Extracting Actionable Medical Data from a Twitter User’s History During a Medical Emergency

Langdon, Theodore

Document

Extracting Actionable Medical Data from a Twitter User’s History During a Medical Emergency 开放存取 Deposited

Analytics

下载PDF文件

下载 Adobe Acrobat Reader

Date Uploaded: 08/19/2023
Date Modified: 08/19/2023

As technology permeates day-to-day life, people have more and more ways to communicate. These forms of communication create a challenge for Public Safety Answering Points (PSAP). Research has shown people are posting on social media for medical help, but PSAPs do not have a way to receive these messages. This research aims to determine if using keywords and filter words can be used to find the actionable calls for help in the midst of the millions of posts made. Actionable is defined as containing enough information to determine the nature of a medical emergency and if is currently occurring or is recent enough that the poster needs help. To determine if this was true, the most prevalent types of voice medical calls to the Cincinnati Fire Department were determined for 2021. A set of keywords and filter words was created for each call type. Then tweets were captured over a period of seventeen hours and filtered using the word lists. The filtering showed there was a valid way to find actionable tweets, and that people were posting such things. By varying the word lists, the signal-to-noise ratio can be adjusted depending on the desires of the agency. as filtering became more strict, the number of missed actionable tweets increased, while the number of incorrectly labeled as actionable decreased.

创建者