Rutgers researchers built a novel multimodal survey platform to increase accuracy in multilingual data collection and reduce selection bias.

A group of researchers, including Dinesh Mendhe, Stephanie Bergren, and XinQi Dong from the Institute for Health, Health Care Policy and Aging Research, have developed a unique survey platform that aims to enhance the precision of multilingual data collection and minimize selection bias. The new research data collection platform, called ADHI, is equipped with various features such as scheduling, actigraphy management, phlebotomy management, participant management, consent management, contact management, family tree architecture, gift card management, and reporting, which facilitate a more efficient and comprehensive data collection process.

Given the ongoing COVID-19 pandemic, secure and distanced data collection platforms are critical for reaching vulnerable populations. Commonly used electronic data collection systems lack a myriad of critical features, including a modern technology stack, new data encryption and security standards, study workflows, and reporting algorithms. Moreover, these systems do not have multilingual mapping functionalities of survey and consent forms. All of these components ultimately increase selection bias while simultaneously reducing the security and quality of the response data. In order to directly address the aforementioned issues, the IFH team has developed a multilingual and highly secure data management platform. ADHI survey is built using stable, tested, and modular programming frameworks and design patterns targeted at accommodating intricately complex structures of polyglot mapping, a large volume of data, encryption, and granular user authorization. The statistical accuracy, along with the multilingual mapping, is the core highlight of this system. The multilingual function of this platform has the ability to eliminate selection biases while creating a well-balanced cross-section of society. Modern survey design workflows and validation checks ultimately prevent data loss and help reduce data collection errors. The platform design was initiated on April 1, 2020, and has been pilot tested for use in multilingual populations. The currently active application version of the system is capable of supporting in-person and telephone interviews, emailing and SMS survey links to every registered participant, building family tree architecture, online consent management, actigraphy management, phlebotomy management, gift card management, contact management and scheduling, participant management, and reporting. This platform also has built-in report functionality. Additional features are being explored to improve the study coordination and monitoring.

Features:

Modular and easily customizable components.
Dynamic one-to-many language mappings for surveys, consents, contact forms, actigraphy, and saliva protocol.
Phlebotomy management.
Saliva collection protocol.
Inbuilt traversable family tree structure for participants, families, and households.
An extensive set of ready-to-use validated Instruments.
Nuanced control over roles and permissions for greater data security and management.
Multilingual consent management.
Actigraphy management.
Gift Card management.
Participant contact history, interview scheduling, and follow-up management.
Easy to view and manage active and inactive participants.
Ability to handle large-size research studies.
User-friendly workflows that simplify complex processes.
Easy to navigate parent and child questionnaires and their relationship.
No programming experience is required.
Built-in functionality for variable coding.
Staff management and quality control protocols.
Faster rendering, reusable components, and virtual DOM (Document object model)
Multisite deployment.
Integrated email and SMS components.

Backups & Storage:

Optional soft-delete functionality is available.
Data is backed up every night. Both full backup and incremental backup options are available.
Ability to keep track of all the changes.

Data consistency & availability:

The application repository layer, service layer, and every database transaction follow the ACID (Atomicity, Consistency, Isolation, and Durability) principle.

Exporting Data & Data Analysis:

Data can be exported in different formats for data analysis, such as CSV, JSON, XML, TSV, Excel, and SQL.
Multiple reports are available within the platform to view participant details, interview details, incompletion reports, completion reports, and other details.

API:

ADHI architecture provides the API (Application programming interface) for integration, access, and retrieval of data without using a view layer/web application.
Loosely coupled components to increase future integrations and rapid development.