SARIMA for predicting the cases numbers of dengue

SARIMA para predição do número de casos de dengue

LETTER TO EDITOR CARTA AO EDITOR

SARIMA for predicting the cases numbers of dengue

SARIMA para predição do número de casos de dengue

Viroj Wiwanitkit

Wiwanitkit House, Bangkhae, Bangkok, Thailand

^{Address to}

Dear Editor:

The recent report by Martinez on predicting the number of cases of dengue based on SARIMA is very informative¹. I have some concerns on this work. First, this work is very similar to another publication by Martinez et al. on using same technique approach for studying². Only a different in setting can be observed. The two works might be a salami publication. Second, the prediction is based on the retrospective data which might not be useful for future prediction in actual life. Due to the rapid change in environmental factors at present, especially for the climate change and global warming, the model might not be effective. The adjustment based on the temperature prediction might be additional helpful. Climatological parameters are required to be implemented in using SARIMA for prediction of the epidemic³.

Received in 12/10/2011

Accepted in 10/01/2012

Response to letter to the editor: simple statistical models can provide good predictions of dengue incidence

Resposta à carta ao editor: modelos estatísticos simples podem trazer boas predições da incidência da dengue

Edson Zangiacomi Martinez; Amaury Lelis Dal Fabbro; Elisângela Aparecida Soares da Silva

Departamento de Medicina Social, Faculdade de Medicina de Ribeirão Preto, Universidade de São Paulo, Ribeirão Preto, SP

^{Address to}

Dear Editor,

We thank Professor Wiwanitkit for his interest in our research on forecast models for dengue incidence^1,2. We are glad for the opportunity to clarify some important points of our research.

First, Professor Wiwanitkit has argued that two articles produced by our research group might be a salami publication. Salami-slicing denotes a type of research misconduct that consists of dividing the results of a research project into a series of articles to maximize the number of publications^3,4, and we strongly disagree that our articles^1,2 are an example of this bad practice. Each of these articles tells its own story, although they present a discussion of the use of the same data analysis strategy. Further, each article deals with different data sets obtained from two different municipalities, evidencing that these localities have different temporal patterns of dengue incidence, and summarizing all these results into a single article would result in a great loss of information and details.

Second, he has stated that the prediction is based on the retrospective data, which might not be useful for future prediction in actual life due to the current rapid change in environmental factors. However, we believe that the high volatility observed in some periods of the time series are primarily due to the introduction and reintroduction of different virus serotypes in a susceptible population, and the results of our articles suggest that the model fits the data adequately, despite the occurrence of this phenomenon within the studied period^1,2. In addition, the out-of-sample predictions generated by the SARIMA models are close to the observed values, suggesting that the model is useful and accurate for forecasting purposes.

REFERENCES

Received in 21/10/2011

Accepted in 10/01/2012

• 1. Martinez EZ, Silva EA, Fabbro AL. A SARIMA forecasting model to predict the number of cases of dengue in Campinas, State of São Paulo, Brazil. Rev Soc Bras Med Trop 2011; 44:436-440.
• 2. Martinez EZ, Silva EA. Predicting the number of cases of dengue infection in Ribeirão Preto, São Paulo State, Brazil, using a SARIMA model. Cad Saude Publica 2011; 27:1809-1818.
• 3. Soebiyanto RP, Adimi F, Kiang RK. Modeling and predicting seasonal influenza transmission in warm regions using climatological parameters. PLoS One 2010; 5:e9450.
• 1. Martinez EZ, Silva EA, Fabbro AL. A SARIMA forecasting model to predict the number of cases of dengue in Campinas, State of São Paulo, Brazil. Rev Soc Bras Med Trop 2011; 44:436-440.
• 2. Martinez EZ, Silva EA. Predicting the number of cases of dengue infection in Ribeirão Preto, São Paulo State, Brazil, using a SARIMA model. Cad Saude Publica 2011; 27:1809-1818.
• 3. Gilbert FJ, Denison AR. Research misconduct. Clin Radiol 2003; 58:499-504.
• 4. Rogers LF. Salami slicing, shotgunning, and the ethics of authorship. AJR Am J Roentgenol 1999; 173:265.

Publication Dates

• Publication in this collection
  27 Feb 2012
• Date of issue
  Feb 2012