Record-level Data Research Articles

Assessing the accuracy of race-and-ethnicity data in the Outcome and Assessment Information Set.

Limitations in the quality of race-and-ethnicity information in Medicare's data systems constrain efforts to assess disparities in care among older Americans. Using demographic information from standardized patient assessments may be an efficient way to enhance the accuracy and completeness of race-and-ethnicity information in Medicare's data systems, but it is critical to first establish the accuracy of these data as they may be prone to inaccurate observer-reported or third-party-based information. This study evaluates the accuracy of patient-level race-and-ethnicity information included in the Outcome and Assessment Information Set (OASIS) submitted by home health agencies. We compared 2017-2022 OASIS-D race-and-ethnicity data to gold-standard self-reported information from the Medicare Consumer Assessment of Healthcare Providers and Systems® survey in a matched sample of 304,804 people with Medicare coverage. We also compared OASIS data to indirect estimates of race-and-ethnicity generated using the Medicare Bayesian Improved Surname and Geocoding (MBISG) 2.1.1 method and to existing Centers for Medicare & Medicaid Services (CMS) administrative records. Compared with existing CMS administrative data, OASIS data are far more accurate for Hispanic, Asian American and Native Hawaiian or other Pacific Islander, and White race-and-ethnicity; slightly less accurate for American Indian or Alaska Native race-and-ethnicity; and similarly accurate for Black race-and-ethnicity. However, MBISG 2.1.1 accuracy exceeds that of both OASIS and CMS administrative data for every racial-and-ethnic category. Patterns of inconsistent reporting of racial-and-ethnic information among people for whom there were multiple observations in the OASIS and Consumer Assessment of Healthcare Providers and Systems (CAHPS) datasets suggest that some of the inaccuracies in OASIS data may result from observation-based reporting that lessens correspondence with self-reported data. When health record data on race-and-ethnicity includes observer-reported information, it can be less accurate than both true self-report and a high-performing imputation approach. Efforts are needed to encourage collection of true self-reported data and explicit record-level data on the source of race-and-ethnicity information.

Mpox in UK households: estimating secondary attack rates and factors associated with transmission, May-November 2022.

We aimed to estimate the secondary attack rate of mpox among UK household contacts and determine factors associated with transmission to inform public health management of contacts, during the global outbreak in 2022. Information was collected via NHS and public health services and included age, gender, place of residence, setting, and type of contact. Aggregate information was summarized for the UK. Record level data was combined for England, Wales and Northern Ireland, and multivariable logistic regression was used to determine factors associated with transmission. The secondary attack rate among UK household mpox contacts was 4% (60/1526). Sexual contact with the index case was associated with a 11-fold increase in adjusted odds of becoming a case in England, Wales, and Northern Ireland (95% CI 5.5-22, p < 0.001). Household contacts outside of London had increased odds compared to London residents (adjusted OR 2.9, 95%CI 1.6-5.4, p < 0.001), while female contacts had reduced odds of becoming a case (aOR: 0.41, 95% CI: 0.15-0.95). We found a low overall secondary attack rate among household mpox contacts with strong evidence of increased transmission risk associated with sexual contact. This evidence will inform the risk assessment of contacts and support prioritization of those with close intimate contact for follow up.

REDCap and the National Mesothelioma Virtual Bank-a scalable and sustainable model for rare disease biorepositories.

Rare disease research requires data sharing networks to power translational studies. We describe novel use of Research Electronic Data Capture (REDCap), a web application for managing clinical data, by the National Mesothelioma Virtual Bank, a federated biospecimen, and data sharing network. National Mesothelioma Virtual Bank (NMVB) uses REDCap to integrate honest broker activities, enabling biospecimen and associated clinical data provisioning to investigators. A Web Portal Query tool was developed to source and visualize REDCap data in interactive, faceted search, enabling cohort discovery by public users. An AWS Lambda function behind an API calculates the counts visually presented, while protecting record level data. The user-friendly interface, quick responsiveness, automatic generation from REDCap, and flexibility to new data, was engineered to sustain the NMVB research community. NMVB implementations enabled a network of 8 research institutions with over 2000 mesothelioma cases, including clinical annotations and biospecimens, and public users' cohort discovery and summary statistics. NMVB usage and impact is demonstrated by high website visits (>150 unique queries per month), resource use requests (>50 letter of interests), and citations (>900) to papers published using NMVB resources. NMVB's REDCap implementation and query tool is a framework for implementing federated and integrated rare disease biobanks and registries. Advantages of this framework include being low-cost, modular, scalable, and efficient. Future advances to NVMB's implementations will include incorporation of -omics data and development of downstream analysis tools to advance mesothelioma and rare disease research. NVMB presents a framework for integrating biobanks and patient registries to enable translational research for rare diseases.

Breaking the Chains of Health Data

Breaking the Chains of Health Data

The pursuit of approaches to federate data to accelerate Alzheimer's disease and related dementia research: GAAIN, DPUK, and ADDI.

There is common consensus that data sharing accelerates science. Data sharing enhances the utility of data and promotes the creation and competition of scientific ideas. Within the Alzheimer's disease and related dementias (ADRD) community, data types and modalities are spread across many organizations, geographies, and governance structures. The ADRD community is not alone in facing these challenges, however, the problem is even more difficult because of the need to share complex biomarker data from centers around the world. Heavy-handed data sharing mandates have, to date, been met with limited success and often outright resistance. Interest in making data Findable, Accessible, Interoperable, and Reusable (FAIR) has often resulted in centralized platforms. However, when data governance and sovereignty structures do not allow the movement of data, other methods, such as federation, must be pursued. Implementation of fully federated data approaches are not without their challenges. The user experience may become more complicated, and federated analysis of unstructured data types remains challenging. Advancement in federated data sharing should be accompanied by improvement in federated learning methodologies so that federated data sharing becomes functionally equivalent to direct access to record level data. In this article, we discuss federated data sharing approaches implemented by three data platforms in the ADRD field: Dementia's Platform UK (DPUK) in 2014, the Global Alzheimer's Association Interactive Network (GAAIN) in 2012, and the Alzheimer's Disease Data Initiative (ADDI) in 2020. We conclude by addressing open questions that the research community needs to solve together.

SAS and R code for probabilistic quantitative bias analysis for misclassified binary variables and binary unmeasured confounders.

Systematic error from selection bias, uncontrolled confounding, and misclassification is ubiquitous in epidemiologic research but is rarely quantified using quantitative bias analysis (QBA). This gap may in part be due to the lack of readily modifiable software to implement these methods. Our objective is to provide computing code that can be tailored to an analyst's dataset. We briefly describe the methods for implementing QBA for misclassification and uncontrolled confounding and present the reader with example code for how such bias analyses, using both summary-level data and individual record-level data, can be implemented in both SAS and R. Our examples show how adjustment for uncontrolled confounding and misclassification can be implemented. Resulting bias-adjusted point estimates can then be compared to conventional results to see the impact of this bias in terms of its direction and magnitude. Further, we show how 95% simulation intervals can be generated that can be compared to conventional 95% confidence intervals to see the impact of the bias on uncertainty. Having easy to implement code that users can apply to their own datasets will hopefully help spur more frequent use of these methods and prevent poor inferences drawn from studies that do not quantify the impact of systematic error on their results.

Local Data Spaces: Leveraging trusted research environments for secure location-based policy research in the age of coronavirus disease-2019

Abstract This work explores the use of Trusted Research Environments for the secure analysis of sensitive, record-level data on local coronavirus disease-2019 (COVID-19) inequalities and economic vulnerabilities. The Local Data Spaces (LDS) project was a targeted rapid response and cross-disciplinary collaborative initiative using the Office for National Statistics’ Secure Research Service for localized comparison and analysis of health and economic outcomes over the course of the COVID-19 pandemic. Embedded researchers worked on co-producing a range of locally focused insights and reports built on secure secondary data and made appropriately open and available to the public and all local stakeholders for wider use. With secure infrastructure and overall data governance practices in place, accredited researchers were able to access a wealth of detailed data and resources to facilitate more targeted local policy analysis. Working with data within such infrastructure as part of a larger research project involved advanced planning and coordination to be efficient. As new and novel granular data resources become securely available (e.g., record-level administrative digital health records or consumer data), a range of local policy insights can be gained across issues of public health or local economic vitality. Many of these new forms of data however often come with a large degree of sensitivity around issues of personal identifiability and how the data is used for public-facing research and require secure and responsible use. Learning to work appropriately with secure data and research environments can open up many avenues for collaboration and analysis.

ICES Data and Analytic Services: Eight Years Young.

ObjectiveIn March 2014, ICES launched Data &amp; Analytic Services (DAS), expanding the access to ICES data and analytics beyond ICES scientists and analytic staff. In eight years, DAS has grown and evolved to increase high quality services offered to an expanding client base of external researchers.&#x0D; ApproachAt the inception of DAS, two services were offered to public sector researchers: data access and analytics. Data access enabled researchers to analyze coded record-level data through a secure virtual environment. Analytics, conducted by DAS staff in ICES analytic environment, provided researchers with risk-cleared summary level reports. In response to growing demand from an increasingly diverse range of researchers, ICES engaged in extensive consultations with internal and external stakeholders to re-evaluate and operationalize new services. Compliance with contractual obligations and Ontario law, organizational capacity to scale up, alignment with ICES’ mission, vision and values, were cornerstones in establishing new offerings.&#x0D; ResultsAnalytic services became available to private sector researchers in June 2016. In March 2017, support for cohort and longitudinal follow-up studies became the newest service offering (researchers provided with a list of applicable individuals defined for the purposes of conducting publicly funded research). As more data assets become available to researchers, requests continue to increase in volume and complexity, particularly of projects seeking to import external data for linkage to ICES data. A second high performance computing virtual environment onboarded researchers September 2021 while the original analytic environment has undergone multiple upgrades, and will soon be fully refreshed. Regular solicitation of feedback has enabled DAS to increase staffing and diversify resources available which improves the client experience at all stages.&#x0D; ConclusionsSince its inception, DAS has expanded from five to thirty personnel, grown and diversified its new and returning client base and has responded to demand for new services. DAS continues to provide high quality services which enable impactful research and is responsive to new opportunities for collaboration and service provision.

Online symptoms self-assessment during COVID-19 pandemic: an analysis of a COVID-19 portal responses from Canada

COVID-19 case was first identified in Canada on January 25, 2020, on a Toronto resident who had travelled to Wuhan China, and not long after, the WHO declared the viral infection a pandemic. Ontario health West created an online self-assessment portal that allowed individuals in the health region and adjourning areas to report any COVID related symptoms. The purpose of this study was to evaluate the utility and usefulness of the Ontario Heath West online COVID-19 self-assessment portal. Record level data obtained from the Ontario Health West self-assessment portal was analyzed. Descriptive statistics using charts and graphs were used to characterize the distribution of responses to the portal. In-depth analysis using correlation, lead-lag analysis, and trend comparison with actual Government of Ontario COVID-19 cases for the region were also conducted. A total of 34,144 distinct responses were recorded on the portal between April 10 and July 29, 2020, with 1,250 (3.7%) responding positively to one of the emergency symptoms questions. Trend analysis showed a peak portal response in May 2020 with a smaller rise subsequently in July 2020, coinciding with the actual COVID-19 peak in the region. The five most reported symptoms on the portal were sore throat (17.2%), headache (12.9%), fatigue (12.3%), digestive problems (12.2%) and cough (9.1%). For four sub-regions, the trend of self-report on the portal positively lagged actual Public Health Ontario reported COVID-19 cases, while for one sub-region, the trend positively led the actual Public Health Ontario reported COVID-19 cases for the area. We found correlation between online COVID-19 self- assessment data and the confirmed COVID-19 cases in the Southwestern region of Ontario. Trends in the COVID-19 associated emergency symptoms reported on the portal also tracked confirmed COVID-19 cases in the community. Peak response to the portal coincided with the peak volume of confirmed cases in Ontario during the first wave of COVID-19 pandemic in Canada, suggesting some consistency between the experiences of portal users and patterns of COVID-19 illness in the community. The portal was a useful tool at the person-level because it provided guidance to individuals about how to access appropriate health services according to the symptoms that they reported and connected them with primary care, reducing unnecessary visit to health facilities for COVID-19 related care.

Data Privacy Protection and Utility Preservation through Bayesian Data Synthesis: A Case Study on Airbnb Listings

ABSTRACT When releasing record-level data containing sensitive information to the public, the data disseminator is responsible for protecting the privacy of every record in the dataset, simultaneously preserving important features of the data for users’ analyses. These goals can be achieved by data synthesis, where confidential data are replaced with synthetic data that are simulated based on statistical models estimated on the confidential data. In this article, we present a data synthesis case study, where synthetic values of price and the number of available days in a sample of the New York Airbnb Open Data are created for privacy protection. One sensitive variable, the number of available days of an Airbnb listing, has a large amount of zero-valued records and also truncated at the two ends. We propose a zero-inflated truncated Poisson regression model for its synthesis. We use a sequential synthesis approach to further synthesize the sensitive price variable. The resulting synthetic data are evaluated for its utility preservation and privacy protection, the latter in the form of disclosure risks. Furthermore, we propose methods to investigate how uncertainties in intruder’s knowledge would influence the identification disclosure risks of the synthetic data. In particular, we explore several realistic scenarios of uncertainties in intruder’s knowledge of available information and evaluate their impacts on the resulting identification disclosure risks.

Common and Unique Barriers to the Exchange of Administrative Healthcare Data in Environmental Public Health Tracking Program.

CDC’s National Environmental Public Health Tracking Program (Tracking Program) receives administrative data annually from 25–30 states to track potential environmental exposures and to make data available for public access. In 2019, the CDC Tracking Program conducted a cross-sectional survey among principal investigators or program managers of the 26 funded programs to improve access to timely, accurate, and local data. All 26 funding recipients reported having access to hospital inpatient data, and most states (69.2%) regularly update data user agreements to receive the data. Among the respondents, 15 receive record-level data with protected health information (PHI) and seven receive record-level data without PHI. Regarding geospatial resolution, approximately 50.0% of recipients have access to the street address or census tract information, 34.6% have access to ZIP code, and 11.5% have other sub-county geographies (e.g., town). Only three states receive administrative data for their residents from all border states. The survey results will help the Tracking Program to identify knowledge gaps and perceived barriers to the use and accessibility of administrative data for the CDC Tracking Program. The information collected will inform the development of resources that can provide solutions for more efficient and timely data exchange.

Interoperability of population-based patient registries

Enabling full interoperability within and between population-based patient-registry domains would open up access to a rich and unique source of health data for secondary data usage. Previous attempts to tackle patient-registry interoperability have met with varying degrees of success, but a unifying solution remains elusive. The purpose of this paper is to show by practical example how a solution is attainable via the implementation of an existing framework based of the concept of federated, semantic metadata registries. One important feature motivating the use of this framework is that it can be implemented gradually and independently within each patient-registry domain. By employing linked open data principles, the framework extends the ISO/IEC 11179 standard to provide both syntactic and semantic interoperability of data elements with the means of specifying automated extraction scripts for retrieval of data from different registry content models. The examples provided address the domain of European population-based cancer registries to demonstrate the feasibility of the approach. One of the examples shows how quick gains are derivable by allowing retrieval of aggregated core data sets. The other examples show how aggregated full sets of data and record-level data might also be retrieved from each local registry. An infrastructure of patient-registry domains adhering to the principles of the framework would provide the semantic contexts and inter-linkage of data necessary for automated search and retrieval of registry data. It would thereby also lay the foundation for making registry data serviceable to artificial intelligence (AI) applications.

Exploring Missouri's New, Innovative, and Multilayered Health Data Suppression Rules

ObjectiveBy the end of this session, users will be able to describe the innovative and multilayered suppression rules that are applied to Missouri's homegrown health data web query system. They will also be able to use the lessons learned and user feedback described in the session to facilitate discussions surrounding the application of suppression to their specific data systems.IntroductionIn Spring 2017, the Missouri Department of Health and Senior Services (MODHSS) launched the Missouri Public Health Information Management System (MOPHIMS) web-based health data platform. Missouri has supported a similar data system since the 1990s, allowing the public, local public health departments, and other stakeholders access to community level birth, death, and hospitalization data (among other datasets). The MOPHIMS system is composed of two separate pieces. Community Data Profiles are topic-, disease-, or demographic-specific reports that contain 15-10 indicators relevant to the report. Because these static reports are developed in-house a multilayered suppression rule is not required. The second piece of MOPHIMS, the Data MICAs, or Missouri Information for Community Assessent, can be used to create customized datasets that slice and dice up to a dozen demographic and system-specific variables to answer complex research questions.The MOPHIMS interface features, among other things, a new and innovative method for addressing confidentiality concerns through the suppression of health data. This pioneering approach integrates multi-level logic that uses inner and outer cell analytics, the use of exempt and conditionally exempt variables, and multiple levels of user access. Moving beyond a simple model of suppressing any values below a certain threshold, MOPHIMS takes a bold step in providing users exceptionally granular data while still protecting citizen privacy.MethodsIn order to implement this new suppression methodology, MODHSS worked with both internal information technology resources (OA-ITSD) and outside contractors to develop the suppression rules utilized in the Data MICAs. Before these meetings began, MODHSS analysts met weekly to determine the overall goals and frames for the rule, knowing that writing the code to implement the complicated and comprehensive vision would be a collaborative and iterative process. Because the MOPHIMS system is homegrown and this specific confidentiality process is not currently utilized (to our knowledge) elsewhere, all of those at the discussion table were required to be innovative, open to criticism, and willing to engage in extremely detailed explanations. A team of users from Missouri’s local public health departments provided feedback throughout this process.A basic description of the process flow that occurs before suppression is applied in MOPHIMS follows. To begin, de-identified record-level data are loaded into online analytical processing (OLAP) cubes and relational databases. No suppression is applied to these back end databases. The information is then aggregated for display on the front end screens of the Data MICAs based on customized user selections. Depending upon which level of access a user has logged in, suppression is then applied to the data output generated using these customized selections. Not only are the rules applied to data tables but also to the MOPHIMS data visualization tools, which include multiple types of charts and maps.ResultsIn addition to the rules themselves, MOPHIMS contains a mechanism that allows users to log in at different levels of access. Public and Registered user levels are free and available to all operators with a valid e-mail address. Partner level access is reserved for epidemiologists at the state and local level who are using the Data MICAs for program planning, evaluation, and grant writing. Because these individuals are required to adhere to the same data dissemination policies as those who create the MOPHIMS system, Partner level access turns off suppression in the MOPHIMS system. Values that would be suppressed at the Public or Registered user levels are shown in italicized, red font. A multi-level approval process is required for individuals to obtain Partner level access to MOPHIMS.ConclusionsMODHSS created an innovative suppression system that allows public health planners to access granular data through customizable queries without risking a confidentiality breach. Users have indicated this is highly preferable to a blanket suppression rule that hides any value under a certain threshold. Additionally, approved MOPHIMS users can view specially formatted values that would otherwise have been suppressed. The flexibility associated with creating a homegrown web query system has allowed the formation and implementation of this multilayered rule, which likely would not have been possible if using an off-the-shelf product. Data disseminators are encouraged to review current confidentiality and suppression rules to determine whether they might be modified to provide more granular data users while still protecting the privacy of citizens.

The Rate Stabilizing Tool: Generating Stable Local-Level Measures of Chronic Disease.

Accurate and precise estimates of local-level epidemiologic measures are critical to informing policy and program decisions, but they often require advanced statistical knowledge, programming/coding skills, and extensive computing power. In response, we developed the Rate Stabilizing Tool (RST), an ArcGIS-based tool that enables users to input their own record-level data to generate more reliable age-standardized measures of chronic disease (eg, prevalence rates, mortality rates) or other population health outcomes at the county or census tract levels. The RST uses 2 forms of empirical Bayesian modeling (nonspatial and spatial) to estimate age-standardized rates and 95% credible intervals for user-specified geographic units. The RST also provides indicators of the reliability of point estimates. In addition to reviewing the RST’s statistical techniques, we present results from a simulation study that illustrates the key benefit of smoothing. We demonstrate the dramatic reduction in root mean-squared error (rMSE), indicating a better compromise between accuracy and stability for both smoothing approaches relative to the unsmoothed estimates. Finally, we provide an example of the RST’s use. This example uses heart disease mortality data for North Carolina census tracts to map the RST output, including reliability of estimates, and demonstrates a subsequent statistical test.

A Toolbox for Research – Population Data BC’s tools for overcoming challenges in the data access process

IntroductionNavigating the regulatory environment of data access is complex. Where there are common challenges, there are opportunities to develop systems and services that can help researchers. These tools are designed to ensure secure and smooth navigation of the process of data access and data analysis.&#x0D; Objectives and ApproachThere is increasing focus on technical solutions to support data requests and analysis. We have responded to existing challenges by developing a suite of tools and services including: MetaData Central, DARonline, AppTracker, Snippets, and REDCap. We will describe the obstacles that motivated each tool, how it was built, and how it addresses the original issue. We will also discuss the overall challenges in building these tools, including technical complexity and getting buy-in from data stewards and researchers. Finally, we will touch on some of the progress we have seen based on these tools, and our plans for future development.&#x0D; ResultsMajor obstacles included providing information to researchers to help form their data request, communications among researchers, data stewards and PopData, general automation of procedures, and sharing among researchers. The tools developed in response include: MetaData Central, AppTracker (for project tracking); DARonline (for data requests); REDCap (for primary data collection); and Snippets (for code sharing). MetaData Central has proven to be a valuable resource for all stakeholders to learn basic data descriptions without needing access to the record level data. This helps in making sure the data requested meets the needs of the research project. AppTracker is helpful for monitoring applications, and reduces the time it takes for us to monitor compliance and provide reports. Other tools are newer and have less evaluation information.&#x0D; Conclusion/ImplicationsUser-friendly, flexible tools are beneficial for making the data access process smoother and for meeting the evolving needs of researchers and data stewards. These tools can create both transparency through a common source of information and efficiency by automating routine processes.

Unconsented linkage between dormant trials and administrative data: practical and regulatory implications

IntroductionHalf of all infants are fed formula milk. However, attrition biases evidence on the long-term safety of formula ingredients. We used unconsented linkage between administrative education and health records of young people who were randomised as infants to formula milks, to determine long-term safety and efficacy.&#x0D; Objectives and ApproachWe used record level data from a series of 9 historical randomised controlled trials (RCTs) conducted in 1982-2002 (n=3,500 participants), which are key to the evidence-base around formula-composition. All later follow-ups are biased by attrition leading to limited evidence around the long-term effects of formula ingredients on cognition and metabolic and cardiovascular health. We sought permissions from data providers and regulatory agencies for unconsented linkage to education and hospital records, as proxy measures for cognitive and health development. We discuss the steps that were implemented to safeguard the participants' privacy and achieve ethical and multi-institutional approval for this project.&#x0D; ResultsAchieving provisional ethical approval took 41 days. Achieving agreement in principle to match trial data to individual level education records took 4 months and 2 weeks, while agreement to match trial data to individual level hospital records is still underway (5.5 months in February 2018). Delays in institutional approval were largely due to unharmonised data security certificates between the two government departments holding the health and education records. Digitising and cleaning all handwritten RCT participant identifiers prior to linkage took 9 months of full-time researcher time. Maintaining separation of identifiers and attribute data required specific secure haven provision. Results on the success of linkage between RCTs and education records will be presented at the conference.&#x0D; Conclusion/ImplicationsWhile directly contributing to the evidence around infant-formula-composition, this project will also act as a proof-of-concept study. Unconsented linkage between dormant RCTs and administrative data could be a novel and cost-effective method to generate evidence on the long-term efficacy and safety of interventions.

Development of Syndrome Definitions for Acute Unintentional Drug and Heroin Overdose

ObjectiveTo develop and evaluate syndrome definitions for the identificationof acute unintentional drug overdose events including opioid, heroin,and unspecified substances among emergency department (ED) visitsin Virginia.IntroductionNationally, deaths due to opioid overdose have continuallyincreased for the past 15 years1. Deaths specifically related to heroinincreased more than four-fold between 2002 and 20142. Hospitalinpatient discharge data provide information on non-fatal overdoses,but include a significant lag in reporting time3. Syndromic ED visitdata provide near real-time identification of public health issues andcan be leveraged to inform public health actions on the emergingthreat of drug overdose.MethodsVirginia Department of Health (VDH) developed two syndromedefinitions in 2014 to capture acute unintentional drug overdoseevents among syndromic ED visit data. Syndrome 1 captured visitsfor overdose, whether or not a specific substance was mentioned.Syndrome 2 captured only visits for heroin overdose. Definitionswere based on free-text terms found within the chief complaintand standardized text or International Classification of Diseases(ICD) codes within the diagnosis field. In 2016, both definitionswere revised to identify additional inclusion and exclusion criteriaaccording to CDC guidance documentation and syndrome definitionsused by other state jurisdictions.Microsoft SQL was used to modify both definitions based on thenewly identified chief complaint and diagnosis criteria. Record leveldata were analyzed for their adherence to established criteria using aniterative evaluation process.The scope of Syndrome 1 (2016) was narrowed from the 2014version by excluding visits for non-opioid substances, heroin, andnon-acute indicators. It included chief complaint and diagnosisterms related to opioids, unspecified substance overdose, narcotics,and Narcan or naloxone, and excluded terms related to suicide,alcohol overdose alone, withdrawal, detoxification, rehab, addiction,constipation, chronic pain, and any specified non-opioid drug ormedication. Syndrome 2 (2016) included chief complaint or diagnosisterms mentioning heroin overdose and excluded suicide, withdrawal,detoxification, rehab, and addiction. Visits with mention of suicide,rehab, or addiction were identified during the evaluation process,resulting in the exclusion of these terms in the revised query.From January 1, 2015 to July 31, 2016, the number of visitscaptured by the revised syndrome definitions was compared to thenumber captured by the 2014 definitions. Correlation coefficientswere calculated using SAS 9.3.ResultsThe revised Syndrome 1 found 4296 fewer ED visits(29% decrease) for acute unintentional drug overdose betweenJanuary 1, 2015 and July 31, 2016 compared to the 2014 definition.Despite the drop in volume, the monthly trends were similar forthe 2014 and 2016 definitions (correlation coefficient = 0.95,p &lt; 0.001). For the same time period, the revised Syndrome 2 definitionreturned 108 fewer visits (6% decrease) for acute unintentional heroinoverdose. The monthly trends were also similar for the 2014 and 2016definitions (correlation coefficient = 0.98, p &lt; 0.001).ConclusionsBoth revised syndrome definitions improved specificity incapturing overdose visits as Syndrome 1 (2016) identified 29% fewervisits and Syndrome 2 (2016) identified 6% fewer visits found to beunrelated to the desired overdose criteria.When developing the revised syndrome definitions, VDH decidedto exclude non-acute drug-related visits. Terms such as addiction,detoxification, rehab, withdrawal, chronic pain, and constipation wereindicative of habitual drug use or abuse instead of acute overdose andwere thus excluded. In narrowing the scope of Syndrome 1, VDHalso identified and excluded visits for specified drug and medicationoverdose. Together, these expanded exclusion criteria resulted ingreater specificity with both updated syndromes.These revised syndrome definitions enable VDH to better trackopioid and heroin overdose trends in near real-time and overextended time periods which can be used to inform public healthactions. Limitations include the inconsistency of diagnosis codingamong syndromic data submitters, which may lead to geographicunderrepresentation of unintentional drug overdose visits based onthe location of health care systems. VDH will continue to evaluate andrefine these overdose syndrome definitions as this emerging healthissue evolves.

The Nascent Pan-Canadian Real-world Health Data Network (PRHDN)

ABSTRACT&#x0D; ObjectiveCanada’s large single payer health systems have created provincial centres with rich, population-wide health and social data holdings that are linkable at the person level. Complementary federal data include a subset of standardized and enriched administrative healthcare data at the Canadian Institute for Health Information, Statistics Canada’s extensive array of survey- and census-based social and geography-based information, and large national datasets that are being created through pan-Canadian initiatives. Unfortunately, these data assets are rarely combined in multi-province or pan-Canadian studies, often because data are not directly comparable from one province to another, or cannot be shared due to legislative or other barriers. There is now growing interest in enabling multi-province studies — even when data are not directly comparable — and in sharing experiences to make more effective use of linked or linkable administrative data across Canada.&#x0D; ApproachNine provincial and national organizations have created a detailed Implementation Plan for the new PRHDN distributed data network. Without requiring that record-level data leave provincial boundaries, the PRHDN will create shared core research data infrastructure: (i) validated algorithms that implement case definitions applicable across provinces, (ii) harmonized common data and (iii) common analytic protocols. The PRHDN will also establish complementary infrastructure including dedicated personnel to assist researchers and decision makers, joined-up training and capacity building sessions, opportunities to share learning and experience related to linking new datasets such as electronic medical records and omic datasets, and forums for knowledge translation and exchange with decision makers.&#x0D; ResultsThe PRHDN is already bringing together expertise from across Canada as researchers, decision makers and data custodians begin to identify opportunities for enhanced use of health data in Canadian research, policy making and practice. A simple PRHDN website has been created (https://www.prhdn.ca/) and more than 200 researchers and policy/decision makers have become members of the PRHDN consortium. Surveys of consortium members are identifying priorities for the first algorithm validation work (to-date the top four priorities are mental health, cardiovascular disease, diabetes and respiratory disease) and the PRHDN Leads Team is functioning as a decision-making body, e.g., in discussions with Statistics Canada.&#x0D; ConclusionsWorking together, provincial and national organizations across Canada have identified concrete steps that can be taken to enable multi-province and pan-Canadian studies based on administrative data within one year of the start of funding. The PRHDN Leads Team is currently discussing the PRHDN vision and Implementation Plan with potential funders.

Institute for Clinical Evaluative Sciences (ICES) Exploratory Data &amp; Analytic Services Private Sector Pilot Project

ABSTRACTObjectivesPrior to the launch of ICES Data &amp; Analytic Services (DAS) in March 2014, only ICES scientists and analysts could access ICES data, and data could only be accessed at physical ICES locations. The DAS infrastructure, which allows public sector researchers to work with coded record level data remotely through a secure virtual environment, together with broader trends including high profile reports that call for increased access to data and the Ontario government’s Open Data initiative, prompted ICES to launch a pilot project to explore potential DAS work with the private sector.&#x0D; ApproachThree mandatory principles were established for all work with the private sector: (i) alignment with ICES’ mission, vision and values; (ii) transparency; (iii) private sector work must not detract from ICES’ research institute work. The pilot included: a jurisdictional scan; informal conversations with private sector organizations to determine potential services/studies of interest; extensive discussions with data partners; the selection and conduct of two pilot studies; focus groups with members of the general public and scientists; external advice on business model options; and an external evaluation of the pilot. No changes to data sharing agreements or ICES processes were required as work with the private sector and public sector are equally allowed under Ontario law.&#x0D; ResultsThe two pilot studies were successfully completed. The first study “The disease burden of gout in Ontario: A real world data retrospective study” was performed by researchers at IMS Brogan (a healthcare analytic services provider) who were provided with access to coded record-level data using the DAS iDAVE environment and performed their own analyses. In the second pilot study, “The impact of adherence to biologics on healthcare resource utilization in rheumatoid arthritis”, Janssen researchers established the research question and study design, and DAS staff and scientists provided advice about data holdings, performed the analyses, and provided Janssen and three government-funded decision making bodies with results tables. Research Ethics Board approval was required for both studies, and both private sector organizations are in the process of publishing findings.&#x0D; ConclusionsICES was able to work with private sector organizations without compromising the three principles. Based on the evaluation of the private sector pilot, and the findings from the focus groups, ICES will begin offering limited analytic services to private sector researchers beginning June 2016 under ICES’ existing corporate structure, and bring recommendations regarding ongoing operations to the ICES Board in June 2017.

Clinical quality registries have the potential to drive improvements in the appropriateness of care.

The provision of timely, relevant and reliable information on patient care to clinicians has been shown to drive improvements in health care quality. Well constructed clinical quality registries collect and report information on both the appropriateness of care (process) in keeping with clinical practice guidelines and the effectiveness of care (outcomes). Notwithstanding the successful establishment of several new registries and improvements in established registries, barriers persist for clinical groups wishing to improve the quality of information and level of participation in registries in Australia. To address these barriers, the Australian Commission on Safety and Quality in Health Care has developed the Framework for Australian Clinical Quality Registries. The Framework describes a mechanism by which government jurisdictions and private hospital groups can authorise and secure record-level data, within high priority clinical domains, to measure, monitor and report the appropriateness and effectiveness of health care. The provision of benchmarked information back to clinicians on the appropriateness and outcomes of care is expected to improve adherence to evidence-based practice and drive improvement in outcomes.