Cyclic patterns of malaria incidence in Burundi

Abstract

Best available descriptions of malaria incidence and mortality dynamics are important to improve and evaluate the implementation of programs to monitor (e.g., remote sensing) and control disease, especially in endemic zones. High-frequency (e.g., semi-annual and seasonal) cycles in malaria incidence have been observed in various countries and they coincide with cycles in the natural environment (e.g., temperature, heliogeophysical activity, etc.). However, neither trend nor cyclical oscillations beyond a 6-month (0.5-year) period for this vector-borne disease were reported in a recent analysis on monthly notifications in Burundi for the years 1997-2003. Since the examination of graphical plots indicated an eventual existence of trans-year (multiannual) variations, we further analyzed the same data in more detail. Here we explore whether low-frequency cycles (beyond seasonality) might exist (e.g., trans-year cycles with periods of 13-24 months or longer). Monthly incidence rate per 100 inhabitants from the Province of Karuzi, Burundi, over the years 1997-2003 was analysed. The exploration of underlying chronomes (time structures) was done by linear and non-linear parametric regression models, autocorrelation, spectral analysis (e.g., fast Fourier transforms) , periodogram regression analysis (PRA) and wavelet transform (WT). By using a periodogram regression analysis, we describe a multicomponent cyclic chronome with periods T>12 months (19 and 86 months, all at P<0.05). Notably, the strongest cyclic pattern in the periodogram of the detrended malaria rates (whereas the peak was suppressed and beyond the semi-annual cycle of 6 months) was ≈1.5-1.6 years (T=19.0 months, R=0.32). A dynamic pattern of "shortening" of the length (period) of the cycles was observed during the pre-epidemic interval (from 8-9 to 5-6 months and from 20-22 to 16-18 months in years 1997-2000) that can be used to anticipate a forthcoming incidence increase and epidemic levels of malaria at a regional level. Indeed, these cycles in malaria incidence correspond to cyclic components of heliogeophysical activity (HGA) such as the sunspot cycle impulses of 0.5-2.0 years as well as the quasi-biennial solar magnetic cycle of 1.5-2.5 years and further, detailed analyses are warranted to investigate such relationships. A multicomponent dynamic cyclical pattern of malaria incidence variations in Burundi (1997Burundi ( -2003 exists thus allowing further, more specific analyses and modelling as well as correlations with similar environmental cycles to be explored. These results might also contribute to better estimates for forecasting, prevention and understanding of malaria dynamics and aetiology.

Assessing temporal associations between environmental factors and malaria morbidity at varying transmission settings in Uganda

“…In terms of vegetation and malaria cases, the present study identified significant lags of 3–23 weeks while in Burundi and Ethiopia, 8 and 12 weeks were significantly correlated to vegetation and malaria [ 27 , 31 ]. As demonstrated in previous studies, malaria transmission and mosquito abundance can be influenced by vegetation cover through provision of an outdoor resting place for the vector [ 14 ].…”

“…More recently however, significant associations were seen at lags starting from 7 and 13 weeks in Kenya [ 26 ]. In the highlands of Burundi, a lag of 4 weeks was significant between temperature and malaria [ 27 ], while lags of 4 and 8 weeks were significant for an area of unstable malaria transmission in Rwanda [ 23 ]. Temperature has been documented to affect survival and growth of both the vector and the parasite in the vector [ 28 – 30 ].…”

Cyclic patterns of cerebral malaria admissions in Papua New Guinea for the years 1987–1996

“…Different statistical methods for time-series analyses and modelling were used, according to a previously described algorithm used to detect cyclicity in similar monthly malaria time-series [18]. Initially, descriptive statistics with linear and nonlinear regression modelling over time were applied (Figs 1–3).…”

“…change over time of the revealed cyclicities ( Figs 6, 7). In particular, PRA was successfully applied earlier in studies on helioepidemiology of various types of cancer and infectious diseases as described previously in more detail elsewhere [13,[18][19][20][21]. A range (spectrum) of the linear coefficients of correlation R was constructed against each of the tested periods T (months) representing the so-called 'periodogram ' (e.g.…”

“…Beyond the particular approach in these previous studies on CM admissions in Papua New Guinea [11,12], other more specific and accurate quantitative approaches in analysing and modelling the number of malaria admissions are needed [13][14][15][16][17] forecasting approach -their interactions should be integrated into the development of all future predictive models. Such additional, appropriate methods, with minimal requirements for a specific algorithm and 95% confidence intervals for the derived estimates and statistical parameters [13][14][15][16], as in the case of malaria admission data exploited here, should be the following: (1) Description, analysis and decomposition of the temporal dynamics; linear and nonlinear regression modelling as well as regression modelling of periodic modes over time and wavelet approach [17,18], with, finally, a reconstruction and derivation of cyclical estimates for modelling of the time-series. (2) Possibly linear and nonlinear regression modelling to independent variables (e.g.…”