A collaboration between the Houston Health Department, Houston Public Works, and Rice University has combined wastewater sampling, laboratory methods, statistical analysis and reporting to manage a comprehensive wastewater monitoring system that targets prominent outbreaks of infectious diseases.
Since the emergence of the COVID-19 pandemic, Houston Public Works has been taking weekly 24-hour composite influent wastewater samples from 38 wastewater treatment plants that serve 2.2 million people to detect SARS-CoV-2, the virus that causes COVID-19.
The work, along with the fast online algorithm that enabled real-time statistical evaluations and changes in virus levels from citywide trends from July 6, 2020, through October 28, 2024, was published in the October 1 issue of Data Science in Science.
“Multivariate nonlinear hierarchical state-space models were used to integrate multiple time-varying data streams to measure wastewater samples for citywide viral load and evaluate SARS-CoV-2 trends as they were increasing or decreasing,” said Katherine Ensor, Rice’s Noah G. Harding Professor of Statistics.
Ensor is a leading expert in state-space modeling and statistical process control frameworks, and is widely recognized for her work in mathematical statistics and methods in computational analysis. She has collaborated closely with the Houston Health Department’s Chief Environmental Science Officer, Loren Hopkins, and the department’s Data Science Division on numerous special projects focusing on public health and well-being since 2010, when Hopkins joined the department.
“Analyzing datasets over long timeframes has provided a greater understanding of how multiple complex variables influence each other and to quantify sources that influence process variation,” said Ensor. “Knowledge gained from years of work has provided an important pathway to optimize Houston’s wastewater monitoring system.”
“A significant result of the historical data and modeling analysis has provided insight into ways to right-size the wastewater monitoring system and reduce cost,” said Lauren Stadler, an associate professor of civil and environmental engineering at Rice.
Stadler and her lab have led efforts to concentrate, extract, sequence, and quantify pathogenic genes in wastewater treatment plant samples. Over the past year, the laboratory processing and analysis of SARS-CoV-2 have been transferred to the Houston Health Department. Meanwhile, Stadler specializes in novel detection and quantification methods for detecting the Human Papillomavirus (HPV), measles, mumps, and rubella viruses, as well as bacterial and fungal pathogens in wastewater samples.
Stadler’s lab also collaborates with the Houston Health Department to analyze for seasonal influenza and respiratory syncytial virus in preK-12 schools in the Houston Independent School District (HISD).
“Historical data from this recent study, which was focused on COVID-19, provided a protocol basis to reduce cost by physically combining samples in equal parts from wastewater treatment plants before laboratory analysis,” said Ensor.
Instead of conducting laboratory analysis of samples from each of the 38 wastewater treatment plants, weekly samples in equal parts from 15 wastewater treatment plants were pooled before analysis.
The pooling study demonstrated that, despite regular collections from large community sources, the team was still able to maintain monitoring coverage across the city.
Julia Schedler, a former Rice Ph.D. student and research scientist under Ensor’s direction who is now an assistant professor of statistics at Cal Poly, worked with Ensor to compare three designs for pooling the sewershed time series based on expert knowledge of the city to best serve the population, geographic median extended Hausdorff distance between sewersheds, clustering based on geographic median extended Hausdorff distance between sewersheds plus additional information gleaned from model observations.
The statisticians found there is strong consistency in citywide SARS-CoV-2 infection trends across the three grouping algorithms, as well as compared to the model using information from each sewershed.
Confident in the analysis, the Houston Health Department brought the pooling into production in November 2024 as a sustainable approach in wastewater-based epidemiology. Rice interns Brandon Fantine, an undergraduate student majoring in statistics with a minor in data science, and Tina C. Li, a graduate student pursuing a professional master’s in electrical and computer engineering with a specialization in AI and data science, helped bring a version of the state-space model into the production process this past summer.
“In addition to city-wide monitoring and reporting, institution-level monitoring through manholes that serve the city jail, schools, and nursing homes is an integral part of our systematic approach to target pathogens at community levels,” said Hopkins.
The strategies created in this study provide pathways to assist with scaling and right-sizing methods in wastewater-based epidemiology for multi-pathogen surveillance that pose significant health risks.
In addition to Ensor, Schedler, Stadler, and Hopkins, manuscript co-authors include analyst Rebecca Schneider and senior analyst Kaavya Domakonda of the Houston Health Department’s Data Science Division, who both oversee daily implementation of the Houston wastewater monitoring system. Rice Ph.D. student Jose Palacio contributed to the statistical code. Schedler manages the GitHub repository https://github.com/hou-wastewater-epi-org.
- Shawn Hutchins, Communications Specialist