Community of Practice Updates

*Slack is a registered trademark and service mark of Slack Technologies, Inc.  **If you have questions about the NSSP CoP, its highly collaborative user groups, the NSSP CoP Slack Workspace (a collaboration platform), or syndromic surveillance, please email syndromic@cste.org.

Policy for Federal Access to NSSP Data

During the November 2022 CoP Monthly meeting, Acting NSSP Lead Karl Soetebier informed the community of efforts to make NSSP data more accessible and to improve collaboration toward common goals.

In early December 2022, Soetebier updated site administrators of NSSP’s ongoing work related to data use. He explained how expanded access to NSSP data during the COVID-19 public health emergency enabled innovation in areas such as trend indicators and classification, anomaly detection, and text mining by age and geography. The ability to work this way routinely, outside the context of a public health emergency, is not permitted by the current data use agreement.

In early 2023, to build on public health response innovations and to continue to enhance data use, CDC has been working on designing a new NSSP agreement that will incorporate lessons learned from the COVID-19 response. The new agreement will enable close collaboration between sites and CDC. The agreement will also enable new innovations and services for sites, maximize responsible use of data and provide timelier synthesis of findings and recommendations, and help respond to the top concerns raised by public health departments in the Federal Access to National Syndromic Surveillance Program Data: Review and Implementation Strategies.

After gathering valuable input, CDC has carefully decided to change from a data use agreement (DUA) to a memorandum of understanding (MOU) with sites to better reflect the collaborative nature of the agreement. By refocusing the DUA as an MOU, the collaborative and participatory goals of the partnership across federal, state, and local public health authorities become more transparent.

To inform community members throughout this process, we will make a new Policy for Federal Access to NSSP Data web page available in spring 2023. It will contain information on the development and long-term benefits we hope to achieve, so stay tuned!

Just released! Review of Federal Access Policies for NSSP Data

In February 2023, The Council of State and Territorial Epidemiologists (CSTE) posted the Federal Access to National Syndromic Surveillance Program Data: Review and Implementation Strategies.

CSTE collaborated with CDC and Texas A&M University School of Public Health to develop this report. They brought together epidemiologists in leadership positions or with decision-making power who represented state, tribal, local or territorial (STLT) public health departments to identify considerations and implementation strategies for permitting federal access to NSSP data.​

For additional context and details about the process and findings, listen to author/contractor Cason Schmit’s presentation for the October 2022 NSSP CoP monthly call. Please direct questions about the report to syndromic@cste.org.

NSSP CoP Monthly Meeting

The National Syndromic Surveillance Program (NSSP) Community of Practice (CoP) met on February 22, 2023. On average, 100 to 120 people participate in these monthly meetings. Recordings for CoP monthly calls are posted in the Knowledge Repository.

NSSP Update

Comparing Overdose ED Data Systems in Michigan: Surprises Abound!

This presentation is an excellent example of how a public health jurisdiction can examine data acquired from different data systems; contrast the strengths, weaknesses, and results of each; and identify data quality improvements. As anyone who works with syndromic data can attest, syndromic data are not always complete and need context and interpretation. What better way to understand your data than by examining it periodically from different perspectives and digging into outliers and anomalies? The results from this analysis have already presented opportunities to collaborate with others to make changes and have informed long-term surveillance goals.

Introduction

Presenter Gabrielle Stroh-Steiner is the lead epidemiologist for the Overdose Data to Action (OD2A) Program in Michigan. She has worked with the program and with syndromic surveillance for about 5 years. In this presentation, she discusses an analysis in which her public health jurisdiction compares three emergency department (ED) data systems used for overdose surveillance and then shares the lessons learned. The reason for initiating this analysis is that Michigan uses each data source for different purposes in overdose surveillance and wanted to make sure data from the different systems aligned and were being reported accurately. Although this analysis was conducted in early 2022, the results and implications hold true today.

Overview of Michigan Data Sources

Michigan uses three ED data sources for overdose surveillance:

• Michigan Health and Hospital Association (MHA)—Data are comprehensive, high quality, slower (6- to 9-month lag and not useful as an alerting mechanism), and missing one key hospital.
• Michigan Syndromic Surveillance System (MSSS)—Data are timely and well understood. Coverage is slightly lower. MSSS onboards each facility.
• MiCelerity, a Michigan-designed system implemented in 2020—Data are timely, identifiable, flexible, and owned by the unit. Since the data source is new, the unit is still learning about its limitations.

MHA data are billing and discharge data collected and purchased quarterly. MHA does all data cleaning and maintenance.

MSSS and MiCelerity are both ADT-based systems. Despite system similarities, a key difference is that MSSS collects data from all state ED visits, whereas MiCelerity only collects substance use-related data. MiCelerity receives data from the health information exchange (HIE); consequently, if an ADT message contains an ICD-10 CM code related to overdose, it will trigger the go-between HIE to send data to the system. MSSS functions a little differently in its collection of ADT data from electronic health records.

Methodology

The analysis compared the most recent year of data available for the three systems: July 2020–June 2021. They looked at 86 facilities that reported to all three data systems. They calculated the frequency of (1) all drug overdose (T36-T50), (2) Opioid overdose (T40.0-T40.4, T40.6), and (3) Unspecified drug overdose (T50.9). Then they compared frequency by time, facility, geography, and demographic groups. They used line graphs, scatter plots, and correlation coefficients.

Anticipated Results…

Going into the analyses, the epidemiologists proposed several hypotheses: They anticipated seeing similar all-drug overdose counts over time across all three data sources. They anticipated lower counts of opioid overdoses in MSSS and MiCelerity than in MHA but comparable counts between MSSS and MiCelerity. In general, they anticipated comparable findings between MSSS and MiCelerity. They also anticipated similar patterns across facilities and counties.

Surprise! The Actual Results…

They looked at number of drug overdose by week in each of the data sources and found the first unexpected finding. Over time, the patterns were similar. MiCelerity and MHA data were similar in magnitude of drug overdose ED visits, whereas the syndromic system, MSSS, had on average about 24% fewer drug overdose ED visits.

They looked at opioid overdoses by week to get a glimpse at how drug-specific overdoses look over time. As expected, MiCelerity and the syndromic system, MSSS, had fewer opioid overdoses on average than MHA. Looking at the correlation across data sources, there was a strong positive correlation across all three.

When they looked at the percent of total drug overdoses unspecified over time in each of the three data sources, a significant number of overdoses in MiCelerity and MSSS were unspecified. MHA had the highest percentage of specified overdoses.

Scatter plots of the overdose counts by facility revealed interesting takeaways. In general, larger facilities with more overdoses in one system showed more in the other system—but they observed more variation than anticipated. Also, some facilities were significant outliers. Initially, the team expected the outliers to be mostly under-reporters in the ADT-based systems (MiCelerity, MSSS) compared with MHA, but that wasn’t always true. They also looked at the correlation between data source by facility, including and excluding outliers. Once the outliers were removed, the correlation improved substantially.

Hot Topics/Open Forum

Anna Frick (AK) kicked off the discussion by asking attendees to contribute more STLT talks—short or long—and to bring thorny issues or questions to the Community. She suggested that if anyone has a partially done project, try tapping the NSSP CoP’s collective mind to make progress.

Several topics were raised:

• Jade Hodge (KS) asked about state hospital association data and experiences with data quality processes and clearinghouses.
• Dave Atrubin (FL) noted that the lines are blurring for reporting and that syndromic data, in part, are starting to be used for reportable disease. This new ground needs more consideration. The Michigan presentation is an excellent way to engage others in this discussion.
• Doug Cretsinger (IA) asked about effective engagement practices for the Community. The Community suggested numerous ways in which this can be achieved, a few of which are shown in Tips for Engaging Staff in SyS.

Tips for Engaging Staff in SyS

Each of us juggles competing priorities. These tips will help you and your team evolve syndromic surveillance (SyS) and promote the use and integration of emergency department data across your organization.

• Join the NSSP Community of Practice (CoP). Attend monthly NSSP CoP meetings. By far, this is the best way to connect with your colleagues and learn different approaches to solving common problems. Find ways to introduce these topics into local meetings.
• Join a subcommittee.
• Open a Slack account—and use it! Slack gives insight into hot topics and is a resource for consulting others and getting feedback fast.
• Engage your communications staff. They can publicize your syndromic successes.
• Check out symposium topics in the Knowledge Repository.
• Integrate syndromic findings into routine meetings.
• Set up ESSENCE office hours once a month to encourage exploration, new tools and practice, auto-reporting, etc., and discuss chat topics.
• Identify grants where syndromic data could be part of the solution. Some response efforts lead to grants.
• Collaborate by design and intent. Start building relationships across program areas. It takes time to engage others and grow knowledge and interest in syndromic data. Send periodic updates about what you’re doing (include data others don’t realize they need). Brainstorm project ideas, explore data gaps, and look for collaboration opportunities. Consider setting up an internal workgroup that meets regularly.
• Invite others to CoP NSSP meetings when the topic is in their purview.
• Identify internal champions and help them pursue agency engagement.