Modeling gradually changing seasonal variation in count data using state space models: a cohort study of hospitalization rates of stroke in atrial fibrillation patients in Denmark from 1977 to 2011

Abstract

BackgroundSeasonal variation in the occurrence of cardiovascular diseases has been recognized for decades. In particular, incidence rates of hospitalization with atrial fibrillation (AF) and stroke have shown to exhibit a seasonal variation. Stroke in AF patients is common and often severe. Obtaining a description of a possible seasonal variation in the occurrence of stroke in AF patients is crucial in clarifying risk factors for developing stroke and initiating prophylaxis treatment.MethodsUsing a dynamic generalized linear model we were able to model gradually changing seasonal variation in hospitalization rates of stroke in AF patients from 1977 to 2011. The study population consisted of all Danes registered with a diagnosis of AF comprising 270,017 subjects. During follow-up, 39,632 subjects were hospitalized with stroke. Incidence rates of stroke in AF patients were analyzed assuming the seasonal variation being a sum of two sinusoids and a local linear trend.ResultsThe results showed that the peak-to-trough ratio decreased from 1.25 to 1.16 during the study period, and that the times of year for peak and trough changed slightly.ConclusionThe present study indicates that using dynamic generalized linear models provides a flexible modeling approach for studying changes in seasonal variation of stroke in AF patients and yields plausible results.

Seasonality of tuberculosis in intermediate endemicity setting dominated by reactivation diseases in Hong Kong

“…Second, we analysed the time evolutions of the annual periodic components with local wavelet power spectrum, to reveal the degree of seasonality regarding the predictability of seasonal variation (i.e., proportion of time-steps over the full time series with significant power at the annual scale) 27 . DGLMs were then fitted to the time series for which wavelet analysis could detect a significant annual seasonality dominating throughout the study period, to further quantify the degree of seasonality in terms of peak-to-trough ratio and identify the peak timing of the seasonality 28 , 29 ( Supplementary Text for full technical details). Sensitivity analysis was conducted by applying the wavelet transform and fitting the regression model to subgroups stratified by various cut-off age.…”

Forecasting emergency department waiting time using a state space representation

“…Similarly, Fukaya et al 21 and Kawamori et al 22 utilized an SS model to measure base body temperature and forecast menstruation periods. Furthermore, Christensen et al 23 employed an SS model to capture seasonal variations in hospitalization rates for stroke. Within a SS framework, we decompose the logarithm of waiting time y t into a conditional mean process 𝜇 t and an error process 𝜀 y t as follows:…”

“…Similarly, Fukaya et al 21 and Kawamori et al 22 utilized an SS model to measure base body temperature and forecast menstruation periods. Furthermore, Christensen et al 23 employed an SS model to capture seasonal variations in hospitalization rates for stroke.…”

Decoding and modelling of time series count data using Poisson hidden Markov model and Markov ordinal logistic regression models

“…5,6 Therefore, researchers use Harmonic Poisson regression in order to model time series data instead of ARIMA models. [7][8][9][10][11][12] However, in such analyses, the assumptions of independence and equi dispersion may be violated. Some researchers used Poisson auto regressive (PAR) models or auto regressive conditional Poisson (ACP) models to handle the autocorrelation and over dispersion in the data.…”