Unsupervised Wisdom: Explore Medical Narratives on Older Adult Falls

Use unsupervised machine learning approaches to extract insights from emergency department narratives about how, when, and why older adults (age 65+) fall. Competition hosted by Centers for Disease Control and Prevention. #health

$70,000 in prizes
3 days left
642 joined

Colorful zoomed-in photo of three people standing side by side and holding hands with crossed arms.

Credit: CDC/NCIPC

Overview

Falls among adults 65 and older is the leading cause of injury-related deaths. Falls can also result in serious injuries to the head and/or broken bones. Some risk factors associated with falls can be reduced through appropriate interventions like treating vision problems, exercising for strength and balance, and removing tripping hazards in your home.

Medical record narratives are a rich yet under-explored source of potential insights about how, when, and why people fall. However, narrative data sources can be difficult to work with, often requiring carefully designed, time-intensive manual coding procedures. Modern machine learning approaches to working with narrative data have the potential to effectively extract insights about older adult falls from narrative medical record data at scale.

The goal in this challenge is to identify effective methods of using unsupervised machine learning to extract insights about older adult falls from emergency department narratives. Insights extracted from medical record narratives can potentially inform interventions for reducing falls.


Prizes and Timeline

Competition End Date:

Oct. 6, 2023, 11:59 p.m. UTC

Place Prize Amount
1st $20,000
2nd $15,000
3rd $10,000
4th $5,000

Bonus prizes of $2,500 each will be awarded for:

  • Most novel approach
  • Most compelling insight
  • Best visualization
  • Best null result*
  • Most helpful shared code
  • Most promising mid-point submissions (up to three awarded)

* Sometimes knowing what didn’t work is just as useful as knowing what did. Please feel free to include null results in your submission. However, if you choose to do this, you must include a hypothesis as to why this was a reasonable exploration, with either a (1) citation of previous academic literature to show where it comes from, or (2) clear, well-defined mechanistic explanation.


Note on prize eligibility: The term Competition Sponsor in the Competition Rules includes the Centers for Disease Control and Prevention. Federal employees acting within the scope of their employment and federally-funded researchers acting within the scope of their funding are not eligible to win a prize in this challenge.


Midpoint Feedback Deadline August 23, 2023 at 11:59:59 PM UTC
Submission Deadline October 6, 2023 at 11:59:59 PM UTC

How to compete

  1. Click the "Compete!" button in the sidebar to enroll in the competition.
  2. Get familiar with the problem through the overview and problem description. You might also want to reference additional resources available on the about page.
  3. Download the data from the data tab.
  4. Explore the data and extract insights.
  5. Share useful code or find code shared by others in the Community Code section.
  6. Consolidate your code into a single notebook and draft an executive summary that highlights your findings and methods. See the submission format section for more details.
  7. (Optional) Click "Midpoint submissions" in the sidebar and upload your executive summary for mid-point feedback by August 23, 2023.
  8. Post questions you have about the challenge to the challenge forum. Questions submitted before August 31, 2023 will be answered at a Q&A event with DrivenData and CDC/NCIPC.
  9. Click “Final submissions” in the sidebar, and then “Make new submission”. You’re in! Have a new technique to add to your solution? Just remove your submission and re-submit by the final deadline.

Challenge sponsored by the Centers for Disease Control and Prevention - National Center for Injury Prevention and Control.

US CDC Logo