A carregar...
Publicação
Early detection of influenza activity using telephone helpline data
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 365.35 KB | Adobe PDF |
Orientador(es)
Resumo(s)
The annual influenza season is a recurring time period characterized by the prevalence of outbreaks of influenza. Early detection of these outbreaks is paramount to allocating resources and effective counter measures.
Methods:
Using data from the Portuguese health helpline we selected 17 triage algorithms that relate with the influenza case definition used by ECDC.
We listed the weekly number of calls for each algorithm and the total number of calls to the counselling line, from week 40 to 20 (seasonal influenza) from 2010-2013.
Data was then compared with data from the sentinel system (standard surveillance in Portugal) with up to 5 weeks lag, using a cross correlation function. Correlations above 0,5 were considered of interest.
In parallel, the MEM - Moving Epidemic Method was applied to determine a baseline threshold for the beginning of the epidemic period.
Results:
We analysed 1.032.724 calls (an average of 10.432 calls each week), and the “cough algorithm” presented the highest correlation (0,83) to data from the sentinel network, with a 3 weeks lag.
Using MEM, we were able to detect the beginning of the epidemic period 3 to 5 weeks in advance, comparing to data from the sentinel network.
Conclusion:
The baseline provided by the selected algorithms allowed, on average, three weeks of advanced warning for seasonal influenza activity.
Using two techniques (cross correlation and MEM) we were able to detect early signs of seasonal influenza activity, specifically with influenza related symptoms, namely coughing.
Descrição
Palavras-chave
Vigilância Epidemiológica Gripe Estados de Saúde e de Doença
Contexto Educativo
Citação
Editora
Instituto Nacional de Saúde Doutor Ricardo Jorge, IP