As A Nurse With Cancer, I Saw The Value Of Patient-Facing Materials

Davis Austria, an emergency and critical care nurse who was treated for head and neck cancer, argues that health systems should test patient-facing medical information for clarity and practical usefulness, not only clinical accuracy. In an October 20, 2026, MedPage Today opinion article, Austria describes how treatment documents left them uncertain about side effects and what to expect, and warns that AI could spread helpful explanations—or repeat information gaps at scale.

Austria writes that, after their diagnosis four years earlier, their radiation oncologist showed them the National Comprehensive Cancer Network (NCCN) treatment recommendations and explained a plan involving 10 chemotherapy rounds and 35 radiation sessions. The document was designed for clinicians. Austria says it did not answer many of the practical questions they had as a patient, such as whether painful swallowing was expected, whether hearing would recover, and how they would learn to eat after a feeding tube was removed.

Austria later downloaded NCCN patient-facing guidance. They write that its advice for radiation burns was to ask a healthcare provider for the best treatment, and that they used Aquaphor on their neck before radiation sessions before finding another document that clarified the timing. Austria says the gaps led them to rely on a Facebook group for people with the same cancer, while their patient portal served mainly as a place to retrieve documents rather than ask questions.

The article also addresses AI in health care. Austria says AI tools can gather information from multiple sources and explain it in response to patients’ questions, but may also provide information that is wrong, incomplete, or misleading. Health systems are also using AI to draft or rewrite patient materials, the article says. Austria points to Epic’s patient-facing virtual assistant, Emmie, which can rewrite after-visit summaries in plainer language, and cites a randomized trial involving 128 patients in Germany: those given AI-generated, patient-oriented discharge summaries scored 9.6 points higher on a patient-activation measure than those given standard physician-written summaries.

Testing Instructions With Patients

The issue Austria raises is whether information intended for patients answers the questions they need to act on, rather than simply presenting accurate clinical content. Instructions that use terms such as “ototoxicity” or abbreviations such as NPO, STAT and PRN may be clear to medical staff but difficult for patients to interpret. Plain-language rewrites could help patients understand care instructions, but readability alone does not establish that the information is complete or safe to follow.

Austria’s argument has added relevance as health systems adopt tools that can produce or revise information at scale. A clear review process could identify gaps before instructions reach patients; without patient testing, the author warns, AI could reproduce those gaps across many documents. The trial cited in the article offers one measure of potential benefit, but its reported result does not by itself establish how every AI-generated document performs or whether the approach improves health outcomes.

From Cancer Care to AI Materials

Austria writes from two perspectives: as a nurse familiar with clinical systems and as a patient who experienced cancer treatment. They say their care team knew them as a clinician, which may have shaped communication and assumptions about what they already understood. The article also refers to a 2024 study in BMC Primary Care about doctors treating fellow doctors; according to Austria, it found clinicians changed how they communicated with medical colleagues who were patients.

The experience Austria recounts occurred four years before the article’s publication. Their central concern is not that clinical guidelines are inaccurate: they describe the clinician-facing recommendations as accurate and evidence-based. Rather, they distinguish between guidance written to help clinicians make treatment decisions and information that helps patients understand symptoms, side effects, and what to do next.

Austria recommends that health systems ask whether materials work for the patients they serve across language, age and social circumstances; whether they have been reviewed for more than medical accuracy; and whether patients have tested their ability to understand and act on the information. The article states that its views are Austria’s alone and do not represent NIH, AIM-AHEAD, Ochsner Health or Xavier University of Louisiana.

“What I did not know was what was normal.”

— Davis Austria, in the MedPage Today opinion article

Limits of the Evidence Cited

The opinion article does not give the name of the German trial or details about its methods beyond the reported sample of 128 patients, the comparison between standard and AI-generated patient-oriented summaries, and the 9.6-point difference in patient activation scores. The score’s scale, the study’s setting and whether the result persisted over time are not specified in the supplied report.

It is also unclear from the article how widely health systems use AI to produce patient materials, how often those materials undergo patient testing, or what safeguards Epic’s Emmie feature applies. Austria’s personal account describes one cancer-treatment experience; it is not evidence that all patients receive the same materials or face the same problems. The author also cautions that AI can be inaccurate or incomplete and says it is not a substitute for a patient’s care team.

Patient Review Before Distribution

Austria’s article does not announce a new policy, deadline or planned study. Its proposed next step is for health systems to ask patients to review instructions before distributing them, and to evaluate materials for clarity, completeness and usefulness alongside medical accuracy. That standard, Austria argues, should apply whether the information is written by staff or produced with AI.

For readers, the article offers no change to current treatment guidance. Patients with questions about symptoms, medication instructions or side effects should contact their care team rather than rely on an AI response or anonymous online advice as a replacement for clinical guidance.

Key Questions

What did Davis Austria argue in the article?

Austria argued that health systems should test patient materials with the people expected to use them, checking whether the information is understandable and actionable as well as medically accurate.

What information did Austria say was missing during cancer treatment?

Austria describes wanting practical guidance about what symptoms and side effects were normal, whether hearing might recover, and how eating could change after a feeding tube was removed.

What did the German trial cited in the article report?

According to Austria’s article, a randomized trial of 128 patients compared standard physician-written discharge summaries with AI-generated versions rewritten for patients. The patient-oriented group scored 9.6 points higher on a patient-activation measure. The supplied report does not provide further study details.

Does Austria recommend using AI instead of a clinician?

No. Austria says AI may help explain health information but can be wrong, incomplete or misleading, and is not a substitute for the clinicians caring for a patient.

What review does Austria recommend for patient materials?

Austria proposes checking whether materials suit the patients a health system serves, reviewing them for more than medical accuracy, and asking actual patients whether they can understand and act on the instructions.

Source: rss