Vaxafe Tutorial

Tutorial of Vaxafe VAE Query and Analysis

This page provides a tutorial for using the features in the Vaxafe website:

The Vaxafe vaccine safety web page includes adverse event data downloaded from the CDC/FDA VAE Reporting System (VAERS). The downloaded data is stored in our customized relational database and processed for users to query and perform statistical data analysis.

VAE Case Query

The VAE case Query generates the information about a specific case of vaccine adverse event (VAE). For example, you can add a VAERS case number to the program and click "Submit":

The following is an example web page with query results:

Conditional VAE Web Query

The Conditional VAE Web Query allows you to search VAE cases based on a specified set of parameters. Users can filter adverse events by vaccine name, vaccination year, US state or territory, etc.:

The following is an example web page with query results:

Statistical data analysis

We can also perform statistical data analysis on the VAE data to identify whether an adverse event is statistically significantly associated with a vaccine. Specifically, you can first choose an adverse event and then choose a vaccine, and then click "Run Test":

The following screenshot shows a typical result page out of the statistical data analysis:

Note: We are in the process of developing more and better visualization of the VAE case report details

Note: Our statistical analysis includes four criteria: (i) the minimal number of case reports is at least 3 and > 0.2% of the total reports for the vaccine, (ii) proportional reporting ratio (PRR) >2, (iii) the Chi-square score >4, and (iv) EBGM score > 2. See references:

Reference for our statistical data analysis pipeline (note: not including EBGM): Sarntivijai S, Xiang Z, Shedden KA, Markel H, Omenn GS, Athey BD, and He Y. Ontology-based combinatorial comparative analysis of adverse events associated with killed and live influenza vaccines. PLoS ONE. 2012, 7(11): e49941. doi:10.1371/journal.pone.0049941. PMID: 23209624. PMCID: PMC3509157.
Reference for the PRR and Chi-square calculations used in our program: Xie J, He Y. Ontology-based vaccine adverse event representation and analysis. Adv Exp Med Biol. 2017;1028:89-103. PMID: 29058218.
PRR in Wiki: https://en.wikipedia.org/wiki/Proportional_reporting_ratio
Chi-squared test in Wiki: https://en.wikipedia.org/wiki/Chi-squared_test
Yates' correctoin of Chi-squared test in Wiki: https://en.wikipedia.org/wiki/Yates%27s_correction_for_continuity
Reference for Chi-square tests: https://www.bmj.com/about-bmj/resources-readers/publications/statistics-square-one/8-chi-squared-tests (Note: We use the shortcut formula introduced in this document, which is also described in the above paper (Xie and He, 2017).
Note: In our analysis, we do not use Yates' correction. The reason is: As described in the above bmj.com paper, "Although there is no precise rule defining the circumstances in which to use Yates’ correction, a common practice is to incorporate it into χ² calculations on tables with a total of under 100 or with any cell containing a value less than 10." Since our VAERS data size is typically big, and we also have the minimal case report cutoff criterion, the requirement for using the correction does not apply to our case. Therefore, we do not use the correction.
References for EBGM:
- https://cran.r-project.org/web/packages/openEBGM/.
- William Dumouchel.Bayesian Data Mining in Large Frequency Tables, with an Application to the FDA Spontaneous Reporting System. Pages 177-190 | Received 01 Jul 1998, Published online: 17 Feb 2012. doi:10.1080/00031305.1999.10474456.
- Package ‘openEBGM’: John Ihrie. EBGM Disproportionality Scores for Adverse Event Data Mining. https://cran.r-project.org/web/packages/openEBGM/openEBGM.pdf.
Cov19VaxKB paper: Philip C. Huang*, Rohit Goru*, Anthony Huffman, Asiyah Yu Lin, Michael F. Cooke, Yongqun He. Cov19VaxKB: A Web-based Integrative COVID-19 Vaccine Knowledge Base. Vaccine: X. Volume 10, April 2022, 100139. https://doi.org/10.1016/j.jvacx.2021.100139.

DISCLAIMER: Before using this website, please ensure that you have read, understand, and agree on the VAERS disclaimer shown at: https://vaers.hhs.gov/data.html. As described in the indicated CDC/FDA VAERS Disclaimer, the VAERS data largely came from voluntary reporting. It has the advantage of nation-wide reporting and can provide some early warning of potential safety issue(s) with a vaccine. However, the VAERS "Reports may include incomplete, inaccurate, coincidental and unverified information". In addition, the VAERS data alone cannot be used to determine the causality of the identified vaccine adverse event. Instead of being causal, it only shows statistical association.

Questions? Please feel free to contact us.