Twitter Sentiment at the Hospital and Patient Level as a Measure of Pediatric Patient Experience

Abstract

Although pediatric cancer survival rates have improved, cancer is still the leading cause of death by disease among children in the United States. With nearly 16,000 new diagnoses each year of children having an average age of six years old, there is still much to be done to improve survival rates among children diagnosed with cancer. Additionally, clinical outcomes can be greatly improved by first understanding the patient perspective and, consequently, there is a need to understand the pediatric cancer patient experience. Interest in the analysis of patient-reported outcome (PRO) measures, including those specific to patient experience, has grown in recent years as attempts are made to understand and assess latent (i.e., unobserved) traits such as quality of life. While most traditional PRO measures involve lengthy and costly surveys that are limited in scope, posts on social media platforms are frequently lauded as more authentic, unbiased measures of patient experience. While many previous studies have utilized Twitter as a source for patient experience data that is not easily captured by traditional PROs, very few studies have examined the use of Twitter data to explore and better understand the experience of pediatric cancer patients specifically. Lexicon-based sentiment analysis of Twitter data was used both to examine the pediatric cancer patient experience as well as to compare the attitudes, perceptions, and overall impressions of individuals with St. Jude experiences to those with more general experiences. Results indicated that patients reflected on their care with more relative negativity when speaking of experiences other than St. Jude. While there were some notable differences between the two comparison groups, the sentiment of both was still mostly positive. Pediatric oncology differs greatly from adult oncological care and, thus, the proposed use of sentiment analysis with patient social media posts serving as authentic PRO data likely differentiates more between medical providers and clinical care settings when assessing adult patient experience. Despite limitations, this exploratory study suggests that Twitter can be utilized as a source for authentic, real-time patient experience data that will improve overall care and better clinical outcomes, even among the pediatric cancer patient population.