Midstream Transformation of Colorado’s Wastewater Data Pipeline

Wastewater surveillance (WWS) emerged during the COVID-19 pandemic as a powerful new public health tool to monitor the burden of COVID across communities. Unlike clinical test-based strategies, wastewater-based estimates of COVID-19 prevalence are not biased by individual testing practices and care-seeking behaviors. Further, changes in disease burden may be captured more rapidly using WWS than other surveillance systems. During the early months of the COVID pandemic, a large range of utilities, laboratories, and public health entities began monitoring wastewater for SARS-CoV-2 in Colorado. Collaboration between many of these early-adopters of WWS resulted in the formation of the Wastewater Collaborative, a group of wastewater utilities, public universities, private laboratories, and local and state public health entities intended to develop WWS for SARS-COV-2. he goal is to improve public health understanding and response to the COVID-19 pandemic. The rapid development of the Wastewater Collaborative necessitated an equally rapid development of data structures to support needs of the constituent organizations. The Phase One data system achieved this goal by leveraging existing relationships within the Colorado university system (Figure 1). Wastewater samples were primarily tested by university labs, analyzed, and compiled using university data resources. COVID-19 case and testing data was sent to the university data partners by the Colorado Department of Public Health and Environment (CDPHE) to contextualize wastewater analysis. The results of these analyses were displayed on a password-protected university-run dashboard. Time-delayed results were also published on a second, public dashboard run by CDPHE using data from the university dashboard. However, as WWS in Colorado matured, important limitations of the Phase 1 data system became evident. These limitations included the difficulty of adding additional utility and laboratory partners, the necessity of aggregating and de-identifying other surveillance data prior to sending it to external data systems, challenges in integrating additional types of data (including mutation and sequencing analysis), and connecting to internal and external data clients (including the CDC's National Wastewater Surveillance System (NWSS)). Because of CDPHE's role in funding and facilitating WWS in Colorado, CDPHE undertook the process of creating and implementing a Phase Two data system, which would achieve the following goals: Support multiple testing laboratories, including public, private, and university laboratories. Support multiple types of sample analysis, including viral RNA quantification, specific mutation quantification, and viral genome sequencing. Support analysis using granular individual-level case, testing, and hospitalization data. Facilitate timely reporting and integration of sample metadata and testing data. Creation of a single data visualization and export platform for internal partners. Expansion of the existing public data visualization platform to accommodate new utilities. Development of data analytic methodologies, including implementation of methodologies developed by NWSS and elsewhere. Support timely dissemination of data and analytics to both CDPHE leadership, local public health agencies, and other local and state policymakers. Facilitate integration of Colorado data with national datasets via NWSS. The Phase Two data system achieved these goals by utilizing CDPHE's pre-existing technology resources and open source data tools (Figure 2). External testing partners are able to input data directly into a shared Google Drive folder in a format of their choosing, and this data is automatically downloaded, parsed, and compiled alongside any CDPHE testing data using a purpose-written script in R-studio. Internal data from CDPHE's other surveillance systems (including individual-level testing, case, and hospitalization data) can be analyzed alongside wastewater testing data (including viral RNA concentration, mutation data, and sequencing data). Data is immediately available for view and download after import on a new private dashboard developed in R-shiny. After a 24-hour viewing period for utilities and local public health agencies, data is also available for view and download on a pre-existing public facing dashboard and uploaded to NWSS. Ultimately, the Phase Two data system has been a significant improvement over the Phase One system, with superior data precision, flexibility in adding new metrics and sites, and the ability to integrate with NWSS and other external data systems. Further, many of the key strengths of the Phase One system are maintained, including data timeliness, fidelity, and visualization tools (Table 1). This data evolution achieved our goals of ensuring the stability and sustainability of WWS for COVID to continue to support the pandemic response.