# Assessing the sensitivity and predictive value of wastewater in detection of Hepatitis A cases in San Diego County

## TL;DR

This study shows wastewater can detect Hepatitis A cases early, with improved accuracy using data aggregation methods.

## Contribution

The study quantifies wastewater surveillance metrics for Hepatitis A detection, including sensitivity and predictive values.

## Key findings

- Wastewater detection sensitivity for HAV shedding cases was 48.1% with raw data.
- Using a 5-sample trimmed centered average increased sensitivity to 84.6%.
- The highest PPV (52.2%) and NPV (74.2%) were achieved with aggregated wastewater data.

## Abstract

Hepatitis A virus (HAV) remains a significant public health concern in the United States. Because infected individuals shed virus through stool, HAV can be detected in wastewater. Shedding occurs prior to the onset of symptoms that lead to clinical diagnosis, highlighting the potential of wastewater as an early case detection tool. This analysis aims to quantify key diagnostic metrics of wastewater surveillance for detecting HAV cases, which have not been previously defined. Utilizing wastewater data from the Point Loma Wastewater Treatment Facility in San Diego County, which serves around 2.2 million people, we assessed the sensitivity, positive predictive value (PPV), and negative predictive value (NPV) of wastewater HAV signals (positive/negative) in identifying shedding cases over a 308-day period. The number of people shedding virus on a given day was estimated through confirmed cases and presumed shedding intervals (2 weeks before and 1 week after symptom onset) and compared to wastewater signals. The sensitivity in detecting at least one shedding case on a given day using observed wastewater signals was 48.1%. Reclassifying the wastewater signal using simple data aggregations yielded sensitivities from 67.3% to 84.6%. Sensitivity increased as more individuals were shedding virus. The highest PPV (52.2%) and NPV (74.2%) were observed when a 5-sample trimmed centered average was used to reclassify the wastewater signal, indicating the utility of this preprocessing method. Conditional on clinical case detection and shedding assumptions, our study demonstrates that wastewater is a promising tool, providing signals that can inform public health surveillance.