Advancing the Use of Emergency Department Syndromic Surveillance Data, New York City, 2012-2016

Recent Advances in the Use of the Mortality Syndromic Surveillance System-New York City, 2015-2020

OBJECTIVE

New York City's automated mortality syndromic surveillance system monitors temporal and spatial patterns in mortality. To describe the use of the syndromic surveillance system, we used the system to find mortality patterns for the 15 leading causes of death and for deaths from rare and reportable diseases in New York City from February 2015 through June 2020. We used results to find aberrations that indicate threats to public health.

METHODS

We used unobserved components models to analyze time series of mortality counts for leading causes of death, historical limits methods for rare and reportable diseases, and SaTScan for temporal-spatial cluster analysis. We obtained data on the number of deaths from the electronic death registry system maintained by the city's Bureau of Vital Statistics.

RESULTS

The mortality syndromic surveillance system detected an increase in the number of deaths from heart disease by April 1, 2020, when 75.0 deaths occurred on March 24, 2020, instead of an expected 45.8 deaths (95% upper prediction limit of 61.0) and an increase in the number of deaths from all causes on March 20, 2020, when 194.0 deaths were observed while 150.1 deaths were expected (95% upper prediction limit of 178.0). The number of deaths from all causes returned to normal the week beginning June 14, 2020, when 990.0 deaths were observed and 998.8 deaths were expected.

PRACTICE IMPLICATIONS

When compared with efforts from New York City to provide yearly vital statistics, the automated mortality syndromic surveillance system can provide timely mortality data with fewer resources and raise the capacity to detect anomalous increases in mortality.

Canadian Wildfire Smoke and Asthma Syndrome Emergency Department Visits in New York City

This study examines the association between the Canadian wildfires that occurred in summer 2023 with emergency department visits for asthma symptoms in New York City.

Evaluation of an Emergency Department Visit Data Mental Health Dashboard

CONTEXT

Local health departments (LHDs) need timely county-level and subcounty-level data to monitor health-related trends, identify health disparities, and inform areas of highest need for interventions as part of their ongoing assessment responsibilities; yet, many health departments rely on secondary data that are not timely and cannot provide subcounty insights.

OBJECTIVE

We developed and evaluated a mental health dashboard in Tableau for an LHD audience featuring statewide syndromic surveillance emergency department (ED) data in North Carolina from the North Carolina Disease Event Tracking and Epidemiologic Collection Tool (NC DETECT).

DESIGN

We developed a dashboard that provides counts, crude rates, and ED visit percentages at statewide and county levels, as well as breakdowns by zip code, sex, age group, race, ethnicity, and insurance coverage for 5 mental health conditions. We evaluated the dashboards through semistructured interviews and a Web-based survey that included the standardized usability questions from the System Usability Scale.

PARTICIPANTS

Convenience sample of LHD public health epidemiologists, health educators, evaluators, and public health informaticians.

RESULTS

Six semistructured interview participants successfully navigated the dashboard but identified usability issues when asked to compare county-level trends displayed in different outputs (eg, tables vs graphs). Thirty respondents answered all questions on the System Usability Scale for the dashboard, which received an above average score of 86.

CONCLUSIONS

The dashboards scored well on the System Usability Scale, but more research is needed to identify best practices in disseminating multiyear syndromic surveillance ED visit data on mental health conditions to LHDs.

Development and Evaluation of Syndromic Surveillance Definitions for Fall- and Hip Fracture-Related Emergency Department Visits Among Adults Aged 65 Years and Older, United States 2017-2018

OBJECTIVE

To develop syndromic surveillance definitions for unintentional fall- and hip fracture-related emergency department (ED) visits among older adults (aged ≥65 years) for use in the Centers for Disease Control and Prevention's National Syndromic Surveillance Program (NSSP) data and compare the percentage of ED visits captured using these new syndromes with ED visits from the Healthcare Cost and Utilization Project Nationwide Emergency Department Sample (HCUP-NEDS), a nationally representative administrative data set.

DESIGN/SETTING

Syndromic definitions were developed using chief complaint terms and discharge diagnosis codes in NSSP data. The percentages of ED visits among older adults related to falls and hip fractures in NSSP were compared with the percentages in HCUP-NEDS in 2017 and 2018.

MEASURES

Prevalence ratios were calculated as the relative difference in the percentage of ED visits related to falls or hip fractures in NSSP compared with HCUP-NEDS. Counts and percentages calculated using HCUP-NEDS were weighted to produce nationally representative estimates. Data were analyzed overall and by sex and age group.

RESULTS

The percentage of ED visits among older adults related to falls in NSSP was 12% less in 2017 (10.81%) and 7% less in 2018 (11.42%) compared with HCUP-NEDS (2017: 12.30%; 2018: 12.26%). The percentage of ED visits among older adults related to hip fractures in NSSP was 41% less in 2017 (0.65%) and 30% less in 2018 (0.76%) compared with HCUP-NEDS (2017: 1.10%; 2018: 1.09%). In both 2017 and 2018, a higher percentage of ED visits among older women and adults aged 85 years or older were related to falls or hip fractures compared with older men and younger age groups across both data sets.

CONCLUSION

A smaller percentage of older adults' ED visits met the falls and hip fracture definitions in NSSP compared with HCUP-NEDS in 2017 and 2018. However, demographic trends remained similar across both data sets.

Effectiveness of early warning systems in the detection of infectious diseases outbreaks: a systematic review

Background

Global pandemics have occurred with increasing frequency over the past decade reflecting the sub-optimum operationalization of surveillance systems handling human health data. Despite the wide array of current surveillance methods, their effectiveness varies with multiple factors. Here, we perform a systematic review of the effectiveness of alternative infectious diseases Early Warning Systems (EWSs) with a focus on the surveillance data collection methods, and taking into consideration feasibility in different settings.

Methods

We searched PubMed and Scopus databases on 21 October 2022. Articles were included if they covered the implementation of an early warning system and evaluated infectious diseases outbreaks that had potential to become pandemics. Of 1669 studies screened, 68 were included in the final sample. We performed quality assessment using an adapted CASP Checklist.

Results

Of the 68 articles included, 42 articles found EWSs successfully functioned independently as surveillance systems for pandemic-wide infectious diseases outbreaks, and 16 studies reported EWSs to have contributing surveillance features through complementary roles. Chief complaints from emergency departments’ data is an effective EWS but it requires standardized formats across hospitals. Centralized Public Health records-based EWSs facilitate information sharing; however, they rely on clinicians’ reporting of cases. Facilitated reporting by remote health settings and rapid alarm transmission are key advantages of Web-based EWSs. Pharmaceutical sales and laboratory results did not prove solo effectiveness. The EWS design combining surveillance data from both health records and staff was very successful. Also, daily surveillance data notification was the most successful and accepted enhancement strategy especially during mass gathering events. Eventually, in Low Middle Income Countries, working to improve and enhance existing systems was more critical than implementing new Syndromic Surveillance approaches.

Conclusions

Our study was able to evaluate the effectiveness of Early Warning Systems in different contexts and resource settings based on the EWSs’ method of data collection. There is consistent evidence that EWSs compiling pre-diagnosis data are more proactive to detect outbreaks. However, the fact that Syndromic Surveillance Systems (SSS) are more proactive than diagnostic disease surveillance should not be taken as an effective clue for outbreaks detection.

Early Detection of SARS-CoV-2 Epidemic Waves: Lessons from the Syndromic Surveillance in Lombardy, Italy

We evaluated the performance of the exponentially weighted moving average (EWMA) model for comparing two families of predictors (i.e., structured and unstructured data from visits to the emergency department (ED)) for the early detection of SARS-CoV-2 epidemic waves. The study included data from 1,282,100 ED visits between 1 January 2011 and 9 December 2021 to a local health unit in Lombardy, Italy. A regression model with an autoregressive integrated moving average (ARIMA) error term was fitted. EWMA residual charts were then plotted to detect outliers in the frequency of the daily ED visits made due to the presence of a respiratory syndrome (based on coded diagnoses) or respiratory symptoms (based on free text data). Alarm signals were compared with the number of confirmed SARS-CoV-2 infections. Overall, 150,300 ED visits were encoded as relating to respiratory syndromes and 87,696 to respiratory symptoms. Four strong alarm signals were detected in March and November 2020 and 2021, coinciding with the onset of the pandemic waves. Alarm signals generated for the respiratory symptoms preceded the occurrence of the first and last pandemic waves. We concluded that the EWMA model is a promising tool for predicting pandemic wave onset.

An Evaluation of Syndromic Surveillance-Related Practices Among Selected State and Local Health Agencies

CONTEXT

Syndromic surveillance consists of the systematic collection and use of near real-time data about health-related events for situational awareness and public health action. As syndromic surveillance programs continue to adopt new technologies and expand, it is valuable to evaluate these syndromic surveillance systems and practices to ensure that they meet public health needs.

OBJECTIVE

This assessment's aim is to provide recent information about syndromic surveillance systems and practice characteristics among a group of state and local health departments.

DESIGN/SETTING

Information was obtained between November 2017 and June 2018 through a telephone survey using an Office of Management and Budget-approved standardized data collection tool. Participants were syndromic surveillance staff from each of 31 state and local health departments participating in the National Syndromic Surveillance Program funded by the Centers for Disease Control and Prevention. Questions included jurisdictional experience, data sources and analysis systems used, syndromic system data processing characteristics, data quality verification procedures, and surveillance activities conducted with syndromic data.

MEASURES

Practice-specific information such as types of systems and data sources used for syndromic surveillance, data quality monitoring, and uses of data for public health situational awareness (eg, investigating occurrences of or trends in diseases).

RESULTS

The survey analysis revealed a wide range of experiences with syndromic surveillance. Participants reported the receipt of data daily or more frequently. Emergency department data were the primary data source; however, other data sources are being integrated into these systems. All health departments routinely monitored data quality. Syndromes of highest priority across the respondents for health events monitoring were influenza-like illness and drug-related syndromes. However, a wide variety of syndromes were reported as priorities across the health departments.

CONCLUSION

Overall, syndromic surveillance was relevantly integrated into the public health surveillance infrastructure. The near real-time nature of the data and its flexibility to monitor different types of health-related issues make it especially useful for public health practitioners. Despite these advances, syndromic surveillance capacity, locally and nationally, must continue to evolve and progress should be monitored to ensure that syndromic surveillance systems and data are optimally able to meet jurisdictional needs.

Epidemiological characteristics of the B.1.526 SARS-CoV-2 variant

To characterize the epidemiological properties of the B.1.526 SARS-CoV-2 (severe acute respiratory syndrome coronavirus 2) variant of interest, here we used nine epidemiological and population datasets and model-inference methods to reconstruct SARS-CoV-2 transmission dynamics in New York City, where B.1.526 emerged. We estimated that B.1.526 had a moderate increase (15 to 25%) in transmissibility, could escape immunity in 0 to 10% of previously infected individuals, and substantially increased the infection fatality risk (IFR) among adults 65 or older by >60% during November 2020 to April 2021, compared to estimates for preexisting variants. Overall, findings suggest that new variants like B.1.526 likely spread in the population weeks before detection and that partial immune escape (e.g., resistance to therapeutic antibodies) could offset prior medical advances and increase IFR. Early preparedness for and close monitoring of SARS-CoV-2 variants, their epidemiological characteristics, and disease severity are thus crucial to COVID-19 (coronavirus disease 2019) response.

Correlation-Based Discovery of Disease Patterns for Syndromic Surveillance

Early outbreak detection is a key aspect in the containment of infectious diseases, as it enables the identification and isolation of infected individuals before the disease can spread to a larger population. Instead of detecting unexpected increases of infections by monitoring confirmed cases, syndromic surveillance aims at the detection of cases with early symptoms, which allows a more timely disclosure of outbreaks. However, the definition of these disease patterns is often challenging, as early symptoms are usually shared among many diseases and a particular disease can have several clinical pictures in the early phase of an infection. As a first step toward the goal to support epidemiologists in the process of defining reliable disease patterns, we present a novel, data-driven approach to discover such patterns in historic data. The key idea is to take into account the correlation between indicators in a health-related data source and the reported number of infections in the respective geographic region. In an preliminary experimental study, we use data from several emergency departments to discover disease patterns for three infectious diseases. Our results show the potential of the proposed approach to find patterns that correlate with the reported infections and to identify indicators that are related to the respective diseases. It also motivates the need for additional measures to overcome practical limitations, such as the requirement to deal with noisy and unbalanced data, and demonstrates the importance of incorporating feedback of domain experts into the learning procedure.

Population-Based Estimates of Coronavirus Disease 2019 (COVID-19)-like Illness, COVID-19 Illness, and Rates of Case Ascertainment, Hospitalizations, and Deaths-Noninstitutionalized New York City Residents, March-April 2020

Using a population-based, representative telephone survey, ~930 000 New York City residents had COVID-19 illness beginning 20 March–30 April 2020, a period with limited testing. For every 1000 persons estimated with COVID-19 illness, 141.8 were tested and reported as cases, 36.8 were hospitalized, and 12.8 died, varying by demographic characteristics.

A Dispatch Screening Tool to Identify Patients at High Risk for COVID-19 in the Prehospital Setting

INTRODUCTION

Emergency medical services (EMS) dispatchers have made efforts to determine whether patients are high risk for coronavirus disease 2019 (COVID-19) so that appropriate personal protective equipment (PPE) can be donned. A screening tool is valuable as the healthcare community balances protection of medical personnel and conservation of PPE. There is little existing literature on the efficacy of prehospital COVID-19 screening tools. The objective of this study was to determine the positive and negative predictive value of an emergency infectious disease surveillance tool for detecting COVID-19 patients and the impact of positive screening on PPE usage.

METHODS

This study was a retrospective chart review of prehospital care reports and hospital electronic health records. We abstracted records for all 911 calls to an urban EMS from March 1-July 31, 2020 that had a documented positive screen for COVID-19 and/or had a positive COVID-19 test. The dispatch screen solicited information regarding travel, sick contacts, and high-risk symptoms. We reviewed charts to determine dispatch-screening results, the outcome of patients' COVID-19 testing, and documentation of crew fidelity to PPE guidelines.

RESULTS

The sample size was 263. The rate of positive COVID-19 tests for all-comers in the state of Massachusetts was 2.0%. The dispatch screen had a sensitivity of 74.9% (confidence interval [CI], 69.21-80.03) and a specificity of 67.7% (CI, 66.91-68.50). The positive predictive value was 4.5% (CI, 4.17-4.80), and the negative predictive value was 99.3% (CI, 99.09-99.40). The most common symptom that triggered a positive screen was shortness of breath (51.5% of calls). The most common high-risk population identified was skilled nursing facility patients (19.5%), but most positive tests did not belong to a high-risk population (58.1%). The EMS personnel were documented as wearing full PPE for the patient in 55.7% of encounters, not wearing PPE in 8.0% of encounters, and not documented in 27.9% of encounters.

CONCLUSION

This dispatch-screening questionnaire has a high negative predictive value but moderate sensitivity and therefore should be used with some caution to guide EMS crews in their PPE usage. Clinical judgment is still essential and may supersede screening status.

Emergency department syndromic surveillance systems: a systematic review

BACKGROUND

Syndromic surveillance provides public health intelligence to aid in early warning and monitoring of public health impacts (e.g. seasonal influenza), or reassurance when an impact has not occurred. Using information collected during routine patient care, syndromic surveillance can be based on signs/symptoms/preliminary diagnoses. This approach makes syndromic surveillance much timelier than surveillance requiring laboratory confirmed diagnoses. The provision of healthcare services and patient access to them varies globally. However, emergency departments (EDs) exist worldwide, providing unscheduled urgent care to people in acute need. This provision of care makes ED syndromic surveillance (EDSyS) a potentially valuable tool for public health surveillance internationally. The objective of this study was to identify and describe the key characteristics of EDSyS systems that have been established and used globally.

METHODS

We systematically reviewed studies published in peer review journals and presented at International Society of Infectious Disease Surveillance conferences (up to and including 2017) to identify EDSyS systems which have been created and used for public health purposes. Search criteria developed to identify "emergency department" and "syndromic surveillance" were applied to NICE healthcare, Global Health and Scopus databases.

RESULTS

In total, 559 studies were identified as eligible for inclusion in the review, comprising 136 journal articles and 423 conference abstracts/papers. From these studies we identified 115 EDSyS systems in 15 different countries/territories across North America, Europe, Asia and Australasia. Systems ranged from local surveillance based on a single ED, to comprehensive national systems. National EDSyS systems were identified in 8 countries/territories: 2 reported inclusion of ≥85% of ED visits nationally (France and Taiwan).

CONCLUSIONS

EDSyS provides a valuable tool for the identification and monitoring of trends in severe illness. Technological advances, particularly in the emergency care patient record, have enabled the evolution of EDSyS over time. EDSyS reporting has become closer to 'real-time', with automated, secure electronic extraction and analysis possible on a daily, or more frequent basis. The dissemination of methods employed and evidence of successful application to public health practice should be encouraged to support learning from best practice, enabling future improvement, harmonisation and collaboration between systems in future.

PROSPERO NUMBER

CRD42017069150 .

Framework for a Community Health Observing System for the Gulf of Mexico Region: Preparing for Future Disasters

The Gulf of Mexico (GoM) region is prone to disasters, including recurrent oil spills, hurricanes, floods, industrial accidents, harmful algal blooms, and the current COVID-19 pandemic. The GoM and other regions of the U.S. lack sufficient baseline health information to identify, attribute, mitigate, and facilitate prevention of major health effects of disasters. Developing capacity to assess adverse human health consequences of future disasters requires establishment of a comprehensive, sustained community health observing system, similar to the extensive and well-established environmental observing systems. We propose a system that combines six levels of health data domains, beginning with three existing, national surveys and studies plus three new nested, longitudinal cohort studies. The latter are the unique and most important parts of the system and are focused on the coastal regions of the five GoM States. A statistically representative sample of participants is proposed for the new cohort studies, stratified to ensure proportional inclusion of urban and rural populations and with additional recruitment as necessary to enroll participants from particularly vulnerable or under-represented groups. Secondary data sources such as syndromic surveillance systems, electronic health records, national community surveys, environmental exposure databases, social media, and remote sensing will inform and augment the collection of primary data. Primary data sources will include participant-provided information via questionnaires, clinical measures of mental and physical health, acquisition of biological specimens, and wearable health monitoring devices. A suite of biomarkers may be derived from biological specimens for use in health assessments, including calculation of allostatic load, a measure of cumulative stress. The framework also addresses data management and sharing, participant retention, and system governance. The observing system is designed to continue indefinitely to ensure that essential pre-, during-, and post-disaster health data are collected and maintained. It could also provide a model/vehicle for effective health observation related to infectious disease pandemics such as COVID-19. To our knowledge, there is no comprehensive, disaster-focused health observing system such as the one proposed here currently in existence or planned elsewhere. Significant strengths of the GoM Community Health Observing System (CHOS) are its longitudinal cohorts and ability to adapt rapidly as needs arise and new technologies develop.

Opportunities and Challenges for Developing Syndromic Surveillance Systems for the Detection of Social Epidemics

This commentary explores the potential and challenges of developing syndromic surveillance systems with the ability to more rapidly detect epidemics of addiction, poverty, housing instability, food insecurity, social isolation and other social determinants of health (SDoH). Epidemiologists tracking SDoH heavily rely on expensive government surveys released annually, delaying for months if not years the timely detection of social epidemics, defined as sudden, rapid or unexpected changes in social determinants of population health. Conversely, infectious disease syndromic surveillance is an effective early warning tool for epidemic diseases using various types of non-traditional epidemiological data from emergency room chief complaints to search query data. Based on such experience, novel social syndromic surveillance systems for early detection of social epidemics with health implications are not only possible but necessary. Challenges to their widespread implementation include incorporating disparate proprietary data sources and database integration. Significantly more resources are critically needed to address these barriers to allow for accessing, integrating and rapidly analyzing appropriate data streams to make syndromic surveillance for social determinants of health widely available to public health professionals.

Designing a Syndromic Bovine Mortality Surveillance System: Lessons Learned From the 1-Year Test of the French OMAR Alert Tool

Between May 2018 and 2019, a syndromic bovine mortality surveillance system (OMAR) was tested in 10 volunteer French départements (French intermediate-level administrative unit) to assess its performance in real conditions, as well as the human and financial resources needed to ensure normal functioning. The system is based on the automated weekly analysis of the number of cattle deaths reported by renderers in the Fallen Stock Data Interchange Database established in January 2011. In our system, every Thursday, the number of deaths is grouped by ISO week and small surveillance areas and then analyzed using traditional time-series analysis steps (cleaning, prediction, signal detection). For each of the five detection algorithms implemented (i.e., the exponentially weighted moving average chart, cumulative sum chart, Shewhart chart, Holt-Winters, and historical limits algorithms), seven detection limits are applied, giving a signal score from 1 (low excess mortality) to 7 (high excess mortality). The severity of excess mortality (alarm) is then classified into four categories, from very low to very high, by combining the signal scores, the relative excess mortality, and the persistence of the signal(s) over the previous 4 weeks. Detailed and interactive weekly reports and a short online questionnaire help pilot départements and the OMAR central coordination cell assess the performance of the system. During the 1-year test, the system showed highly variable sensitivity among départements. This variability was partly due not only to the demographic distribution of cattle (very few signals in low-density areas) but also to the renderer's delay in reporting to the Fallen Stock Data Interchange Database (on average, only 40% of the number of real deaths had been transmitted within week, with huge variations among départements). As a result, in the pilot départements, very few alarms required on-farm investigation and excess mortality often involved a small number of farms already known to have health or welfare problems. Despite its perfectibility, the system nevertheless proved useful in the daily work of animal health professionals for collective and individual surveillance. The test is still ongoing for a second year in nine départements to evaluate the effectiveness of the improvements agreed upon at the final meeting.

Monitoring Emergency Department Visits From Puerto Rico in the Aftermath of Hurricane Maria Using Syndromic Surveillance - New York City, 2017

OBJECTIVE

Syndromic surveillance has been useful for routine surveillance on a variety of health outcomes and for informing situational awareness during public health emergencies. Following the landfall of Hurricane Maria in 2017, the New York City (NYC) Department of Health and Mental Hygiene (DOHMH) implemented an enhanced syndromic surveillance system to characterize related emergency department (ED) visits.

METHODS

ED visits with any mention of specific key words ("Puerto," "Rico," "hurricane," "Maria") in the ED chief complaint or Puerto Rico patient home Zip Code were identified from the DOHMH syndromic surveillance system in the 8-week window leading up to and following landfall. Visit volume comparisons pre- and post-Hurricane Maria were performed using Fisher's exact test.

RESULTS

Analyses identified an overall increase in NYC ED utilization relating to Puerto Rico following Hurricane Maria landfall. In particular, there was a small but significant increase in visits involving a medication refill or essential medical equipment. Visits for other outcomes, such as mental illness, also increased, but the differences were not statistically significant.

CONCLUSIONS

Gaining this situational awareness of medical service use was informative following Hurricane Maria, and, following any natural disaster, the same surveillance methods could be easily established to aid an effective emergency response.

Analysis of Unstructured Text-Based Data Using Machine Learning Techniques: The Case of Pediatric Emergency Department Records in Nicaragua

Free-text information is still widely used in emergency department (ED) records. Machine learning techniques are useful for analyzing narratives, but they have been used mostly for English-language data sets. Considering such a framework, the performance of an ML classification task of a Spanish-language ED visits database was tested. ED visits collected in the EDs of nine hospitals in Nicaragua were analyzed. Spanish-language, free-text discharge diagnoses were considered in the analysis. Five-hundred random forests were trained on a set of bootstrap samples of the whole data set (1,789 ED visits) to perform the classification task. For each one, after having identified optimal parameter value, the final validated model was trained on the whole bootstrapped data set and tested. The classification accuracies had a median of 0.783 (95% CI [0.779, 0.796]). Machine learning techniques seemed to be a promising opportunity for the exploitation of unstructured information reported in ED records in low- and middle-income Spanish-speaking countries.

Natural language generation for electronic health records

One broad goal of biomedical informatics is to generate fully-synthetic, faithfully representative electronic health records (EHRs) to facilitate data sharing between healthcare providers and researchers and promote methodological research. A variety of methods existing for generating synthetic EHRs, but they are not capable of generating unstructured text, like emergency department (ED) chief complaints, history of present illness, or progress notes. Here, we use the encoder–decoder model, a deep learning algorithm that features in many contemporary machine translation systems, to generate synthetic chief complaints from discrete variables in EHRs, like age group, gender, and discharge diagnosis. After being trained end-to-end on authentic records, the model can generate realistic chief complaint text that appears to preserve the epidemiological information encoded in the original record-sentence pairs. As a side effect of the model’s optimization goal, these synthetic chief complaints are also free of relatively uncommon abbreviation and misspellings, and they include none of the personally identifiable information (PII) that was in the training data, suggesting that this model may be used to support the de-identification of text in EHRs. When combined with algorithms like generative adversarial networks (GANs), our model could be used to generate fully-synthetic EHRs, allowing healthcare providers to share faithful representations of multimodal medical data without compromising patient privacy. This is an important advance that we hope will facilitate the development of machine-learning methods for clinical decision support, disease surveillance, and other data-hungry applications in biomedical informatics.

Natural language generation for electronic health records

One broad goal of biomedical informatics is to generate fully-synthetic, faithfully representative electronic health records (EHRs) to facilitate data sharing between healthcare providers and researchers and promote methodological research. A variety of methods existing for generating synthetic EHRs, but they are not capable of generating unstructured text, like emergency department (ED) chief complaints, history of present illness, or progress notes. Here, we use the encoder-decoder model, a deep learning algorithm that features in many contemporary machine translation systems, to generate synthetic chief complaints from discrete variables in EHRs, like age group, gender, and discharge diagnosis. After being trained end-to-end on authentic records, the model can generate realistic chief complaint text that appears to preserve the epidemiological information encoded in the original record-sentence pairs. As a side effect of the model's optimization goal, these synthetic chief complaints are also free of relatively uncommon abbreviation and misspellings, and they include none of the personally identifiable information (PII) that was in the training data, suggesting that this model may be used to support the de-identification of text in EHRs. When combined with algorithms like generative adversarial networks (GANs), our model could be used to generate fully-synthetic EHRs, allowing healthcare providers to share faithful representations of multimodal medical data without compromising patient privacy. This is an important advance that we hope will facilitate the development of machine-learning methods for clinical decision support, disease surveillance, and other data-hungry applications in biomedical informatics.

Perspective on Improving Environmental Monitoring of Biothreats

For more than a decade, the United States has performed environmental monitoring by collecting and analyzing air samples for a handful of biological threat agents (BTAs) in order to detect a possible biological attack. This effort has faced numerous technical challenges including timeliness, sampling efficiency, sensitivity, specificity, and robustness. The cost of city-wide environmental monitoring using conventional technology has also been a challenge. A large group of scientists with expertise in bioterrorism defense met to assess the objectives and current efficacy of environmental monitoring and to identify operational and technological changes that could enhance its efficacy and cost-effectiveness, thus enhancing its value. The highest priority operational change that was identified was to abandon the current concept of city-wide environmental monitoring because the operational costs were too high and its value was compromised by low detection sensitivity and other environmental factors. Instead, it was suggested that the focus should primarily be on indoor monitoring and secondarily on special-event monitoring because objectives are tractable and these operational settings are aligned with likelihood and risk assessments. The highest priority technological change identified was the development of a reagent-less, real-time sensor that can identify a potential airborne release and trigger secondary tests of greater sensitivity and specificity for occasional samples of interest. This technological change could be transformative with the potential to greatly reduce operational costs and thereby create the opportunity to expand the scope and effectiveness of environmental monitoring.

Google search patterns monitoring the daily health impact of heatwaves in England: How do the findings compare to established syndromic surveillance systems from 2013 to 2017?

One of the implications of climate change is a predicted increase in frequent and severe heatwaves. The impact of heatwaves on the health of the population is captured through real-time syndromic healthcare surveillance systems monitored daily in England during the summer months. Internet search data could potentially provide improved timeliness and help to assess the wider population health impact of heat by capturing a population sub-group who are symptomatic but do not seek healthcare. A retrospective observational study was carried out from June 2013 to September 2017 in England to compare daily trends in validated syndromic surveillance heat-related morbidity indicators against symptom-based heatwave related Google search terms. The degree of correlation was determined with Spearman correlation coefficients and lag assessment was carried out to determine timeliness. Daily increases in frequency in Google search terms during heatwave events correlated well with validated syndromic indicators. Correlation coefficients between search term frequency and syndromic indicators from 2013 to 2017 were highest with the telehealth service NHS 111 (range of 0.684-0.900 by search term). Lag analysis revealed a similar timeliness between the data sources, suggesting Google data did not provide a delayed or earlier signal in the context of England's syndromic surveillance systems. This work highlights the potential benefits for countries which lack established public health surveillance systems to monitor heat-related morbidity and the use of internet search data to assess the wider population health impact of exposure to heat.

Public and Population Health Informatics: The Bridging of Big Data to Benefit Communities

Objective:  To summarize the recent public and population health informatics literature with a focus on the synergistic “bridging” of electronic data to benefit communities and other populations.

Methods:  The review was primarily driven by a search of the literature from July 1, 2016 to September 30, 2017. The search included articles indexed in PubMed using subject headings with (MeSH) keywords “public health informatics” and “social determinants of health”. The “social determinants of health” search was refined to include articles that contained the keywords “public health”, “population health” or “surveillance”.

Results:  Several categories were observed in the review focusing on public health's socio-technical infrastructure: evaluation of surveillance practices, surveillance methods, interoperable health information infrastructure, mobile health, social media, and population health. Common trends discussing socio-technical infrastructure included big data platforms, social determinants of health, geographical information systems, novel data sources, and new visualization techniques. A common thread connected these categories of workforce, governance, and sustainability: using clinical resources and data to bridge public and population health.

Conclusions:  Both medical care providers and public health agencies are increasingly using informatics and big data tools to create and share digital information. The intent of this “bridging” is to proactively identify, monitor, and improve a range of medical, environmental, and social factors relevant to the health of communities. These efforts show a significant growth in a range of population health-centric information exchange and analytics activities.