Forecasting Non-Stationary Diarrhea, Acute Respiratory Infection, and Malaria Time-Series in Niono, Mali
2007

Forecasting Diarrhea, ARI, and Malaria in Mali

Sample size: 333990 publication 10 minutes Evidence: high

Author Information

Author(s): Medina Daniel C., Findley Sally E., Guindo Boubacar, Doumbia Seydou

Primary Institution: Columbia University

Hypothesis

Can the multiplicative Holt-Winters method accurately forecast non-stationary time-series data for diarrhea, acute respiratory infection, and malaria in Niono, Mali?

Conclusion

The multiplicative Holt-Winters forecasting method is effective for predicting disease trends in Niono, Mali, and could enhance public health interventions.

Supporting Evidence

  • 92% of the forecasts generated had mean absolute percentage errors around 25%.
  • The method accommodates seasonal and inter-annual fluctuations.
  • Forecasts could improve infectious disease management in the district of Niono.
  • The MHW method adapts to changes in disease dynamics over time.
  • Forecasting accuracy was assessed using mean absolute percentage error (MAPE).
  • The study highlights the importance of age-specific forecasts for disease management.
  • Consultation rates for diarrhea, ARI, and malaria accounted for 59.2% of all consultations.
  • The MHW method does not require disease-specific tailoring for forecasting.

Takeaway

This study shows a way to predict when people might get sick from diarrhea, respiratory infections, or malaria, helping doctors prepare better.

Methodology

The study used a longitudinal retrospective analysis from 1996 to 2004, applying the multiplicative Holt-Winters method to forecast disease consultation rates.

Potential Biases

Potential bias due to missing data and the method of interpolation used.

Limitations

Missing monthly records were interpolated, which may introduce bias in consultation rates.

Participant Demographics

The study focused on the population of Niono, Mali, with a projected population of 278,741 in 2004.

Statistical Information

P-Value

p<0.001

Confidence Interval

95% prediction intervals were estimated using bootstrapping.

Statistical Significance

p<0.001

Digital Object Identifier (DOI)

10.1371/journal.pone.0001181

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication