NUR 305 Module 7 Final Project Poster example

Listening Notes: Evaluating Ambient AI Documentation for Medical-Surgical Nurses

[Student Name]

Southern New Hampshire University

NUR 305: Information Management and Patient Care Technologies

Final Project Poster Presentation

[Instructor Name]

[Date]

The organization, setting and figures below are a composite written as a model document. No real employer, client, colleague or patient is described.

Panel 1: The Problem

Nurses on the 30-bed medical-surgical unit spend a large part of each shift documenting.

Admission assessments, shift assessments and patient education are typed after the conversation, often hours later.

Research on documentation burden describes lost time with patients, duplicate entry and contribution to burnout (Moy et al., 2021).

Question: can ambient AI documentation return time to patient care without harming the accuracy of the record?

Speaker notes: Our unit's nurses told us in a short survey that documentation is the part of the shift they would most like to change. A scoping review of documentation burden found that it has been measured in many ways, including time in the record and self-reported burden, and that nurses and physicians both carry it. We asked whether a technology now spreading through physician practice could help nurses, and what it would take to use it safely.

Panel 2: What the Technology Does

A secure microphone, usually a smartphone app, records the conversation between nurse and patient with the patient's consent.

Software transcribes the conversation and a language model drafts structured documentation, such as an admission history or education record.

The nurse reviews, edits and signs the draft; nothing enters the record without that review.

Audio is deleted after the draft is produced, under the organization's retention policy.

Speaker notes: The key idea is that the device listens so the nurse does not have to type during the conversation. The draft appears in the record for review, and the nurse remains the author. For nurses, the most promising uses are admission interviews and discharge teaching, where much of what is said needs to be documented.

Panel 3: What the Evidence Shows

Large-scale physician use: a health system rollout to thousands of physicians reported time saved on documentation and positive clinician feedback, with occasional errors in drafted notes (Tierney et al., 2024).

Nursing evidence is still early; published nursing writing is mostly commentary and small pilots (Topaz, 2025).

No study yet shows effects on patient outcomes in inpatient nursing.

Speaker notes: The strongest data come from physician practice. In a large health system, ambient scribes were adopted quickly by physicians, who reported spending less time on notes and more attention on patients, and reviewers found that most drafted notes were accurate but that errors, including statements that were not said in the visit, did occur. Nursing writing on the subject raises the question of whether nurses can trust these drafts and emphasizes that nurses must stay accountable for what is signed. We therefore treat the evidence as promising for time savings and uncertain for accuracy in nursing documentation.

Panel 4: Nursing Safeguards

Consent: the patient agrees to recording each time, and can decline without any effect on care.

Verification: the nurse reads every drafted entry before signing; high-risk fields such as allergies and advance directives are confirmed directly with the patient.

Privacy: recording stops when visitors or roommates share sensitive information not intended for the record.

Accountability: the signed note is the nurse's own work, exactly as with a typed entry.

Speaker notes: These safeguards connect to themes from the whole course: data integrity, privacy and the meaning of the nurse's signature. A drafted entry that is signed without review is no better than a copied-forward assessment. The nurse must remain the author of the record, and the technology must stay a drafting tool.

Panel 5: Proposed Pilot and Measures

Scope: eight volunteer nurses use ambient documentation for admission interviews for 12 weeks.

Time measure: minutes from admission arrival to completed admission documentation, compared with the prior 12 weeks.

Accuracy measure: a weekly audit of 10 drafted admissions for errors, including content not stated by the patient.

Experience measures: nurse survey on burden and trust; patient comments on being recorded.

Speaker notes: We propose starting small, with admissions only and a group of volunteers, because admission interviews are long, structured and heavily documented. We would compare time to completed admission documentation with the previous quarter and audit drafts for errors every week. If the error audit finds any content that was never said, we would review it with the vendor and the informatics team before continuing.

Panel 6: Barriers We Expect

Trust: nurses may doubt drafts after a single visible error.

Workflow: reviewing a long draft can take as long as typing if it is poorly organized.

Equipment and cost: devices, licenses and integration with the record require funding.

Patient comfort: some patients will not want a conversation recorded.

Speaker notes: We expect these barriers because they appear in the physician experience and in the nursing commentary. Trust is the most important. We plan to share the weekly accuracy audit results with the pilot nurses so that trust is based on evidence rather than impressions, and to ask them which parts of the draft took longest to review so the template can be improved.

Panel 7: Recommendation

Pilot, do not adopt: ambient AI documentation is promising for reducing nursing documentation time, but evidence in nursing is limited.

Proceed only with consent, full nurse review and weekly accuracy audits.

Decide on expansion after 12 weeks using the time, accuracy and experience measures.

Speaker notes: Our recommendation is a careful pilot rather than a unit-wide rollout. The technology addresses a real problem, but the evidence comes mainly from physicians, and the risks to the accuracy of the record are real. A short, measured pilot will tell our unit whether the time saved is worth the review it requires. If the pilot succeeds, the next step would be discharge teaching, where the same tool could document what was taught and how the patient responded to teach-back questions. If it fails on accuracy, the unit will still have learned something valuable: exactly which kinds of nursing content language models draft poorly, which can guide future vendor choices.

References

Moy, A. J., Schwartz, J. M., Chen, R., Sadri, S., Lucas, E., Cato, K. D., & Rossetti, S. C. (2021). Measurement of clinical documentation burden among physicians and nurses using electronic health records: A scoping review. Journal of the American Medical Informatics Association, 28(5), 998-1008. https://doi.org/10.1093/jamia/ocaa325

Tierney, A. A., Gayre, G., Hoberman, B., Mattern, B., Ballesca, M., Kipnis, P., Liu, V., & Lee, K. (2024). Ambient artificial intelligence scribes to alleviate the burden of clinical documentation. NEJM Catalyst, 5(3). https://doi.org/10.1056/CAT.23.0404

Topaz, M. (2025). Invisible scribes: Can nurses trust ambient AI for clinical documentation? The Journal of Continuing Education in Nursing, 56(9), 358-359. https://doi.org/10.3928/00220124-20250814-03