Epicentre Epidemiologie Epidemiology

In July 2008, the government of Guinea Bissau officially declared a cholera epidemic and put in place a response plan. As there is often an inter-epidemic period after a large outbreak such as this, it is important to take the opportunity to learn lessons from previous outbreaks to better assess risks in the future. Analysis of available historical data on cholera epidemics in Guinea Bissau can help to develop a statistical model for risk assessment. Use of these models can help to ensure that there is a timely response of cholera outbreaks, thereby avoiding geographic spread, by identifying “signals” or indicators of when an outbreak is likely to occur or to predict the magnitude of an outbreak in an early stage.
The general objective of this study was to implement a time series and a spatial analysis in order to help to the timely identification of cholera outbreaks adapted to each region of Guinea Bissau. The specific objectives of the study were: (i) to describe the historical cholera data and epidemics in Guinea-Bissau identifying long term trends and the seasonal pattern; (ii) to obtain a predictive model of the seasonal risk of finding cholera cases in each region of the country; (iii) to calculate a threshold to detect epidemics with potential to affect more than 0.5% of the population
We described the historical cholera data in Guinea-Bissau in terms of time and place. The data were analyzed on a daily basis and then aggregated in epidemiological weeks as defined by the WHO. Attack rates and case fatality ratios were computed for each epidemic. A sequential description of the outbreaks was performed.
Multviare analysis using a Poisson regression model was used both to assess the secular trend and the seasonal pattern. Three different criteria were analyzed to define situation with high potential to produce large outbreaks: : (i) the number of regions affected; (ii) the value of the decimal logarithm of the number of cases; and (iii) the observation of inflexion points on the logarithm curve. All three criteria have been associated with high burden epidemics.
There is not a long-term defined trend in the occurrence of cholera epidemics in the country over the period investigated here. On the contrary, there is a seasonal pattern of cholera in the country:  the risk of occurrence of cholera cases increases from April and is maximum in mid-September. There is a spatial pattern of cholera incidence in Guinea-Bissau: the most affected areas are the capital (Bissau), Biombo region and the Bijagos Islands. Other areas highly affected are São Domingos, Nhacra in Oio, Tite in Quinara and Bedanda and Catio in Tombali. The analysis provides a sufficient number of elements to create a decision framework to define situations that lead to high burden epidemics.