Fernández-Niño Julián Alfredo, Cárdenas-Cárdenas Luz Mery, Hernández-Ávila Juan Eugenio, Palacio-Mejía Lina Sofía, Castañeda-Orjuela Carlos Andrés
Centro de Información para Decisiones en Salud Pública (CENIDSP), Instituto Nacional de Salud Pública, Cuernavaca, Morelos, México.
Biomedica. 2015 Dec 4;36(0):44-55. doi: 10.7705/biomedica.v36i0.2869.
Dengue has a seasonal behavior associated with climatic changes, vector cycles, circulating serotypes, and population dynamics. The wavelet analysis makes it possible to separate a very long time series into calendar time and periods. This is the first time this technique is used in an exploratory manner to model the behavior of dengue in Colombia.
To explore the annual seasonal dengue patterns in Colombia and in its five most endemic municipalities for the period 2007 to 2012, and for roughly annual cycles between 1978 and 2013 at the national level.
We made an exploratory wavelet analysis using data from all incident cases of dengue per epidemiological week for the period 2007 to 2012, and per year for 1978 to 2013. We used a first-order autoregressive model as the null hypothesis.
The effect of the 2010 epidemic was evident in both the national time series and the series for the five municipalities. Differences in interannual seasonal patterns were observed among municipalities. In addition, we identified roughly annual cycles of 2 to 5 years since 2004 at a national level.
Wavelet analysis is useful to study a long time series containing changing seasonal patterns, as is the case of dengue in Colombia, and to identify differences among regions. These patterns need to be explored at smaller aggregate levels, and their relationships with different predictive variables need to be investigated.
登革热具有与气候变化、病媒周期、流行血清型和种群动态相关的季节性特征。小波分析能够将很长的时间序列分解为日历时间和周期。这是首次以探索性方式使用该技术对哥伦比亚登革热的行为进行建模。
探索2007年至2012年期间哥伦比亚及其五个登革热流行最严重的城市的年度季节性登革热模式，以及1978年至2013年期间全国大致的年度周期。
我们使用了2007年至2012年期间每个流行病学周以及1978年至2013年期间每年的所有登革热发病病例数据进行探索性小波分析。我们使用一阶自回归模型作为原假设。
2010年疫情的影响在全国时间序列和五个城市的序列中都很明显。各城市之间观察到年际季节性模式存在差异。此外，我们在国家层面确定了自2004年以来大致2至5年的年度周期。
小波分析有助于研究包含变化季节性模式的长时间序列，如哥伦比亚登革热的情况，并识别不同地区之间的差异。这些模式需要在更小的汇总层面进行探索，并且需要研究它们与不同预测变量之间的关系。
