Twitter is a social networking service, founded in 2006, and now with over 400 million monthly active users. This study was designed to assess the utility of Twitter as a tool for the assessment and comparison of healthcare systems.

Java and R programs were used to capture, analyze, and compare Tweets posted about one academic university hospital system and two community hospital systems in a large metro area in the Southeastern United States throughout 2016. Natural language processing with sentimental analysis was used in conjunction with an object clustering algorithm to categorize and quantify positive versus negative posts regarding care received at these facilities. An analysis of the positive and negative tweets for the academic facility was completed.

Among the 12,579 Tweets captured for 2016, 89.4% were associated with the academic medical center, 9.7% with community hospital #1, and 0.9% with community hospital #2. Relative to annual patient visits, the use of Twitter was 0.31% for the academic hospital, 0.34% for community hospital #1, and 0.03% for community hospital #2. A majority of Tweets were positive for all three hospital systems. Hospital system #2 received the highest percentage of positive and the lowest percentage of negative tweets (Table 1). The subject of positive comments concerning the academic facility were hospital staff, hospital facilities, and hospital services, while the negative comments focused on appointments, wait time, and emergency room visits.

This study demonstrates the utility of Twitter to determine the overall sentiment of a segment of users of competing healthcare systems. In our initial analysis, we were able to compare the largest competitors of our academic system and to assess the topics of the positive and negative comments made by our healthcare system users. Further study and development of computing techniques are needed to fully maximize the potential this tool offers.