GRAPHICAL ABSTRACT

Key Message

• This study used text mining of self-reported symptoms following COVID-19 vaccination to identify previously unrecognized adverse reactions.

• Our findings enhance the understanding of vaccine side effects and provide valuable insights for future research and surveillance system improvement.

INTRODUCTION

To address the unprecedented coronavirus disease 2019 (COVID-19) pandemic and mitigate its public health impact, the development, approval, and distribution of safe and effective vaccines were rapidly accelerated [1]. Consequently, many countries have established and implemented monitoring systems to ensure COVID-19 vaccine safety [1-3]. In Korea, both a web-based passive surveillance system and a text message-based active surveillance system have been deployed for vaccine safety monitoring [4,5]. Unlike systems relying solely on electronic health records, active surveillance proactively collects self-reported symptoms or newly developed health conditions from vaccinated individuals during a predefined period, thereby enabling near real-time safety monitoring [3]. These systems have made substantial contributions to identifying unexpected or undesirable adverse reactions, providing a basis for anticipating adverse events related to vaccines for future COVID-19 variants or other viral infections.

Since its emergence, severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) has undergone continuous and significant evolution, necessitating ongoing vaccination with updated COVID-19 vaccines as a key public health strategy [6]. The Korea Disease Control and Prevention Agency (KDCA) releases annual vaccination plans to prevent the spread of COVID-19. Beginning with the 2023-2024 season, this plan has recommended concurrent administration of COVID-19 and seasonal influenza vaccines for individuals aged 65 years and older [7-9]. Numerous surveillance studies have evaluated vaccine safety and contributed to reducing public hesitancy [10-12]. Monitoring through spontaneous reporting systems can rapidly identify emergent safety signals and novel adverse reactions of concern [1]. In Korea, as the COVID-19 vaccine program was integrated with the influenza vaccination campaign, an active, text message-based surveillance system was introduced. This system encouraged participants to report any adverse reactions experienced—beyond those typically captured within the first 7 days after vaccination—via text message. By enhancing the questionnaire, this approach broadened the scope of adverse event investigation [13]. Text-based data thus offer valuable insights into adverse reactions that may not have been previously reported. Text mining methods can extract hidden knowledge by identifying patterns in large volumes of unstructured text [14]. While several studies have assessed adverse reactions to the coadministration of COVID-19 and influenza vaccines [15,16], ongoing vigilance in monitoring adverse reactions in the general population remains vital for comprehensive safety assessment.

Analyses of social media data, such as those from Twitter, related to COVID-19, have often focused on tracking the evolution of emotions and sentiments throughout the pandemic [17-19], as well as on the analysis of topics and information sharing [20-22]. In the context of COVID-19 vaccination, text analysis has examined reasons for vaccine hesitancy [23], and sentiment analysis has been conducted regarding the vaccine itself [24]. Methods such as sentiment analysis and topic modeling have been used to investigate public perceptions. However, few studies have applied text analysis specifically to identify potential adverse reactions.

Therefore, in this study, we conducted a text mining analysis to identify potential adverse reactions following COVID-19 vaccination using records collected across 2 seasons (2023-2024 and 2024-2025). Because vaccine types varied between seasons, we analyzed each season separately and further examined whether adverse reaction reporting differed according to concomitant influenza vaccination.

MATERIALS AND METHODS

Data source

This study used anonymized data collected by the KDCA. The KDCA conducted active surveillance through the Immunization Registry Information System (IRIS) following administration of the COVID-19 XBB.1.5 vaccine during the 2023-2024 season and the COVID-19 JN.1 vaccine during the 2024-2025 season [25]. Further details on the surveillance system are described elsewhere [4,5]. The 2023-2024 vaccination campaign, which began on October 19, 2023, initially targeted individuals aged 65 years and older, immunocompromised individuals aged 12 to 64, and residents and workers in vulnerable facilities such as nursing hospitals and nursing homes. The campaign was subsequently expanded to include all individuals aged 12 to 64 beginning November 1 [7]. The 2024-2025 vaccination campaign, which commenced on October 11, 2024, focused on individuals aged 65 years and older, immunocompromised patients, hospitalized individuals, and residents and workers at vulnerable facilities [8]. A text-message (short message service, SMS)-based surveillance survey was administered to vaccine recipients who consented to participate via smartphone, monitoring adverse reactions and health status for 7 days post-vaccination. Surveillance was planned to collect data from 10,000 respondents per vaccine manufacturer [12].

Collected data included sex, age, vaccination date, vaccine manufacturer, and information regarding concurrent influenza vaccination. Post-vaccination outcome data comprised any health issues, fever, adverse reactions, interference with daily activities, and healthcare facility visits. Questionnaires on adverse reactions were divided into categories for local and systemic reactions, with responses captured through closed-ended questions. Respondents could report multiple adverse reactions. Additionally, an open-ended question allowed participants to describe any other local or systemic health issues experienced after vaccination in free-text form. For this study, we analyzed text data obtained from the open-ended question. Newly identified adverse reactions were defined as symptoms not included in official national reports [26] or in the predefined local and systemic reactions listed in the questionnaire.

The first survey period (2023-2024 season) included data on adverse reactions in individuals who received the COVID-19 vaccine (10,099 Pfizer-BioNTech and 10,083 Moderna) between October 19, 2023, and November 6, 2023, and who responded to text messages at least once within 7 days post-vaccination. The second survey period (2024-2025 season) included data from individuals vaccinated between October 11, 2024, and November 30, 2024 (10,002 Pfizer-BioNTech, 10,003 Moderna, and 2,906 Novavax). Thus, for text mining analysis, the analytical dataset consisted of 2,608 records (i.e., total submissions, including multiple entries from the same individual) from 1,864 individuals in the first period and 2,054 records from 1,515 individuals in the second period. Data on concurrent COVID-19 and seasonal influenza vaccinations were also obtained from the IRIS. Variables including sex, age, vaccine manufacturer, and concurrent influenza vaccination were considered in the analysis.

Data preprocessing for text mining

Text preprocessing is essential in text mining to remove noise and extract meaningful information [27]. The specific procedures used in this study are described in the Supplementary Material 1.

Data analysis

Descriptive analysis

A basic descriptive analysis was conducted on 1,864 (survey period 1) and 1,515 (survey period 2) COVID-19 vaccine recipients. Categorical variables are presented as frequencies and percentages, while continuous variables are reported as means with standard deviations and medians with interquartile ranges. We also evaluated the impact of concurrent administration of the seasonal influenza vaccine on basic characteristics using the chi-square test, t-test, and Mann-Whitney U test.

Word frequencies

The frequency of adverse reaction-related words was categorized as very common (≥10%), common (≥1 to <10%), or uncommon (≥0.1 to <1.0%) [28]. In addition, the daily reporting percentage for adverse reactions was calculated. Additional details are provided in the Supplementary Material 1.

Semantic network

To uncover patterns in the co-occurrence of adverse reactions, we conducted both co-occurrence and phi-coefficient analyses, visualizing the results to facilitate interpretation. In network analysis, linguistic units are represented as nodes, while relationships observed in actual language use are depicted as links. Keywords with high centrality are considered core terms and can be characterized by degree centrality, betweenness centrality, or closeness centrality. Degree centrality reflects how directly a keyword is connected to other keywords. Betweenness centrality indicates the extent to which a node functions as an intermediary between other nodes. Closeness centrality is based on the distance between nodes and illustrates how close a node is to others. These centrality metrics can be quantified [29]. To examine the usage context of specific words, network analysis included terms such as “injection.”

The co-occurrence network visualized relationships using lines, with the thickness of each line representing the frequency of paired term appearances, regardless of order. The network visualized relationships among keywords reported more than ten times, with node size corresponding to degree centrality.

The phi-coefficient quantifies the degree to which pairs of terms co-occur relative to their individual frequencies and is equivalent to the correlation coefficient for binary variables [30]. For the correlation matrix, we visualized only term pairs with a correlation >0.1, with line thickness representing the phi-correlation coefficient. The Infomap algorithm was applied to group nodes, which were assigned the same color to denote group membership.

All analyses were performed using R version 4.3.1 (R Foundation for Statistical Computing, Vienna, Austria), with R packages including “KoNLP,” “tidyr,” “tidytext,” and “ggraph.”

Ethics statement

All participants provided informed consent to be included in the surveillance database. The study protocol was reviewed and approved by the Institutional Review Board of Ewha Womans University (No. ewha-202401-0011-01).

RESULTS

Among vaccinated individuals, 1,864 (9.2%) and 1,515 (6.6%) responded during the first and second survey periods, respectively, and were included in the analysis. In the first survey period, 65.7% of respondents were male, with a mean age of 68.11 years, and each participant reported adverse reactions an average of 1.4 times. Additionally, 38.2% (n=712) of respondents received the seasonal influenza vaccine concurrently. In the second survey period, 63.0% of the 1,515 respondents were male, with a mean age of 70.98 years. Each individual reported an average of approximately 1.4 adverse reactions, and 81.2% (n=1,230) received the seasonal influenza vaccine at the same time. Compared to those who received only the COVID-19 vaccine, individuals who received both COVID-19 and influenza vaccines concurrently had a higher proportion of persons over 65 years of age, and most received the Moderna vaccine (Table 1).

During the 2023-2024 season, a total of 1,864 individuals reported 2,608 adverse reactions via text messages following vaccination, while in the 2024-2025 season, 1,515 individuals reported a total of 2,054 adverse reactions. Adverse reactions were most frequently reported the day after vaccination, with reporting decreasing over time. There were no significant differences in the percentage of adverse reactions reported within the first 7 days after vaccination between groups stratified by concurrent influenza vaccination status (Supplementary Material 2).

Table 2 displays the terms reported by more than 1% of individuals who reported adverse reactions through text messages. The most frequently reported term was “body aches” (21.9%), with other cold-related symptoms also frequently cited. Sleep-related symptoms, such as “drowsiness” (2.5%) and “sleep disturbance” (1.7%), were also observed. Additionally, terms such as “chest tightness” (1.4%), “lumbago” (1.4%), “shortness of breath” (1.4%), “indigestion” (1.3%), “heart palpitations” (1.3%), and “loss of appetite” (1.1%) were noted. In the second survey period, “body aches” (18.5%) remained the most frequently reported term, while “lumbago” (1.5%), “drowsiness” (1.4%), and “indigestion” (1.3%) were also commonly reported.

Although the frequencies were classified as uncommon (<1.0%), eye-related adverse reactions such as “vision abnormality” (0.8%), “eye pain” (0.4%), and “dry eyes” (<0.1%) were documented. Ear-related reactions, including “tinnitus” (0.4%), “earache” (0.3%), and “ear fullness” (0.2%), were also reported. From the fifth day after vaccination, 4 individuals reported “stomatitis” (0.2%) as an adverse reaction. Other oral cavity-related adverse reactions included “gum pain” (0.2%), “dry mouth” (0.2%), “gum swelling” (0.1%), and “tooth pain” (0.1%). Additional reported symptoms included “hoarseness” (0.7%), “blushing” (0.6%), “increased blood pressure (BP)” (0.4%), and “brain fog” (0.4%). “Memory impairments” were reported by 3 individuals (0.2%), all of whom had received only the COVID-19 vaccine. In the second survey period, terms related to the eyes (“vision abnormality,” “eye fatigue,” “eyelid,” “blepharospasm”) and ears (“tinnitus,” “earache,” “ear fullness”) were again reported. Other reactions included “hoarseness” (0.9%), “blushing” (0.5%), “increased BP” (0.7%), and “brain fog” (0.5%) (Supplementary Material 3).

Terms related to frequently reported adverse reactions showed differences in ranking between groups, but the reported percentages were generally similar. However, some terms differed between groups: “sore throat” and “palpitations” were reported than when both COVID-19 and influenza vaccines were given together. In the second survey period, the frequencies of “nasal congestion” and “increased BP” also differed (Figure 1). Regarding differences in reporting by sex, the terms “headache,” “cold sweat,” “shortness of breath,” “vomiting,” and “indigestion” were reported more often by females than males. In the second survey period, “neck pain,” “nausea,” and “joint pain” were more frequently reported by women, whereas “fatigue,” “drowsiness,” and “sleep disturbance” were more commonly reported by males (Supplementary Material 4). The frequency of reported “joint pain” varied across age groups and was higher in younger participants (Supplementary Material 5).

The reporting percentages for specific, frequently reported terms either remained constant, decreased, or increased over the 7 days following vaccination. The term “body aches” was most frequently reported the day after vaccination and tended to decline thereafter. In contrast, terms such as “sputum,” “cough,” “runny nose,” “sore throat,” and “cold” tended to increase as time progressed post-vaccination. Similar trends were seen in the second survey period (Figure 2).

Figure 3 presents the analysis of co-occurring terms. “injection” and “headache” were often reported alongside other terms. “Injection” was primarily reported together with “pain,” “myalgia,” and “swelling,” whereas “headache” was commonly co-reported with “fatigue,” “body aches,” and “dizziness.” This indicates that various systemic adverse reactions were frequently reported together. In the second survey period, “body aches” and “fatigue” were also commonly reported with other symptoms. Results from the phi-coefficient analysis for pairs of terms are shown in Supplementary Material 6. In the first survey period, “injection” and “pain” showed the highest correlation coefficients, followed by “abdominal pain” and “diarrhea,” and then “cough” and “sputum.” In the second survey period, the pair “injection” and “pain” again showed the strongest correlation.

DISCUSSION

Using a text mining approach, we identified various common (≥1%) adverse reactions among individuals vaccinated against COVID-19 over 2 survey seasons, including reactions not commonly reported in the existing literature. Furthermore, the reporting percentages for most identified adverse reactions remained consistent, regardless of concurrent influenza vaccination.

In early 2021, COVID-19 vaccination campaigns began across multiple countries, including the United States [23]. The efficacy of these vaccines has been demonstrated through both randomized controlled trials and observational studies [31]. However, given the unprecedented speed of vaccine approval, our understanding of potential adverse events remains incomplete. Despite the passage of considerable time, COVID-19 vaccination programs and safety monitoring continue, aiming to address emerging variants and prevent outbreaks. Most studies have primarily focused on identifying a limited spectrum of severe adverse reactions associated with these vaccines. Text analytics provide an opportunity to uncover a broader range of both actual and potential concerns by analyzing keywords in large datasets [32]. Thus, we utilized text analysis to gain a more comprehensive understanding of vaccine safety. As a result, we identified pain in greater detail and captured a wide range of symptoms, including respiratory symptoms, sleep disturbances, lumbago, and indigestion.

Adverse reactions following vaccination can vary depending on the surveillance and reporting system, but local reactions, fatigue, headache, myalgia, and joint pain are commonly reported [11,12,33]. Sex-related and age-related differences in adverse event reporting have also been noted [11,34,35], potentially due to variations in immune response and behavioral factors [34]. However, direct comparisons with our study are limited by differences in the scope of reported reactions, vaccine types, and study populations. Some studies have reported temporary adverse events not previously documented in the literature. In line with earlier findings [15,36], symptoms related to colds were detected, as well as reactions such as sleep disturbance, indigestion, lumbago, and loss of appetite. Although earlier vaccine safety surveillance prompted investigations into symptoms like chest pain, dizziness, and generalized redness, our analysis revealed additional adverse reactions not previously emphasized [12].

Our text analysis also identified uncommon adverse reactions (≥0.1 to <1.0%) involving the eyes, ears, and oral cavity. Despite differences in vaccine types between seasons, a range of eye-related and ear-related symptoms was reported consistently across both years. Although no widespread vision-related adverse events have been reported for the COVID-19 vaccine, isolated cases of dry eyes and transient vision loss (amaurosis fugax) were described after the first and second doses of BNT162b2 mRNA (Pfizer-BioNTech, Comirnaty) in healthcare workers in Italy [34]. In China, monitoring of inactivated COVID-19 vaccines (WIV04 and HB02) found blurred vision in 2 patients after the second dose, and bloodshot or congested eyes in several cases; other uncommon reactions such as palpitations, oral infection, and irregular menstruation were also noted [35]. A study conducted in the United Kingdom among individuals vaccinated between December 8, 2020, and May 17, 2021, similarly reported adverse reactions involving the eyes and ears [37]. However, most reports focus only on local and systemic reactions, making it challenging to obtain information on various adverse reactions.

For practical reasons, the coadministration of the COVID-19 vaccine and the seasonal influenza vaccine was recommended [7,38]. Regarding the safety of simultaneous vaccination, previous studies found that individuals receiving both the seasonal influenza vaccine and a COVID-19 mRNA booster reported a slightly higher frequency of systemic reactions (8% with Pfizer-BioNTech booster and 11% with Moderna booster) compared to those who received only the COVID-19 vaccine, though these reactions were mild and transient [37,39]. The TACTIC study, a randomized clinical trial, showed that coadministration of influenza and COVID-19 vaccines resulted in lower quantitative and functional antibody responses compared to the COVID-19 vaccine alone [15]. However, no differences in adverse reactions were observed [15], which is consistent with previous findings [16,33].

In our study, during the 0-7 days after vaccination, sore throat and palpitations were reported more frequently among those who received only the COVID-19 vaccine compared to those who received both vaccines simultaneously; other adverse reactions were similar between groups. However, this pattern was not observed in the 2024-2025 survey. It is important to note that spontaneous reporting of adverse events enables cross-sectional assessment of actual experiences, but may result in underreporting or missing data [40,41]. Therefore, caution is warranted in drawing causal conclusions. Enhanced investigation systems may facilitate more comprehensive lists of adverse reactions, allowing differences to be verified in the future.

The advantage of text analysis lies in its ability to reveal prominent topics within text data and to clarify semantic relationships and meanings. We observed that reporting patterns for adverse reactions after vaccination varied over time, with cold-related symptoms increasing as time progressed (Figure 2). This may provide insights into temporal variations in symptom presentation. Our results are consistent with previous studies indicating that the pattern of symptom occurrence varies over time [37]. Reporting patterns also showed that pain-related and cold-related adverse reactions were frequently reported together (Figure 3), suggesting co-occurrence among vaccinated individuals. While 1 study described co-occurring COVID-19 symptoms using social media data [42], there is a lack of research on co-occurring adverse events after COVID-19 vaccination. Our analysis of text-based, self-reported data enabled the identification of a range of health problems by organ system and may provide further insight into potential adverse reactions.

When interpreting the results of this study, in addition to the limitations noted in previous research [3], several further considerations should be taken into account. Because responses frequently contained mixed descriptions, local and systemic adverse reactions were not analyzed separately. This may be due to unclear wording of survey questions or variations in participant understanding. Additionally, because symptoms were not verified using clinical diagnostic criteria, there may be concerns regarding the accuracy of reported adverse reactions. As data were collected through self-report, issues such as underreporting and missing data are possible. Since only text responses were analyzed, adverse reactions captured through closed-ended questions may be underestimated. However, most respondents answered both the closed-ended questions and provided text responses, likely resulting in some redundancy. Despite extensive efforts in preprocessing for text analysis, the accuracy of results may have been affected by inaccurate or irrelevant responses, as well as reports reflecting underlying medical history rather than new adverse events. Furthermore, because the survey primarily included older adults and immunocompromised individuals prioritized by the national immunization plan, findings should be interpreted in the context of this specific population. Caution is also warranted, as results are limited to respondents who consented to text message-based surveillance, were not linked to electronic health records, and may not capture the full spectrum of adverse events, reflecting only short-term reactions within the first 7 days after vaccination.

Nevertheless, we were able to explore in detail potential adverse reactions beyond those previously documented by employing a text mining approach. Analysis of data from both survey seasons revealed consistent reporting of respiratory symptoms, sleep disturbances, lumbago, and indigestion, as well as a variety of eye, ear, and oral cavity-related adverse reactions following COVID-19 vaccination. These findings highlight the need for ongoing and vigilant monitoring of such reactions after vaccination. We also investigated differences in adverse reactions reported in text responses according to the coadministration of influenza vaccine. This approach provided a broader understanding of concurrent adverse reactions, as visualized through continuous phi-coefficient and co-occurrence analyses.

Timely public health responses to safeguard population health are especially crucial during a pandemic. Active surveillance systems for vaccine safety that communicate directly with vaccine recipients allow for the rapid detection of safety signals. However, these systems also face the challenge of identifying potentially hidden adverse reactions within free-text data. As such, it is necessary to develop automated systems for classifying text-based data, and the findings of this study may serve as a foundation for such efforts.

This study explored adverse reactions based on self-reported symptoms using text mining of data collected through an active surveillance system for COVID-19 vaccines. Notably, commonly identified adverse reactions included respiratory symptoms, sleep disturbances, lumbago, and indigestion. Our findings contribute to a deeper understanding of vaccine adverse reactions and provide valuable information for improving future adverse event data management and surveillance systems.

Supplementary materials

NOTES

Conflict of interest

The authors have no conflicts of interest to declare for this study.

Hyesook Park has been the associate editor of the Epidemiology and Health since 2021. She was not involved in the review process.

Funding

This study received financial support from the COVID-19 Vaccination Program (6400-6434-300, 6400-6434-303), which is administered by the Korea Disease Control and Prevention Agency.

Acknowledgements

We would like to express our sincere gratitude to all the participants who took part in the survey for this study.

Author contributions

Conceptualization: Park HS, Lee HA, Park B. Data curation: Kim CH, Kim Y, Park HJ, Jun S, Lee HL. Formal analysis: Lee HA, Park B, Kim CH, Kim Y. Funding acquisition: Park HS. Methodology: Lee HA, Park B. Project administration: Park HJ, Jun S, Lee HL, Kwon SL, Heo Y, Lee HM. Visualization: Lee HA, Park B, Kim CH, Kim Y. Writing – original draft: Lee HA, Park B. Writing – review & editing: Lee HA, Park B, Kim CH, Kin Y, Park HJ, Jun S, Lee HL, Kwon SL, Heo Y, Lee HM, Park HS.

Figure 1.

Differences in frequent terms for adverse reactions reported via text during the first 7 days following coronavirus disease 2019 (COVID-19) vaccination, according to with or without a concurrent influenza vaccine: (A) survey period 1 (October 19, 2023 to November 6, 2023) and (B) survey period 2 (October 11, 2024 to November 30, 2024). Gray background indicates newly identified adverse reactions via text analysis. *p<0.05.

Figure 2.

Percentages of daily reports of frequently reported terms during the first 7 days following coronavirus disease 2019 vaccination according to survey period: (A) survey period 1 (October 19, 2023 to November 6, 2023) and (B) survey period 2 (October 11, 2024 to November 30, 2024). The percentages were calculated using the frequency of reported terms, focusing on those who provided adverse reactions via text on each day.

Figure 3.

Co-occurrence network diagram for terms related to adverse reactions reported via text during the first 7 days following coronavirus disease 2019 vaccination according to survey period: (A) survey period 1 (October 19, 2023 to November 6, 2023) and (B) survey period 2 (October 11, 2024 to November 30, 2024). Node size indicates centrality and edge thickness indicates the frequency of term pairs. It is organized around key terms from pairs that have been reported more than 10 times.

REFERENCES

• 1. Genov G. Safety monitoring of COVID-19 vaccines: perspective from the European medicines agency. Clin Pharmacol Ther 2023;113:1223-1234. https://doi.org/10.1002/cpt.2828ArticlePubMedPMC
• 2. Yamaguchi T, Iwagami M, Ishiguro C, Fujii D, Yamamoto N, Narisawa M, et al. Safety monitoring of COVID-19 vaccines in Japan. Lancet Reg Health West Pac 2022;23:100442. https://doi.org/10.1016/j.lanwpc.2022.100442ArticlePubMedPMC
• 3. Myers TR, Marquez PL, Gee JM, Hause AM, Panagiotakopoulos L, Zhang B, et al. The v-safe after vaccination health checker: active vaccine safety monitoring during CDC’s COVID-19 pandemic response. Vaccine 2023;41:1310-1318. https://doi.org/10.1016/j.vaccine.2022.12.031ArticlePubMedPMC
• 4. Kim S, Hwang I, Ko M, Kwon Y, Lee YK. Safety monitoring of COVID-19 vaccination among adolescents aged 12 to 17 years old in the Republic of Korea. Osong Public Health Res Perspect 2022;13:230-237. https://doi.org/10.24171/j.phrp.2022.0122ArticlePubMedPMC
• 5. Lee YK, Kwon Y, Heo Y, Kim EK, Kim SY, Cho H, et al. Safety monitoring of COVID-19 vaccines: February 26, 2021, to June 4, 2022, Republic of Korea. Clin Exp Pediatr 2023;66:415-423. https://doi.org/10.3345/cep.2022.00815ArticlePubMedPMC
• 6. Grant R, Sacks JA, Abraham P, Chunsuttiwat S, Cohen C, Figueroa JP, et al. When to update COVID-19 vaccine composition. Nat Med 2023;29:776-780. https://doi.org/10.1038/s41591-023-02220-yArticlePubMed
• 7. Korea Disease Control and Prevention Agency (KDCA). Press release: announcement of COVID-19 vaccination plan for 23-24 season; 2023 [cited 2024 Mar 20]. Available from: https://www.kdca.go.kr/board/board.es?mid=a20501010000&bid=0015&list_no=723542&cg_code=&act=view&nPage=2 (Korean)
• 8. Korea Disease Control and Prevention Agency (KDCA). Press release: announcement of COVID-19 vaccination plan for 24-25 season; 2024 [cited 2025 Feb 7]. Available from: https://www.kdca.go.kr/board/board.es?mid=a20501010000&bid=0015&list_no=726095&act=view (Korean)
• 9. Kwon SL, Kim SY, Song M, Lee HM, Ban SH, Lee MS, et al. Assessing the determinants of influenza and COVID-19 vaccine co-administration decisions in the elderly. Hum Vaccin Immunother 2024;20:2346966. https://doi.org/10.1080/21645515.2024.2346966ArticlePubMedPMC
• 10. Bettinger JA, Sadarangani M, De Serres G, Valiquette L, Vanderkooi OG, Kellner JD, et al. The Canadian National Vaccine Safety Network: surveillance of adverse events following immunisation among individuals immunised with the COVID-19 vaccine, a cohort study in Canada. BMJ Open 2022;12:e051254. https://doi.org/10.1136/bmjopen-2021-051254ArticlePubMedPMC
• 11. Montalti M, Soldà G, Di Valerio Z, Salussolia A, Lenzi J, Forcellini M, et al. ROCCA observational study: early results on safety of Sputnik V vaccine (Gam-COVID-Vac) in the Republic of San Marino using active surveillance. EClinicalMedicine 2021;38:101027. https://doi.org/10.1016/j.eclinm.2021.101027ArticlePubMedPMC
• 12. Park B, Lee HA, Kim Y, Kim CH, Park H, Jun S, et al. Active surveillance for safety monitoring of XBB.1.5-containing COVID-19 mRNA vaccines in Korea. J Korean Med Sci 2024;39:e309. https://doi.org/10.3346/jkms.2024.39.e309ArticlePubMedPMC
• 13. Korea Disease Control and Prevention Agency (KDCA). Causality assessment of adverse events following COVID-19 vaccination; 2022 [cited 2024 Apr 9]. Available from: https://www.kovasc.com/bbs/book/show?sid=69 (Korean)
• 14. Back Y. Text-mining using R. 1st ed. . Paju: Hanul Akademi; 2017. (Korean).
• 15. Dulfer EA, Geckin B, Taks EJ, GeurtsvanKessel CH, Dijkstra H, van Emst L, et al. Timing and sequence of vaccination against COVID-19 and influenza (TACTIC): a single-blind, placebo-controlled randomized clinical trial. Lancet Reg Health Eur 2023;29:100628. https://doi.org/10.1016/j.lanepe.2023.100628ArticlePubMedPMC
• 16. Lazarus R, Baos S, Cappel-Porter H, Carson-Stevens A, Clout M, Culliford L, et al. Safety and immunogenicity of concomitant administration of COVID-19 vaccines (ChAdOx1 or BNT162b2) with seasonal influenza vaccines in adults in the UK (ComFluCOV): a multicentre, randomised, controlled, phase 4 trial. Lancet 2021;398:2277-2287. https://doi.org/10.1016/S0140-6736(21)02329-1ArticlePubMedPMC
• 17. Storey VC, O’Leary DE. Text analysis of evolving emotions and sentiments in COVID-19 Twitter communication. Cognit Comput 2022;1-24. https://doi.org/10.1007/s12559-022-10025-3ArticlePubMed
• 18. Pandey D, Pradhan B. Understanding COVID-19 response by twitter users: a text analysis approach. Heliyon 2022;8:e09994. https://doi.org/10.1016/j.heliyon.2022.e09994ArticlePubMedPMC
• 19. Jalil Z, Abbasi A, Javed AR, Badruddin Khan M, Abul Hasanat MH, Malik KM, et al. COVID-19 related sentiment analysis using state-of-the-art machine learning and deep learning techniques. Front Public Health 2022;9:812735. https://doi.org/10.3389/fpubh.2021.812735ArticlePubMedPMC
• 20. Jeon E, Oh S, Cho Y. Keyword analysis in Korean articles related to COVID-19 using topic modeling. Korean J Public Health 2022;59:1-11. (Korean). https://doi.org/10.17262/KJPH.2022.06.59.1.1Article
• 21. Kwon N, Oh J, Ha E. Topic and trends of public perception and sentiments of COVID-19 pandemic in South Korea: a text mining approach. Ewha Med J 2022;45:46-54. (Korean). https://doi.org/10.12771/emj.2022.45.2.46Article
• 22. Song SH. Issue analysis applying text mining on COVID-19 [dissertation]. Seoul: Seoul National University; 2022. (Korean).
• 23. Qorib M, Oladunni T, Denis M, Ososanya E, Cotae P. Covid-19 vaccine hesitancy: text mining, sentiment analysis and machine learning on COVID-19 vaccination Twitter dataset. Expert Syst Appl 2023;212:118715. https://doi.org/10.1016/j.eswa.2022.118715ArticlePubMedPMC
• 24. Shim JG, Ryu KH, Lee SH, Cho EA, Lee YJ, Ahn JH. Text mining approaches to analyze public sentiment changes regarding COVID-19 vaccines on social media in Korea. Int J Environ Res Public Health 2021;18:6549. https://doi.org/10.3390/ijerph18126549ArticlePubMedPMC
• 25. Kim J, Song M, Ahn S, Kwon SL. Twenty years of progress and the way forward: immunization registry information system in Korea. J Korean Med Sci 2024;39:e119. https://doi.org/10.3346/jkms.2024.39.e119ArticlePubMedPMC
• 26. Korea Disease Control and Prevention Agency (KDCA). Report on adverse events following COVID-19 vaccination (week 216); 2025 [cited 2025 May 8]. Available from: https://dportal.kdca.go.kr/pot/bbs/BD_selectBbs.do?q_bbsSn=1018&q_bbsDocNo=20250424172438878&q_clsfNo=-1 (Korean)
• 27. Hacking C, Verbeek H, Hamers JP, Sion K, Aarts S. Text mining in long-term care: exploring the usefulness of artificial intelligence in a nursing home setting. PLoS One 2022;17:e0268281. https://doi.org/10.1371/journal.pone.0268281ArticlePubMedPMC
• 28. Ministry of Food and Drug Safety. General consideration on labeling for drug products - content and format; 2022 [cited 2024 Mar 3]. Available from https://www.mfds.go.kr/brd/m_1059/view.do?seq=14898&srchFr=&srchTo=&srchWord=MaPP&srchTp=0&itm_seq_1=0&itm_seq_2=0&multi_itm_seq=0&company_cd=&company_nm=&Data_stts_gubun=C9999&page=3 (Korean)
• 29. Lee SS. Network analysis methods. Seoul: Nonhyeong; 2012. p. 207-268 (Korean).
• 30. Zhang Z. Text mining for social and behavioral research using R: a case study on teaching evaluation; 2018 [cited 2024 Mar 20]. Available from: https://books.psychstat.org/textmining
• 31. Rahmani K, Shavaleh R, Forouhi M, Disfani HF, Kamandi M, Oskooi RK, et al. The effectiveness of COVID-19 vaccines in reducing the incidence, hospitalization, and mortality from COVID-19: a systematic review and meta-analysis. Front Public Health 2022;10:873596. https://doi.org/10.3389/fpubh.2022.873596ArticlePubMedPMC
• 32. Park C, Chung C. Text network analysis: sharing meanings among stakeholders of the policy through a socio-cognitive network analysis. J Gov Stud 2013;19:73-108. (Korean).
• 33. Moro PL, Zhang B, Ennulat C, Harris M, McVey R, Woody G, et al. Safety of co-administration of mRNA COVID-19 and seasonal inactivated influenza vaccines in the vaccine adverse event reporting system (VAERS) during July 1, 2021-June 30, 2022. Vaccine 2023;41:1859-1863. https://doi.org/10.1016/j.vaccine.2022.12.069ArticlePubMedPMC
• 34. Ripabelli G, Tamburro M, Buccieri N, Adesso C, Caggiano V, Cannizzaro F, et al. Active surveillance of adverse events in healthcare workers recipients after vaccination with COVID-19 BNT162b2 vaccine (Pfizer-BioNTech, Comirnaty): a cross-sectional study. J Community Health 2022;47:211-225. https://doi.org/10.1007/s10900-021-01039-3ArticlePubMedPMC
• 35. Hu M, Guo W, Liu L, Yang Y, Xu Q, Cheng F, et al. Safety monitoring in inactivated COVID-19 vaccines by clinical pharmacists from a single center in China. Front Immunol 2022;13:882919. https://doi.org/10.3389/fimmu.2022.882919ArticlePubMedPMC
• 36. Hause AM, Zhang B, Yue X, Marquez P, Myers TR, Parker C, et al. Reactogenicity of simultaneous COVID-19 mRNA booster and influenza vaccination in the US. JAMA Netw Open 2022;5:e2222241. https://doi.org/10.1001/jamanetworkopen.2022.22241ArticlePubMedPMC
• 37. Canas LS, Österdahl MF, Deng J, Hu C, Selvachandran S, Polidori L, et al. Disentangling post-vaccination symptoms from early COVID-19. EClinicalMedicine 2021;42:101212. https://doi.org/10.1016/j.eclinm.2021.101212ArticlePubMedPMC
• 38. Centers for Disease Control and Prevention (CDC). Interim clinical considerations for use of COVID-19 vaccines in the United States [cited 2024 Mar 3]. Available from: https://www.cdc.gov/covid/hcp/vaccine-considerations/index.html?CDC_AA_refVal=https%3A%2F%2Fwww.cdc.gov%2Fvaccines%2Fcovid-19%2Fclinical-considerations%2Finterim-considerations-us.html
• 39. Centers for Disease Control and Prevention (CDC). Getting a flu vaccine and other recommended vaccines at the same time [cited 2024 Mar 3]. Available from: https://www.cdc.gov/flu/vaccines/coadministration.html?CDC_AAref_Val=https://www.cdc.gov/flu/prevent/coadministration.htm
• 40. Vaccine Adverse Event Reporting System (VAERS). Guide to interpreting VAERS data; 2021 [cited 2025 Feb 7]. Available from: https://vaers.hhs.gov/data/dataguide.html
• 41. European Medicines Agency. EudraVigilance - European database of suspected adverse drug reaction reports; 2021 [cited 2025 Feb 7]. Available from: https://www.adrreports.eu/en/
• 42. Wu J, Wang L, Hua Y, Li M, Zhou L, Bates DW, et al. Trend and co-occurrence network of COVID-19 symptoms from large-scale social media data: infoveillance study. J Med Internet Res 2023;25:e45419. https://doi.org/10.2196/45419ArticlePubMedPMC

Figure & Data

References

Citations

Citations to this article as recorded by