Background: Disease surveillance is a critical function of public health, provides essential information about disease burden, clinical and epidemiologic parameters of disease, and is an important element to effective and timely case and contact tracing. The COVID-19 pandemic has demonstrated the essential role these functions have to preserve public health. Syndromic surveillance, electronic laboratory reporting in the meaningful use program, and the growth of the National Healthcare Safety Network (NHSN) have created linkages between hospitals, commercial labs, and public health that can collect and organize data, often through EHR and order workflows, to improve the timeliness and completeness of reporting. In theory, the standard data formats and exchange methods provided by meaningful use should enable rapid healthcare data exchange in the setting of disruptive healthcare events like a pandemic. In reality, access to data remains challenging, and even if available, often lack conformity to regulated standards.
Objective: We sought to use regulated interoperability standards already in production to generate regional bed capacity awareness, enhance the capture of epidemiological risk factors and clinical variables among COVID-19 tested patients, and reduce the administrative burden of reporting for stakeholders in a manner that could be replicated by other public health agencies.
Methods: Following a public health order mandating data submission, we developed technical infrastructure to combine multiple data feeds from electronic health record systems. We measured the completeness of each feed, and the match rate between feeds.
Results: A cloud-based environment was created that received data from electronic lab reporting, consolidated clinical data architecture, and bed capacity data feeds from sites. Data governance was planned from the project beginning to aid in consensus and principles for data use. 88,906 total persons from CCDA data among 14 facilities, and 408,741 persons from ELR records among 88 facilities, were submitted. Fields routinely absent from ELR feeds included travel histories, clinical symptoms, and comorbidities. CCDA data provided an improvement in the quality of data available for surveillance and was highly complete with <5% for all records types with the exception of patient cell phone. 90.1% of records could be matched between CCDA and ELR feeds.
Conclusions: We describe the development of a city-wide public health data hub for the surveillance of COVID-19 infection. We were able to assess the completeness of existing ELR feeds, augment these feeds with CCDA documents, establish secure transfer methods for data exchange, develop cloud-based architecture to enable secure data storage and analytics, and produced meaningful dashboards for the monitoring of capacity and disease burden. We see this public health and clinical data registry as an informative example of the power of common standards across electronic records, and a potential template for future extension of the use of standards to improve public health surveillance.