NR 588AI Week 8 Board Executive Briefing Example

Executive Briefing to the Board of Trustees: Continue the Readmission Risk Model With a Stronger Equity Floor, Two Added Nurses and a Named Assurance, or Suspend It

Student Name

College of Nursing, Chamberlain University

NR 588AI: Advanced Nursing Leadership in Artificial Intelligence-Integrated Complex Systems

Instructor Name

Month Day, Year

Executive Briefing to the Board of Trustees: Continue the Readmission Risk Model With a Stronger Equity Floor, Two Added Nurses and a Named Assurance, or Suspend It

The Position

Management asks the Board to approve continuing the readmission risk model for 2028, with the equity floor raised, two transitional care nurses added at about $210,000 a year and quarterly assurance signed by the System Patient Safety Officer. The alternative is to suspend it and return to the previous selection criteria.

Exposure the System Cannot Yet Measure

Some patients are invisible to this tool's data: people without a fixed address, people readmitted to hospitals outside our region and patients whose race, ethnicity or language was not recorded, about 8% of discharges. We cannot say how well the tool serves them. This is stated first because it is the exposure most likely to surprise the Board if raised from outside.

Exposure to Patients

In 2027 the tool selected 41% of adult patients later readmitted within 30 days, against 33% under the old criteria. For patients who prefer Spanish or have no primary care provider, it selected fewer, even after a correction, a pattern researchers have documented in other care management tools (Obermeyer et al., 2019). One such patient, not called after discharge in June, was readmitted on day six and spent time in intensive care. A review found several decisions by the system, not the vendor, that contributed.

Promises Made

The system's published principles for artificial intelligence state that tools will be monitored for fairness and that clinicians remain in control of care. The June case showed both promises were kept only in part: monitoring reports went unread for three months, and training overstated what the list could do. Both have been corrected.

Money Committed

The contract costs about $140,000 a year to 2029, with a right to end at any time if conditions are breached. The two added nurses would cost about $210,000 a year. Suspending the tool saves the contract cost but not the nursing need, because the old criteria select fewer of the patients who are readmitted.

Could the System Explain Its Decision?

If asked by a regulator or a journalist why a Spanish-speaking patient was not called, the system can now show: an equity review commissioned before go-live with a consequence agreed in advance, a correction applied, an accountability review after the June case and the changes made. It could not have shown this in May. Char et al. (2018) warn that the ethical risks of clinical models often lie in how institutions use them; the Board's assurance should rest on that use.

Assurance, With a Name

Ongoing fairness monitoring after deployment is the approach published guidance recommends (Rajkomar et al., 2018).

Every quarter the System Patient Safety Officer will put in writing, over a signature, that no watched group has fallen more than 8 points behind the system-wide capture figure, that each vendor report arrived and was read, and that the model has not changed without notice. Anything that cannot be confirmed will be named, along with the action taken.

Background, Compressed

The tool ranks adult inpatients daily by readmission risk; the top 20%, plus a floor for two groups, get a nurse call within 48 hours. It was approved in November 2026 with conditions, went live in March 2027 and is overseen by the patient safety officer.

Options

Continue with the raised floor and two nurses (recommended). Expected to bring capture for the two groups close to the overall rate, at a total cost of about $350,000 a year.

Continue without added nurses. The raised floor would push the list beyond the team's capacity, and some listed patients would not be called.

Suspend. Return to the old criteria, which miss more readmitted patients overall, though with no known group-level gap because they were never measured by group.

What Management Asks the Board to Decide

Approve continuation for 2028 with the conditions above, and the $210,000 for two nurses in the 2028 budget. The Board Quality Committee will receive the first signed assurance in April 2028 and a full equity review in January 2029.

What Would Lead Management to Recommend Suspension

Two quarters in a row, after the change, with any watched group more than 8 points short of the system-wide capture figure, or a further case in which a review finds the tool's configuration contributed to harm.

References

Char, D. S., Shah, N. H., & Magnus, D. (2018). Implementing machine learning in health care: Addressing ethical challenges. New England Journal of Medicine, 378(11), 981-983. https://doi.org/10.1056/NEJMp1714229

Obermeyer, Z., Powers, B., Vogeli, C., & Mullainathan, S. (2019). Dissecting racial bias in an algorithm used to manage the health of populations. Science, 366(6464), 447-453. https://doi.org/10.1126/science.aax2342

Rajkomar, A., Hardt, M., Howell, M. D., Corrado, G., & Chin, M. H. (2018). Ensuring fairness in machine learning to advance health equity. Annals of Internal Medicine, 169(12), 866-872. https://doi.org/10.7326/M18-1990