Overview
National Alzheimer's Coordinating Center (NACC) envisions a more streamlined and efficient ADRCs data accessing process...
There are a total of 37 Alzheimer’s Disease Research Centers (ADRCs) funded by NIH (National Institutes of Health) across 26 states in the United States. The ADRCs first independently collect various data types and then deposit data to multiple data centers, including NCRAD, ADGC, NIAGADS, SCAN, and NACC (see below for the acronyms table).
In this project, NACC is our client and their need was to create a self-service Data Front Door (DFD) which connects data from all data centers and makes ADRC data more interoperable, harmonized, and fair for all users.
Our goal was to assist NACC to understand user’s need and their current pain points to provide an essential foundation for future development.
My Role
I worked with one nursing Ph.D. student and one biomedical informatics Ph.D. student to conduct research throughout the 10-week project. We collaborated with National Alzheimer's Coordinating Center to construct a REDCap survey at the Spring 2022 Alzheimer's Disease Genetic Consortium conference to understand the data access experiences of different stakeholders. We also conducted user interviews to explore user pain points in depth. My team’s work provided a foundation and directions for future full-scale design requirements investigation.
Problem
Currently, data users who need to access data from ADRCs must go to multiple data center sites separately to check the availability of data and request data access repeatedly.
The three main issues that users have are
1
Inefficient
Researchers and others who need data have to spend time looking for the right data on multiple websites and contact or log into multiple sites to access data.
2
Inconvenient
Researchers use multimodal data to conduct research, they need to retrieve different types of data from different sites before they can get a sense of the cohort size.
3
Unprocessed
This is some text inside of a div block.Too raw for people who only want to get some metadata, statistics, and data visualized.
Why is it important to have a data front door?
A well-designed data front door that provides clear and simple ways for users to achieve their goals will facilitate easy use of ADRC's data by researchers. It supports researchers in conducting impactful research and advancing the field.
Research and survey
To further understand how users experience the system. We conducted three semi-structured interviews with a graduate student, an associate professor, and a senior professor.
The purpose of the interviews was to understand the users' struggles and desires so that we could propose appropriate solutions and designs for the data front door.
Interview
During the interview, we will focus on three aspects:
Use cases (e.g., how interviewees requested data from the three data centers)
Current pain points ( e.g., difficulties in retrieving data or viewing data)
Needs and wants ( e.g., interviewees’ desired process for accessing data)
User Persona
We consolidated the interviewees' insights and information into three personas. These three personas summarize interviewees' wants and needs, frustrations, and platform evaluations.
Below are the main barriers and needs that we consolidated from all the interviews and put them in a chart.
Survey
At the 2022 Spring ADRC meeting, NACC set up a booth at the conference to distribute the REDCap survey to participants. The survey captures information regarding people’s experience with NACC, NIAGADS, and NCRAD, ease of use rating, ideal process of retrieving data, barriers and etc.
User Needs and Feature Recommendations from the survey
1. Single-access portal
The survey asked about participants’ perceptions of the usefulness of developing a single portal for users to search, visualize, and access NACC, NCRAD, and NIAGADS.
Users agreed that having a single portal would be very and extremely useful for their work.
2. Interested data stream and features
The survey asked about participants’ level of interest in connecting additional data streams using NACC IDs.
Among the 6 participants, most of them were interested in having Digital Biomarkers and Digital Neuropathology data streams.
3. Data harmonization and integration
Multiple participants mentioned data integration across several questions.
Models
In addition to interviews and surveys, we experienced the data access process on different platforms ourselves.
We summarized our own experience with the information from the research and created two models: a use case diagram and a flow chart.These two models to demonstrate people's frustrations and current pain points.
Use Case Diagram
The "Search Data" use case links to different types of datasets (imaging, genetic, biospecimen, longitudinal clinical, and neuropathology).These datasets are then linked to NACC, NIAGAD, and NCRAD.
It illustrates that when users are browsing data and need to be multimodal, they may need to log into multiple sites to get the data they need.
flowchart
This diagram illustrates the process of browsing the data and the decision-making of users.
If users need multi-modal, they need to consider which data they want to use and check which data centers contain that type of data. Users need to access it separately and do the same process of requesting data and downloading data for each data center.
From this flowchart, we can see that there are many redundancies involved in the process, and that is what we are trying to solve in this project.
Overall Design suggestions and Potential Barriers
Combining insights from interviews and surveys, we collated four features in the future DFD prototype
A unified, single-access data interface
Data integration and harmonization could be the potential barriers to achieving a unified data portal.
Self-service data access
There are voices from both interviewees and survey takers requesting self-service data access in the DFD feature. It will increase the efficiency of data retrieval and ensure the users have the most updated data.
However, the potential barrier to this feature is data privacy concerns. NACC has to ensure that the users will not be able to trace back to the individual patients or data providers.
Data integration with a unique identifier
Three of the clinician researchers in the survey complained about the inconsistencies across data in ARDCs and NACC. They suggested DFD consider adding a globally unique identifier (Global Unique ID or NACC unique ID) to allow users to better identify their data and sample information.
A potential barrier to this feature is that the NACC and the other data centers do not have enough data scientists or experts to perform data harmonization.
Metadata reporting and Cloud Analysis
Both interviewees and survey participants stated that the DFD should allow users to probe the data (quality, time span, etc) before downloading. There could also be an option to run a simple initial analysis (e.g. quality checks, simple statistics) on Cloud before downloading.
Next Steps
After reporting our user analysis results to our client, we summarized five plans for the DFD project:
Revision of the survey
There are some unmatch responses in the RedCap survey that might affect the quality of the data collected. When asked about the ease of use across different databases, several participants responded “0” to the question.
Implication: we need to RedCap survey to allow participants to have options on: “no experience using this database”.
Recruit more participants to take the RedCap survey
We have limited sample size (n=6) which provided constructive responses to the project. Three of the nine participants who took the survey did not finish the questionnaire or did not provide valid responses.
Implication: it is crucial for the NACC to recruit more participants (data users, researchers, clinicians) to take the survey.
Allow Graduate student perspective
In two of our interviews, the interviewees mentioned that they have limited experiences in data application process since they usually asked their graduate students to run the data request process.
Implication: we considered it might be useful to recruit some graduate students’ perspectives since they are more familiar with the data application process.
Focus group
NACC is planning to conduct some focus groups after gathering more insights from the survey.
Implication: Interview questions and discussion topics should be drafted after a deeper dive into the RedCap survey.
Minimum Viable product (MVP) vs. Ideal version:
NACC has stressed the importance of collating the minimum viable product versus the idea version for the prototype.
Implication: we should be asking the interviewees to prioritize/or card sort different features for the DFD prototype in the focus groups.
.jpg)
