How to read a 300-page hospital referral packet in minutes

A hospital referral packet for one new admission can run 150 to 300 pages.¹ It arrives as a mix of discharge summaries, progress notes, medication lists, therapy evaluations, lab results and scanned forms, usually in no particular order. Somewhere inside it is the one line that changes how the first shift goes: an oxygen need on exertion, a new anticoagulant, a pressure injury, a swallow precaution.

This guide covers what matters most in a referral packet, a practical order for reading it, and how cited AI summaries are changing admissions and the first shift.

Why are referral packets so hard to review?

Volume is only part of it. Packets are built for the hospital's records, not for the nurse who has to care for the resident tonight. Research on hospital-to-SNF transitions has repeatedly found communication barriers between sending and receiving providers, including differences in access to information.³ And only about 55% of hospitals can electronically find, send, receive and integrate patient information, so much of it still arrives as scanned pages.²

Meanwhile acuity keeps climbing. Average comorbidities rose from 3.4 to nearly 5 per patient between 2019 and 2024.² More conditions means more pages, and more places for something important to hide.

What should a nurse look for first in a hospital referral packet?

Experienced admissions nurses read for risk first and history second. A practical first pass:

• Code status and advance directives. Confirm they match across documents.
• Discharge medication list. Reconcile it against the hospital MAR. Flag new anticoagulants, insulin, opioids and antibiotics with stop dates.
• Allergies and isolation needs. Including any multidrug-resistant organisms.
• Respiratory needs. Oxygen at rest or on exertion, flow rates, CPAP or BiPAP.
• Skin and wounds. Pressure injuries, surgical sites, dressing orders.
• Swallow and diet. Texture, liquid consistency and aspiration precautions.
• Mobility and fall risk. Weight-bearing status, therapy orders and recent falls.
• Cognition and behavior. Delirium during the stay, dementia, sitter or safety needs.
• Lines, drains and devices. PICC lines, catheters, drains and removal plans.
• Pending items. Lab results still outstanding and follow-up appointments.

What is the fastest order to read a referral packet?

1. Discharge summary first. It frames the stay and should list the active problems.
2. Medication reconciliation next. Compare the discharge list with the last days of the MAR.
3. Most recent nursing and therapy notes. They show how the patient is actually doing now, not on admission.
4. Search, don't scroll. Look for high-risk terms: oxygen, fall, wound, isolation, NPO, anticoagulant, DNR.
5. Write down what you couldn't confirm. Those are the questions for the hospital before transfer.

How do AI summaries change referral review?

AI can now read an entire packet and produce a short summary organized around the checklist above. In one reported example, packets of 150 to 300 pages were condensed to about 30, and referral review time fell from roughly 25 minutes to 10.¹

Speed alone is not enough for clinical use. The summary has to be verifiable. The standard to hold any tool to: every statement links to the page it came from, so a nurse can check "oxygen on exertion" against page 17 in one click instead of trusting a paraphrase.

Why this matters for the first shift, not just admissions

Admissions teams read packets to decide yes or no. The floor nurse reads them to keep the resident safe tonight, often with five other residents on the assignment. When the packet's key facts carry forward into the shift handoff, with their sources attached, the incoming nurse starts the shift knowing what the admissions team knew.

That is what ChartRail does: a private AI reads the referral packets and nursing notes before the shift, and hands each nurse what changed, with every line traced to its source page.

Frequently asked questions