The situation
Cureatr's pharmacist-led medication reconciliation service existed to satisfy a specific quality requirement: reviewing a patient's current and discharge medications within 30 days of a hospital discharge. But every one of those visits depended on a single document — the hospital's discharge summary — and getting a copy of it into a pharmacist's hands was, by a wide margin, the most broken part of the process.
When we surveyed our own pharmacists on the least efficient part of preparing for a patient call, this ranked as their single biggest pain point. Discharge summaries arrived from three different places — an automated data feed, a manual hospital request, or the patient themselves — with no consistent way to know which one applied to a given patient, or whether we even had the document at all.
Diagnosing three separate problems
As with most workflow problems that look like "people just need to work faster," the real issues were structural, not effort-based. I broke this down into three distinct problems worth solving separately, because each had a different owner and a different fix:
-
User experience
When a discharge summary did arrive automatically, there was no way to flag which document in a long list was the right one — pharmacists were re-hunting for it every time they prepped for a call.
-
Manual process risk
When a discharge summary didn't arrive automatically, requesting it from the hospital was a fully manual process — tracked in a spreadsheet, and requested by sharing patient information directly in Slack.
-
Missing data
As a product team, we had no visibility into how often we actually had a discharge summary on hand, or which acquisition path — automated feed, hospital request, or patient — was carrying the load.
Designing the fix
I led the requirements gathering directly with the three groups who touched this process differently — clinic staff prepping for visits, the medical records team chasing down documents from hospitals, and our own product team trying to measure any of it — and turned that into a solution with a few connected pieces:
An at-a-glance status banner now sits next to every patient's discharge details in the product, showing exactly where things stood: no one had looked into it yet, the patient had it on hand, it had been located and "pinned" from the incoming document feed, it had been requested and was in a queue, or it had been manually uploaded after a hospital fax. Pharmacists could pin the correct document straight from the existing document list — no more re-opening the same folder on every visit.
For summaries that didn't arrive automatically, a single click replaced the Slack message and spreadsheet entirely — triggering an automated, prioritized request queue in our internal systems that the medical records team could work from directly, with faxed documents uploaded straight back into the same patient record everyone else was already looking at.
What changed
The efficiency gains showed up at every step of the process:
Beyond the direct savings, the new visibility into acquisition data — automated feed versus hospital request versus patient on hand — gave us a real benchmark for the first time, and directly informed the design of a follow-on service we piloted afterward. Just as importantly, clinical staff no longer had to choose between doing their job efficiently and handling patient information carefully — the safer path and the faster path became the same path.
Figures reflect internal performance metrics generalized for public sharing; specific vendor names and internal system names have been omitted or simplified.