Background

​Every week, our Patient Access leadership team was tasked with identifying actionable insights from thousands of patient survey comments — but doing so required hours of manual review. Supervisors had to sift through large volumes of free-text data just to find a handful of comments relevant to our department, often spending significant time filtering out unrelated feedback. Each campus had its own approach, resulting in inconsistent reviews, duplicated effort, and a lack of centralized visibility. The manual nature of the process also made it difficult to determine whether concerns were isolated incidents or widespread trends requiring systemic change.

Challenges

We needed a way to standardize, scale, and accelerate how we processed patient feedback — particularly for Patient Access–specific issues like registration, check-in, scheduling, insurance processing, and first-touch experiences. Without a consistent, data-driven framework, insights were anecdotal at best, and decisions around process improvements lacked the backing of real-time, department-specific sentiment. In short, we had the data — but not the infrastructure to learn from it efficiently or meaningfully.

Solution

To solve this, I designed and implemented a custom patient comment analyzer using the OpenAI GPT API and Python. The tool automatically ingests thousands of weekly survey comments and uses natural language processing to categorize each entry based on a taxonomy of Patient Access–related themes — such as wait times, check-in experience, insurance handling, and staff interactions. Each comment is also scored for sentiment (positive, neutral, negative), allowing us to quickly quantify both volume and tone at scale.

Once categorized and scored, the data is funneled into a dashboard that enables flexible, high-impact reporting. Users can filter comments by theme, sentiment, patient type, and campus, enabling supervisors to focus their attention only on the most relevant feedback. What once took hours of manual reading is now a streamlined, targeted review process that’s both consistent and scalable. The system is also updated regularly to evolve alongside patient priorities and operational goals.

Impact

This project has fundamentally transformed how we listen to our patients. With AI doing the heavy lifting, we’re not just reading more feedback — we’re actually understanding it. We can now spot persistent issues quickly, identify emerging concerns before they escalate, and track how patient sentiment changes over time and across locations. Most importantly, we’re using this data to drive meaningful improvements in how patients experience care. From reducing frustration at check-in to celebrating what’s working well, we now have a feedback loop that helps us meet patients where they are and respond more proactively to their needs. It’s no longer just about processing data — it’s about honoring patient voices and delivering a better, more informed care experience system-wide.