Clustering of equine grass sickness cases in the United Kingdom: a study considering the effect of position-dependent reporting on the space-time K-function.

This study focuses on identifying patterns in the clustering of Equine Grass Sickness (EGS) cases in the United Kingdom. The researchers also investigated whether uneven reporting of this disease across locations and time affects the validity of a particular statistical testing method known as the space-time K-function.

Equine Grass Sickness (EGS)

• EGS is a debilitating and often fatal disease that affects horses, and is largely associated with pasture grazing. Its cause is unknown, but there are indications that a bacterium known as Clostridium botulinum type C could be implicated.
• The condition is more common in certain parts of the UK, with the highest incidence believed to occur in Scotland. Seasonal variation in the disease is apparent, with most cases appearing between the months of April and July.

Methodology and Data

• The research team used data from EGS cases reported in England and Wales in 1999 and 2000, which were confirmed through histological (microscopic tissue) examination. These reports were gathered from various veterinary diagnostic centres across the UK.
• The data collected did not encompass all EGS cases, and the ratio of reported cases to actual number of cases would have fluctuated across different locations and times. This variance in reporting forms a key part of the study.

Space-Time K-Function and Reporting Variance

• The researchers examined the possible effect of uneven EGS case reporting on the space-time K-function, a statistical test that is used to detect and analyse clusters in point data across both space and time.
• The goal was to identify whether spatial-temporal properties of a ‘thinned’ point process (which represents reduced or ‘thinned out’ data) could be appropriately explored using this function, given the discrepancies in case reporting.

Findings

• The study concluded that uneven EGS reporting, dependent on location and time, did not invalidate the usage of the Monte Carlo test with the space-time interaction feature of the K-function.
• Substantial evidence was found suggesting the occurrence of EGS clustering in both space and time (P < 0.001), which suggests that localized factors such as contagious transmission, local climate or pasture management practices could significantly influence the disease occurrence.