NU710 Unit 5 Data Confidentiality Analysis Example

A Spreadsheet on a Home Laptop and a Chatbot for the Call Notes: A Unit 5 Data Confidentiality Analysis of a DNP Project's Patient Data

[Student Name]

Purdue University Global

NU710: Advanced Health Information and Technologies

Unit 5 Data Confidentiality Analysis

[Instructor Name]

[Date]

The project, hospital, people and data are composite and illustrative; the rules and sources cited are real.

The Data and Who Needed It

The project needed patient-level data for three purposes: to know whom to call after discharge, to measure returns within 30 days and to report results to the steering group and, later, in the DNP project paper. The first purpose needs identifiers, since a call needs a name and a number. The second needs identifiers only long enough to link discharges with returns. The third needs none at all.

Use One: The Spreadsheet on a Personal Laptop

The student, who is also the unit educator, exported a daily list from the record into a spreadsheet to track calls and returns, and kept it on a personal laptop to work on at home. The spreadsheet held 104 patients' names, medical record numbers, dates of birth, phone numbers, discharge dates and glucose values.

Under the HIPAA Privacy Rule, using protected health information for quality assessment and improvement within the organization falls under health care operations, which do not require patient authorization. But the rule's minimum necessary standard limits use to what the purpose requires, and the Security Rule's administrative, physical and technical protections follow electronic health data onto any device that holds it. A personal laptop outside the hospital's management, with no encryption verified by the hospital, is the kind of device involved in many reported breaches. The ethical obligation is the same: the ANA Code of Ethics treats privacy and confidentiality as part of the nurse's duty to protect the patient (American Nurses Association [ANA], 2015).

The project's purposes could have been met inside the hospital's systems: a report in the record listing patients due for calls, and a secure shared drive for the tracking sheet, accessible only to the educator and the two call nurses.

Use Two: The Chatbot and the Call Notes

The second use was more serious. A staff nurse, trying to help, copied a week of follow-up call notes into a free public AI chatbot and asked it to summarize common problems. The notes included first names, ages, medication doses and details such as "lives with son who drinks." The nurse removed last names but not much else.

The hospital has signed no business associate agreement with the company running that chatbot, so pasting protected health information into it is a disclosure outside the organization. Depending on the service's terms, the text may be stored, reviewed by people or used to train models. Removing last names does not make data de-identified. Under the Privacy Rule, de-identification requires either removing a defined list of identifiers, including all elements of dates except year and any other unique characteristic, or a qualified expert's documented judgment that the chance of anyone being recognized is negligible (U.S. Department of Health and Human Services [HHS], 2012). Notes that combine a first name, an age, a unit and a family situation can identify a patient to anyone who knows the community.

The nurse removed the surname and left the patient.

Abuhammad (2025) argues that nurses, as key custodians of health information, often lack training and institutional support for confidentiality in an era of AI tools, and calls for education and policy reform. That describes this case. The nurse was not careless in intent; nobody had told the unit what was allowed.

Use Three: Reporting Results

For the steering group and the eventual project paper, the project needs aggregate results: counts, rates and run charts. These require no identifiers. The DNP project will also need a determination from the hospital and the university of whether it is quality improvement or research, because research use of identifiable data would require review under the Common Rule or a waiver. The pilot's design, a practice change measured against local baseline with no generalizable research question, points toward quality improvement, but the determination is not the student's to make alone.

Why Nobody Stopped It

The three uses grew one at a time, each a reasonable shortcut on a busy day. The educator exported the list because the record had no report for patients due for calls. The nurse used the chatbot because the summary was needed for a meeting the next morning and nobody had time to read 40 notes. These are system gaps, the same kind the course has examined in alerts and risk scores. A confidentiality failure that comes from missing tools will recur unless the tools are supplied.

Ethical Weighing

None of the three uses was meant to harm patients, and the data were used for their benefit. But confidentiality is not only about intent. Patients starting insulin share details about money, family and drinking because they trust the hospital to keep them. A breach would harm the specific patients exposed and, more broadly, the trust that makes honest follow-up calls possible.

Safeguards the Project Will Adopt

1. Keep identifiable data inside the hospital. Replace the laptop spreadsheet with a record-based report and a tracking sheet on a secure drive, restricted to three named staff. The laptop file will be deleted and the deletion confirmed with the privacy officer.

2. Report the chatbot disclosure. Notify the privacy officer, who will assess whether it is a reportable breach, and follow the hospital's process.

3. Use only approved AI tools. Any summary of call notes will be done in a hospital-approved tool with a business associate agreement, or by hand.

4. De-identify before analysis. Results for reporting will be produced from a data set with direct identifiers removed and dates reduced to month, with a key held separately by the analyst.

5. Obtain a quality improvement determination from the hospital and the university before any results leave the organization.

6. Teach the unit. A ten-minute huddle on what data may be used where, with a one-page guide, for every nurse who makes follow-up calls.

Conclusion

The project's data use was well meant and poorly protected. Identifiable data on a personal laptop exceeded what the purpose needed and sat outside the hospital's safeguards; call notes pasted into a public chatbot were disclosed outside the organization and were not de-identified by removing surnames. Keeping identifiable data inside approved systems, using only covered tools, de-identifying for analysis, reporting the disclosure and teaching staff would bring the project in line with HIPAA and with the confidentiality nurses owe patients.

References

Abuhammad, S. (2025). Strengthening ethical practices of patient data confidentiality and sharing among nurses in the artificial intelligence-driven healthcare era. Sage Open Nursing, 11, 23779608251398113. https://doi.org/10.1177/23779608251398113

American Nurses Association. (2015). Code of ethics for nurses with interpretive statements. American Nurses Association.

U.S. Department of Health and Human Services. (2012). Guidance regarding methods for de-identification of protected health information in accordance with the Health Insurance Portability and Accountability Act (HIPAA) Privacy Rule. https://www.hhs.gov/hipaa/for-professionals/special-topics/de-identification/index.html