Abstract
Background: Trillium Health Partners (THP), one of Canada’s largest leading academically-affiliated health centres in Ontario, Canada, recognized a critical gap in patient experience (PE) data collection methods. Traditionally reliant on paper-based and telephone surveys, THP's approach was increasingly seen as insufficient in capturing the full diversity of patient feedback. This limitation was particularly significant within the context of integrated care, where timely, comprehensive, and accurate patient feedback is essential for informed decision-making and continuous quality improvement. As healthcare systems globally move towards more integrated models of care, the ability to gather diverse and actionable PE data becomes even more crucial. Without it, hospitals risk missing vital insights that could enhance patient care.
Approach: THP embarked on a project to overhaul its PE data collection system. The goal was to develop a more efficient and inclusive approach to capture a broader range of patient experiences and facilitate rapid quality improvements across different care settings. The new system integrated multiple data collection formats, including email, QR codes, and tablets, making it easier for patients to provide feedback. The transition to this new system resulted in a substantial increase in response rates while also expanding the scope and diversity of the feedback collected. Using PE survey data collected and stored on Qualtrics, we conducted a pilot study to develop an approach to mining PE data for rapid quality improvements. This approach leveraged advanced tools such as the Qualtrics dashboard, StatsIQ for statistical analysis, TextIQ for qualitative analysis via natural language processing, and N-Vivo for open-ended response analysis.
Results: The pilot study analysis focused on four medicine units—two with the highest and two with the lowest overall PE scores based on patient ratings of their overall experience on a 0-10 scale. The team identified critical enablers and barriers to PE, such as emotional support (fears, anxieties, worries) and engagement for units with high and low overall PE scores. Regression models were created for each medicine unit to provide more nuanced insight into the drivers of PE in each unit. Analysis of open-ended responses revealed that positive experiences were related to effective communication and humane treatment, while negative experiences were due to a lack of communication and operational inefficiencies. This presentation will discuss how the PE data mining approach was developed and the findings of this pilot on promoting collaborative, rapid, data-driven quality improvement.
Implications: The lessons learned from this pilot study underscore the importance of leveraging PE data (both closed and open-ended text data) for rapid and evidence-based quality improvement. We developed a PE data mining approach to learn from a broad range of patient experiences and facilitate rapid, actionable insights that could be immediately applied to improve care. The following steps involve refining the tools, protocols, and processes developed during this project to guide rapid quality improvement efforts across other clinical programs and functions.