Electronic Surveillance System For Early Notification Of Community-based Epidemics Research Articles

A Framework for Developing and Assessing Custom Case Definitions: A Demonstration Applied to Opioid Overdose in Maryland.

Public health epidemiologists monitor data sources for disease outbreaks and other events of public health concern, but manual review of records to identify cases of interest is slow and labor-intensive and may not reflect evolving data practices. To automatically identify cases from electronic data sources, epidemiologists must use "case definitions" or formal logic that captures the criteria used to identify a record as a case of interest. To establish a methodology for development and evaluation of case definitions. A logical evaluation framework to approach case definitions will allow jurisdictions the flexibility to implement a case definition tailored to their goals and available data. Case definition development is explained as a process with multiple logical components combining free-text and categorical data fields. The process is illustrated with the development of a case definition to identify emergency medical services (EMS) call records related to opioid overdoses in Maryland. The Maryland Department of Health (MDH) installation of the Electronic Surveillance System for Early Notification of Community-Based Epidemics (ESSENCE), which began capturing EMS call records in ESSENCE in 2019 to improve statewide coverage of all-hazards health issues. We describe a case definition evaluation framework and demonstrate its application through development of an opioid overdose case definition to be used in MDH ESSENCE. We show the iterative process of development, from defining how a case can be identified conceptually to examining each component of the conceptual definition and then exploring how to capture that component using available data. We present a framework for developing and qualitatively assessing case definitions and demonstrate an application of the framework to identifying opioid overdose incidents from MDH EMS data. We discuss guidelines to support jurisdictions in applying this framework to their own data and public health challenges to improve local surveillance capability.

Heterogeneity and Interoperability in Local Public Health Information Systems.

Heterogeneity and Interoperability in Local Public Health Information Systems.

Comparing Automated Cluster Detection Methods for Carbapenem-Resistant Enterobacteriaceae (CRE): Rule-Based Versus Statistical

Background: The Tennessee (TN) Department of Health (TDH) has been identifying clusters of reportable conditions using the Electronic Surveillance System for Early Notification of Community-Based Epidemics (ESSENCE), a cluster detection method using space-time scan permutation statistics based on patient ZIP code. CRE are reportable in Tennessee; isolate submission is required for carbapenemase (CP) production and resistance mechanism (eg, KPC gene) testing. The Council for Outbreak Response: Healthcare-Associated Infections (HAI) and Antimicrobial-Resistant (AR) Pathogens (CORHA) released proposed thresholds of reporting CRE to public health. Thresholds vary by healthcare facility type and regional epidemiology. The TDH HAI/AR program currently runs a daily automated SAS code using the CORHA reporting threshold to help public health identify suspect KPC clusters. We evaluated our rule-based CORHA method against 2 space-time statistic-based methods for KPC cluster detection in Tennessee. Methods: Simulations for each cluster detection method were performed using retrospective CP-CRE surveillance data for 2018. Simulations were conducted using (1) CORHA reporting thresholds by facility case count to flag clusters of 2 or more cases within 28 days, (2) ESSENCE using patient residence ZIP code and the earliest of collection date or symptom onset date as is used for other reportable conditions in Tennessee, and (3) a modified space-time statistical method using SaTScan in which reporting facility, rather than a geographic location, was used as space variable to detect within-facility clusters within 1–28 days. We compared the number and overlap of cases and clusters identified with each method. Univariate logistic regression with CORHA flagging as predictor and flagging by each ESSENCE or CORHA method as outcome variables, were used to compare cases tagged by each method pair, respectively. Results: Of 183 KPC CP-CRE cases, 54 (30.6%) were flagged as part of suspect clusters by at least 1 method. Simulations generated 16 alerts (36 cases) using CORHA, 10 clusters (25 cases) using modified SaTScan, and 10 clusters (20 cases) using standard ESSENCE protocol. Among KPC CP-CRE cases flagged by CORHA, 12 (33.3%) were also flagged by modified SaTScan and 2 (5%) by ESSENCE. A case flagged using CORHA method has 5.15 (95% CI, 2.10–12.64) times higher odds of also being flagged by the modified SaTScan method compared to cases not flagged by CORHA. Conclusions: An algorithm based on CORHA thresholds for reporting CRE to public health had strong agreement with modified SaTScan, a space-time method. We intend to explore the extension of the time interval for ESSENCE.Funding: NoneDisclosures: None

Jurisdictional Etiquette Workgroup: An offshoot of a Syndromic Surveillance Consortium

Objective● To demonstrate the importance of a cross-jurisdictional etiquette workgroup in the Texas Southeast region that leverages on the Syndromic Surveillance Consortium● To promote data sharing and communicate the findings of disease to assist rapid investigation and data sharingKey words:ESSENCE (Electronic Surveillance System for Early notification of Community-based Epidemics)IntroductionSyndromic data is shifting the way surveillance has been done traditionally. Most recently, surveillance has gone beyond city limits and county boundary lines. In southeast Texas, a regional consortium of public health agencies and stakeholders in the 13-County area governs the local ESSENCE system. The Houston Health Department, (HHD) is responsible for deploying ESSENCE to the entire region.To effectively monitor the health of the region’s population, a need arose to establish clear guidelines for disease investigation and data sharing triggered by syndromic surveillance across the area. Since Houston’s instance of ESSENCE serves all 13 counties, the consortium instituted a cross- jurisdictional etiquette group. The purpose of the group is to determine the standard protocol for responding to ESSENCE alerts and best practices for data sharing and use among consortium members.MethodsTo achieve these goals, it was determined that a smaller group of stakeholders besides governing officials is needed to provide guidance for regional data sharing and use. The etiquette group was established in the first quarter of 2018 and it included four consortium representatives from the 6/5 south region of Texas. Their first meeting tackled issues relating to data sharing.ResultsThe following products emerged from the activities of the etiquette group within 3 months of its existence:● Publication/presentation guidance/policy to avoid duplication of efforts and misrepresentation of jurisdiction.● Procedure for alert responses●. Instructions for within-systems management of alerts;●. Instructions for events/times of interest (e.g., political convention, Olympics);● Instructions of syndromes of interest/syndrome-specific policies;● Instructions for changing the syndrome definition;● Notification procedures for identification of a single case of reportable disease/important free text element within data.ConclusionsCross jurisdictional workgroups can influence rapid investigations of disease, protect patient health information and promote privacy and data security and confidentiality by establishing set rules/guidelines for data exchange. All 13-counties in the region rely on these guidelines as a standard for responsibly accessing, using and sharing data in the Texas Southeast ESSENCE system.Lessons Learned:● As the etiquette group continues to evolve, there is need for more resources to help foster data use and sharing among jurisdictional partners.● Partner engagement is limited due to ongoing process of configuring the new system ESSENCE.● Since disease has no boundaries, allocation of jurisdictional responsibilities for responding to alerts should be operationalized● Continuous training is essential to ensure all system users adhere to the protocols in place for meaningful data use and data sharing

Monitoring Sexual Violence Visits in Emergency Department Data to Improve Public Health

ObjectiveTo describe characteristics of sexual violence emergency department visits in Washington State.IntroductionAlthough sexual violence is a pressing public health and safety issue, it has historically been challenging to monitor population trends with precision. Approximately 31% of incidents of sexual violence are reported to law enforcement and only 5% lead to an arrest1, making the use of law enforcement data challenging. Syndromic surveillance data from emergency departments provides an opportunity to use care-seeking to more accurately surveil sexual violence without introducing additional burdens on either patients or healthcare providers.MethodsUsing the National Syndromic Surveillance Program (NSSP) Electronic Surveillance System for Early Notification of Community-based Epidemics (ESSENCE) platform, staff from the Rapid Health Information Network (RHINO) program at the Washington State Department of Health created a syndrome definition for sexual violence in emergency department data using a combination of ICD-10 diagnostic codes and chief complaint terms likely to be used for sexual violence visits.Staff executed the query using both the Chief Complaint Query Validation and Facility Location (Full Details) data sources in the NSSP ESSENCE platform. Validation occurred by examining visits from 2017 using the Original Chief Complaint, Diagnosis Combo, and Original Triage Notes fields in the data details output to determine if a visit was a true positive for sexual violence. Staff then used the R Studio platform to create n-grams to analyze commonly occurring terms in the data.RHINO staff collaborated with colleagues in the agency's Injury and Violence Prevention Section to better understand the trends observed in the data and the utility of using syndromic surveillance to inform public health practice.ResultsThe query identified 1,550 visits for sexual violence in 2017. Female patients were disproportionately represented (87.16%), with female patients aged 10-29 years making up 47.03% of captured visits. Overall, patients 10-29 years of age represented 52.90% of all identified visits. Older patients, particularly older men were less represented.Among the captured visits, staff analyzed a sample of 347 visits and found that 88.76% were correctly identified as being related to sexual violence. Although triage notes are an optional field for Washington State syndromic reporting and only present in approximately 40% of records, analysis of the triage notes also provided contextual details on the time (36.89%) and place (18.44%) of the incident, and the identity of the assailant (17.29%).Among patients 10-29 years old, several increases in the percentage of emergency department visits for sexual violence were observed in conjunction with secondary and post-secondary school term breaks, as well as at the beginning of Autumn and end of Spring terms at most Washington State universities. The trend was present in both patients 10-19 years and 20-29 years, with a stronger signal in the 10-19 year age group. The pattern was not present in either older or younger patients.ConclusionsThe seasonal trend associated with the academic calendar in patients 10-29 years of age is consistent with other data on teen and campus dating violence2,3 and provides another piece of information to validate and inform the work of social service groups serving adolescents and young adults.Syndromic data is particularly well suited to translating surveillance into actionable public health—having additional data to support the hypotheses of state Rape Prevention and Education programs has the potential to encourage greater participation from universities and other academic institutions to improve consent and sexual violence response programs. Regardless of whether patients are affiliated with a specific institution, the overall safety of the communities where students live is of importance to academic institutions. Similarly, yearly increases in visits during annual mass gatherings may be of use in communicating with event organizers strategies for reducing incidents of violence during that period.Indexing the query for sexual violence within the NSSP ESSENCE platform allows surveillance practitioners to quickly and easily monitor emergency department visits for sexual violence using a standardized methodology. As a national platform, NSSP ESSENCE facilitates collaboration across borders between local, state, Tribal and Urban Indian, and national public health agencies. This improved availability and performance of the query. Additionally, the point-and-click nature of the ESSENCE platform makes using syndromic data more accessible for local health and social service staff who are not trained in epidemiology. The ease of collaboration between partners using the platform also makes it well suited to work which span state, local, and Tribal, and Urban Indian health needs. Data and query sharing increases the likelihood that the data will be actionable and therefore positively influence public health.

Influenza laboratory testing and its application in timely Department of Defense biosurveillance

ObjectiveTo describe influenza laboratory testing and results in the Military Health System and how influenza laboratory results may be used in DoD Electronic Surveillance System for Early Notification of Community-based Epidemics (ESSENCE)IntroductionTimely influenza data can help public health decision-makers identify influenza outbreaks and respond with preventative measures. DoD ESSENCE has the unique advantage of ingesting multiple data sources from the Military Health System (MHS), including outpatient, inpatient, and emergency department (ED) medical encounter diagnosis codes and laboratory-confirmed influenza data, to aid in influenza outbreak monitoring. The Influenza-like Illness (ILI) syndrome definition includes ICD-9 or ICD-10 codes that may increase the number of false positive alerts. Laboratory-confirmed influenza data provides an increased positive predictive value (PPV). The gold standard for influenza testing is molecular assays or viral culture. However, the tests may take 3-10 days to result. Rapid influenza diagnostic tests (RIDTs) have a lower sensitivity, but the timeliness of receiving a result improves to within <15 minutes. We evaluate the utility of RIDTs for routine ILI surveillance.MethodsAdministrative medical encounters for ILI and influenza laboratory-confirmed data were analyzed from the MHS from June 2013 – September 2017 (Figure 1). The medical encounters and laboratory data include outpatient, inpatient, and ED data. The ILI syndrome case definition is a medical encounter during the study period with an ICD-9 or ICD-10 codes in any diagnostic position (ICD-9 codes = 79.99, 382.9, 460, 461.9, 465.8, 465.9, 466.0, 486, 487.0, 487.1, 487.8, 488, 490, 780.6, or 786.2; ICD-10 codes = B97.89, H66.9, J00, J01.9, J06.9, J09, J09.X, J10, J10.0, J10.1, J10.2, J10.8, J11, J11.0, J11.1, J11.2, J11.8, J12.89, J12.9, J18, J20.9, J40, R05, R50.9). The ILI dataset was limited to care provided in the MHS as laboratory data is only available for direct care. We describe influenza laboratory testing practices in the MHS. We aggregated the ILI encounters and RIDT positive results into daily counts and generated a weekly Pearson’s correlation.ResultsInfluenza tests are ordered throughout the year; the mean weekly percentage of ILI encounters in which an influenza laboratory test is ordered is 5.62%, with a range from 0.68% in the off season to 19.2% during peak influenza activity. The mean weekly percentage of positive influenza laboratory results among all ILI encounters is 0.82%, with a range from 0.01% to 5.73% (Figure 2). The percent of ILI encounters in which a test is ordered increases as the influenza season progresses. Influenza laboratory tests conducted in the MHS include RIDTs, PCR, culture, and DFA. Among all influenza tests ordered in the MHS, 66.0% were RIDTs, 22.7% were PCR, and 11.3% were viral culture. Often, a confirmatory test is ordered following a RIDT; 20% of RIDTs have follow-up tests. The mean timeliness of influenza test result data in the MHS was 11.26 days for viral culture, 2.94 days for PCR, and 0.11 days for RIDTs. The RIDT results were moderately correlated with ILI encounters for the entire year (mean weekly Pearson correlation coefficient rho=0.60, 95% CI: 0.55, 0.66, Figure 3). During the influenza season, the mean weekly Pearson correlation coefficient increases to rho=0.75, 95% CI: 0.70, 0.79.ConclusionsThe DoD has the unique advantage of access to the electronic health record and laboratory tests and results of all MHS beneficiaries. This analysis provides evidence for increased utilization of positive RIDTs in ESSENCE. The moderate correlation between the ILI syndrome and positive RIDTs may be associated with ICD-10 codes included in the ILI syndrome definition that contribute to false positive influenza cases. Ongoing research is focused on improving this ILI syndrome definition using ICD-10 codes. Rapid influenza diagnostic tests provide more timely results than other influenza test types. In conjunction with ILI medical encounter data, positive RIDT data provides a more complete and timely picture of the true burden of influenza on the MHS population for early warning of influenza outbreaks.

Using Evaluation to Inform the BioSense Platform: Results from a 2018 Survey

ObjectiveTo assess the present status of utility, functionality, usability and user satisfaction of the BioSense Platform.IntroductionSince 2015, CDC’s Division of Health Informatics and Surveillance staff have conducted evaluations to provide information on the utility, functionality, usability and user satisfaction associated with the National Syndromic Surveillance Program’s BioSense Platform tools. The BioSense Platform tools include: 1) Access and Management Center (AMC), a tool that enables site administrators to manage users and data permissions; 2) Electronic Surveillance System for Early Notification of Community-based Epidemics (ESSENCE), a software application that enables syndromic surveillance related data visualization and analysis; 3) Adminer, a tool that allows users to access site data on the datamart; and 4) Rstudio, an application that can be used for data analysis and visualization. The evaluation findings have informed activities that led to improvements in functionality, development or procurement of platform associated tools, and development of resource materials. In May 2018, NSSP conducted an evaluation with eight jurisdictions that participated in the first user acceptance testing (UAT) evaluations in 2015. The purpose of the evaluation was to assess the present status of utility, functionality, usability and user satisfaction of the tools on the BioSense Platform, and delineate progress since 2015.MethodsCDC’s evaluation framework and utilization-focused evaluation were used to inform and engage stakeholders, develop the evaluation questions, metrics, and methodology. Eight selected jurisdictions participated in an online, Epi-Info survey that captured quantitative and qualitative information. Prior to the survey, participants received a presentation about the evolution of the BioSense Platform since 2015, and were provided an overview of components to evaluate. The participants were asked to assess the following key areas based on use of the BioSense Platform within the past 30 days: 1) the utility, functionality and usability of the AMC, ESSENCE, Adminer and Rstudio; 2) how well the enhanced data flow has enabled them to conduct syndromic surveillance activities; 3) usefulness of the quick start guides. Additionally, participants were asked to provide suggestions for other improvements to the BioSense Platform, and to indicate their overall satisfaction. Descriptive statistics were generated and thematic analysis was conducted to identify themes from qualitative responses.ResultsOverall, participant’s responses remained positive about the utility, functionality, usability and overall satisfaction of the BioSense Platform. Participants indicated using the BioSense Platform regularly (e.g. daily, weekly and/or monthly) within those 30 days. Certain functions have been used more than others across the various tools to conduct syndromic surveillance, with at least 50% of participants reporting use. These included creating data access rules, viewing and verifying raw and processed data, running time series, conducting free-text queries, and assessing data details and total ER visit counts by hospital, county/region, or state. The challenges ranged from tool performance to user interpretation of the function. Participants reported that the enhanced data flow improved their data quality and helped identify issues. Although participants scored ESSENCE to have average usability per the system usability scale (SUS score=63.5 in 2018), the BioSense Platform and its tools were reported as useful by 88% of participants. Further, participants continue to be comfortable using the AMC, however creating data access rules that are outside of simple use cases continue to be a challenge. Participants comfort level with Adminer improved from 2016 to 2018 with all participants reporting comfortable in using the tool. The use of each tool’s quick start guides varied. Of those who used the guides, all of the participants agreed that the Adminer and Data Dictionary guides were useful. There was a smaller number of participants agreeing that the other guides were useful. Lastly, participants provided recommendations to improving the BioSense Platform. The most frequent recommendations were improving the data access control architecture, and sharing aggregate data with hospitals in their state.ConclusionsThe development and operationalization of the BioSense Platform and associated tools has been in an environment of continuing advancements in technology and changing public health needs and priorities. Up-to-date evaluation activities have helped to ensure that BioSense is best suited to address these challenges and meet the syndromic surveillance needs of users. Overall, the findings outlined above indicate that the functionality and utility of BioSense are well suited to meet user needs.ReferencesBangor A, Kortm P, Miller J. Determining What Individual SUS Scores Mean: Adding an Adjective Rating Scale. Journal of Usability Studies. 2009; 4:114-123.

Streamlined Development of Analytic Fusion Capability for Health Surveillance

Streamlined Development of Analytic Fusion Capability for Health Surveillance

Impact of the NSSP’s transition to ESSENCE on chief complaint field-based syndromes

Objective: To evaluate the effect and implications of changing the chief complaint field during the National Syndromic Surveillance Program (NSSP) transition from BioSense 2.0 analytical tools to BioSense Platform – ESSENCEIntroduction: In January 2017, the NSSP transitioned their BioSense analytical tools to Electronic Surveillance System for Early Notification of Community-Based Epidemics (ESSENCE). The chief complaint field in BioSense 2.0 was a concatenation of the record’s chief complaint, admission reason, triage notes, and diagnostic impression. Following the transition to ESSENCE, the chief complaint field was comprised of the first chief complaint entered or the first admission reason, if the chief complaint was blank. Furthermore, the ESSENCE chief complaint field was electronically parsed (i.e., the original chief complaint text was altered to translate abbreviations and remove punctuation). This abstract highlights key findings from Maricopa County Department of Public Health’s evaluation of the new chief complaint field, its impact on heat-related illness syndromic surveillance, and implications for ongoing surveillance efforts.Methods: For this evaluation, we used the heat-related illness query recommended in Council of State and Territorial Epidemiologists’ (CSTE)2016 Guidance Document for Implementing Heat-Related Illness Syndromic Surveillance. Before the transition, we used BioSense 2.0’s, phpMyAdmin analytical tool to generate a list of patients who visited Maricopa County emergency departments or inpatient hospitals between 5/1/2016 – 9/30/2016 due to heat-related illness. After the transition, we used the CC and DD Category “Heat-related Illness, v1” in ESSENCE, which was based on the CSTE heat-related illness query, to generate a list of patients for the same time period. We compared the line-lists and time-series trends from phpMyAdmin and ESSENCE.Results: The phpMyAdmin analytical tool identified 785 heat-related illness records with the query (Figure). 642 (82%) of these heat-related illness records were also captured by ESSENCE. Reasons for 143 (18%) records not being identified by ESSENCE included: the patient’s admission reason field contained keywords that were not available in the ESSENCE chief complaint field (n=94, 66%); data access changed, which disabled access to patients who resided in zip codes that crossed a county border (30, 21%); discrepancies between ESSENCE parsing and text in the original chief complaint (11, 8%); heat-related illness discharge diagnoses were removed by the facility after the phpMyAdmin line-list for heat-related illness was extracted (7, 5%); and one record was undetermined. Conversely, ESSENCE captured 36 additional heat-related illness records, not previously captured by phpMyAdmin. Reasons included: a query exclusion term was located in the patient’s admission reason but not the ESSENCE chief complaint field (16, 44%); a heat-related illness discharge diagnosis code was added by the facility after the data were extracted by phpMyAdmin (4, 11%); and 16 (44%) were undetermined. Time-series trend evaluation revealed a significant correlation between the two surveillance tools (Pearson coefficient = 0.97, p < 0.01).Conclusions: Though the data trends over time were not significantly affected by changes in the chief complaint field, differences in the field’s composition have important implications for syndromic surveillance practitioners. Free-text queries designed to search the chief complaint field in ESSENCE may not retrieve records previously identified with BioSense 2.0 analytical tools, which may limit individual case-finding capacity. The elimination of admission reason from the chief complaint field in ESSENCE has the greatest effect on case-finding capacity. Furthermore, surveillance reports produced by ESSENCE cannot be directly compared to reports that were previously published with data from BioSense 2.0. These limitations may be addressed if ESSENCE creates a feature that allows users to easily query fields (e.g., admission reason) in addition to the chief compliant field.

Maricopa County’s use of NSSP ESSENCE to detect cases during a hepatitis A outbreak

ObjectiveTo demonstrate the utility of the National Syndromic Surveillance Program’s (NSSP) version of the Electronic Surveillance System for Early Notification of Community-based Epidemics (ESSENCE) for case detection during a 2017 outbreak of hepatitis A virus (HAV) infection among persons experiencing homelessness in Maricopa County, Arizona.IntroductionOn 3/29/2017, the Maricopa County Department of Public Health (MCDPH) received three reports of confirmed HAV infection from an onsite clinic at Campus A that assists individuals experiencing homelessness, a population at risk for HAV transmission. To identify the scope of the problem, the department initiated rapid HAV infection case detection using NSSP ESSENCE.MethodsMCDPH created a myESSENCE dashboard that searched for chief complaint keywords and discharge diagnosis codes from 15 Maricopa County emergency department and inpatient hospital records using three separate queries: (1) HAV infection; (2) hepatitis virus infection; and (3) HAV infection symptoms combined with terms for homelessness. The dashboard was reviewed retrospectively for the 90 days prior to the initial report of the cluster (12/28/2016 – 3/29/2017). Based on this review, MCDPH epidemiologists decided to use the first query (HAV infection-specific) only, because reviewing the line list for all three queries was resource-intensive and resulted in duplicate cases. The query was monitored every weekday morning from 3/30/2017 to the close of the outbreak on 7/22/2017.ResultsFrom 3/30/2017 – 7/22/2017, MCDPH identified 37 potential HAV infection cases in ESSENCE. Eleven cases were classified as outbreak cases, while the other 26 patients lacked recent HAV infection symptoms, laboratory confirmation, or association with Campus A. All 11 outbreak cases’ records included the ICD-CM-10 code B15.9 (Hepatitis A without hepatic coma), and 3 records included the code Z59 (Problems related to housing and economic circumstances). The HAV infection-specific query in ESSENCE identified 11 (73%) of the 15 total outbreak cases; however, all cases were reported in MEDSIS prior to a being flagged in ESSENCE. On average, cases were reported to MEDSIS 9 days earlier than identified in ESSENCE (range, 0 - 21 days). Of note, ESSENCE helped identify the outbreak index case, previously lost to follow-up, as an individual experiencing homelessness. This information helped MCDPH perform additional follow-up, which revealed that the individual had arrived from San Diego, CA, a city with an ongoing outbreak of HAV infection in their homeless and drug-using population. This epidemiologic link was identified on 3/30/2017, over a month before the Centers for Disease Control and Prevention confirmed a match by genome sequencing between the index case and the neighboring outbreak on 5/5/2017.ConclusionsUse of ESSENCE identified most of the HAV infection cases from this outbreak among individuals experiencing homelessness but no sooner than traditional surveillance methods. However, use of ESSENCE allowed for identification of the outbreak index case, leading to epidemiological linkage to outbreak origin approximately one month prior to molecular laboratory confirmation.

Evaluation of an arboviral syndrome query used in Maricopa County, Arizona

ObjectiveTo evaluate Arizona’s arboviral syndromic surveillance protocol in Maricopa County.IntroductionTimely identification of arboviral disease is key to prevent transmission in the community, but traditional surveillance may take up to 14 days between specimen collection and health department notification. Arizona state and county health agencies began monitoring National Syndromic Surveillance Program BioSense 2.0 data for patients infected with West Nile virus (WNV), St. Louis encephalitis virus (SLEV), chikungunya, or dengue virus in August 2015. Zika virus was added in April 2016. Our novel methods were presented at the International Society for Disease Surveillance 2015 Annual Conference. [1] Twice per week, we queried patient records from 15 Maricopa County BioSense-enrolled emergency department and inpatient hospitals for chief complaint keywords and discharge diagnosis codes. Our “Case Investigation Decision Tree” helped us determine whether records had a high or low degree of evidence for arboviral disease and necessitated further investigation. This study evaluated how Arizona’s protocol for conducting syndromic surveillance compared to traditional arboviral surveillance in terms of accuracy and timeliness in Maricopa County from August 2015 through December 2016.MethodsWe followed guidelines from the Centers for Disease Control and Prevention (CDC) to evaluate two major attributes of the protocol: accuracy [measured as positive predictive value (PPV) and sensitivity] and timeliness. [2] Arizona’s Medical Electronic Disease Surveillance Intelligence System (MEDSIS) was considered the “gold standard” system and BioSense was the test system. PPV was calculated as the proportion of records identified by BioSense that were reported to MEDSIS, regardless of final case classification. Sensitivity was the proportion of confirmed or probable cases in MEDSIS identified by BioSense. Though not all MEDSIS cases were seen at BioSense-reporting facilities, the sensitivity demonstrates how each query contributed to arboviral surveillance overall. We assessed timeliness in two ways: (1) the difference between the date when keywords or diagnosis codes were first identified by BioSense and the date the same patient was first reported to MEDSIS; and (2) the difference between the date the BioSense record was first reviewed by the Maricopa County Department of Public Health (MCDPH) syndromic surveillance team and the date the same patient was first investigated through MEDSIS by the MCDPH disease investigators. We assessed whether timeliness was affected by the method in which a record was identified in BioSense (i.e., chief complaint keyword or discharge diagnosis code).ResultsThe arboviral syndromic surveillance queries identified 62 records during the evaluation period (Table). For each arboviral query, the proportion of BioSense records that were also reported through MEDSIS ranged from 25.0% to 32.4%, except chikungunya, which had a PPV of 0%. BioSense records that had a high degree of evidence for arboviral disease tended to have a higher PPV compared to those with low evidence. BioSense records that were not already reported to MEDSIS met neither clinical nor exposure criteria for the arboviral diseases and were not deemed a public health risk. The sensitivities of the WNV and Zika queries to detect confirmed or probable cases in MEDSIS were 8.2% and 5.6%, respectively, while SLEV, chikungunya, and dengue queries were 0%. On average, patients were reported to MEDSIS 7 days prior to BioSense identifying keywords or diagnosis codes. In addition, MEDSIS cases were investigated by MCDPH disease investigators 10 days prior to MCDPH syndromic surveillance team review of BioSense records, on average. The average time between MEDSIS report date and BioSense identification date was shorter for BioSense records identified by chief complaint keywords than by diagnosis codes (4 and 8 days after MEDSIS, respectively).ConclusionsArizona’s arboviral syndromic surveillance protocol provided MCDPH with situational awareness, but BioSense data were not available more quickly than traditional mandated reporting. Through this process, we reviewed patient records that mentioned arboviral diseases and confirmed that these reportable conditions were captured in our traditional surveillance system. The decision tree was effective at prioritizing records for further investigation. Timeliness may be improved by updating the queries to include more chief complaint keywords and reviewing BioSense more than twice per week. MCDPH plans to evaluate Arizona’s updated arboviral syndromic surveillance protocol, which was adapted for BioSense Platform’s Electronic Surveillance System for Early Notification of Community-based Epidemics (ESSENCE).

Dengue Surveillance in Veterans Affairs Healthcare Facilities, 2007–2010

BackgroundAlthough dengue is endemic in Puerto Rico (PR), 2007 and 2010 were recognized as epidemic years. In the continental United States (US), outside of the Texas-Mexico border, there had not been a dengue outbreak since 1946 until dengue re-emerged in Key West, Florida (FL), in 2009–2010. The objective of this study was to use electronic and manual surveillance systems to identify dengue cases in Veterans Affairs (VA) healthcare facilities and then to clinically compare dengue cases in Veterans presenting for care in PR and in FL.MethodologyOutpatient encounters from 1/2007–12/2010 and inpatient admissions (only available from 10/2009–12/2010) with dengue diagnostic codes at all VA facilities were identified using VA's Electronic Surveillance System for Early Notification of Community-based Epidemics (ESSENCE). Additional case sources included VA data from Centers for Disease Control and Prevention BioSense and VA infection preventionists. Case reviews were performed. Categorical data was compared using Mantel-Haenszel or Fisher Exact tests and continuous variables using t-tests. Dengue case residence was mapped.FindingsTwo hundred eighty-eight and 21 PR and FL dengue cases respectively were identified. Of 21 FL cases, 12 were exposed in Key West and 9 were imported. During epidemic years, FL cases had significantly increased dengue testing and intensive care admissions, but lower hospitalization rates and headache or eye pain symptoms compared to PR cases. There were no significant differences in clinical symptoms, laboratory abnormalities or outcomes between epidemic and non-epidemic year cases in FL and PR. Confirmed/probable cases were significantly more likely to be hospitalized and have thrombocytopenia or leukopenia compared to suspected cases.ConclusionsDengue re-introduction in the continental US warrants increased dengue surveillance and education in VA. Throughout VA, under-testing of suspected cases highlights the need to emphasize use of diagnostic testing to better understand the magnitude of dengue among Veterans.

Using n-Grams for Syndromic Surveillance in a Turkish Emergency Department Without English Translation: A Feasibility Study.

IntroductionSyndromic surveillance is designed for early detection of disease outbreaks. An important data source for syndromic surveillance is free-text chief complaints (CCs), which are generally recorded in the local language. For automated syndromic surveillance, CCs must be classified into predefined syndromic categories. The n-gram classifier is created by using text fragments to measure associations between chief complaints (CC) and a syndromic grouping of ICD codes.ObjectivesThe objective was to create a Turkish n-gram CC classifier for the respiratory syndrome and then compare daily volumes between the n-gram CC classifier and a respiratory ICD-10 code grouping on a test set of data.MethodsThe design was a feasibility study based on retrospective cohort data. The setting was a university hospital emergency department (ED) in Turkey. Included were all ED visits in the 2002 database of this hospital. Two of the authors created a respiratory grouping of International Classification of Diseases, 10th Revision ICD-10-CM codes by consensus, chosen to be similar to a standard respiratory (RESP) grouping of ICD codes created by the Electronic Surveillance System for Early Notification of Community-based Epidemics (ESSENCE), a project of the Centers for Disease Control and Prevention. An n-gram method adapted from AT&T Labs’ technologies was applied to the first 10 months of data as a training set to create a Turkish CC RESP classifier. The classifier was then tested on the subsequent 2 months of visits to generate a time series graph and determine the correlation with daily volumes measured by the CC classifier versus the RESP ICD-10 grouping.ResultsThe Turkish ED database contained 30,157 visits. The correlation (R2) of n-gram versus ICD-10 for the test set was 0.78.ConclusionThe n-gram method automatically created a CC RESP classifier of the Turkish CCs that performed similarly to the ICD-10 RESP grouping. The n-gram technique has the advantage of systematic, consistent, and rapid deployment as well as language independence.

Enhancing spatial detection accuracy for syndromic surveillance with street level incidence data

BackgroundThe Department of Defense Military Health System operates a syndromic surveillance system that monitors medical records at more than 450 non-combat Military Treatment Facilities (MTF) worldwide. The Electronic Surveillance System for Early Notification of Community-based Epidemics (ESSENCE) uses both temporal and spatial algorithms to detect disease outbreaks. This study focuses on spatial detection and attempts to improve the effectiveness of the ESSENCE implementation of the spatial scan statistic by increasing the spatial resolution of incidence data from zip codes to street address level.MethodsInfluenza-Like Illness (ILI) was used as a test syndrome to develop methods to improve the spatial accuracy of detected alerts. Simulated incident clusters of various sizes were superimposed on real ILI incidents from the 2008/2009 influenza season. Clusters were detected using the spatial scan statistic and their displacement from simulated loci was measured. Detected cluster size distributions were also evaluated for compliance with simulated cluster sizes.ResultsRelative to the ESSENCE zip code based method, clusters detected using street level incidents were displaced on average 65% less for 2 and 5 mile radius clusters and 31% less for 10 mile radius clusters. Detected cluster size distributions for the street address method were quasi normal and sizes tended to slightly exceed simulated radii. ESSENCE methods yielded fragmented distributions and had high rates of zero radius and oversized clusters.ConclusionsSpatial detection accuracy improved notably with regard to both location and size when incidents were geocoded to street addresses rather than zip code centroids. Since street address geocoding success rates were only 73.5%, zip codes were still used for more than one quarter of ILI cases. Thus, further advances in spatial detection accuracy are dependant on systematic improvements in the collection of individual address information.