TL;DR
Turn your wind-down time into listening time
- Thousands of audiobooks, podcasts and originals
- Listen on your phone, tablet or Echo — also offline
- Cancel anytime
In a MedPage Today opinion article published Oct. 20, 2026, nurse and cancer survivor Davis Austria describes how clinician-focused guidance failed to answer practical questions during treatment. Austria argues that health systems should have patients test educational materials, including AI-generated versions, for clarity and usefulness as well as medical accuracy.
Nurse and cancer survivor Davis Austria is urging health systems to test medical instructions with patients before distributing them, arguing that clinically accurate materials can still fail to answer the questions patients need to manage care. In an Oct. 20 MedPage Today opinion article, Austria says the challenge is becoming more pressing as providers use artificial intelligence to generate or rewrite patient information.
Austria, who identifies as an emergency and critical care nurse and studies evaluation of AI-generated patient information, recounts being diagnosed with head and neck cancer four years ago. The radiation oncologist showed Austria National Comprehensive Cancer Network (NCCN) treatment recommendations calling for 10 rounds of chemotherapy and 35 radiation sessions, then sent Austria home with the document. Austria describes it as accurate and evidence-based, but written for clinicians rather than patients.
The information did not address many of Austria’s immediate concerns: what treatment side effects were normal, whether swallowing pain suggested a problem, whether hearing might recover, and how eating would work after a feeding tube was removed. Austria says patient-facing NCCN material advised asking a provider about treatment for radiation burns, but did not resolve a practical question about when to use Aquaphor. Austria later found another document with timing guidance and says they had been applying it at the wrong time.
Unable to get specific answers from the documents or a static patient portal, Austria turned to a Facebook group for people with the same cancer, where suggestions came from anonymous users. Austria contrasts that experience with AI tools that can gather medical information and respond in accessible language, while warning that they may be wrong, incomplete or misleading and cannot replace clinicians.
Testing Instructions With Patients
Austria’s account points to a gap between information that meets a clinical standard and information a patient can understand and use. Discharge instructions and treatment materials shape what people do between appointments, including when they seek help and how they follow care plans. If guidance leaves key questions unanswered, patients may turn to informal online sources or remain unsure about what to expect.
The issue also applies to AI-generated information. Austria argues that a tool can reproduce a communication gap at scale if health systems assess only whether its output is medically accurate. Materials should also be checked for clarity, completeness and usefulness, and tested with the patients who will rely on them across differences in language, age and social circumstances.
Austria cites a randomized trial involving 128 patients in Germany. The study compared standard physician-written discharge summaries with AI-generated versions rewritten for patients; those receiving the patient-oriented versions scored 9.6 points higher on a measure of patient activation, or readiness to manage their care. The result, as Austria presents it, supports testing patient-focused communication, but does not by itself establish that AI-written instructions improve health outcomes in every setting.
patient education materials testing toolkit
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
From Clinical Guidance to Plain Language
The NCCN recommendations Austria received were designed to guide oncologists’ treatment decisions. Austria distinguishes those clinician-facing guidelines from materials meant for patients, saying the former could be sound while still failing to explain what treatment might feel like or what to do in specific situations.
Austria also refers to a 2024 study in BMC Primary Care of doctors treating fellow doctors. The study found that clinicians changed how they communicated with colleagues who were patients, including making assumptions about what they already knew. Austria says being recognized as a medical professional shaped the interaction during the cancer diagnosis, even though the immediate need was for support as a patient.
The article describes health systems’ growing use of AI to create or rewrite patient information. Austria points to Epic’s patient-facing virtual assistant, Emmie, which includes a feature that rewrites after-visit summaries in plainer language. Austria’s examples of medical shorthand include “NPO,” “STAT” and “PRN,” terms that can be familiar to health workers but unclear to patients.
““What is happening to me, and what happens next?””
— Davis Austria, in the MedPage Today opinion article
AI-generated patient information review
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Limits of the Evidence and Tools
Austria’s article is an opinion piece that combines a personal account with references to research and health-system products. It does not report a new clinical study or provide data showing how widely AI-generated patient materials are being used, how often they contain errors, or whether the tools consistently improve understanding across patient groups.
The German trial described in the article involved 128 patients and measured patient activation; the source material does not provide further details about the study’s design, duration or effects on clinical outcomes. Austria’s account also does not establish how common the communication problems described are among people receiving cancer care. AI responses can vary, and Austria explicitly warns that they may be inaccurate or incomplete.
medical discharge instruction guides
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Patient Testing Before Wider Use
Austria calls on health systems to ask three questions before hiring writers or adopting AI tools for patient education: Does the material work for the patients the system serves, across language, age and social circumstances? Has it been reviewed for more than medical accuracy? Have actual patients tested whether they can understand and act on it?
The article does not announce a policy change, deadline or next formal step by a health system. Austria’s recommendation is that patient testing happen before materials reach patients. The broader question is whether health providers using AI or other tools will build that review into their processes and evaluate the resulting materials with the people expected to use them.
cancer treatment patient education resources
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
What is the news in Davis Austria’s article?
Austria published an opinion article urging health systems to test patient-facing medical information with patients. The piece draws on Austria’s experience receiving cancer treatment and argues that materials can be accurate yet fail to answer patients’ practical questions.
What did Austria say was missing from the cancer materials?
Austria says the documents did not answer specific concerns about side effects, what symptoms were normal, hearing recovery or eating after a feeding tube was removed. Austria also describes finding guidance about Aquaphor timing only in a separate document.
Does Austria say AI is a substitute for clinicians?
No. Austria says AI can gather information and explain it in language tailored to questions, but may also be wrong, incomplete or misleading. The article says AI is not a substitute for the clinicians caring for a patient.
What study does Austria cite?
Austria cites a randomized trial of 128 patients in Germany that compared standard physician-written discharge summaries with AI-generated patient-oriented versions. Patients receiving the rewritten summaries scored 9.6 points higher on a patient-activation measure, according to the article.
What does Austria want health systems to do?
Austria recommends checking materials for usefulness and clarity as well as medical accuracy, and asking patients to test whether they can understand and act on the information before it is distributed.
Source: rss
Halloween Picks
halloween
As an affiliate, we earn on qualifying purchases.
