Citation

Jomah ND, Ojo JF, Odigie EA, Olugasa BO. Afr. J. Med. Med. Sci. 2014; 43 Suppl: 87-95.

Copyright

(Copyright © 2014, Spectrum Books)

DOI

unavailable

PMID

26949785

Abstract

OBJECTIVE: The post-civil war records of dog bite injuries (DBI) and rabies-like-illness (RLI) among humans in Liberia is a vital epidemiological resource for developing a predictive model to guide the allocation of resources towards human rabies control. Whereas DBI and RLI are high, they are largely under-reported. The objective of this study was to develop a time model of the case-pattern and apply it to derive predictors of time-trend point distribution of DBI-RLI cases.

METHODS: A retrospective 6 years data of DBI distribution among humans countrywide were converted to quarterly series using a transformation technique of Minimizing Squared First Difference statistic. The generated dataset was used to train a time-trend model of the DBI-RLI syndrome in Liberia. An additive detenninistic time-trend model was selected due to its performance compared to multiplication model of trend and seasonal movement. Parameter predictors were run on least square method to predict DBI cases for a prospective 4 years period, covering 2014-2017.

RESULTS: The two-stage predictive model of DBI case-pattern between 2014 and 2017 was characterised by a uniform upward trend within Liberia's coastal and hinterland Counties over the forecast period.

CONCLUSION: This paper describes a translational application of the time-trend distribution pattern of DBI epidemics, 2008-2013 reported in Liberia, on which a predictive model was developed. A computationally feasible two-stage time-trend permutation approach is proposed to estimate the time-trend parameters and conduct predictive inference on DBI-RLI in Liberia.

Keywords: Animal Bites; Dog Bites

Language: en