES SAEs, and reasons for differences explored.

METHODS

This study delineated reasons a priori for how SAEs derived using EHRs may differ from surveys by comparing each strategy's case classification criteria and reviewing the literature. Hypertension, diabetes, obesity, asthma, and smoking EHR-based SAEs for 2021 in all ZCTAs and municipalities in Massachusetts were estimated with Bayesian mixed effects modeling and poststratification in the summer/fall of 2023. These SAEs were compared to BRFSS PLACES SAEs published by the U.S. Centers for Disease Control and Prevention.

RESULTS

Mean prevalence was higher in EHR data versus BRFSS in both municipalities and ZCTAs for all outcomes except asthma. ZCTA and municipal symmetric mean absolute percentages ranged from 12.0 to 38.2% and 13.1 to 39.8%, respectively. There was greater variability in EHR-based SAEs versus BRFSS PLACES in both municipalities and ZCTAs.

CONCLUSIONS

EHR-based SAEs tended to be higher than BRFSS and more variable. Possible explanations include detection of undiagnosed cases and over-classification using EHR data, and under-reporting within BRFSS. Both EHR and survey-based surveillance have strengths and limitations that should inform their preferred uses in public health surveillance.

Electronic Case Reporting Development, Implementation, and Expansion in the United States

INTRODUCTION

The COVID-19 pandemic highlighted the need for a nationwide health information technology solution that could improve upon manual case reporting and decrease the clinical and administrative burden on the US health care system. We describe the development, implementation, and nationwide expansion of electronic case reporting (eCR), including its effect on public health surveillance and pandemic readiness.

METHODS

Multidisciplinary teams developed and implemented a standards-based, shared, scalable, and interoperable eCR infrastructure during 2014-2020. From January 27, 2020, to January 7, 2023, the team conducted a nationwide scale-up effort and determined the number of eCR-capable electronic health record (EHR) products, the number of reportable conditions available within the infrastructure, and technical connections of health care organizations (HCOs) and jurisdictional public health agencies (PHAs) to the eCR infrastructure. The team also conducted data quality studies to determine whether HCOs were discontinuing manual case reporting and early results of eCR timeliness.

RESULTS

During the study period, the number of eCR-capable EHR products developed or in development increased 11-fold (from 3 to 33), the number of reportable conditions available increased 28-fold (from 6 to 173), the number of HCOs connected to the eCR infrastructure increased 143-fold (from 153 to 22 000), and the number of jurisdictional PHAs connected to the eCR infrastructure increased 2.75-fold (from 24 to 66). Data quality reviews with PHAs resulted in select HCOs discontinuing manual case reporting and using eCR-exclusive case reporting in 13 PHA jurisdictions. The timeliness of eCR was <1 minute.

PRACTICE IMPLICATIONS

The growth of eCR can revolutionize public health case surveillance by producing data that are more timely and complete than manual case reporting while reducing reporting burden.

Periodontal Disease, Tooth Loss, and Systemic Conditions: An Exploratory Study

BACKGROUND

Although systemic medical conditions are associated with periodontitis and tooth loss, large-scale studies that include less prevalent systemic conditions are needed. The purpose of the study was to investigate the link between periodontal disease and tooth loss with systemic medical conditions in a large and diverse population.

METHODS

Dental charts of adult patients who had attended the dental clinics seeking dental therapy of the universities contributing data to the BigMouth network and accepted the protocol of the study were included. Dental Procedure Codes and Current Procedural Terminology procedures were utilised to identify patients with and without periodontitis. Data were extracted from patients' electronic health records including demographic characteristics, dental procedural codes, and self-reported medical conditions as well as the number of missing teeth.

RESULTS

A total of 108,307 records were ultimately included in the analysis; 42,377 of them included a diagnosis of periodontitis. The median age of the included population was 47.0 years, and 55.2% were female. Older and male individuals were significantly more likely to be in the periodontitis group and have higher number of missing teeth. A number of systemic conditions are associated with periodontitis and a higher number of missing teeth. High blood pressure, smoking, drug use, and diabetes were all found to be significant. Other significant conditions were anaemia, lymphoma, glaucoma, dialysis, bronchitis, sinusitis hepatitis, and asthma.

CONCLUSIONS

Within the limitations of this retrospective study that utilised the BigMouth dental data repository, the association of a number of systemic conditions such as smoking, diabetes, and hypertension with periodontitis and tooth loss has been confirmed. Additional connections have been highlighted for conditions that are not commonly reported in the literature.

Characterizing opioid overdose hotspots for place-based overdose prevention and treatment interventions: A geo-spatial analysis of Rhode Island, USA

OBJECTIVE

Examine differences in neighborhood characteristics and services between overdose hotspot and non-hotspot neighborhoods and identify neighborhood-level population factors associated with increased overdose incidence.

METHODS

We conducted a population-based retrospective analysis of Rhode Island, USA residents who had a fatal or non-fatal overdose from 2016 to 2020 using an environmental scan and data from Rhode Island emergency medical services, State Unintentional Drug Overdose Reporting System, and the American Community Survey. We conducted a spatial scan via SaTScan to identify non-fatal and fatal overdose hotspots and compared the characteristics of hotspot and non-hotspot neighborhoods. We identified associations between census block group-level characteristics using a Besag-York-Mollié model specification with a conditional autoregressive spatial random effect.

RESULTS

We identified 7 non-fatal and 3 fatal overdose hotspots in Rhode Island during the study period. Hotspot neighborhoods had higher proportions of Black and Latino/a residents, renter-occupied housing, vacant housing, unemployment, and cost-burdened households. A higher proportion of hotspot neighborhoods had a religious organization, a health center, or a police station. Non-fatal overdose risk increased in a dose responsive manner with increasing proportions of residents living in poverty. There was increased relative risk of non-fatal and fatal overdoses in neighborhoods with crowded housing above the mean (RR 1.19 [95 % CI 1.05, 1.34]; RR 1.21 [95 % CI 1.18, 1.38], respectively).

CONCLUSION

Neighborhoods with increased prevalence of housing instability and poverty are at highest risk of overdose. The high availability of social services in overdose hotspots presents an opportunity to work with established organizations to prevent overdose deaths.

Causal Effects of Stochastic PrEP Interventions on HIV Incidence Among Men Who Have Sex With Men

Antiretroviral preexposure prophylaxis (PrEP) is highly effective in preventing human immunodeficiency virus (HIV) infection, but uptake has been limited and inequitable. Although interventions to increase PrEP uptake are being evaluated in clinical trials among men who have sex with men (MSM), those trials cannot evaluate effects on HIV incidence. Estimates from observational studies of the causal effects of PrEP-uptake interventions on HIV incidence can inform decisions about intervention scale-up. We used longitudinal electronic health record data from HIV-negative MSM accessing care at Fenway Health, a community health center in Boston, Massachusetts, from January 2012 through February 2018, with 2 years of follow-up. We considered stochastic interventions that increased the chance of initiating PrEP in several high-priority subgroups. We estimated the effects of these interventions on population-level HIV incidence using a novel inverse-probability weighted estimator of the generalized g-formula, adjusting for baseline and time-varying confounders. Our results suggest that even modest increases in PrEP initiation in high-priority subgroups of MSM could meaningfully reduce HIV incidence in the overall population of MSM. Interventions tailored to Black and Latino MSM should be prioritized to maximize equity and impact.

Using Electronic Health Record System to Establish a National Patient's Registry : Lessons learned from the Cancer Registry in Iran

BACKGROUND

In Iran, the Integrated Electronic Health Record system, called SEPAS, has been established to store all patient encounters of individuals referring to healthcare facilities.

OBJECTIVE

We aimed to develop a model for cleaning SEPAS and applying its data in other databases.

METHODS

We used cancer data from SEPAS as the sample. We developed a guideline to identify codes for cancer-related diagnoses and services in the database. Furthermore, we searched the SEPAS database based on ICD-10 and the diagnosis description in English and Farsi in an Excel sheet. We added codes and descriptions of pharmaceuticals and procedures to the list. We applied the above database and linked it to the patient records to identify cancer patients. A dashboard was designed based on this information for every cancer patient.

RESULTS

We selected 5,841 diagnostic codes and phrases, 9,300 cancer pharmaceutics codes, and 452 codes from cancer-specific items related to the diagnostic procedures and treatment methods. Linkage of this list to the patient list generated a database of about 197,164 cancer patients for linkage in the registry database.

CONCLUSIONS

Patient registries are one of the most important sources of information in healthcare systems. Data linkage between Electronic Health Record Systems (EHRs) and registries, despite its challenges, is profitable. EHRs can be used for case finding in any patient registry to reduce the time and cost of case finding.

MENDS-on-FHIR: Leveraging the OMOP common data model and FHIR standards for national chronic disease surveillance

Objective

The Multi-State EHR-Based Network for Disease Surveillance (MENDS) is a population-based chronic disease surveillance distributed data network that uses institution-specific extraction-transformation-load (ETL) routines. MENDS-on-FHIR examined using Health Language Seven's Fast Healthcare Interoperability Resources (HL7® FHIR®) and US Core Implementation Guide (US Core IG) compliant resources derived from the Observational Medical Outcomes Partnership (OMOP) Common Data Model (CDM) to create a standards-based ETL pipeline.

Materials and Methods

The input data source was a research data warehouse containing clinical and administrative data in OMOP CDM Version 5.3 format. OMOP-to-FHIR transformations, using a unique JavaScript Object Notation (JSON)-to-JSON transformation language called Whistle, created FHIR R4 V4.0.1/US Core IG V4.0.0 conformant resources that were stored in a local FHIR server. A REST-based Bulk FHIR $export request extracted FHIR resources to populate a local MENDS database.

Results

Eleven OMOP tables were used to create 10 FHIR/US Core compliant resource types. A total of 1.13 trillion resources were extracted and inserted into the MENDS repository. A very low rate of non-compliant resources was observed.

Discussion

OMOP-to-FHIR transformation results passed validation with less than a 1% non-compliance rate. These standards-compliant FHIR resources provided standardized data elements required by the MENDS surveillance use case. The Bulk FHIR application programming interface (API) enabled population-level data exchange using interoperable FHIR resources. The OMOP-to-FHIR transformation pipeline creates a FHIR interface for accessing OMOP data.

Conclusion

MENDS-on-FHIR successfully replaced custom ETL with standards-based interoperable FHIR resources using Bulk FHIR. The OMOP-to-FHIR transformations provide an alternative mechanism for sharing OMOP data.

An open source corpus and automatic tool for section identification in Spanish health records

BACKGROUND

Electronic Clinical Narratives (ECNs) store valuable individual's health information. However, there are few available open-source data. Besides, ECNs can be structurally heterogeneous, ranging from documents with explicit section headings or titles to unstructured notes. This lack of structure complicates building automatic systems and their evaluation.

OBJECTIVE

The aim of the present work is to provide the scientific community with a Spanish open-source dataset to build and evaluate automatic section identification systems. Together with this dataset, the purpose is to design and implement a suitable evaluation measure and a fine-tuned language model adapted to the task.

MATERIALS AND METHODS

A corpus of unstructured clinical records, in this case progress notes written in Spanish, was annotated with seven major section types. Existing metrics for the presented task were thoroughly assessed and, based on the most suitable one, we defined a new B2 metric better tailored given the task.

RESULTS

The annotated corpus, as well as the designed new evaluation script and a baseline model are freely available for the community. This model reaches an average B2 score of 71.3 on our open source dataset and an average B2 of 67.0 in data scarcity scenarios where the target corpus and its structure differs from the dataset used for training the LM.

CONCLUSION

Although section identification in unstructured clinical narratives is challenging, this work shows that it is possible to build competitive automatic systems when both data and the right evaluation metrics are available. The annotated data, the implemented evaluation scripts, and the section identification Language Model are open-sourced hoping that this contribution will foster the building of more and better systems.

Unraveling a blockchain-based framework towards patient empowerment: A scoping review envisioning future smart health technologies

The COVID-19 pandemic shows us how crucial patient empowerment can be in the healthcare ecosystem. Now, we know that scientific advancement, technology integration, and patient empowerment need to be orchestrated to realize future smart health technologies. In that effort, this paper unravels the Good (advantages), Bad (challenges/limitations), and Ugly (lacking patient empowerment) of the blockchain technology integration in the Electronic Health Record (EHR) paradigm in the existing healthcare landscape. Our study addresses four methodically-tailored and patient-centric Research Questions, primarily examining 138 relevant scientific papers. This scoping review also explores how the pervasiveness of blockchain technology can help to empower patients in terms of access, awareness, and control. Finally, this scoping review leverages the insights gleaned from this study and contributes to the body of knowledge by proposing a patient-centric blockchain-based framework. This work will envision orchestrating three essential elements with harmony: scientific advancement (Healthcare and EHR), technology integration (Blockchain Technology), and patient empowerment (access, awareness, and control).

Linking emergency care and police department data to strengthen timely information on violence-related paediatric injuries

BACKGROUND

Combined ED and police department (PD) data have improved violence surveillance in the UK, enabling significantly improved prevention. We sought to determine if the addition of emergency medical service (EMS) data to ED data would contribute meaningful information on violence-related paediatric injuries beyond PD record data in a US city.

METHODS

Cross-sectional data on self-reported violence-related injuries of youth treated in the ED between January 2015 and September 2016 were combined with incidents classified by EMS as intentional interpersonal violence and incidents in which the PD responded to a youth injury from a simple or aggravated assault, robbery or sexual offence. Nearest neighbour hierarchical spatial clustering detected areas in which 10 or more incidents occurred during this period (hotspots), with the radii of the area being 1000, 1500, 2000 and 3000 ft. Overlap of PD incidents within ED&EMS hotspots (and vice versa) was calculated and Spearman's r tested statistical associations between the data sets, or ED&EMS contribution to PD violence information.

RESULTS

There were 935 unique ED&EMS records (ED=381; EMS=554). Of these, 877 (94%) were not in PD records. In large hotspots >2000 ft, ED&EMS records identified one additional incident for every three in the PD database. ED and EMS provided significant numbers of incidents not reported to PD. Significant correlations of ED&EMS incidents in PD hotspots imply that the ED&EMS incidents are as pervasive across the city as that reported by PD. In addition, ED and EMS provided unique violence information, as ED&EMS hotspots never included a majority (>50%) of PD records. Most (676/877; 77%) incidents unique to ED&EMS records were within 1000 ft of a school or park.

CONCLUSIONS

Many violence locations in ED and EMS data were not present in PD records. A combined PD, ED and EMS database resulted in new knowledge of the geospatial distribution of violence-related paediatric injuries and can be used for data-informed and targeted prevention of violence in which children are injured-especially in and around schools and parks.

Using latent class analysis to inform the design of an EHR-based national chronic disease surveillance model

The Multi-state EHR-based Network for Disease Surveillance (MENDS) developed a pilot electronic health record (EHR) surveillance system capable of providing national chronic disease estimates. To strategically engage partner sites, MENDS conducted a latent class analysis (LCA) and grouped states by similarities in socioeconomics, demographics, chronic disease and behavioral risk factor prevalence, health outcomes, and health insurance coverage. Three latent classes of states were identified, which inform the recruitment of additional partner sites in conjunction with additional factors (e.g. partner site capacity and data availability, information technology infrastructure). This methodology can be used to inform other public health surveillance modernization efforts that leverage timely EHR data to address gaps, use existing technology, and advance surveillance.

Machine Learning-Enabled Clinical Information Systems Using Fast Healthcare Interoperability Resources Data Standards: Scoping Review

Background

Machine learning-enabled clinical information systems (ML-CISs) have the potential to drive health care delivery and research. The Fast Healthcare Interoperability Resources (FHIR) data standard has been increasingly applied in developing these systems. However, methods for applying FHIR to ML-CISs are variable.

Objective

This study evaluates and compares the functionalities, strengths, and weaknesses of existing systems and proposes guidelines for optimizing future work with ML-CISs.

Methods

Embase, PubMed, and Web of Science were searched for articles describing machine learning systems that were used for clinical data analytics or decision support in compliance with FHIR standards. Information regarding each system's functionality, data sources, formats, security, performance, resource requirements, scalability, strengths, and limitations was compared across systems.

Results

A total of 39 articles describing FHIR-based ML-CISs were divided into the following three categories according to their primary focus: clinical decision support systems (n=18), data management and analytic platforms (n=10), or auxiliary modules and application programming interfaces (n=11). Model strengths included novel use of cloud systems, Bayesian networks, visualization strategies, and techniques for translating unstructured or free-text data to FHIR frameworks. Many intelligent systems lacked electronic health record interoperability and externally validated evidence of clinical efficacy.

Conclusions

Shortcomings in current ML-CISs can be addressed by incorporating modular and interoperable data management, analytic platforms, secure interinstitutional data exchange, and application programming interfaces with adequate scalability to support both real-time and prospective clinical applications that use electronic health record platforms with diverse implementations.

Patterns in the Development of Pediatric Allergy

OBJECTIVES

Describe clinical and epidemiologic patterns of pediatric allergy using longitudinal electronic health records (EHRs) from a multistate consortium of US practices.

METHODS

Using the multistate Comparative Effectiveness Research through Collaborative Electronic Reporting EHR database, we defined a cohort of 218 485 children (0-18 years) who were observed for ≥5 years between 1999 and 2020. Children with atopic dermatitis (AD), immunoglobulin E-mediated food allergy (IgE-FA), asthma, allergic rhinitis (AR), and eosinophilic esophagitis (EoE) were identified using a combination of diagnosis codes and medication prescriptions. We determined age at diagnosis, cumulative incidence, and allergic comorbidity.

RESULTS

Allergic disease cumulative (and peak age of) incidence was 10.3% (4 months) for AD, 4.0% (13 months) for IgE-FA, 20.1% (13 months) for asthma, 19.7% (26 months) for AR, and 0.11% (35 months) for EoE. The most diagnosed IgE-FAs were peanut (1.9%), egg (0.8%), and shellfish (0.6%). A total of 13.4% of children had ≥2 allergic conditions, and respiratory allergies (ie, asthma, AR) were commonly comorbid with each other, and with other allergic conditions.

CONCLUSIONS

We detail pediatric allergy patterns using longitudinal, health care provider-based data from EHR systems across multiple US states and varied pediatric practice types. Our results support the population-level allergic march progression and indicate high rates of comorbidity among children with food and respiratory allergies.

Improving Diagnostics with Deep Forest Applied to Electronic Health Records

An electronic health record (EHR) is a vital high-dimensional part of medical concepts. Discovering implicit correlations in the information of this data set and the research and informative aspects can improve the treatment and management process. The challenge of concern is the data sources' limitations in finding a stable model to relate medical concepts and use these existing connections. This paper presents Patient Forest, a novel end-to-end approach for learning patient representations from tree-structured data for readmission and mortality prediction tasks. By leveraging statistical features, the proposed model is able to provide an accurate and reliable classifier for predicting readmission and mortality. Experiments on MIMIC-III and eICU datasets demonstrate Patient Forest outperforms existing machine learning models, especially when the training data are limited. Additionally, a qualitative evaluation of Patient Forest is conducted by visualising the learnt representations in 2D space using the t-SNE, which further confirms the effectiveness of the proposed model in learning EHR representations.

Factors Influencing the Effectiveness of Adopting Electronic Medical Record-Based Reporting Systems for Notifiable Disease Surveillance: A Quantitative Analysis

The coronavirus disease 2019 (COVID-19) pandemic has led to greater attention being given to infectious disease surveillance systems and their notification functionalities. Although numerous studies have explored the benefits of integrating functionalities with electronic medical record (EMR) systems, empirical studies on the topic are rare. The current study assessed which factors influence the effectiveness of EMR-based reporting systems (EMR-RSs) for notifiable disease surveillance. This study interviewed staff from hospitals with a coverage that represented 51.39% of the notifiable disease reporting volume in Taiwan. Exact logistic regression was employed to determine which factors influenced the effectiveness of Taiwan's EMR-RS. The results revealed that the influential factors included hospitals' early participation in the EMR-RS project, frequent consultation with the information technology (IT) provider of the Taiwan Centers for Disease Control (TWCDC), and retrieval of data from at least one internal database. They also revealed that using an EMR-RS resulted in more timely, accurate, and convenient reporting in hospitals. In addition, developing by an internal IT unit instead of outsourcing EMR-RS development led to more accurate and convenient reporting. Automatically loading the required data enhanced the convenience, and designing input fields that may be unavailable in current databases to enable physicians to add data to legacy databases also boosted effectiveness of the reporting system.

Causal Evaluation of Post-Marketing Drugs for Drug-induced Liver Injury from Electronic Health Records

Drug-induced liver injury (DILI) is one of the most common and serious adverse drug reactions that can lead to acute liver failure and death. Detection of DILI and causal estimation of drug-hepatotoxicity association are of great importance for patient safety. This paper proposes a framework for causal estimation of post-marketing drugs for DILI from real-world electronic health record (EHR) data. Randomized clinical trials were replicated at scale by automatically generating different user and non-user cohorts for each potential drug, and average treatment effects (ATEs) of drugs were estimated using targeted maximum likelihood estimation. Ten years of real-world EHRs were used to validate the framework. Of all 1199 single-ingredient drugs analyzed, 7 novel and 7 known drug-hepatotoxicity associations were found to be causal.

ARCH: Large-scale Knowledge Graph via Aggregated Narrative Codified Health Records Analysis

Objective

Electronic health record (EHR) systems contain a wealth of clinical data stored as both codified data and free-text narrative notes, covering hundreds of thousands of clinical concepts available for research and clinical care. The complex, massive, heterogeneous, and noisy nature of EHR data imposes significant challenges for feature representation, information extraction, and uncertainty quantification. To address these challenges, we proposed an efficient Aggregated naRrative Codified Health (ARCH) records analysis to generate a large-scale knowledge graph (KG) for a comprehensive set of EHR codified and narrative features.

Methods

The ARCH algorithm first derives embedding vectors from a co-occurrence matrix of all EHR concepts and then generates cosine similarities along with associated p -values to measure the strength of relatedness between clinical features with statistical certainty quantification. In the final step, ARCH performs a sparse embedding regression to remove indirect linkage between entity pairs. We validated the clinical utility of the ARCH knowledge graph, generated from 12.5 million patients in the Veterans Affairs (VA) healthcare system, through downstream tasks including detecting known relationships between entity pairs, predicting drug side effects, disease phenotyping, as well as sub-typing Alzheimer's disease patients.

Results

ARCH produces high-quality clinical embeddings and KG for over 60,000 EHR concepts, as visualized in the R-shiny powered web-API (https://celehs.hms.harvard.edu/ARCH/). The ARCH embeddings attained an average area under the ROC curve (AUC) of 0.926 and 0.861 for detecting pairs of similar EHR concepts when the concepts are mapped to codified data and to NLP data; and 0.810 (codified) and 0.843 (NLP) for detecting related pairs. Based on the p -values computed by ARCH, the sensitivity of detecting similar and related entity pairs are 0.906 and 0.888 under false discovery rate (FDR) control of 5%. For detecting drug side effects, the cosine similarity based on the ARCH semantic representations achieved an AUC of 0.723 while the AUC improved to 0.826 after few-shot training via minimizing the loss function on the training data set. Incorporating NLP data substantially improved the ability to detect side effects in the EHR. For example, based on unsupervised ARCH embeddings, the power of detecting drug-side effects pairs when using codified data only was 0.15, much lower than the power of 0.51 when using both codified and NLP concepts. Compared to existing large-scale representation learning methods including PubmedBERT, BioBERT and SAPBERT, ARCH attains the most robust performance and substantially higher accuracy in detecting these relationships. Incorporating ARCH selected features in weakly supervised phenotyping algorithms can improve the robustness of algorithm performance, especially for diseases that benefit from NLP features as supporting evidence. For example, the phenotyping algorithm for depression attained an AUC of 0.927 when using ARCH selected features but only 0.857 when using codified features selected via the KESER network[1]. In addition, embeddings and knowledge graphs generated from the ARCH network were able to cluster AD patients into two subgroups, where the fast progression subgroup had a much higher mortality rate.

Conclusions

The proposed ARCH algorithm generates large-scale high-quality semantic representations and knowledge graph for both codified and NLP EHR features, useful for a wide range of predictive modeling tasks.

Predicting Visit Cost of Obstructive Sleep Apnea Using Electronic Healthcare Records With Transformer

BACKGROUND

Obstructive sleep apnea (OSA) is growing increasingly prevalent in many countries as obesity rises. Sufficient, effective treatment of OSA entails high social and financial costs for healthcare.

OBJECTIVE

For treatment purposes, predicting OSA patients' visit expenses for the coming year is crucial. Reliable estimates enable healthcare decision-makers to perform careful fiscal management and budget well for effective distribution of resources to hospitals. The challenges created by scarcity of high-quality patient data are exacerbated by the fact that just a third of those data from OSA patients can be used to train analytics models: only OSA patients with more than 365 days of follow-up are relevant for predicting a year's expenditures.

METHODS AND PROCEDURES

The authors propose a translational engineering method applying two Transformer models, one for augmenting the input via data from shorter visit histories and the other predicting the costs by considering both the material thus enriched and cases with more than a year's follow-up. This method effectively adapts state-of-the-art Transformer models to create practical cost prediction solutions that can be implemented in OSA management, potentially enhancing patient care and resource allocation.

RESULTS

The two-model solution permits putting the limited body of OSA patient data to productive use. Relative to a single-Transformer solution using only a third of the high-quality patient data, the solution with two models improved the prediction performance's [Formula: see text] from 88.8% to 97.5%. Even using baseline models with the model-augmented data improved the [Formula: see text] considerably, from 61.6% to 81.9%.

CONCLUSION

The proposed method makes prediction with the most of the available high-quality data by carefully exploiting details, which are not directly relevant for answering the question of the next year's likely expenditure. Clinical and Translational Impact Statement: Public Health- Lack of high-quality source data hinders data-driven analytics-based research in healthcare. The paper presents a method that couples data augmentation and prediction in cases of scant healthcare data.

An evaluation of existing text de-identification tools for use with patient progress notes from Australian general practice

INTRODUCTION

Digitized patient progress notes from general practice represent a significant resource for clinical and public health research but cannot feasibly and ethically be used for these purposes without automated de-identification. Internationally, several open-source natural language processing tools have been developed, however, given wide variations in clinical documentation practices, these cannot be utilized without appropriate review. We evaluated the performance of four de-identification tools and assessed their suitability for customization to Australian general practice progress notes.

METHODS

Four tools were selected: three rule-based (HMS Scrubber, MIT De-id, Philter) and one machine learning (MIST). 300 patient progress notes from three general practice clinics were manually annotated with personally identifying information. We conducted a pairwise comparison between the manual annotations and patient identifiers automatically detected by each tool, measuring recall (sensitivity), precision (positive predictive value), f1-score (harmonic mean of precision and recall), and f2-score (weighs recall 2x higher than precision). Error analysis was also conducted to better understand each tool's structure and performance.

RESULTS

Manual annotation detected 701 identifiers in seven categories. The rule-based tools detected identifiers in six categories and MIST in three. Philter achieved the highest aggregate recall (67%) and the highest recall for NAME (87%). HMS Scrubber achieved the highest recall for DATE (94%) and all tools performed poorly on LOCATION. MIST achieved the highest precision for NAME and DATE while also achieving similar recall to the rule-based tools for DATE and highest recall for LOCATION. Philter had the lowest aggregate precision (37%), however preliminary adjustments of its rules and dictionaries showed a substantial reduction in false positives.

CONCLUSION

Existing off-the-shelf solutions for automated de-identification of clinical text are not immediately suitable for our context without modification. Philter is the most promising candidate due to its high recall and flexibility however will require extensive revising of its pattern matching rules and dictionaries.

Work-related injuries and attendance at a Canadian regional emergency department

BACKGROUND

Monitoring trends in the burden of illness and injury attributable to work is key in assessing occupational health hazards; however, New Brunswick does not participate in the Canadian National Ambulatory Care Reporting System which itself does not collect details of occupation and industry.

AIMS

We set out to determine the proportion of emergency department attendances that were attributable to a work-related cause. We also wanted to evaluate the recording of occupation in the electronic health record system, and to describe the characteristics of patients with a work-related presentation.

METHODS

A retrospective observational study over a 1-year period was conducted using an administrative database obtained from Canadian Emergency Department Information System. Descriptive statistics are used to present the analysis of categorical and continuous data.

RESULTS

A total of 49 365 patients were included for analysis. Two per cent of patients presented with a self-reported work-related condition. Health care and social assistance, construction, retail trade and manufacturing were the most common industries reported by patients.

CONCLUSIONS

This study found the rate of work-related medical conditions to be substantially less than expected, and that occupation was not captured for any patients presenting to the emergency department with a work-related condition, despite a field being available in the electronic health record registration system. We were able to analyse the industry sectors for work-related presentations. The recording and coding of occupation and industry would significantly benefit occupational epidemiology in emergency medicine as well as potentially improving patient outcomes and health system efficiencies.

Using electronic health record data to link families: an illustrative example using intergenerational patterns of obesity

OBJECTIVE

Electronic health record (EHR) data are a valuable resource for population health research but lack critical information such as relationships between individuals. Emergency contacts in EHRs can be used to link family members, creating a population that is more representative of a community than traditional family cohorts.

MATERIALS AND METHODS

We revised a published algorithm: relationship inference from the electronic health record (RIFTEHR). Our version, Pythonic RIFTEHR (P-RIFTEHR), identifies a patient's emergency contacts, matches them to existing patients (when available) using network graphs, checks for conflicts, and infers new relationships. P-RIFTEHR was run on December 15, 2021 in the Northwestern Medicine Electronic Data Warehouse (NMEDW) on approximately 2.95 million individuals and was validated using the existing link between children born at NM hospitals and their mothers. As proof-of-concept, we modeled the association between parent and child obesity using logistic regression.

RESULTS

The P-RIFTEHR algorithm matched 1 157 454 individuals in 448 278 families. The median family size was 2, the largest was 32 persons, and 247 families spanned 4 generations or more. Validation of the mother-child pairs resulted in 95.1% sensitivity. Children were 2 times more likely to be obese if a parent is obese (OR: 2.30; 95% CI, 2.23-2.37).

CONCLUSION

P-RIFTEHR can identify familiar relationships in a large, diverse population in an integrated health system. Estimates of parent-child inheritability of obesity using family structures identified by the algorithm were consistent with previously published estimates from traditional cohort studies.

Transforming Public Health Data Systems to Advance the Population's Health

Policy Points Accurate and reliable data systems are critical for delivering the essential services and foundational capabilities of public health for a 21st -century public health infrastructure. Chronic underfunding, workforce shortages, and operational silos limit the effectiveness of America's public health data systems, with the country's anemic response to COVID-19 highlighting the results of long-standing infrastructure gaps. As the public health sector begins an unprecedented data modernization effort, scholars and policymakers should ensure ongoing reforms are aligned with the five components of an ideal public health data system: outcomes and equity oriented, actionable, interoperable, collaborative, and grounded in a robust public health system.

Automating Case Reporting of Chlamydia and Gonorrhea to Public Health Authorities in Illinois Clinics: Implementation and Evaluation of Findings

BACKGROUND

Chlamydia and gonorrhea cases continue to rise in Illinois, increasing by 16.4% and 70.9% in 2019, respectively, compared with 2015. Providers are required to report both chlamydia and gonorrhea, as mandated by public health laws. Manual reporting remains a huge burden; 90%-93% of cases were reported to Illinois Department of Public Health (IDPH) via electronic laboratory reporting (ELR), and the remaining were reported through web-based data entry platforms, faxes, and phone calls. However, cases reported via ELRs only contain information available to a laboratory facility and do not contain additional data needed for public health. Such data are typically found in an electronic health record (EHR). Electronic case reports (eCRs) were developed and automated the generation of case reports from EHRs to be reported to public health agencies.

OBJECTIVE

Prior studies consolidated trigger criteria for eCRs, and compared with manual reporting, found it to be more complete. The goal of this project is to pilot standards-based eCR for chlamydia and gonorrhea. We evaluated the throughput, completeness, and timeliness of eCR compared to ELR, as well as the implementation experience at a large health center-controlled network in Illinois.

METHODS

For this study, we selected 8 clinics located on the north, west, and south sides of Chicago to implement the eCRs; these cases were reported to IDPH. The study period was 52 days. The centralized EHR used by these clinics leveraged 2 of the 3 case detection scenarios, which were previously defined as the trigger, to generate an eCR. These messages were successfully transmitted via Health Level 7 electronic initial case report standard. Upon receipt by IDPH, these eCRs were parsed and housed in a staging database.

RESULTS

During the study period, 183 eCRs representing 135 unique patients were received by IDPH. eCR reported 95% (n=113 cases) of all the chlamydia cases and 97% (n=70 cases) of all the gonorrhea cases reported from the participating clinical sites. eCR found an additional 14 (19%) cases of gonorrhea that were not reported via ELR. However, ELR reported an additional 6 cases of chlamydia and 2 cases of gonorrhea, which were not reported via eCR. ELR reported 100% of chlamydia cases but only 81% of gonorrhea cases. While key elements such as patient and provider names were complete in both eCR and ELR, eCR was found to report additional clinical data, including history of present illness, reason for visit, symptoms, diagnosis, and medications.

CONCLUSIONS

eCR successfully identified and created automated reports for chlamydia and gonorrhea cases in the implementing clinics in Illinois. eCR demonstrated a more complete case report and represents a promising future of reducing provider burden for reporting cases while achieving greater semantic interoperability between health care systems and public health.

Leveraging Electronic Health Record Data for Timely Chronic Disease Surveillance: The Multi-State EHR-Based Network for Disease Surveillance

CONTEXT

Electronic health record (EHR) data can potentially make chronic disease surveillance more timely, actionable, and sustainable. Although use of EHR data can address numerous limitations of traditional surveillance methods, timely surveillance data with broad population coverage require scalable systems. This report describes implementation, challenges, and lessons learned from the Multi-State EHR-Based Network for Disease Surveillance (MENDS) to help inform how others work with EHR data to develop distributed networks for surveillance.

PROGRAM

Funded by the Centers for Disease Control and Prevention (CDC), MENDS is a data modernization demonstration project that aims to develop a timely national chronic disease sentinel surveillance system using EHR data. It facilitates partnerships between data contributors (health information exchanges, other data aggregators) and data users (state and local health departments). MENDS uses query and visualization software to track local emerging trends. The program also uses statistical and geospatial methods to generate prevalence estimates of chronic disease risk measures at the national and local levels. Resulting data products are designed to inform public health practice and improve the health of the population.

IMPLEMENTATION

MENDS includes 5 partner sites that leverage EHR data from 91 health system and clinic partners and represents approximately 10 million patients across the United States. Key areas of implementation include governance, partnerships, technical infrastructure and support, chronic disease algorithms and validation, weighting and modeling, and workforce education for public health data users.

DISCUSSION

MENDS presents a scalable distributed network model for implementing national chronic disease surveillance that leverages EHR data. Priorities as MENDS matures include producing prevalence estimates at various geographic and subpopulation levels, developing enhanced data sharing and interoperability capacity using international data standards, scaling the network to improve coverage nationally and among underrepresented geographic areas and subpopulations, and expanding surveillance of additional chronic disease measures and social determinants of health.

Secondary data for global health digitalisation

Substantial opportunities for global health intelligence and research arise from the combined and optimised use of secondary data within data ecosystems. Secondary data are information being used for purposes other than those intended when they were collected. These data can be gathered from sources on the verge of widespread use such as the internet, wearables, mobile phone apps, electronic health records, or genome sequencing. To utilise their full potential, we offer guidance by outlining available sources and approaches for the processing of secondary data. Furthermore, in addition to indicators for the regulatory and ethical evaluation of strategies for the best use of secondary data, we also propose criteria for assessing reusability. This overview supports more precise and effective policy decision making leading to earlier detection and better prevention of emerging health threats than is currently the case.

Information Technology Systems for Infection Control in German University Hospitals-Results of a Structured Survey a Year into the Severe Acute Respiratory Syndrome Coronavirus 2 Pandemic

BACKGROUND

 Digitalization is playing a major role in mastering the current coronavirus 2019 (COVID-19) pandemic. However, several outbreaks of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in German hospitals last year have shown that many of the surveillance and warning mechanisms related to infection control (IC) in hospitals need to be updated.

OBJECTIVES

 The main objective of the following work was to assess the state of information technology (IT) systems supporting IC and surveillance in German university hospitals in March 2021, almost a year into the SARS-CoV-2 pandemic.

METHODS

 As part of the National Research Network for Applied Surveillance and Testing project within the Network University Medicine, a cross-sectional survey was conducted to assess the situation of IC IT systems in 36 university hospitals in Germany.

RESULTS

 Among the most prominent findings were the lack of standardization of IC IT systems and the predominant use of commercial IC IT systems, while the vast majority of hospitals reported inadequacies in the features their IC IT systems provide for their daily work. However, as the pandemic has shown that there is a need for systems that can help improve health care, several German university hospitals have already started this upgrade independently.

CONCLUSIONS

 The deep challenges faced by the German health care sector regarding the integration and interoperability of IT systems designed for IC and surveillance are unlikely to be solved through punctual interventions and require collaboration between educational, medical, and administrative institutions.

State-level metabolic comorbidity prevalence and control among adults age 50-plus with diabetes: estimates from electronic health records and survey data in five states

BACKGROUND

Although treatment and control of diabetes can prevent complications and reduce morbidity, few data sources exist at the state level for surveillance of diabetes comorbidities and control. Surveys and electronic health records (EHRs) offer different strengths and weaknesses for surveillance of diabetes and major metabolic comorbidities. Data from self-report surveys suffer from cognitive and recall biases, and generally cannot be used for surveillance of undiagnosed cases. EHR data are becoming more readily available, but pose particular challenges for population estimation since patients are not randomly selected, not everyone has the relevant biomarker measurements, and those included tend to cluster geographically.

METHODS

We analyzed data from the National Health and Nutritional Examination Survey, the Health and Retirement Study, and EHR data from the DARTNet Institute to create state-level adjusted estimates of the prevalence and control of diabetes, and the prevalence and control of hypertension and high cholesterol in the diabetes population, age 50 and over for five states: Alabama, California, Florida, Louisiana, and Massachusetts.

RESULTS

The estimates from the two surveys generally aligned well. The EHR data were consistent with the surveys for many measures, but yielded consistently lower estimates of undiagnosed diabetes prevalence, and identified somewhat fewer comorbidities in most states.

CONCLUSIONS

Despite these limitations, EHRs may be a promising source for diabetes surveillance and assessment of control as the datasets are large and created during the routine delivery of health care.

TRIAL REGISTRATION

Not applicable.

Real-world data for precision public health of noncommunicable diseases: a scoping review

BACKGROUND

Global public health action to address noncommunicable diseases (NCDs) requires new approaches. NCDs are primarily prevented and managed in the community where there is little investment in digital health systems and analytics; this has created a data chasm and relatively silent burden of disease. The nascent but rapidly emerging area of precision public health offers exciting new opportunities to transform our approach to NCD prevention. Precision public health uses routinely collected real-world data on determinants of health (social, environmental, behavioural, biomedical and commercial) to inform precision decision-making, interventions and policy based on social position, equity and disease risk, and continuously monitors outcomes - the right intervention for the right population at the right time. This scoping review aims to identify global exemplars of precision public health and the data sources and methods of their aggregation/application to NCD prevention.

METHODS

The Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for scoping reviews (PRISMA-ScR) was followed. Six databases were systematically searched for articles published until February 2021. Articles were included if they described digital aggregation of real-world data and 'traditional' data for applied community, population or public health management of NCDs. Real-world data was defined as routinely collected (1) Clinical, Medication and Family History (2) Claims/Billing (3) Mobile Health (4) Environmental (5) Social media (6) Molecular profiling (7) Patient-centred (e.g., personal health record). Results were analysed descriptively and mapped according to the three horizons framework for digital health transformation.

RESULTS

Six studies were included. Studies developed population health surveillance methods and tools using diverse real-world data (e.g., electronic health records and health insurance providers) and traditional data (e.g., Census and administrative databases) for precision surveillance of 28 NCDs. Population health analytics were applied consistently with descriptive, geospatial and temporal functions. Evidence of using surveillance tools to create precision public health models of care or improve policy and practice decisions was unclear.

CONCLUSIONS

Applications of real-world data and designed data to address NCDs are emerging with greater precision. Digital transformation of the public health sector must be accelerated to create an efficient and sustainable predict-prevent healthcare system.

Profiling Disease and Economic Burden in CRSwNP Using Machine Learning

Purpose

Chronic rhinosinusitis with nasal polyps (CRSwNP) is associated with high healthcare resource utilization (HRU) and economic cost; however, heterogeneity of clinical burden among patients with differing clinical characteristics has not been fully elucidated. Here, an unsupervised machine learning approach supported by clinical validation identified distinct clusters of patients with CRSwNP and compared healthcare burden.

Patients and Methods

This retrospective analysis identified adult patients with ≥2 claims for CRSwNP and date of first diagnosis (index date) between January 2015 and June 2019 from a healthcare database. Patients were required to have enrollment in the database 6-months pre- and 12-months post-index. Patients were assigned to clusters using latent class analysis. All-cause and nasal polyp (NP)-related HRU and costs were compared between clusters.

Results

Among 12,807 patients, 5 clusters were identified: cluster 1: no surgery/low comorbidity/low medication use (n = 4076); cluster 2: no surgery/low comorbidity/high medication use (n = 2201); cluster 3: no surgery/high comorbidity/high medication use (n = 2093); cluster 4: surgery/low comorbidity/moderate medication use (n = 3168); cluster 5: surgery/high comorbidity/high medication use (n = 1269). All-cause HRU was similar across clusters. NP-related HRU was highest in the surgical clusters (clusters 4 and 5). All-cause costs were similar in clusters 1–3 ($15,833–$17,461) and highest in clusters 4 ($31,083) and 5 ($31,103), driven by outpatient costs. Total NP-related costs were also highest for clusters 4 and 5 ($14,193 and $16,100, respectively).

Conclusion

Substantial heterogeneity exists in clinical and economic burden among patients with CRSwNP. Machine learning offers a novel approach to better understand the diverse, complex burden of illness in CRSwNP.

Early Detection of COVID-19 outbreaks using Textual Analysis of Electronic Medical Records

Purpose

Our objective was to develop a tool promoting early detection of COVID-19 cases by focusing epidemiological investigations and PCR examinations during a period of limited testing capabilities.

Methods

We developed an algorithm for analyzing medical records recorded by healthcare providers in the Israeli Defense Forces. The algorithm utilized textual analysis to detect patients presenting with suspicious symptoms and was tested among 92 randomly selected units. Detection of a potential cluster of patients in a unit prompted a focused epidemiological investigation aided by data provided by the algorithm.

Results

During a month of follow up, the algorithm has flagged 17 of the units for investigation. The subsequent epidemiological investigations led to the testing of 78 persons and the detection of eight cases in four clusters that were previously gone unnoticed. The resulting positive test rate of 10.25% was five time higher than the IDF average at the time of the study. No cases of COVID-19 in the examined units were missed by the algorithm.

Conclusions

This study depicts the successful development and large scale deployment of a textual analysis based algorithm for early detection of COVID-19 cases, demonstrating the potential of natural language processing of medical text as a tool for promoting public health.

Artificial intelligence-enabled decision support in nephrology

Kidney pathophysiology is often complex, nonlinear and heterogeneous, which limits the utility of hypothetical-deductive reasoning and linear, statistical approaches to diagnosis and treatment. Emerging evidence suggests that artificial intelligence (AI)-enabled decision support systems - which use algorithms based on learned examples - may have an important role in nephrology. Contemporary AI applications can accurately predict the onset of acute kidney injury before notable biochemical changes occur; can identify modifiable risk factors for chronic kidney disease onset and progression; can match or exceed human accuracy in recognizing renal tumours on imaging studies; and may augment prognostication and decision-making following renal transplantation. Future AI applications have the potential to make real-time, continuous recommendations for discrete actions and yield the greatest probability of achieving optimal kidney health outcomes. Realizing the clinical integration of AI applications will require cooperative, multidisciplinary commitment to ensure algorithm fairness, overcome barriers to clinical implementation, and build an AI-competent workforce. AI-enabled decision support should preserve the pre-eminence of wisdom and augment rather than replace human decision-making. By anchoring intuition with objective predictions and classifications, this approach should favour clinician intuition when it is honed by experience.

Associations between Florida counties' COVID-19 case and death rates and meaningful use among Medicaid providers: Cross-sectional ecologic study

Although the Health Information Technology for Economic and Clinical Health (HITECH) Act has accelerated adoption of Electronic Health Records (EHRs) among Medicaid providers, only half achieved Meaningful Use. Furthermore, Meaningful Use' impact on reporting and/or clinical outcomes remains unknown. To address this deficit, we assessed the difference between Medicaid providers who did and did not achieve Meaningful Use regarding Florida county-level cumulative COVID-19 death, case and case fatality rates (CFR), accounting for county-level demographics, socioeconomic and clinical markers, and healthcare environment. We found that cumulative incidence rates of COVID-19 deaths and CFRs were significantly different between the 5025 Medicaid providers not achieving Meaningful Use and the 3723 achieving Meaningful Use (mean 0.8334/1000 population; SD = 0.3489 vs. mean = 0.8216/1000; SD = 0.3227, respectively) (P = .01). CFRs were .01797 and .01781, respectively, P = .04. County-level characteristics independently associated with increased COVID-19 death rates and CFRs include greater concentrations of persons of African American or Black race, lower median household income, higher unemployment, and higher concentrations of those living in poverty and without health insurance (all P < .001). In accordance with other studies, social determinants of health were independently associated with clinical outcomes. Our findings also suggest that the association between Florida counties' public health outcomes and Meaningful Use achievement may have had less to do with using EHRs for reporting of clinical outcomes and more to do with using EHRs for coordination of care-a key measure of quality. The Florida Medicaid Promoting Interoperability Program which incentivized Medicaid providers towards achieving Meaningful Use, has demonstrated success regarding both rates of adoption and clinical outcomes. Because the Program ends in 2021, we support programs such as HealthyPeople 2030 Health IT which address the remaining half of Florida Medicaid providers who have not yet achieved Meaningful Use.

Clinical Network for Big Data and Personalized Health: Study Protocol and Preliminary Results

The use of secondary hospital-based clinical data and electronical health records (EHR) represent a cost-efficient alternative to investigate chronic conditions. We present the Clinical Network Big Data and Personalised Health project, which collects EHRs for patients accessing hospitals in Central-Southern Italy, through an integrated digital platform to create a digital hub for the collection, management and analysis of personal, clinical and environmental information for patients, associated with a biobank to perform multi-omic analyses. A total of 12,864 participants (61.7% women, mean age 52.6 ± 17.6 years) signed a written informed consent to allow access to their EHRs. The majority of hospital access was in obstetrics and gynaecology (36.3%), while the main reason for hospitalization was represented by diseases of the circulatory system (21.2%). Participants had a secondary education (63.5%), were mostly retired (25.45%), reported low levels of physical activity (59.6%), had low adherence to the Mediterranean diet and were smokers (30.2%). A large percentage (35.8%) were overweight and the prevalence of hypertension, diabetes and hyperlipidemia was 36.4%, 11.1% and 19.6%, respectively. Blood samples were retrieved for 8686 patients (67.5%). This project is aimed at creating a digital hub for the collection, management and analysis of personal, clinical, diagnostic and environmental information for patients, and is associated with a biobank to perform multi-omic analyses.

Estimating prevalence of child and youth mental disorder and mental health-related service contacts: a comparison of survey data and linked administrative health data

AIMS

Prevalence estimates of child and youth mental disorder and mental health-related service contacts are needed for policy formulation, research, advocacy and resource allocation. Our aim is to compare prevalence estimates of child and youth mental disorder and mental health-related service contacts derived from general population survey data v. linked administrative health data.

METHODS

Provincially representative 2014 Ontario Child Health Study data were linked to administrative health records for 5563 children and youth aged 4-17 in Ontario. Emotional disorders (mood and anxiety) and attention-deficit/hyperactivity disorder were assessed using a standardised diagnostic interview in the survey and using diagnostic codes in administrative health data. Physician-based mental health-related service contacts were assessed using parent self-reports from the survey and administrative data related to mental health-related diagnostic codes. Prevalence estimates were calculated and compared based on one-sample z-tests and ratios of survey data to administrative data-based prevalence. Sensitivity, specificity and agreement between classifications were compared using κ. Prevalence estimates were calculated by age, sex and geography sub-groups and consistent group differences across data source were counted.

RESULTS

Disorder prevalence and service contact estimates were significantly higher in survey data in all cases, except for mood disorder. Ratios of survey data to administrative data-based prevalence varied, ranging from 0.80 (mood) to 11.01 (attention-deficit/hyperactivity disorder). Specificity was high (0.98-1.00), sensitivity was low (0.07-0.41) and agreement ranged from slight (κ = 0.13) to moderate (κ = 0.46). Out of 18 sub-group difference comparisons, half were non-significant in either data source. In the remaining nine comparisons, the only significant differences between groups that were consistent across data source were for sex-based differences (attention-deficit/hyperactivity disorder and service contacts). There were no consistent age- or geography-based differences in prevalence across data sources.

CONCLUSIONS

Our findings suggest that conclusions drawn about prevalence, service contacts and sub-group differences in these estimates are dependent on data source. Further research is needed to understand who and what is being captured by each source. Researchers should conduct data linkage where possible to access and compare multiple sources of information.

Evaluating peripheral intravascular catheter insertion, maintenance and removal practices in small hospitals using a standardized audit tool

AIM

The aim of this study was to evaluate clinical practice about peripheral intravenous catheter (PIVC) insertion, maintenance and removal in a cohort of Victorian hospitals.

DESIGN

A standardized PIVC audit tool was developed, and results from point prevalent surveys were conducted.

METHODS

Hospitalized patients requiring a PIVC insertion were eligible for audit. Audit data submitted between 2015 and 2019 were extracted for the current study.

RESULTS

3566 PIVC insertions in 15 Victorian public hospitals were evaluated. 57.6% of PIVCs were inserted in wards, 18.7% in operating theatres and 11.6% in Emergency Departments (ED). 45.2% were inserted by nurses and 38.2% by medical staff. The preferred site for insertion was the dorsum of the hand and forearm (58.8%). 22.6% did not report a visual infusion phlebitis score at least daily, and 48% did not document a daily dressing assessment. Reasons for PIVC removal included no longer required (63%) and phlebitis (4.8%). No bloodstream infections were reported.

Hospitalisation with injection-related infections: Validation of diagnostic codes to monitor admission trends at a tertiary care hospital in Melbourne, Australia

INTRODUCTION

Injection-related infections (IRI) cause morbidity and mortality in people who inject drugs. Hospital administrative datasets can be used to describe hospitalisation trends, but there are no validated algorithms to identify injecting drug use and IRIs. We aimed to validate International Classification of Diseases (ICD) codes to identify admissions with IRIs and use these codes to describe IRIs within our hospital.

METHODS

We developed a candidate set of ICD codes to identify current injecting drug use and IRI and extracted admissions satisfying both criteria. We then used manual chart review data from 1 January 2017 to 30 April 2019 to evaluate the performance of these codes and refine our algorithm by selecting codes with a high-positive predictive value (PPV). We used the refined algorithm to describe trends and outcomes of people who inject drugs with an IRI at Alfred Hospital, Melbourne from 2008 to 2020.

RESULTS

Current injecting drug use was best predicted by opioid-related disorders (F11), 80% (95% confidence interval [CI] 74-85%), and other stimulant-related disorders (F15), 82% (95% CI 70-90%). All PPVs were ≥67% to identify specific IRIs, and ≥84% for identifying any IRI. Using these codes over 12 years, IRIs increased from 138 to 249 per 100 000 admissions, and skin and soft tissues infections (SSTI) were the most common (797/1751, 46%).

DISCUSSION AND CONCLUSION

Validated ICD-based algorithms can inform passive surveillance systems. Strategies to reduce hospitalisation with IRIs should be supported by early intervention and prevention, particularly for SSTIs which may represent delayed access to care.

Precision Public Health for Non-communicable Diseases: An Emerging Strategic Roadmap and Multinational Use Cases

Non-communicable diseases (NCDs) remain the largest global public health threat. The emerging field of precision public health (PPH) offers a transformative opportunity to capitalize on digital health data to create an agile, responsive and data-driven public health system to actively prevent NCDs. Using learnings from digital health, our aim is to propose a vision toward PPH for NCDs across three horizons of digital health transformation: Horizon 1—digital public health workflows; Horizon 2—population health data and analytics; Horizon 3—precision public health. This perspective provides a high-level strategic roadmap for public health practitioners and policymakers, health system stakeholders and researchers to achieving PPH for NCDs. Two multinational use cases are presented to contextualize our roadmap in pragmatic action: ESP and RiskScape (USA), a mature PPH platform for multiple NCDs, and PopHQ (Australia), a proof-of-concept population health informatics tool to monitor and prevent obesity. Our intent is to provide a strategic foundation to guide new health policy, investment and research in the rapidly emerging but nascent area of PPH to reduce the public health burden of NCDs.

Minnesota Electronic Health Record Consortium COVID-19 Project: Informing Pandemic Response Through Statewide Collaboration Using Observational Data

OBJECTIVE

Robust disease and syndromic surveillance tools are underdeveloped in the United States, as evidenced by limitations and heterogeneity in sociodemographic data collection throughout the COVID-19 pandemic. To monitor the COVID-19 pandemic in Minnesota, we developed a federated data network in March 2020 using electronic health record (EHR) data from 8 multispecialty health systems.

MATERIALS AND METHODS

In this serial cross-sectional study, we examined patients of all ages who received a COVID-19 polymerase chain reaction test, had symptoms of a viral illness, or received an influenza test from January 3, 2016, through November 7, 2020. We evaluated COVID-19 testing rates among patients with symptoms of viral illness and percentage positivity among all patients tested, in aggregate and by zip code. We stratified results by patient and area-level characteristics.

RESULTS

Cumulative COVID-19 positivity rates were similar for people aged 12-64 years (range, 15.1%-17.6%) but lower for adults aged ≥65 years (range, 9.3%-10.7%). We found notable racial and ethnic disparities in positivity rates early in the pandemic, whereas COVID-19 positivity was similarly elevated across most racial and ethnic groups by the end of 2020. Positivity rates remained substantially higher among Hispanic patients compared with other racial and ethnic groups throughout the study period. We found similar trends across area-level income and rurality, with disparities early in the pandemic converging over time.

PRACTICE IMPLICATIONS

We rapidly developed a distributed data network across Minnesota to monitor the COVID-19 pandemic. Our findings highlight the utility of using EHR data to monitor the current pandemic as well as future public health priorities. Building partnerships with public health agencies can help ensure data streams are flexible and tailored to meet the changing needs of decision makers.

Using inpatient electronic medical records to study influenza for pandemic preparedness

BACKGROUND

We assessed the ability to identify key data relevant to influenza and other respiratory virus surveillance in a large-scale US-based hospital electronic medical record (EMR) dataset using seasonal influenza as a use case. We describe characteristics and outcomes of hospitalized influenza cases across three seasons.

METHODS

We identified patients with an influenza diagnosis between March 2017 and March 2020 in 140 US hospitals as part of the US FDA's Sentinel System. We calculated descriptive statistics on the presence of high-risk conditions, influenza antiviral administrations, and severity endpoints.

RESULTS

Among 5.1 million hospitalizations, we identified 29,520 hospitalizations with an influenza diagnosis; 64% were treated with an influenza antiviral within 2 days of admission, and 25% were treated >2 days after admission. Patients treated >2 days after admission had more comorbidities than patients treated within 2 days of admission. Patients never treated during hospitalization had more documentation of cardiovascular and other diseases than treated patients. We observed more severe endpoints in patients never treated (death = 3%, mechanical ventilation [MV] = 9%, intensive care unit [ICU] = 26%) or patients treated >2 days after admission (death = 2%, MV = 14%, ICU = 32%) than in patients treated earlier (treated on admission: death = 1%, MV = 5%, ICU = 23%, treated within 2 days of admission: death = 1%, MV = 7%, ICU = 27%).

CONCLUSIONS

We identified important trends in influenza severity related to treatment timing in a large inpatient dataset, laying the groundwork for the use of this and other inpatient EMR data for influenza and other respiratory virus surveillance.

Multi-channel Fusion LSTM for Medical Event Prediction using HERs

Automatic medical event prediction (MEP), e.g. diagnosis prediction, medication prediction, using electronic health records (EHRs) is a popular research direction in health informatics. In many cases, MEP relies on the determinations from different types of medical events, which demonstrates the heterogeneous nature of EHRs. However, most existing methods for MEP fail to distinguishingly model the type of event that is highly associated with the prediction task, i.e. task-wise event, which usually plays a more significant role than other events. In this paper, we proposed a Long Short-Term Memory network (LSTM)-based method for MEP, named Multi-Channel Fusion LSTM (MCF-LSTM), which models the correlations between different types of medical events using multiple network channels. To this end, we designed a task-wise fusion module, in which a gated network is applied to select how much information can be transferred between events. Furthermore, the irregular temporal interval between adjacent medical visits is also modeled in an individual channel, which is combined with other events in a unified manner. We compared MCF-LSTM with state-of-the-art methods on four MEP tasks on two public datasets: MIMIC-III and eICU. Experimental results show that MCF-LSTM achieves promising results on AUC(receiver operating characteristic curve), AUPR (area under the precision-recall curve), and top-k recall, and outperforms other methods with high stability.