The Care Stack

Your EHR is not your data strategy

Victor Wilson · · 4 min read

Most health systems spent years and a lot of money putting in their EHR. It runs clinical care and much of the revenue cycle. It comes with a reporting database and a set of analytics tools, and those tools have gotten much better over time.

So it’s natural to ask: why do we need anything else? Can’t the EHR be our data strategy?

I understand the appeal. But in my experience, the answer is no, and the reason has little to do with the EHR vendor.

To be fair, EHR reporting is strong at what it was built for. If the question is about patients, encounters, orders, or what happened during a visit, the EHR’s reporting database is usually the best place to answer it. The data is detailed, and many reports come ready to use.

For clinical operations and quality reporting, that’s often enough. If your team uses those tools well, keep using them.

EHR vendors are also moving fast on AI. Epic says more than 85% of its customers use Epic AI, and in March it previewed Agent Factory, a no-code tool for building agents, along with Curiosity, medical models trained on its Cosmos data. The same month, Microsoft opened Dragon Copilot to partner-built agents. For work that happens inside the EHR, tools like these will be hard to beat.

What it can’t see

The trouble starts when you ask the questions a CFO or COO asks.

What does it cost us to deliver a hip replacement, including the implant and the OR time? Which units are relying most on overtime and contract labor? Are we being paid what our contracts say? Which supply categories are drifting off contract?

None of those questions can be answered from the EHR alone. The data lives somewhere else:

  • Finance: the general ledger, budgets, and cost accounting.
  • Supply chain: purchasing, invoices, contracts, and the item master, usually in an ERP.
  • Workforce: HR, payroll, timekeeping, and scheduling systems.
  • Claims and payments: remittances from payers, denials, and what was actually collected.
  • Contracts: payer agreements and vendor agreements, often stored as documents.

The EHR knows a nurse was assigned to a patient. Payroll knows what that nurse cost. The EHR knows an implant was used. Supply chain knows what you paid for it, and whether that matched the contract. The interesting answers live where these sources meet.

The same is true for AI agents. An agent can only see what it can reach. One that runs inside the EHR is well placed to draft a note or prepare a referral. Ask it why cost per case went up on a service line, and it needs the general ledger, the invoices and payroll. If those don’t sit somewhere it can reach, it can’t give you the full answer, however good the model is.

Some teams try to pull finance or supply data into the EHR’s reporting environment. Sometimes that works for a narrow need. As a general approach, it tends to struggle.

The EHR’s reporting database is designed around the EHR’s own data model. Bringing in outside data means bending it into shapes it wasn’t meant for. You also tie your analytics to one vendor’s roadmap and release schedule, which may not match your priorities.

There’s a quieter problem too. When the EHR team owns all reporting, every request competes with clinical work for the same small group of people. Clinical needs usually win, as they should. Finance and supply chain wait.

What a data strategy actually is

A data strategy isn’t a tool choice. It’s a set of decisions about how your organization answers its most important questions.

It covers which questions matter most. It covers where data from each source lands so it can be combined. It names who owns the definitions, so that “cost per case” means one thing everywhere. And it lays out how the sources connect: shared keys for providers, locations, cost centers, patients, and dates, so that a record in payroll can be lined up with a record in the EHR.

Now it also covers what your AI agents can read, and who decides that.

The EHR is a central part of that strategy. It is probably your richest single source. But it sits alongside the others, feeding a shared place where they can meet. In most health systems I’ve seen, that shared place is a separate data platform, and the EHR’s reporting tools keep doing what they do well.

A simple test

Here’s a way to check where you stand.

Write down the ten questions your executive team asks most often. Not reports, questions. Next to each one, list every system that holds part of the answer.

If most of your questions only need the EHR, your current setup may be fine. If most of them need three or four systems, and today someone answers them by exporting spreadsheets and stitching them together by hand, then your EHR is doing its job and your data strategy is missing. Start with the question on that list that’s worth the most money, and build the connections it needs first.

Discussion

    Join the discussion