[Construction and application of joinpoint regression model for series cumulative data]

Area of Science:

• Epidemiology
• Biostatistics
• Public Health

Background:

• Joinpoint Regression (JPR) is a statistical method for analyzing trends in rates.
• Analyzing cumulative disease incidence data presents unique challenges for trend estimation.
• Accurate trend analysis is crucial for public health interventions and disease surveillance.

Purpose of the Study:

• To develop and validate a JPR model specifically for series cumulative incidence data.
• To compare the performance of the cumulative JPR model against traditional weekly data analysis.
• To assess the model's utility in analyzing dengue fever incidence trends.

Main Methods:

• Constructed a JPR model based on the principle of Poisson distribution additivity for cumulative data.
• Analyzed weekly notifiable dengue fever incidence and cumulative data from Guangdong province (2008-2017).
• Evaluated model performance using Mean Squared Errors (MSE) and Mean Absolute Percentage Error (MAPE).

Main Results:

• The logarithmic linear JPR model using cumulative incidence data showed lower MSE and MAPE compared to weekly data analysis (except in 2015).
• The JPR model demonstrated significantly improved fitting accuracy for trend analysis of series cumulative data.
• The model proved effective in capturing trend changes and predicting staged cumulative incidence.

Conclusions:

• The developed JPR model enhances the accuracy of trend analysis for cumulative disease incidence data.
• This approach is valuable for analyzing historical trends and forecasting future staged incidence.
• The model provides a robust tool for epidemiological surveillance and public health planning.