[Zeng-Ding phenomenon: further demonstration and studies on its predictive value in epidemic of measles and scarlet fever].
Mots clés
Abstrait
OBJECTIVE
To demonstrate further the existence of Zeng-Ding phenomenon in disease epidemic and to explore the relationship between it and the time series in different kinds of diseases.
METHODS
Incidence data of notifiable communicable diseases during 1975 to 1996 were collected. Time series of measles and scarlet fever incidence during 1975 to 1995, 1980 to 1995, 1985 to 1995 and 1990 to 1995 were established. Correlation analysis was conducted between monthly cumulative percentage and predictive ratio of increase to decrease in incidence rate at the best cut-off point. Prediction was studied based on the constructed extrapolation model.
RESULTS
Correlation analysis showed that 98.3% (232/236) of the coefficients of correlation were negative (R < 0), indicating further the existence of Zeng-Ding phenomenon in disease epidemic. There was significant difference in coefficients of correlation between the four time series, which accounted for 63.8%, 54.2%, 44.1% and 35.0%, respectively; and for 73.3%, 56.7%, 36.7% and 36.7%, respectively, in measles, and for 53.6%, 51.8%, 51.7% and 33.3%, respectively, in scarlet fever. It showed that Zeng-Ding phenomenon correlated with the time series and the kinds of diseases. Prediction from extrapolation model showed that there was significant difference in predictive agreement between two time series of 1975 to 1995 (65.5%) and 1985 to 1995 (37.0%) with chi 2 of 4.54 and P < 0.05, indicating a trend that predictive agreement increased with prolonging of time series, and a trend that predictive agreement for scarlet fever increased with decreasing of coefficients of correlation.
CONCLUSIONS
Predictive value of the incidence data can be evaluated by their source and length of time series.