A simpler model to help understand the source of Campylobacter infection has been developed by researchers.

The team compared the asymmetric island model which uses genetic data and evolutionary processes to a newly developed genetic-free model with results published in the Journal of the Royal Society.

Modeling the epidemiological data and genetic information of microbial isolates from cases provides a methodology for tracing the source of infection. Currently, scientists use a range of models to help distinguish between different animal sources of human infection.

Campylobacter is the most frequently notified enteric disease in New Zealand, according to the Ministry of Health.

Jing Liao, from Massey University, said in recent years the figures have reduced, thanks in part to guidance provided by statistical models but New Zealand’s rate is still high compared to international standards.

“Understanding the source of infection, including drinking contaminated water, eating undercooked animal food products, or handling food products contaminated by animal feces, is essential for the implementation of control measures,” she said.

“By modeling the potential sources and pathways of infections, you gain the ability to identify wher the highest risks are and wher interventions will do the most good. It provides timely and accurate data on high-risk areas for public health officials to guide efforts. Simpler non-genetic models can help to test model assumptions that underpin the more complex genetic models, but they may not perform well under certain conditions.”

In 2005 to 2007, modeling helped to identify high-rates of campylobacteriosis linked to poultry. This allowed the New Zealand Food Safety Authority (now Ministry for Primary Industries) and the poultry industry to adopt measures such as improving slaughter and processing, to reduce contamination.

Researchers looked at whether complex genetic models yield superior attribution results compared to simpler, non-genetic models, and if they can be improved to include risk factor information on individual human cases.

They used surveillance data on campylobacteriosis from New Zealand’s Manawatū region from 2005 to 2014 and compared the asymmetric island model with an adapted Dirichlet model, a genetic-free model, to determine whether it could guide management with the same success.

The model differed from the asymmetric island one as it did not model pathogenic evolution but inferred the sampling distribution of genotypes directly from the observed count data.

The data contained 1,460 isolates taken from human cases, and 2,128 isolates sampled from chicken carcasses, cattle, sheep, environmental water and wild birds.

Scientists found the simpler model was as robust as the more complex one for identifying the source of common human strains of Campylobacter but did not perform as well for rare strains. However, if most infections are caused by highly observed strains, the simpler model may work and is quicker to implement.

Both models demonstrated a clear effect of rurality on attribution: cases in rural areas were more likely to have originated from ruminants, while those in main urban centers were more likely to be of poultry origin.

“For the Dirichlet and asymmetric Island models, the linear and categorical models of rurality show broadly the same trend, suggesting that the additional flexibility given by the categorical model is not required and that the shift in attribution as rurality changes is adequately modeled by a linear trend on the logit scale,” said researchers.

The asymmetric Island model attributed 7 percent of human cases across all rurality levels to sources other than poultry, ruminants, and water and gives a small attribution to water in highly rural areas, while the Dirichlet model indicates these sources are unimportant.

Future work may include adapting models with additional risk factors such as age, occupation, and contact with animals.