Near-term forecasting of companion animal tick paralysis incidence: An iterative ensemble model

Abstract

Tick paralysis resulting from bites from Ixodes holocyclus and I. cornuatus is one of the leading causes of emergency veterinary admissions for companion animals in Australia, often resulting in death if left untreated. Availability of timely information on periods of increased risk can help modulate behaviors that reduce exposures to ticks and improve awareness of owners for the need of lifesaving preventative ectoparasite treatment. Improved awareness of clinicians and pet owners about temporal changes in tick paralysis risk can be assisted by ecological forecasting frameworks that integrate environmental information into statistical time series models. Using an 11-year time series of tick paralysis cases from veterinary clinics in one of Australia’s hotspots for the paralysis tick Ixodes holocyclus, we asked whether an ensemble model could accurately forecast clinical caseloads over near-term horizons. We fit a series of statistical time series (ARIMA, GARCH) and generative models (Prophet, Generalised Additive Model) using environmental variables as predictors, and then combined forecasts into a weighted ensemble to minimise prediction interval error. Our results indicate that variables related to temperature anomalies, levels of vegetation moisture and the Southern Oscillation Index can be useful for predicting tick paralysis admissions. Our model forecasted tick paralysis cases with exceptional accuracy while preserving epidemiological interpretability, outperforming a field-leading benchmark Exponential Smoothing model by reducing both point and prediction interval errors. Using online particle filtering to assimilate new observations and adjust forecast distributions when new data became available, our model adapted to changing temporal conditions and provided further reduced forecast errors. We expect our model pipeline to act as a platform for developing early warning systems that can notify clinicians and pet owners about heightened risks of environmentally driven veterinary conditions.

Mortality, incidence and seasonality of canine and feline patients treated with tick antiserum in three far North Queensland veterinary clinics from 2000 to 2020

“…In our study, the months with the equal highest mean of tick antiserum treatments were August and September. By comparison, October was the busiest tick paralysis month in three other studies 10,16,27 with a peak of October and November being described in another. 9 Our study found that Spring was still the busiest season, but the Spring percentage (45%) was lower compared to previous studies (56%-74%).…”

“…In these years, the broad inclusion criteria for survival compared with mortality may have favoured a disproportionately higher survival percentage of known outcomes. Finally, certain analyses fell outside of the scope of this study, and potential future uses of the database could include integration into correlative ecological niche models such as those used in other studies, 17,28 or ensemble models of environmental predictors 27 . This may improve our understanding of the effects of more detailed environmental variables on tick paralysis incidence, and increase the accuracy of tick paralysis incidence predictive models.…”

Dynamic generalised additive models (DGAMs) for forecasting discrete ecological time series

“…In recent years, there has been increased interest in using time‐series models for uncertainty interval estimation as opposed to point predictions, a trend that lends well to Bayesian inference (Gelman et al, 2017; Makridakis et al, 2020). This is particularly relevant for ecological forecasts, where point estimates are less important for making informed decisions than are conditional probability statements (Clark et al, 2022; Dietze, 2017; Dietze et al, 2018; White et al, 2019).…”

A Multi‐Model Ensemble of Baseline and Process‐Based Models Improves the Predictive Skill of Near‐Term Lake Forecasts

“…This type of MME may be able to overcome the limitations of individual process and baseline models that are unable to consistently forecast all environmental conditions with high accuracy across space (i.e., multiple depths in a lake), time (i.e., different seasons within a year), and forecast horizons. Implementation in other disciplines has overwhelmingly found that MMEs often produce more skillful forecasts, on average, than individual model forecasts (Atiya, 2020; Clark et al., 2022; Humphries et al., 2018; Velázquez et al., 2011). Using MMEs also leads to greater diversity in forecast predictions, potentially increasing decision‐making success (Boettiger, 2022).…”

“…Despite the urgent need for freshwater forecasts, however, the optimal modeling approach for developing forecasts remains unresolved across different spatial and temporal scales. One promising forecasting approach that has emerged from other disciplines is multi‐model ensembles (MMEs), in which more than one model is used to simultaneously forecast the same variable into the future (Chandler, 2013; Clark et al., 2022; Humphries et al., 2018; Kirtman et al., 2014; Long et al., 2021; Velázquez et al., 2011). To date, MMEs have not been applied to near‐term freshwater forecasting (reviewed by Lofton et al., 2023), motivating the need to understand how an MME forecast performs relative to individual models, as well as how the structure of the different models in the MME influences forecast performance.…”