SlicerDicer is a self-service data exploration tool built into Epic, the electronic health record system used by many hospitals and health systems. It lets physicians, department managers, researchers, and other staff run their own searches across large patient populations without needing to submit a request to IT or the analytics team. Think of it as a way to quickly answer questions like “How many of our diabetic patients are over 65?” or “What percentage of patients with a certain diagnosis were readmitted within 30 days?” using point-and-click filters and visual charts.
What SlicerDicer Actually Does
SlicerDicer is designed to be a starting point for exploring a question, not a tool for final analysis or direct patient care decisions. You pick a population (say, all patients seen in your department), apply filters (age range, diagnosis, lab results, medications), and instantly see counts and visualizations. You can then refine those filters on the fly, narrowing or broadening your search until the data tells you something useful.
The tool pulls from Caboodle, which is Epic’s data warehouse. Data is refreshed nightly, so what you see in SlicerDicer reflects the previous day’s information rather than real-time records. This one-day delay is worth keeping in mind if you need up-to-the-minute data for clinical decisions.
A key feature is that the default view presents de-identified, aggregate data. You see counts and trends, not individual patient names or charts. To protect privacy, searches that return fewer than 10 patients won’t show an exact number. You’ll only see that 10 or fewer records matched. The exception is when you search within your own patient panel, where exact counts are visible.
Who Uses It and Why
SlicerDicer serves several different roles across a health system. Clinicians use it to investigate hunches about their patient populations, spotting patterns in diagnoses, outcomes, or demographics. Department managers use it for operational questions: tracking volumes, understanding patient mix, or identifying trends over time. You can view up to 10 different populations simultaneously, making side-by-side comparisons straightforward.
Researchers find it particularly valuable for early-stage work. Because the data is de-identified at the aggregate level, exploring preliminary hypotheses in SlicerDicer typically does not require institutional review board (IRB) approval. This makes it a fast way to check whether a research question is worth pursuing before investing in a full study. If the numbers look promising, you can use those preliminary findings to support a funding proposal or justify a formal data request. SlicerDicer should not, however, be used for actual research study analysis.
How It Differs From Reporting Workbench
Epic offers another self-service tool called Reporting Workbench, and the two serve different purposes. SlicerDicer is for exploration: quick, visual, aggregate-level. It helps you understand the shape of a population. Reporting Workbench is for extraction: it produces tabular, record-level datasets you can export and work with. If you need a list of specific patients who meet certain criteria, or you want to send a bulk message to a group of patients, Reporting Workbench is the right tool.
The data sources differ, too. SlicerDicer queries Caboodle (the data warehouse) with a one-day delay. Reporting Workbench pulls directly from Chronicles, Epic’s live production database, so data is available as soon as it’s entered. The trade-off is that Reporting Workbench reports are typically built from administrator-created templates rather than the freeform, drag-and-drop exploration SlicerDicer offers.
How to Access It
SlicerDicer lives inside Epic’s Hyperspace application. You can find it by clicking the Epic button, navigating to Reports, and selecting SlicerDicer. Your organization controls which data models and populations you can see based on your role and security settings, so not every user has identical access. Some organizations offer multiple “session types” or data models tailored to specific departments or use cases.
Most health systems include any patient who has had an appointment or admission since their Epic go-live date. At Johns Hopkins, for example, that includes every patient seen since April 2013.
Natural Language Querying
Epic has been working with technology partners to bring natural language capabilities to SlicerDicer. The goal is to let users type questions in plain English rather than building queries with filters and menus. As Epic’s vice president of data and analytics, Phil Lindemann, described it, the feature would let clinical leaders “explore data in a conversational and intuitive way” to find trends. This functionality is still evolving, but it signals where the tool is headed: making population-level data accessible to users with no analytics training at all.
Limitations to Keep in Mind
SlicerDicer is powerful for what it’s designed to do, but it has clear boundaries. It does not allow record-level data extracts, meaning you cannot pull a spreadsheet of individual patient details. The one-day data lag means it’s not suitable for time-sensitive clinical workflows. And because it’s built for exploration rather than proof, it shouldn’t be treated as a rigorous analytical tool for establishing causality or making treatment decisions.
Organizations also vary in how they configure SlicerDicer. The filters, data models, and populations available to you depend on what your Epic team has set up and what security access your role includes. If you’re not seeing the data you expect, your organization’s Epic support team or analytics group can help clarify what’s available and how to request additional access.