INTRODUCTION
Seasonal infectious diseases have long posed significant public health challenges in the Republic of Korea. Influenza, norovirus, severe fever with thrombocytopenia syndrome (SFTS), and tsutsugamushi disease each follow distinct epidemiological patterns influenced by seasonal and environmental factors. However, the COVID-19 pandemic has introduced a profound disruption to global disease transmission dynamics. This study seeks to quantify the long-term patterns of these infections between 2005 and 2023, while analyzing how COVID-19 impacted their seasonality and outbreak intensity. By integrating classical time-series forecasting through the SARIMAX model with advanced deep learning via LSTM networks, and combining them in a hybrid SARIMAX-LSTM model, this research delivers robust predictive insights. Meteorological data and change point detection were employed to enhance the understanding of outbreak shifts. The goal is not only to understand historical trends but also to forecast and prepare for future outbreaks, particularly in a post-pandemic era.
EPIDEMIOLOGICAL TRENDS ACROSS TWO DECADES
The study spans 18 years, offering a comprehensive view of infectious disease seasonality in South Korea. Influenza and norovirus displayed predictable annual patterns until 2020, while SFTS and tsutsugamushi disease showed more static endemic behaviors. The onset of COVID-19 disrupted these trends, significantly reducing the incidence of influenza-like illnesses and norovirus infections. Interestingly, vector-borne diseases such as SFTS and tsutsugamushi disease did not follow this decline, suggesting that their transmission mechanisms were less impacted by pandemic-related public health interventions such as mask-wearing and social distancing. This divergence reinforces the need to categorize infectious diseases based on both their transmission modes and vulnerability to behavioral changes within the population.
IMPACT OF COVID-19 ON SEASONAL DISEASES
The global pandemic triggered a cascade of health policy shifts that influenced not only SARS-CoV-2 but also other respiratory and gastrointestinal diseases. The study reveals that the incidence of influenza sharply declined during the pandemic and has not returned to pre-pandemic levels even in recent years. Norovirus followed a different trajectory—while it dropped in 2020, it rebounded to prior levels soon after. This suggests that viral shedding and environmental persistence may play crucial roles in post-pandemic resurgence. On the other hand, the relatively stable patterns of SFTS and tsutsugamushi disease throughout the pandemic period imply limited impact from urban-centric COVID-19 interventions, further underlining the importance of ecological and vector-focused disease modeling.
ADVANCED MODELING TECHNIQUES FOR OUTBREAK DETECTION
In this study, SARIMAX, LSTM, and a novel SARIMAX-LSTM hybrid model were implemented to predict infectious disease trends. The SARIMAX model leverages temporal patterns and incorporates exogenous variables such as meteorological data. Meanwhile, LSTM neural networks process long-term dependencies in sequential data. Combining these methodologies produced a hybrid model that demonstrated superior performance in outbreak forecasting. This approach enabled the identification of subtle trend shifts and seasonal fluctuations across diseases. The application of change point detection (CPD) techniques further strengthened the models by highlighting statistically significant deviations in disease incidence, especially during the COVID-19 outbreak period. These insights are essential for early warning systems and targeted resource allocation.
DIFFERENTIAL RECOVERY AND RESURGENCE POTENTIAL
Post-COVID-19 disease dynamics have shown that not all infections recover at the same pace or pattern. Influenza, for instance, experienced a steep decline but remains suppressed compared to pre-pandemic baselines, hinting at altered immunity or sustained behavior changes. Conversely, norovirus quickly resumed its cyclical outbreaks. Predictive modeling suggests that influenza-like illness (ILI) outbreaks may resurge in the near future due to waning immunity and lowered population-level exposure. These findings stress the need to monitor evolving trends rather than assume automatic recovery of seasonal patterns. Disease-specific strategies—such as continued surveillance, targeted vaccination, and behavior-based interventions—must be developed in anticipation of these shifts.
PUBLIC HEALTH IMPLICATIONS AND STRATEGIC RECOMMENDATIONS
The differential effects of the COVID-19 pandemic on infectious disease incidence call for disease-specific public health responses. For respiratory and gastrointestinal viruses like influenza and norovirus, adaptive interventions—including enhanced surveillance during transition seasons and targeted public awareness campaigns—are critical. Vector-borne diseases, which remained stable throughout the pandemic, require sustained ecological monitoring and regional preparedness. Predictive analytics using hybrid SARIMAX-LSTM models and real-time meteorological data can serve as early warning systems. These tools empower healthcare agencies to preempt outbreaks, optimize resource allocation, and reduce the healthcare burden. Moving forward, preparedness must factor in both historical data and the new post-pandemic normal.
๐ Visit: https://infectious-diseases-conferences.pencis.com
๐ฉ Contact: infectioussupport@pencis.com
๐
Nominate Now: https://infectious-diseases-conferences.pencis.com/award-nomination/?ecategory=Awards&rcategory=Awardee
๐ Register: https://infectious-diseases-conferences.pencis.com/award-registration/
HASHTAGS
#InfectiousDiseases, #SeasonalDiseases, #SARIMAXModel, #LSTMForecasting, #HybridModels, #OutbreakDetection, #COVID19Impact, #InfluenzaTrends, #NorovirusSurveillance, #SFTSAnalysis, #TsutsugamushiDisease, #ChangePointDetection, #PublicHealthResearch, #EpidemiologyKorea, #DeepLearningModels, #PandemicInfluence, #DiseasePrediction, #MeteorologicalData, #HealthSurveillance, #DataDrivenHealthPolicy,