Worldwide Index of Serotype-Specific Pneumococcal Antibody Responses (WISSPAR): A curated database of clinical trial data

Introduction

Pneumococcal disease, caused by Streptococcus pneumoniae, includes common, noninvasive infections such as otitis media, and severe, life-threatening infections including pneumonia, sepsis and meningitis Pneumococcal disease causes hundreds of thousands of deaths per year worldwide, most of these occurring in children, older adults, and people with immunocompromising conditions. The burden of pneumococcal disease is greatest in low- and middle- income countries, but there are high burdens of disease in infants and older adults in both high- and low-income settings^1,2.

Pneumococcal conjugate vaccines (PCVs) have been shown to have high efficacy against disease and mortality³. Vaccines targeting different serotypes and used with various dosing schedules have been employed in a wide array of populations globally^4,5. With several new pneumococcal vaccines recently licensed and in development, there is a need to compare PCVs and ensure maintenance of vaccine effectiveness^6–9. New PCVs are typically licensed based on comparisons of immunogenicity with existing vaccines. Therefore, there is a need to be able to readily compare the immunogenicity of different PCVs across studies. These immunogenicity data are used by regulators and by national immunization technical advisory groups (NITAGs) to determine which vaccine products, dosing schedules, and age groups will be included in immunization policy recommendations for their respective populations.

Immunogenicity data for pneumococcal vaccine trials should be compared and interpreted with caution, and the web interface has prominent statements about the caveats of the data. Data on immunogenicity alone cannot be used to infer differences in effectiveness between vaccines. These data need to be combined with information on the protective concentration of antibodies required to protect against each serotype in different populations for meaningful comparisons¹⁰. Despite these cautions, pulling immunogenicity data together into a single location will make this necessary process of comparison easier and less labor-intensive.

Review and synthesis of the data to inform these policy decisions can be labor intensive and cumbersome. Streamlining this process and making immunogenicity data available in a format that can be readily searched and extracted has value for policy makers and for researchers interested in vaccine efficacy, effectiveness, impact, and optimization. WISSPAR is a resource that provides an interactive platform to summarize and compare the immunogenicity data from clinical trials of pneumococcal vaccines.

Methods

Dashboard contents

Public-facing WISSPAR content is organized into Data Dashboards, Analysis Tools, and Resources (Table 1). The Data Dashboards provide customizable visualizations of the serotype-specific immunogenicity data in the Graphical Overview (Figure 1), and summary information from the included trials in the Clinical Trials Overview. The analysis tools allow users to efficiently pinpoint trials of interest in the Look up a trial tool, and export custom .csv files for downstream analyses using the Immunogenicity Data Export tool. There are community resources available, including a blog to keep users abreast of updates and demonstrate WISSPAR capabilities example analyses. An example of how WISSPAR can be used to make comparisons of serotype-specific differences in immunogenicity between three vaccine products is published on the WISSPAR blog¹¹.

Table 1. Descriptions of page content for the Worldwide Index of Serotype-Specific Antibody Responses, WISSPAR.

WISSPAR has a suite of tools to summarize, compare, and export immunogenicity data from clinical trials of pneumococcal vaccines.

WISSPAR Website Pages	Content Summary
PUBLIC DATA DASHBOARDS
Graphical View	An interactive Shiny app that allows users to visualize measurements of serotype-specific immunogenicity and export images. Available variables include:   •  Vaccine   •  Serotypes   •  Age category   •  Schedule   •  Doses received and timing   •  Trial phase   •  Sponsor Users can visualize measurements of IgG Geometric Mean Concentrations (GMCs), Opsonophagocytic Activity assays (OPAs), and GMC and OPA ratios.
Clinical Trials Overview	Filterable listing of clinical trials included on WISSPAR, allowing users to search for particular clinical trials or criteria of interest, including:   •  Vaccines   •  Schedule   •  Manufacturers   •  Age Category   •  Ethnicity   •  Continent   •  Immunocompromised Groups   •  Gender
ANALYSIS TOOLS
Look up a trial	Quick-search function that allows users to access complete trial information by entering the NCT ID from clinicaltrials.gov
Immunogenicity Data Export	Highly customizable data export tool where users can create customized datasets for downstream analyses. Variables include:
	Filters	Fields
	•  Income group   •  Trial phase   •  Vaccine   •  Age category   •  Immunocompromised group   •  Responsible party	•  Clinical trial data   •  Study eligibility   •  Study location   •  Outcome Measures   •  Outcome Overview
RESOURCES
Blog	Link to WISSPAR blog posts, which includes updates from the team, examples of how to use WISSPAR to compare trial data, and more.
Videos	Video walk-through of WISSPAR features and functions

Figure 1. Data visualization app on the Worldwide Index of Serotype-Specific Antibody Responses, WISSPAR.

WISSPAR’s data visualization app is based in R Shiny and allows users to generate customizable, downloadable figures showing immunogenicity measurements stratified by serotype, vaccine product, vaccine manufacturer, trial sponsor, and vaccination schedule.

Data hosting and database structure

The WISSPAR platform consists of a few different services hosted on Digital Ocean (Figure 2). The backend database is PostgreSQL (V13) with connection pools set up for security and guaranteed availability; there are also automated backups for recovery. The web application is written in Golang (1.16) for the backend connecting to the database and integrating with the clinicaltrials.gov website. On the front end of the web application, the javascript framework is SvelteJS to allow for quick component loading times and better code organization over traditional javascript. We used GitLab for our software version control, when a change is made in the repository on GitLab it is detected by our Digital Ocean repository and automatically deployed. We also have fail safes built in so that if a code build fails to deploy it will revert to the last working version without any downtime to the web application. Interactive plots, with filters and selectors, was created in RShiny and deployed through the shiny.io server.

Figure 2. Organizational Schematic of the Worldwide Index of Serotype-Specific Antibody Responses, WISSPAR.

The data import process from clinicaltrials.gov and the sources of data and the data types used in WISSPAR are described here.