Mapping high rate clusters of animal contact related human Salmonella enterica single state outbreaks in the United States, 2009 to 2022. A spatial epidemiological approach to inform public health surveillance

This study utilizes a spatial epidemiological approach to analyze 104 animal contact-related *Salmonella* outbreaks in the U.S. from 2009 to 2022, identifying significant high-risk clusters in the Mountain West, Midwest, and Northeast to guide targeted public health interventions.

The Gist

Imagine the United States as a giant, bustling neighborhood where people and animals live side-by-side. Sometimes, a sneaky germ called Salmonella (specifically the non-typhoidal kind) jumps from animals to humans, causing sickness. This paper is like a detective story where researchers used a high-tech "magic map" to figure out exactly where these germs are hiding and why.

Here is the story of their investigation, broken down into simple parts:

1. The Mystery: Why are we getting sick?

Salmonella is a common germ that usually makes people sick after eating bad food. But this study focused on a different way people get sick: touching animals. Whether it's petting a cow at a fair, holding a baby chick, or playing with a pet turtle, these interactions can pass the germ to humans.

The researchers wanted to know: Are these "animal-touch" sicknesses happening randomly everywhere, or are they clustering in specific neighborhoods?

2. The Tools: The "Germ Radar"

To solve this, the team didn't just count cases; they used a spatial epidemiological approach. Think of this as a weather radar for germs.

• The Map: They looked at data from 2009 to 2022 from the CDC (the country's health police).
• The Smoothing: Some states are huge with few people, while others are small and crowded. To make a fair comparison, they used a "smoothing" technique. Imagine pouring water over a bumpy landscape; the water fills the low spots and levels out the high peaks, giving a clearer picture of where the "flood" (outbreaks) is actually happening.
• The Clustering: They used special math to find "hot spots" (areas with way more sickness than expected) and "cold spots" (areas that are surprisingly safe).

3. The Findings: Where are the Hot Zones?

The "Germ Radar" lit up three main areas in the US that are consistently risky:

• The Mountain West (Wyoming, Montana, Colorado, Idaho, Utah): This is the biggest hotspot. It's like a "Germ Super-Highway."
  • Why? These areas have lots of farms and ranches. People here often touch cattle, pigs, and sheep. The study found that cattle are a major source of the trouble here.
• The Midwest (Ohio, Wisconsin, Iowa, Minnesota, Nebraska): Another busy zone.
  • Why? This is the heart of American livestock. People here get sick from touching poultry (baby chicks and ducks) and farm animals at fairs and veterinary clinics.
• The Northeast (Vermont, New Hampshire): A smaller but persistent hotspot.
  • Why? Here, the trouble comes from backyard chickens and reptiles (like pet turtles and lizards).

The "Cold Spots": Interestingly, the Southern US generally had fewer of these specific outbreaks. It's not that the South is germ-free, but the type of animal contact that causes these specific outbreaks is less common there.

4. The Culprits: Who is the "Bad Guy"?

The study identified three main groups of animals acting as the "villains" in this story:

1. Mammals: Cows, pigs, horses, and even small pets like mice and ferrets.
2. Birds: Baby chicks, ducklings, turkeys, and backyard chickens. (These are very popular with kids and often cause outbreaks at schools or fairs).
3. Reptiles: Turtles, lizards, and snakes. (Even though they look cool, they carry Salmonella very easily).

5. The Villain's Hideouts: Where does the contact happen?

It's not just on big farms. The "germ transfer" happens in many places:

• Farms and Feed Stores: Where people buy animal food.
• Fairs and Petting Zoos: Where kids get to hug the animals.
• Veterinary Clinics: Where sick animals are treated.
• Backyards and Homes: Where people keep pet reptiles or chickens.

6. The Lesson: What should we do?

The researchers are saying, "We can't stop the germs entirely, but we can stop the spread if we know where to look."

• Targeted Education: Instead of telling everyone in the whole country the same thing, we should send specific warnings to the Mountain West about cows, to the Midwest about baby chicks, and to the Northeast about pet turtles.
• Better Hygiene at Fairs: Petting zoos and fairs need to be extra clean and teach kids to wash their hands immediately after touching an animal.
• Watch the Reptiles: If you have a pet turtle, remember it's a germ factory. Wash your hands after every touch!

The Bottom Line

This paper is like a GPS for public health. It tells us that Salmonella outbreaks from animals aren't random; they have a pattern. By knowing that the Mountain West, Midwest, and Northeast are the "danger zones," health officials can put their resources there to keep people safe, just like putting extra fire hydrants in the neighborhood with the most fire risk.

Technical Summary

1. Problem Statement

Nontyphoidal Salmonella enterica (NTS) is a leading cause of gastroenteric infections globally. While foodborne transmission is the primary route, animal contact-related outbreaks are a significant and growing public health concern in the United States. Despite existing literature on NTS, there is a lack of comprehensive, national-level analyses examining the longitudinal spatial and spatio-temporal clustering of animal-contact-associated outbreaks. Previous studies often focused on specific serovars, short timeframes, or limited geographic regions. Understanding the geographic distribution and clustering of these outbreaks is critical for designing targeted local public health prevention and intervention programs.

2. Methodology

The study employed a robust geospatial analytical framework using data from the National Outbreak Reporting System (NORS) covering the period 2009–2022.

• Data Scope: The analysis focused on single-state outbreaks (where exposure occurred in one state) within the 48 contiguous U.S. states and the District of Columbia. Multi-state outbreaks were excluded to isolate local exposure dynamics.
• Data Processing:
  • State-level population data were sourced from the U.S. Census Bureau.
  • Administrative boundaries were projected using the USA Contiguous Albers Equal Area Conic projection.
  • Incidence Rates (IRs) were calculated per 1 million person-years (MPY).
• Analytical Techniques:
  1. Disease Mapping:
     • Choropleth Maps: Visualized raw IRs for two periods (2009–2015 and 2016–2022).
     • Spatial Empirical Bayes (SEB) Smoothing: Applied to stabilize IRs in states with small populations using 1st-order Queen contiguity weights.
     • Empirical Bayes Kriging (EBK): Used for spatial interpolation to create continuous isopleth maps of outbreak risk.
  2. Global Clustering:
     • Incremental Spatial Autocorrelation (ISA): Used Moran's I to identify the distance band where spatial clustering was most significant.
  3. Local Clustering:
     • Hot Spot Analysis (Getis-Ord Gi):* Identified statistically significant clusters of high (hot spots) and low (cold spots) IRs.
     • Cluster and Outlier Analysis (Local Moran's I): Detected High-High (HH), Low-Low (LL), High-Low (HL), and Low-High (LH) spatial patterns.
  4. Scan Statistics:
     • SaTScan (Discrete Poisson Model): Conducted purely spatial and space-time retrospective analyses to detect clusters with significantly higher-than-expected outbreak rates.
     • Parameters included a maximum spatial cluster size of 10% of the population and a temporal window of 50% of the study period.

3. Key Contributions

• First National Spatio-Temporal Analysis: This is one of the first studies to map the longitudinal spatial and spatio-temporal clustering of animal-contact NTS outbreaks across the entire contiguous U.S. over a 14-year period.
• Multi-Method Framework: The study integrates disease mapping, smoothing, interpolation, and multiple cluster detection algorithms (Global/Local Moran's I, Getis-Ord Gi*, and Scan Statistics) to provide a comprehensive risk assessment.
• Granular Exposure Characterization: The research links specific geographic clusters to distinct animal reservoirs (mammals, birds, reptiles) and exposure settings (farms, veterinary clinics, fairgrounds, residential), moving beyond broad national averages.

4. Key Results

• Overall Incidence: A total of 104 single-state outbreaks were reported. The mean annual incidence rate was 0.02 outbreaks per million person-years.
• Primary Reservoirs: The most common animal contacts were mammals (cattle, pigs, sheep, horses), birds (backyard chickens, ducklings, turkeys), and reptiles (turtles, lizards).
• Exposure Settings: Outbreaks occurred in diverse settings including farms, fairgrounds, agricultural feed stores, veterinary clinics, and residential homes.
• Spatial Clustering Findings:
  • Mountain West: Consistently identified as a high-risk region (Wyoming, Montana, Idaho, Colorado, Utah).
    • Drivers: High livestock density (cattle), contact with poultry and reptiles, and exposure at farms and fairgrounds.
  • Midwest & Northeast: Identified as secondary high-risk clusters (e.g., Ohio, West Virginia, Vermont, New Hampshire, Wisconsin).
    • Drivers: Contact with livestock and backyard poultry; settings included veterinary clinics, educational institutions, and child daycares.
  • Low-Risk Regions: The Southern and Southwestern regions consistently showed the lowest outbreak rates and clustering.
• Temporal Trends:
  • 2009–2015: No significant global spatial autocorrelation was found.
  • 2016–2022: Significant global clustering emerged at a distance of 671.38 km, with a peak Moran's I of 0.252. High-high clusters became more pronounced in the Mountain West.
• Scan Statistics:
  • Purely Spatial: Identified a major high-rate cluster (RR = 5.51) covering Wyoming, Montana, Idaho, Utah, Colorado, South Dakota, and Nebraska.
  • Space-Time: Detected four significant high-rate clusters. The most significant (RR = 16.50) occurred in Vermont and New Hampshire (2013–2017), primarily linked to reptiles and backyard poultry. Another major cluster (RR = 8.26) was found in the Mountain West (2012–2018).

5. Significance and Implications

• Targeted Surveillance: The identification of persistent high-risk clusters in the Mountain West, Midwest, and Northeast allows public health authorities to allocate resources and surveillance efforts more efficiently.
• Intervention Strategies: The findings suggest that prevention programs must be region-specific:
  • Mountain West: Focus on cattle farm biosecurity, fairground sanitation, and public education regarding livestock contact.
  • Midwest/Northeast: Emphasis on backyard poultry safety, reptile handling education, and infection control in veterinary clinics and educational settings.
• Policy Recommendations: The study advocates for:
  • Enhanced public education on zoonotic risks in non-farm settings (e.g., petting zoos, residential backyards).
  • Improved sanitation protocols at agricultural fairs and feed stores.
  • Regular surveillance of animal populations in high-risk counties.
• Limitations: The study notes potential underreporting due to voluntary state reporting and the confounding effect of the COVID-19 pandemic (2019–2021) on reporting rates. It also excludes multi-state outbreaks, which represent a separate but significant burden.

In conclusion, this study demonstrates that animal-contact NTS outbreaks in the U.S. are not randomly distributed but exhibit distinct, persistent spatiotemporal clustering. These patterns are driven by regional variations in livestock density, animal types, and human-animal interaction settings, necessitating geographically tailored public health interventions.