Hostname: page-component-74656f9884-qr5l4 Total loading time: 0 Render date: 2026-11-05T04:37:58.263Z Has data issue: false hasContentIssue false
  • English
  • Français

Transmission-route–specific trends in communicable diseases in Taiwan: an interrupted time-series analysis

Published online by Cambridge University Press:  04 November 2026

Jie Yu Lau Open the ORCID record for Jie Yu Lau [Opens in a new window] ,
Wen-Chi Pan  and
Dar-Der Ji
Jie Yu Lau
Affiliation:
Institute of Public Health, National Yang Ming Chiao Tung University , Taiwan International Health Program, National Yang Ming Chiao Tung University , Taiwan
Wen-Chi Pan
Affiliation:
International Health Program, National Yang Ming Chiao Tung University , Taiwan Institute of Environmental and Occupational Health Sciences, National Yang Ming Chiao Tung University , Taiwan
Dar-Der Ji*
Affiliation:
International Health Program, National Yang Ming Chiao Tung University , Taiwan Department of Tropical Medicine, National Yang Ming Chiao Tung University , Taiwan
*
Corresponding author: Dar-Der Ji; Email: ddji@nycu.edu.tw
Rights & Permissions [Opens in a new window]

Abstract

The COVID-19 pandemic substantially altered patterns of communicable disease (CD) transmission through implementation of non-pharmaceutical interventions and changes in healthcare utilization. However, long-term trends remain insufficiently characterized. We conducted an ecological time-series study using national surveillance data of notifiable CD in Taiwan from January 2017 to December 2023. Diseases were categorized into transmission-route groups, and temporal trends were assessed using segmented interrupted time-series models with autoregressive integrated moving average error structures to estimate phase-specific level and slope changes. A total of 240556 cases were reported, with more than 96% locally acquired. Overall incidence declined during the pandemic phase (2020–2022) and increased in 2023, exceeding pre-pandemic levels. This increase was largely influenced by dengue-related vector-borne disease activity. Substantial heterogeneity was observed across transmission pathways: airborne and droplet diseases declined markedly with partial recovery, sexually transmitted and blood-borne diseases increased consistently, and vector-borne diseases showed an outbreak-driven increase in 2023. Interrupted time-series analysis identified significant negative level changes at the onset of the pandemic and positive slope changes in the post-pandemic phase. These findings suggest transmission-route–specific temporal patterns during pandemic and post-pandemic phases and underscore the need for adaptive surveillance and prevention strategies in the post-pandemic era.

Keywords

communicable diseasesCOVID-19interrupted time-series analysissurveillanceTaiwan

Information

Type
Original Paper
Information
Epidemiology & Infection , Volume 154 , 2026 , e146
Check for updates
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0), which permits unrestricted re-use, distribution and reproduction, provided the original article is properly cited.
Copyright
© The Author(s), 2026. Published by Cambridge University Press

Highlights

  • • National notifiable disease trends differed by transmission pathway.

  • • Airborne and droplet diseases declined markedly during the pandemic.

  • • Sexually transmitted and blood-borne diseases increased steadily.

  • • The 2023 increase was largely driven by dengue outbreak dynamics.

  • • Findings support adaptive surveillance by transmission pathway.

Introduction

The COVID-19 pandemic substantially altered patterns of human mobility, healthcare utilization, international travel, and infection prevention practices worldwide. As of 2024, more than 700 million cases and 6 million deaths have been reported globally [1]. Although the World Health Organization declared the end of the COVID-19 public health emergency of international concern in May 2023 [Reference Cheng2], SARS-CoV-2 continues to circulate, and the broader consequences of the pandemic on public health systems remain evident.

Beyond the direct impact of SARS-CoV-2, multiple countries have reported pronounced changes in the incidence of other communicable diseases (CDs) during the pandemic period. Significant declines in respiratory and gastrointestinal infections were reported across multiple settings during 2020–2021 [Reference Ullrich3–Reference Deursen8]. However, trends differed by transmission route, with heterogeneous patterns reported for vector-borne and sexually transmitted infections [Reference Ullrich3, Reference Adegbija4]. Most published analyses have focused on the early pandemic phase, with limited evaluation of longer-term trends extending into the post-pandemic period.

Taiwan provides a distinctive context for examining these dynamics. The country maintains a centralized national notifiable disease surveillance system with comprehensive monthly reporting across 73 communicable diseases [9]. During the early pandemic phase, Taiwan implemented timely public health measures and maintained relatively low levels of indigenous COVID-19 transmission [Reference Tang, Lai and Chao10, Reference Lai, Lee and Hsueh11]. Concurrently, declines in several notifiable infectious diseases were observed in 2020 compared with pre-pandemic levels [Reference Lai12]. However, the longer-term evolution of communicable disease patterns across pandemic and post-pandemic phases has not been comprehensively evaluated.

Interrupted time-series (ITS) analysis provides a rigorous framework for assessing temporal changes in longitudinal surveillance data while accounting for secular trends, seasonality, and autocorrelation [Reference Zhou13, Reference Schaffer, Dobbins and Pearson14]. Rather than estimating direct causal effects of specific public health interventions, ITS can be applied to characterize phase-specific shifts in disease incidence corresponding to major epidemiological transitions.

In this study, we analysed national monthly surveillance data on notifiable communicable diseases in Taiwan from 2017 to 2023. Using interrupted time-series models with autoregressive integrated moving average (ARIMA) error structures, we aimed to characterize long-term temporal trends across pre-pandemic, pandemic, and post-pandemic phases, and assess transmission-route–specific heterogeneity in disease dynamics across these phases.

Methods

Study design and data source

This study was an ecological time-series analysis based on national surveillance data from Taiwan. Monthly incidence data for notifiable communicable diseases were obtained from the Statistics of Communicable Diseases and Surveillance System (SCDSS) maintained by the Taiwan Centers for Disease Control (CDC). The surveillance system encompasses 73 legally notifiable communicable diseases, excluding COVID-19 and acute hepatitis of unspecified aetiology.

The study period extended from 1 January 2017 to 31 December 2023. For analytical purposes, the timeline was categorized into three epidemiological phases: pre-pandemic (January 2017–December 2019), pandemic (January 2020–December 2022), and post-pandemic (January 2023–December 2023). These temporal segments correspond to major epidemiological and policy phases of the COVID-19 pandemic in Taiwan and were used to examine phase-specific changes in CD dynamics.

The Taiwan CDC classifies notifiable CD by transmission pathway. For this analysis, diseases were grouped into five transmission-route categories: airborne and droplet diseases (ADDs), contact transmission diseases, food- and water-borne diseases (FWBDs), vector-borne diseases (VBDs), and sexually transmitted and blood-borne diseases (STBBDs). Diseases were assigned to these categories according to the primary surveillance-defined classification used by the Taiwan CDC. For diseases with multiple potential transmission pathways, the predominant or surveillance-defined route was used for grouping. For selected analyses, vector-borne, food- and water-borne, and contact diseases were combined (VFCD) to reflect similarities in non-respiratory or environmentally mediated transmission (Supplementary Table 1).

The population under surveillance comprised the entire resident population of Taiwan, ~23 million individuals, during the study period. HIV infection and AIDS cases among foreign nationals were excluded in accordance with national reporting standards. Age groups were categorized as <1, 1–4, 5–14, 15–24, 25–39, 40–64, and ≥65 years. Geographic regions were defined according to six administrative divisions designated by the Taiwan CDC.

Statistical analysis

Descriptive analyses were conducted to summarize annual case counts and proportional distributions by transmission route, sex, age group, and geographic region. Year-specific percentage changes in annual incidence during the pandemic (2020–2022) and post-pandemic (2023) phases were calculated relative to the mean annual incidence observed during the pre-pandemic period (2017–2019). These descriptive comparisons are presented in Supplementary Table 2. All statistical analyses were performed using R 4.5.0 (R Foundation for Statistical Computing, Vienna, Austria). Statistical significance was defined as a two-sided p-value <0.05.

A continuous interrupted time-series (ITS) framework was applied to monthly incidence data from January 2017 through December 2023, comprising 84 monthly time points. ITS is a regression-based approach used to examine whether a time series changes after a predefined interruption point. In this study, the interruption points divided the surveillance period into epidemiological phases and allowed estimation of both immediate and gradual changes in monthly incidence. Two temporal segmentation points were defined a priori: January 2020, corresponding to the onset of the COVID-19 pandemic phase in Taiwan, and January 2023, corresponding to the post-pandemic phase following large-scale societal reopening. To assess the robustness of the findings to alternative breakpoint specifications, sensitivity analyses were conducted using March 2020 as an alternative pandemic-phase breakpoint and October 2022 as an alternative reopening-phase breakpoint. In addition, dengue-exclusion sensitivity analyses were performed for notifiable CD and VFCD using the primary breakpoint specification to evaluate the influence of dengue outbreak dynamics on the post-pandemic increase.

Segmented regression models incorporating autoregressive integrated moving average (ARIMA) error structures were used to estimate changes in level (immediate step change) and slope (gradual change) across epidemiological phases. In this framework, a level change, also referred to as a step change, represents an immediate change in incidence at the interruption point, whereas a slope change, also referred to as a ramp change, represents a gradual change in the monthly trend after the interruption point. This modelling approach permits characterization of temporal shifts while accounting for underlying secular trends, seasonal variation, and autocorrelation commonly observed in infectious disease surveillance data. ARIMA error structures were used to account for correlation between observations over time, including seasonal patterns and autocorrelation in monthly surveillance data. The analysis was designed to describe phase-specific temporal dynamics rather than to estimate causal effects of specific public health interventions. ITS models were implemented in R using the forecast and stats packages.

ARIMA model identification followed the Box–Jenkins methodology. ARIMA error structures were selected separately for each outcome and phase-specific model. Stationarity and the need for non-seasonal or seasonal differencing were assessed using visual inspection of time-series plots, autocorrelation function (ACF) and partial autocorrelation function (PACF) patterns, and the differencing order suggested during ARIMA identification. Both non-seasonal ARIMA (p, d, q) and seasonal ARIMA (p, d, q)(P, D, Q) with a 12-month seasonal period were considered. Model selection was guided by the Akaike information criterion (AIC), corrected AIC (AICc), and Bayesian information criterion (BIC), with preference given to parsimonious models minimizing information criterion values.

Model adequacy was evaluated through residual diagnostics, including visual inspection of residual time-series plots, examination of autocorrelation function (ACF) plots, and Ljung–Box tests for residual autocorrelation. Models were deemed adequate when residuals approximated white noise and Ljung–Box test results were non-significant (p > 0.05). Diagnostic plots are presented in the Supplementary Materials (Supplementary Figure 1).

Relative percentage changes were calculated by dividing the estimated level or slope change by the model-estimated expected incidence at the corresponding interruption point. Relative changes for slope estimates were expressed as approximate monthly percentage changes relative to the expected level at the breakpoint.

Results

Overall surveillance patterns

Between January 2017 and December 2023, a total of 240,556 cases of notifiable communicable diseases were reported in Taiwan, including 3,560 imported cases. Locally acquired infections accounted for more than 96% of reported cases annually throughout the study period (Table 1). Total annual case numbers declined during the pandemic phase (2020–2022) compared with the pre-pandemic period (2017–2019), followed by an increase to 57,372 cases in 2023, exceeding pre-pandemic levels. Imported cases exhibited a similar temporal pattern, peaking in 2019, declining markedly in 2020–2021, and partially rebounding in 2023. Across most years, males represented the majority of reported cases (76–82%), largely reflecting patterns observed in sexually transmitted and blood-borne diseases (STBBD). Individuals aged 25–39 years consistently accounted for the largest proportion of cases, although demographic distributions shifted in 2023 in association with the dengue outbreak.

Table 1.

The descriptive characteristics of notifiable communicable diseases pre-pandemic, pandemic, and post-pandemic in Taiwan from 2017 to 2023

A table detailing notifiable communicable disease characteristics in Taiwan from 2017 to 2023, categorized by case type, transmission, agent, sex, locality, and age. See long description.
Table 1. Long description

The table is organized into three main time periods: Pre-pandemic (2017, 2018, 2019), Pandemic (2020, 2021), and Post-pandemic (2022, 2023).

* Indigenous cases: Remained high throughout, peaking in 2023 at 56,882 (99.15 percent).

* Imported cases: Peaked in 2019 at 1,142 (3.51 percent) and dropped significantly during the pandemic to a low of 85 (0.30 percent) in 2021.

* Transmission pathway: STBBD (sexually transmitted and blood-borne diseases) was the dominant pathway until 2022, but VBD (vector-borne diseases) surged to 27,334 cases (47.64 percent) in 2023. ADD (airborne and droplet diseases) showed a steady decline from 37.71 percent in 2017 to 15.30 percent in 2023.

* Causative agent: Bacteria were the primary agent for most years (peaking at 88.02 percent in 2022), but Viruses surged to 53.76 percent in 2023. Parasites remained below 1.3 percent, and Prions were at 0.

* Sex: Males consistently represented the majority of cases, ranging from 67.86 percent to 81.84 percent.

* Locality: The Taipei Area was the primary hotspot until 2022. In 2023, the Southern Area saw a massive increase to 25,956 cases (45.24 percent).

* Age: The 25 to 39 years age group consistently had the highest case percentages until 2023, when the 40 to 64 years group increased to 31.39 percent.

N.A.: not available; STBBD: sexually transmitted and blood-borne diseases; ADD: airborne and droplet diseases; FWBD: food- and water-borne diseases; VBD: vector-borne diseases.

Trends by transmission pathway

Substantial heterogeneity was observed across transmission-route categories (Supplementary Table 2). STBBD demonstrated a sustained upward trajectory throughout the study period, primarily driven by increasing cases of syphilis and gonorrhoea. In contrast, ADD declined markedly during the pandemic phase, with only partial recovery observed in 2023. VBD decreased during 2020–2022 but increased sharply in 2023, largely reflecting dengue outbreak dynamics. FWBD and contact transmission diseases exhibited moderate declines during the pandemic phase, followed by partial increases in 2023. The overall rise in total communicable disease incidence observed in 2023 was predominantly attributable to vector-borne disease activity rather than uniform increases across transmission routes. A descriptive decomposition showed that dengue fever accounted for 92.1% of the increase in notifiable CD in 2023 compared with the mean annual number of cases during 2020–2022, supporting that the apparent overall increase was predominantly dengue-driven rather than a uniform increase across communicable diseases.

Indigenous sexually transmitted diseases

Among STBBD with more than 100 annual cases, syphilis and gonorrhoea demonstrated consistent increases over time (Figure 1). Syphilis cases increased from 8799 in 2020 to 9941 in 2023, while gonorrhoea rose from 4209 cases in 2018 to 8257 in 2023. Temporary declines were observed during peaks of community COVID-19 transmission in mid-2021 and mid-2022, followed by subsequent rebounds. In contrast, HIV infection and AIDS cases declined steadily from 2017 to 2023. Amoebiasis decreased during the early pandemic phase and gradually increased from 2021 onward. An increasing proportion of indigenous cases relative to imported cases was observed in later years of the study period.

Figure 1.

The indigenous cases of important sexually transmitted diseases were over 100 cases per year from 2017 to 2023. (a) Cases of Syphilis; (b) cases of Gonorrhoea; (c) cases of HIV infection; (d) cases of AIDS; (e) cases of Amoebiasis.

Five bar charts showing annual incidence of Syphilis, Gonorrhea, H I V, A I D S, and Amoebiasis from 2017 to 2023. See long description.
Figure 1. Long description

A multi-panel figure containing five bar charts, each with Year on the x-axis from 2017 to 2023 and Incidence of Cases on the y-axis.

* Panel A. Indigenous Cases of Syphilis. The y-axis ranges from 0 to 10000. Cases remain high and relatively stable, starting near 10000 in 2017, dipping slightly to a low in 2020, and returning to approximately 10000 by 2023.

* Panel B. Indigenous Cases of Gonorrhea. The y-axis ranges from 0 to 8000. There is a significant upward trend, starting around 4500 in 2017 and increasing steadily to over 8000 by 2023.

* Panel C. Indigenous Cases of HIV. The y-axis ranges from 0 to 2500. There is a consistent downward trend, starting at 2500 in 2017 and decreasing every year to just under 1000 in 2023.

* Panel D. Indigenous Cases of AIDS. The y-axis ranges from 0 to 1000. Similar to HI V, there is a steady decline from approximately 1400 in 2017 to roughly 600 in 2023.

* Panel E. Indigenous Cases of Amoebiasis. The y-axis ranges from 0 to 200. The data shows a U-shaped curve, starting near 190 in 2017, dropping to a low of approximately 125 in 2020, and rising back to nearly 200 by 2023.

Interrupted time-series analysis

Segmented regression analysis identified significant changes in level and slope across epidemiological phases (Table 2 and Figure 2). For notifiable CD, a significant immediate reduction in level was observed at the transition to the pandemic phase (step: −224.82, 95% CI: −412.27 to −37.38), corresponding to a −7.5% relative change compared with the expected level at the breakpoint. The subsequent slope change during 2020–2022 was negative but not statistically significant (ramp: −7.20, 95% CI: −15.62 to 1.21; relative change: −0.24% per month), suggesting relatively stable trends after the initial decline.

Table 2.

ITS-ARIMA modelled the incidence of CD with various transmission pathways from 2017 to 2022 (pre-pandemic and pandemic) and 2020 to 2023 (pandemic and post-pandemic) in Taiwan

A statistical table showing I T S A R I M A model results for communicable disease incidence in Taiwan across two periods: 2017 to 2022 and 2020 to 2023. See long description.
Table 2. Long description

The table contains 10 columns: Transmission pathways, ITS-ARIMA model, Ljung-Box Test p-value, AI C, AICc, BI C, Step (95% CI), Step relative change (95% CI), Ramp (95% CI), and Ramp relative change (95% CI).

Key data rows include:

* Notifiable CD 2017 to 2022: ARIMA (0,0,1) (1,0,0). Step relative change is minus 7.5%.

* Notifiable CD 2020 to 2023: ARIMA (3,0,1). Step relative change is minus 79.02% and Ramp relative change is 28.05%.

* STBBD 2017 to 2022: ARIMA (0,0,0) (1,0,0). Ramp relative change is 0.34%.

* STBBD 2020 to 2023: ARIMA (1,0,0) (1,0,0). Step relative change is 2.26%.

* ADD 2017 to 2022: ARIMA (0,0,1) (0,0,2). Step relative change is minus 10.80% and Ramp relative change is minus 1.03%.

* ADD 2020 to 2023: ARIMA (0,1,0) (1,0,0). Step relative change is 26.31%.

* VFCD 2017 to 2022: ARIMA (2,0,1). Step relative change is minus 34.16% and Ramp relative change is minus 1.33%.

* VFCD 2020 to 2023: ARIMA (2,0,1) (1,0,0). Step relative change is minus 1,501.29% and Ramp relative change is 555.34%.

Statistical significance is indicated by asterisks: * for p < 0.05, ** for p < 0.01, and *** for p < 0.001.

ITS: interrupted time series; ARIMA: autoregressive integrated moving average; AIC: Akaike information criterion; AICc: Akaike information criterion corrected; BIC: Bayesian information criterion; Notifiable CD: notifiable communicable diseases; STBBD: sexually transmitted and blood-borne diseases; ADD: airborne and droplet diseases; VFCD: vector-borne, food- and water-borne, and contact diseases. The significance of p-value was *:<0.05; **:<0.01 *** < 0.001.

AIC, AICc, and BIC are presented descriptively for the selected models and were used for model selection and parsimony assessment, not for formal comparison across disease categories.

Figure 2.

Interrupted time series (ITS) analysis of notifiable communicable diseases in Taiwan by transmission pathway. Panels depict observed and predicted monthly incidence pre-pandemic, pandemic, and post-pandemic. (a)–(d) Trends from January 2017 to December 2022 (before to during COVID-19), with the vertical dashed line and arrow indicating the implementation of COVID-19 non-pharmaceutical interventions (NPIs) in January 2020. (e)–(h) Trends from January 2020 to December 2023 (during to post-COVID-19), with the dashed line and arrow indicating the relaxation of major NPIs in January 2023. Disease categories include notifiable communicable diseases (CD; a, e), sexually transmitted and blood-borne diseases (STBBD; (b, f)), airborne and droplet diseases (ADD; c, g), and vector-borne, food- and water-borne, and contact diseases (VFCD; d, h). Blue solid lines represent observed values, red dashed lines represent ARIMA-predicted counterfactual values, and pink shaded bands represent 95% prediction intervals around the predicted trends.

Eight line graphs showing interrupted time series analysis of communicable diseases in Taiwan across pre-pandemic, pandemic, and post-pandemic periods. See long description.
Figure 2. Long description

Eight line graphs arranged in two columns. The left column (A to D) covers 2017 to 2022, and the right column (E to H) covers 2020 to 2024. Each graph plots monthly incidence (blue solid line) against ARIMA-predicted counterfactual values (red dashed line) with a pink 95 percent prediction interval.

* Panel A and E: CD (Communicable Diseases). Panel A shows a drop in incidence following NPI implementation in January 2020. Panel E shows a massive spike in incidence following NPI removal in January 2023.

* Panel B and F: STBBD (Sexually Transmitted and Blood-Borne Diseases). These show relatively stable trends with minor fluctuations around the predicted values both during and after NPI implementation.

* Panel C and G: ADD (Airborne and Droplet Diseases). Panel C shows a significant and sustained decrease in incidence below the predicted counterfactual after January 2020. Panel G shows a gradual return toward predicted levels after January 2023.

* Panel D and H: VFCD (Vector-borne, Food- and Water-borne, and Contact Diseases). Panel D shows a decrease after January 2020. Panel H shows a dramatic, sharp peak in incidence shortly after the NPI removal in early 2023.

In all panels, a vertical dashed line with a black arrow indicates the intervention point: NPI implementation in January 2020 for the left column, and NPI removal in January 2023 for the right column.

At the transition to the post-pandemic phase, notifiable CD showed a negative level change at the breakpoint (step: −1933.34, 95% CI: −2786.93 to −1079.75; relative change: −79.02%) but a strong positive slope change (ramp: 682.22, 95% CI: 560.87 to 811.57; relative change: 28.05%). The negative step estimate should be interpreted cautiously as model re-anchoring at the phase boundary rather than as a true abrupt epidemiological decline at reopening. The significant positive slope change was consistent with the rapid increase in reported cases during 2023.

Transmission-route–specific analyses demonstrated substantial heterogeneity. STBBD showed a significant positive slope change during the pandemic phase (ramp: 5.32, 95% CI: 0.55 to 10.10; relative change: 0.34%), indicating continued monthly increases despite reductions in other transmission categories. No significant level or slope changes were observed for STBBD during the post-pandemic phase.

In contrast, ADD showed significant negative level and slope changes during the pandemic phase (step: −135.52, 95% CI: −226.25 to −44.79; relative change: −10.80%; ramp: −12.90, 95% CI: −16.90 to −8.89; relative change: −1.03%), consistent with sustained declines across 2020–2022. During the post-pandemic phase, ADD showed positive but non-significant level and slope changes, suggesting no clear rebound to pre-pandemic levels.

VFCD also demonstrated significant reductions in both level and slope during the pandemic phase (step: −40.94, 95% CI: −75.42 to −6.45; relative change: −34.16%; ramp: −1.60, 95% CI: −3.10 to −0.09; relative change: −1.33%). During the post-pandemic phase, VFCD showed a negative level change at the breakpoint but a pronounced positive slope change thereafter (step: −1638.89, 95% CI: −2290.00 to −987.78; ramp: 606.24, 95% CI: 493.46 to 719.02). Similar to the total CD model, the negative step estimate likely reflected model re-anchoring at the phase boundary, whereas the positive slope change corresponded to the marked increase during 2023.

Overall, ITS estimates demonstrated heterogeneous temporal patterns across transmission-route categories. Sensitivity analyses using March 2020 and October 2022 as alternative breakpoints yielded broadly similar transmission-route–specific patterns to the main analysis, although the magnitude and statistical significance of some estimates varied across model specifications (Supplementary Table 3 and Supplementary Figure 2). Dengue-exclusion sensitivity analyses showed that the magnitude of the post-pandemic increase was attenuated after excluding dengue fever, particularly for notifiable CD and VFCD (Supplementary Table 4 and Supplementary Figure 3). These findings indicated that the apparent increase in total communicable disease incidence in 2023 was substantially influenced by dengue outbreak dynamics rather than representing a uniform post-pandemic increase across transmission-route categories.

Discussion

This study identified substantial transmission-route–specific heterogeneity in temporal trends of notifiable CD in Taiwan from 2017 to 2023. Overall incidence declined during the pandemic phase and increased in 2023, although this increase was largely driven by dengue-related vector-borne disease activity. Airborne and droplet diseases showed sustained reductions during 2020–2022 with only partial recovery thereafter. In contrast, sexually transmitted and blood-borne diseases exhibited continued upward trends throughout the study period. Vector-borne diseases declined during the pandemic but increased sharply in 2023, largely driven by outbreak dynamics rather than uniform resurgence. These findings suggest that temporal changes in CD incidence varied by transmission route and may reflect a combination of pandemic-related disruptions, changes in healthcare utilization and surveillance, and disease-specific outbreak dynamics. For the post-pandemic phase, slope changes were considered more informative than negative level changes when interpreting temporal trends, particularly for notifiable CD and VFCD.

The pronounced reduction in airborne and droplet diseases during 2020–2022 temporally coincided with the implementation of non-pharmaceutical interventions (NPIs), including universal masking, social distancing, border control, and contact tracing. Similar collateral declines in respiratory and gastrointestinal infections have been reported across multiple countries during early pandemic phases [Reference Ullrich3, Reference Adegbija4, Reference Galvin15–Reference Poole18]. In Taiwan, sustained mask use and high public adherence to infection control measures may have contributed to the observed changes in level and slope in interrupted time-series models [Reference Launay7, Reference Deursen8, Reference Tan19]. However, reduced healthcare utilization and altered health-seeking behaviour during peak outbreak periods may also have contributed to decreased case detection, particularly for diseases with mild or self-limiting presentations [Reference Tan19, 20].

Vector-borne diseases demonstrated a distinct temporal pattern. The decline observed during the pandemic phase is consistent with reduced international travel and limited importation of dengue and chikungunya cases under stringent border control policies [Reference Ullrich3]. In contrast to reports from Japan and Australia, where increased outdoor activity was associated with elevated risk of vector-borne diseases [Reference Adegbija4, Reference Hirae, Hoshina and Koga5], no sustained increase was observed in Taiwan during 2020–2022. The sharp increase in 2023 was temporally associated with border reopening and increased travel to endemic regions [Reference Wang and Tsay21–23]. The concentration of dengue transmission in southern and Kao-Ping regions, where climatic conditions favour Aedes aegypti propagation, together with dengue-exclusion sensitivity analyses, suggests that the 2023 increase was primarily driven by localized dengue outbreak dynamics rather than a generalized increase across communicable diseases in the post-pandemic period [Reference Eder24].

Sexually transmitted and blood-borne diseases exhibited continued upward trends throughout the study period. Syphilis and gonorrhoea increased consistently, with transient declines corresponding to peak domestic COVID-19 waves, likely reflecting disruptions in healthcare access and testing services [Reference Lai, Lee and Hsueh11, Reference Chan25]. The sustained positive slope observed in ITS models suggests that transmission dynamics of these infections were less sensitive to population-wide NPIs than respiratory infections. In contrast, HIV and AIDS cases declined gradually, which may be associated with expanded uptake of pre-exposure prophylaxis and targeted prevention strategies implemented before the pandemic [Reference DiNenno26–Reference Huang28], as well as temporary reductions in screening during COVID-19 [Reference Wu29]. In the absence of individual-level behavioural and testing data, these findings should be interpreted as temporal associations rather than direct evidence of behavioural change.

The observed heterogeneity across transmission pathways highlights the differential impact of pandemic-related disruptions according to transmission mechanisms, mobility patterns, and healthcare access. Respiratory and environmentally mediated infections showed temporal declines broadly consistent with the period of population-level NPIs, although changes in healthcare-seeking behaviour, diagnostic practices, and reporting patterns may also have contributed [Reference Ullrich3, Reference Adegbija4, Reference Galvin15–Reference Tan19]. In contrast, sexually transmitted infections may be more strongly influenced by social network structures and targeted screening programmes [Reference de Vries30–Reference Orellana33]. Outbreak-prone diseases such as dengue may additionally reflect importation pressure and climatic conditions, particularly following the relaxation of travel restrictions [Reference Wang and Tsay21–Reference Eder24]. These distinctions have important implications for the design of adaptive surveillance and prevention strategies in the post-pandemic context. These findings may inform prioritization of surveillance resources and targeted intervention strategies in future public health responses to large-scale disruptions.

Several strengths enhance the robustness of the present study. The use of comprehensive nationwide surveillance data spanning seven consecutive years enabled population-level assessment across before pandemic, pandemic, and post-pandemic phases. In addition, the application of segmented interrupted time-series models with ARIMA error structures allowed characterization of both immediate and gradual changes in incidence while accounting for seasonality and autocorrelation, thereby strengthening the internal validity of temporal inferences.

Nevertheless, several limitations should be considered. First, as an ecological study, causal relationships between COVID-19–related NPIs and changes in disease incidence cannot be established, and findings should be interpreted as temporal associations. The fixed interruption points should be interpreted as simplified markers of broader epidemiological phases, as segmented models may not fully capture phased policy changes or heterogeneous behavioural responses. Counterfactual estimates also rely on the assumption that pre-existing trends would have continued, which may be affected by concurrent changes in prevention policies, including STI prevention policies and pre-exposure prophylaxis (PrEP) uptake, healthcare access, testing behaviour, travel patterns, and climate variability. Second, reliance on passive surveillance data may introduce reporting bias that varies across disease categories, particularly during periods of reduced healthcare utilization. Reduced access to healthcare and screening during peak pandemic periods may have contributed to underdiagnosis of diseases with mild or self-limiting presentations, whereas diseases subject to intensified public health surveillance, such as dengue, may have been less affected. Third, vector-borne diseases are characterized by episodic, outbreak-driven dynamics; consequently, slope estimates during the post-pandemic phase may be disproportionately influenced by single-year outbreak events. Fourth, transmission-route classification may be subject to misclassification, as some communicable diseases involve multiple transmission pathways, potentially obscuring disease-specific heterogeneity within aggregated route categories.

Conclusion

In conclusion, our findings showed that airborne and droplet diseases declined markedly during the pandemic phase, whereas sexually transmitted and blood-borne diseases continued to increase throughout the study period, and the 2023 increase in vector-borne diseases was largely attributable to a dengue outbreak. These findings suggest that changes in notifiable communicable disease incidence during and after the COVID-19 pandemic were not uniform across transmission pathways, but reflected differences in transmission mechanisms, mobility patterns, healthcare access, surveillance practices, and outbreak dynamics. Strengthening transmission-route–specific surveillance and prevention strategies will be essential for mitigating future outbreaks and enhancing public health preparedness in the post-pandemic era.

Supplementary material

The supplementary material for this article can be found at http://doi.org/10.1017/S0950268826102295.

Data availability statement

The data analysed in this study are publicly available from the Statistics of Communicable Diseases and Surveillance System (SCDSS) provided by the Taiwan Centers for Disease Control (CDC). https://www.cdc.gov.tw/En/Report/MonthList/14567.

Acknowledgements

The authors prepared the manuscript and used OpenAI’s ChatGPT and Grammarly to assist with language editing and minor proofreading. The authors remain fully responsible for the content, analysis, and interpretation.

Author contributions

All authors conceived the study. J.Y.L. curated data, performed analysis, visualized, drafted, reviewed, and edited the manuscript. W.-C.P. performed analysis, validated, reviewed, and edited the manuscript. D.-D.J. validated, drafted, reviewed, and edited the manuscript. All authors reviewed and approved the final manuscript.

Funding statement

This study was supported by the National Science and Technology Council, Taiwan (NSTC-113-2321-B-A49-011).

Competing interests

The authors declare no competing interests.

Ethical standards

The study was reviewed and approved by the Institutional Review Board (IRB) of National Yang Ming Chiao Tung University (NYCU112082AE).

References

World Health Organization (2024) WHO COVID-19 Dashboard. https://data.who.int/dashboards/covid19/cases (accessed 23 August 2024).Google Scholar
Cheng, K, et al. (2023) WHO declares the end of the COVID-19 global health emergency: Lessons and recommendations from the perspective of ChatGPT/GPT-4. International Journal of Surgery 109(9), 2859–2862. https://doi.org/10.1097/JS9.0000000000000521.CrossRefGoogle ScholarPubMed
Ullrich, A, et al. (2021) Impact of the COVID-19 pandemic and associated non-pharmaceutical interventions on other notifiable infectious diseases in Germany: An analysis of national surveillance data during week 1-2016 – Week 32-2020. Lancet Regional Health Europe 6, 100103. https://doi.org/10.1016/j.lanepe.2021.100103.CrossRefGoogle ScholarPubMed
Adegbija, O, et al. (2021) Notifiable diseases after implementation of COVID-19 public health prevention measures in Central Queensland, Australia. Communicable Diseases Intelligence 45. https://doi.org/10.33321/cdi.2021.45.11.CrossRefGoogle ScholarPubMed
Hirae, K, Hoshina, T and Koga, H (2023) Impact of the COVID-19 pandemic on the epidemiology of other communicable diseases in Japan. International Journal of Infectious Diseases 128, 265–271. https://doi.org/10.1016/j.ijid.2023.01.013.CrossRefGoogle ScholarPubMed
Li, H, et al. (2022) Comparison of 19 major infectious diseases during COVID-19 epidemic and previous years in Zhejiang: Implications for prevention measures. BMC Infectious Diseases 22(1), 296. https://doi.org/10.1186/s12879-022-07301-w.CrossRefGoogle Scholar
Launay, T, et al. (2021) Common communicable diseases in the general population in France during the COVID-19 pandemic. PLoS One 16(10), e0258391. https://doi.org/10.1371/journal.pone.0258391.CrossRefGoogle ScholarPubMed
Deursen, B, et al. (2022) A sharp decrease in reported non-COVID-19 notifiable infectious diseases during the first wave of the COVID-19 epidemic in the Rotterdam region, the Netherlands: A descriptive study. BMC Infectious Diseases 22(1), 208. https://doi.org/10.1186/s12879-022-07209-5.CrossRefGoogle Scholar
Taiwan Centers for Disease Control (2024) Taiwan National Infectious Disease Statistics System. https://nidss.cdc.gov.tw/en/nndss/Cdcwnh07?id=061 (accessed 27 August 2024).Google Scholar
Tang, HJ, Lai, CC and Chao, CM (2022) The collateral effect of COVID-19 on the epidemiology of airborne/droplet-transmitted notifiable infectious diseases in Taiwan. Antibiotics 11(4), 478. https://doi.org/10.3390/antibiotics11040478.CrossRefGoogle ScholarPubMed
Lai, CC, Lee, PI and Hsueh, PR (2023) How Taiwan has responded to COVID-19 and how COVID-19 has affected Taiwan, 2020–2022. Journal of Microbiology, Immunology and Infection 56(3), 433–441. https://doi.org/10.1016/j.jmii.2023.04.001.CrossRefGoogle ScholarPubMed
Lai, CC, et al. (2021) The impact of the coronavirus disease 2019 epidemic on notifiable infectious diseases in Taiwan: A database analysis. Travel Medicine and Infectious Disease 40, 101997. https://doi.org/10.1016/j.tmaid.2021.101997.CrossRefGoogle ScholarPubMed
Zhou, Q, et al. (2023) Interrupted time series analysis using the ARIMA model of the impact of COVID-19 on the incidence rate of notifiable communicable diseases in China. BMC Infectious Diseases 23(1), 375. https://doi.org/10.1186/s12879-023-08229-5.CrossRefGoogle ScholarPubMed
Schaffer, AL, Dobbins, TA and Pearson, SA (2021) Interrupted time series analysis using autoregressive integrated moving average (ARIMA) models: A guide for evaluating large-scale health interventions. BMC Medical Research Methodology 21(1), 58. https://doi.org/10.1186/s12874-021-01235-8.CrossRefGoogle ScholarPubMed
Galvin, CJ, et al. (2020) COVID-19 preventive measures showing an unintended decline in infectious diseases in Taiwan. International Journal of Infectious Diseases 98, 18–20. https://doi.org/10.1016/j.ijid.2020.06.062.CrossRefGoogle ScholarPubMed
Huang, HY, et al. (2021) Impact of the COVID-19 nonpharmaceutical interventions on influenza and other respiratory viral infections in New Zealand. Nature Communications 12(1), 1001. https://doi.org/10.1038/s41467-021-21157-9.CrossRefGoogle ScholarPubMed
Chen, B, et al. (2021) Changes in incidence of notifiable infectious diseases in China under the prevention and control measures of COVID-19. Frontiers in Public Health 9, 728768. https://doi.org/10.3389/fpubh.2021.728768.CrossRefGoogle Scholar
Poole, S, et al. (2020) Physical distancing in schools for SARS-CoV-2 and the resurgence of rhinovirus. Lancet Respiratory Medicine 8(12), e92–e93. https://doi.org/10.1016/S2213-2600(20)30502-6.CrossRefGoogle ScholarPubMed
Tan, HMJ, et al. (2021) The impact of COVID-19 pandemic on the health-seeking behaviour of an Asian population with acute respiratory infections in a densely populated community. BMC Public Health 21(1), 1196. https://doi.org/10.1186/s12889-021-11200-1.CrossRefGoogle Scholar
Ministry of Health and Welfare (2020) Infection Control Policies for Medical Facilities and Relevant Institutions. https://covid19.mohw.gov.tw/en/cp-4787-53905-206.html (accessed 3 September 2024).Google Scholar
Wang, YY and Tsay, PK (2022) Risk factors associated with passengers with imported dengue fever at international airports in Taiwan. International Journal of Environmental Research and Public Health 19(17), 11096. https://doi.org/10.3390/ijerph191711096.CrossRefGoogle ScholarPubMed
University of Nebraska Medical Center (2023) Thailand Reports a Nearly 500% Increase in Dengue in 2023. https://www.unmc.edu/healthsecurity/transmission/2023/06/14/thailand-reports-a-nearly-500-increase-in-dengue-in-2023/ (accessed 3 September 2024).Google Scholar
World Health Organization (2023) Dengue – Global Situation. https://www.who.int/emergencies/disease-outbreak-news/item/2023-DON498 (accessed 3 September 2024).Google Scholar
Eder, M, et al. (2018) Scoping review on vector-borne diseases in urban areas: Transmission dynamics, vectorial capacity and co-infection. Infectious Diseases of Poverty 7(1), 90. https://doi.org/10.1186/s40249-018-0475-7.CrossRefGoogle Scholar
Chan, TC, et al. (2022) Effectiveness of controlling COVID-19 epidemic by implementing soft lockdown policy and extensive community screening in Taiwan. Scientific Reports 12(1), 12053. https://doi.org/10.1038/s41598-022-16011-x.CrossRefGoogle ScholarPubMed
DiNenno, EA, et al. (2022) HIV testing before and during the COVID-19 pandemic—United States, 2019–2020. Morbidity and Mortality Weekly Report 71(25), 820–824. https://doi.org/10.15585/mmwr.mm7125a2.CrossRefGoogle Scholar
Stewart, J and Baeten, JM (2022) HIV pre-exposure prophylaxis and sexually transmitted infections: Intersection and opportunity. Nature Reviews Urology 19(1), 7–15. https://doi.org/10.1038/s41585-021-00527-4.CrossRefGoogle ScholarPubMed
Huang, HY, et al. (2024) In the fight against HIV/AIDS: The arduous implementation of government-funded pre-exposure prophylaxis programme in Taiwan. Sexually Transmitted Infections 100(4), 216–221. https://doi.org/10.1136/sextrans-2023-055917.CrossRefGoogle ScholarPubMed
Wu, PY, et al. (2022) Impact of coronavirus disease 2019 on the HIV testing and health care delivery at a university hospital in Taiwan, 2020–2021. Journal of Microbiology, Immunology and Infection 55(6), 1005–1012. https://doi.org/10.1016/j.jmii.2022.10.001.CrossRefGoogle Scholar
de Vries, DC, et al. (2024) Sexual distancing during 5 periods around and during the COVID-19 pandemic, Amsterdam, the Netherlands. Sexually Transmitted Diseases 51(1), 72–80. https://doi.org/10.1097/OLQ.0000000000001888.CrossRefGoogle ScholarPubMed
Nerlander, L, et al. (2024) Sharp increase in gonorrhoea notifications among young people, EU/EEA, July 2022 to June 2023. Eurosurveillance 29(10), 2400113. https://doi.org/10.2807/1560-7917.ES.2024.29.10.2400113.CrossRefGoogle ScholarPubMed
Suzuki, Y, et al. (2020) Association between syphilis incidence and dating app use in Japan. JMA Journal 3(2), 109–117. https://doi.org/10.31662/jmaj.2019-0033.Google ScholarPubMed
Orellana, CB, et al. (2024) Sexually transmitted infections and dating app use. Mathematical Biosciences and Engineering 21(3), 3999–4035. https://doi.org/10.3934/mbe.2024177.CrossRefGoogle ScholarPubMed
Figure 0

Table 1. The descriptive characteristics of notifiable communicable diseases pre-pandemic, pandemic, and post-pandemic in Taiwan from 2017 to 2023Table 1. Long description.

Figure 1

Figure 1. The indigenous cases of important sexually transmitted diseases were over 100 cases per year from 2017 to 2023. (a) Cases of Syphilis; (b) cases of Gonorrhoea; (c) cases of HIV infection; (d) cases of AIDS; (e) cases of Amoebiasis.Figure 1. Long description.

Figure 2

Table 2. ITS-ARIMA modelled the incidence of CD with various transmission pathways from 2017 to 2022 (pre-pandemic and pandemic) and 2020 to 2023 (pandemic and post-pandemic) in TaiwanTable 2. Long description.

Figure 3

Figure 2. Interrupted time series (ITS) analysis of notifiable communicable diseases in Taiwan by transmission pathway. Panels depict observed and predicted monthly incidence pre-pandemic, pandemic, and post-pandemic. (a)–(d) Trends from January 2017 to December 2022 (before to during COVID-19), with the vertical dashed line and arrow indicating the implementation of COVID-19 non-pharmaceutical interventions (NPIs) in January 2020. (e)–(h) Trends from January 2020 to December 2023 (during to post-COVID-19), with the dashed line and arrow indicating the relaxation of major NPIs in January 2023. Disease categories include notifiable communicable diseases (CD; a, e), sexually transmitted and blood-borne diseases (STBBD; (b, f)), airborne and droplet diseases (ADD; c, g), and vector-borne, food- and water-borne, and contact diseases (VFCD; d, h). Blue solid lines represent observed values, red dashed lines represent ARIMA-predicted counterfactual values, and pink shaded bands represent 95% prediction intervals around the predicted trends.Figure 2. Long description.

Supplementary material: File

Lau et al. supplementary material

Lau et al. supplementary material
Download Lau et al. supplementary material(File)
File 5.3 MB