| Course | MHA FPX 5068 Leadership, Management, and Meaningful Use of Healthcare Technology |
|---|---|
| Assessment | Assessment 3 |
| Paper type | Technology governance plan |
| Length | About 1,110 words, 5 pages |
| Format | APA 7 student paper |
| School | Capella University |
| Program | Master of Health Administration (MHA) |
| Updated | September 2026 |
Free sample paper for MHA FPX 5068 Assessment 3
Governing Artificial Intelligence Tools in a Health System: Decision Rights, Training Obligations, Data Ownership, and Misuse
Student Name
Capella University
MHA FPX 5068: Leadership, Management, and Meaningful Use of Healthcare Technology
Instructor Name
Month Day, Year
Introduction
A composite regional health system, here called Summit Valley Health, uses a growing number of artificial intelligence tools: a sepsis prediction model in its electronic health record, an ambient documentation tool that drafts clinical notes from recorded conversations, a scheduling tool that predicts no-shows, and several tools built into vendor products. Staff also have access to general-purpose generative tools on the internet. Until now, each tool has been approved by whichever department wanted it. This paper proposes a governance structure.
Frameworks and Rules
Two sources guide the structure. The National Institute of Standards and Technology (2023) published a voluntary framework for managing risks of artificial intelligence, organized around four functions: govern, map, measure, and manage. It emphasizes that governance must assign accountability, define roles, and establish processes for identifying and managing risks across a tool's life. Federal health information technology rules finalized in 2024 require developers of certified health information technology to make information about predictive decision support tools available to users, including their intended use, the data used to develop them, and known risks (Office of the National Coordinator for Health Information Technology, 2024). This transparency gives health systems information they need to evaluate tools but does not do the evaluation for them.
Decision Rights by Role
Decision rights are assigned to roles rather than departments, so that it is clear who decides.
| Decision | Decides | Must be consulted |
|---|---|---|
| Approve a new clinical tool for use | Artificial intelligence governance committee, chaired by the chief medical information officer | Clinical specialty leaders affected; chief nursing officer |
| Approve a new administrative tool | Same committee, with the chief operating officer as co-chair for that decision | Department leader requesting the tool |
| Approve use of patient data to train or test a tool | Chief privacy officer | Chief information security officer; legal counsel |
| Suspend a tool for safety concerns | Chief medical information officer, acting alone if urgent | Committee reviews at next meeting |
| Set local thresholds, such as a prediction score that triggers an alert | Clinical owner of the tool | Nursing and physician users |
| Approve general-purpose generative tools for staff use | Committee | Chief information security officer; chief privacy officer |
The committee will include the chief medical information officer, chief nursing informatics officer, chief privacy officer, chief information security officer, a health equity leader, a legal representative, and two frontline clinicians. Assigning each decision to a role removes the ambiguity that let tools be adopted without anyone being accountable for their effects.
Training as an Obligation
Training on artificial intelligence tools will be required, not optional, for anyone who uses them. Clinicians using the sepsis prediction model must complete a short module explaining what the score predicts, how accurate it is at Summit Valley, and what it should not be used for, such as ruling out sepsis in a patient with a low score. Clinicians using the ambient documentation tool must complete training on reviewing drafted notes, since errors in drafts become errors in the record if signed without review. Access to each tool will be granted only after training is recorded, and refreshers will be required when a tool changes significantly.
Data Ownership Assigned
Each tool has a named clinical or business owner responsible for its performance, and data used by tools are assigned owners as well. Patient data remain governed by the health system's existing data stewardship rules; the privacy officer approves any use of patient data to develop or test a tool. Data produced by tools, such as sepsis scores and drafted notes, become part of the record or operational data and are owned by the relevant data steward. Contracts with vendors must state that Summit Valley retains ownership of its patient data, that vendors may not use it to train products for other customers without approval, and that data will be returned or destroyed at contract end.
Misuse Addressed
Governance must address misuse directly. Three forms are most likely. The first is entering patient information into public generative tools that are not covered by a business associate agreement, which would expose protected health information. The policy prohibits this, the information security team will block the most common public tools on the health system's network, and approved tools with appropriate agreements will be made available so that staff have a safe alternative. The second is over-reliance: treating a tool's output as a decision rather than an input, such as signing an ambient note without reading it. Audits of a sample of signed notes each month will check for errors. The third is use outside intended purpose, such as applying the no-show prediction to decide which patients get appointments, which could reduce access for groups already facing barriers. The committee's health equity leader will review each tool's use for this risk.
Performance Monitoring
Each tool's owner will report performance to the committee at least twice a year, including accuracy at Summit Valley, how often clinicians override or ignore it, any safety events associated with it, and whether its performance differs across patient groups. Tools whose performance declines, for example because the patient population or clinical practice has changed, will be reviewed and recalibrated or retired.
Communicating the Structure
Governance works only if staff know it exists. The structure will be announced by the chief medical officer and chief nursing officer together, with a short explanation of why it was created. A page on the staff intranet will list approved tools, the owner of each, the training required, and how to request a new tool or report a concern. Department leaders will be asked to route any vendor offering an artificial intelligence feature to the committee before a contract is signed, which closes the path by which several current tools arrived without review.
The Governance Reviewed
The governance structure itself will be reviewed after one year and every two years after that. The review will ask whether tools are being approved in reasonable time, whether suspended tools were handled well, whether training requirements are being met, and whether the structure has kept pace with new kinds of tools. Reviewing the governance itself prevents it from becoming a bottleneck or, just as risky, a formality.
Conclusion
Summit Valley's governance structure assigns decision rights by role, makes training an obligation, assigns data ownership, addresses the most likely forms of misuse, monitors tool performance, and reviews itself. It follows the national risk management framework and makes use of the transparency that federal rules now require of developers.
References
National Institute of Standards and Technology. (2023). Artificial intelligence risk management framework (AI RMF 1.0) (NIST AI 100-1). U.S. Department of Commerce.
Office of the National Coordinator for Health Information Technology. (2024). Health data, technology, and interoperability: Certification program updates, algorithm transparency, and information sharing. Federal Register, 89(6), 1192.
MHA FPX 5068 Assessment 3 instructions, in plain terms
The final assessment asks you to design governance for a health care technology after implementation. You identify what must be governed, assign decision rights, define training obligations, assign data ownership, address misuse, set performance monitoring and plan communication and review. For artificial intelligence tools, governance must also cover validation before use, bias across patient groups and transparency with clinicians and patients about when a tool influenced a decision. Frameworks such as the NIST AI Risk Management Framework and federal guidance on algorithms in health care support the plan. The plan should name roles and committees rather than general responsibilities. It closes the course's thread on technology leadership, so it should connect to the goals and adoption work from earlier assessments.
How the MHA-FPX5068 Assessment 3 example is put together
Decision rights, data ownership, training, accountability. The opening states what has to be governed after implementation, which is more than most organizations anticipate. A decision rights block names who approves changes, who arbitrates conflicts between groups and who can override. A data block assigns ownership of accuracy for each significant data set. A training block sets out the ongoing obligation: new starters, refreshers, competence checks and who verifies them. An accountability block covers what happens when the system is misused or bypassed. A review block states how the governance itself gets revisited. The closing names what the arrangement deliberately does not cover. Every element names a role rather than a committee, and privacy obligations are cited wherever they bear on access or on records.
Reading the MHA FPX 5068 Assessment 3 scoring guide
The scoring guide rates the scope of governance, the decision rights, the training, the data ownership, the misuse policy, the monitoring and the review. Decision rights reach Distinguished when roles are named for approving new tools, changing existing ones, resolving conflicts and overriding a tool's output. Training scores higher when treated as ongoing, since tools and staff both change over time. Data ownership earns credit when assigned to roles for inputs, outputs and the models themselves. Misuse should be defined, such as using a tool outside its validated population, with a response. Monitoring for performance drift and bias shows responsibility for AI tools. Communication and a scheduled review of the governance itself complete the plan. APA and writing are graded.
MHA FPX 5068 Assessment 3 help: mistakes that cost marks
Students often assume governance ends at go-live, when most problems with AI tools appear later. Another common mistake is assigning responsibility to committees without naming who actually decides. Training is a one-time event, so new staff never learn the tool's limits. Data ownership is unclear, especially for model inputs and outputs shared with a vendor. Misuse is not defined, so no one knows when a clinician has relied on a tool improperly. Monitoring for drift and bias is missing, even though patient populations shift. The governance structure itself is never reviewed. A one-page chart of who approves, monitors and decides for each AI tool makes the plan usable by busy leaders.
Get a MHA FPX 5068 Assessment 3 example written to your instructions
Send the Assessment 3 instructions and the MHA FPX 5068 scoring guide from your courseroom, with the evaluation and adoption plan your earlier assessments produced. We write a custom example against those exact criteria and return it in 24 to 48 hours. The first custom sample is free, and naming decision rights by role is what turns a governance proposal into something that would actually function. The paper above is an original model document written by our desk, not a submitted student paper and not an official Capella University document.
More MHA FPX 5068 papers and MHA samples
- MHA FPX 5068 Assessment 1: Meaningful Use of Health Care Technology
- MHA FPX 5068 Assessment 2: Winning Clinician Adoption of Remote Blood Pressure Monitoring
- MHA FPX 5001 Assessment 3: Professional Development Plan
- MHA FPX 5064 Assessment 3: Evaluating Referral Management Vendors
- MHA FPX 5006 Assessment 3: Cost-Benefit Analysis
- MHA FPX 5042 Assessment 3: Personal Leadership Targets Drawn From the Team's Needs
MHA FPX 5068 Assessment 3 questions, answered
What actually needs governing after go live?
Configuration changes, conflicts between what different groups need, data quality, access rights and training. Each of those generates decisions weekly in a live clinical system. Naming who makes each one is what prevents the default answer, which is whoever shouts loudest or whoever the vendor speaks to.
Why is training an ongoing obligation?
Because of turnover. In many healthcare settings a substantial share of the people using a system within a year were not employed at implementation. Training treated as a launch event leaves those staff learning from colleagues, which propagates workarounds and errors quietly through the organization.
How should misuse be handled?
With a stated route rather than a warning. What happens when somebody shares a login, bypasses an alert or documents late affects both safety and compliance. Naming who is told, what is recorded and what follows turns a governance document into something that functions when it is actually needed.
What needs governing after an AI tool goes live in MHA FPX 5068 Assessment 3?
Changes to the tool, who may use it and for which patients, training for new staff, data sharing, monitoring of accuracy and bias, handling of errors or misuse and periodic review of whether the tool is still worth using.
Where can I find a free MHA FPX 5068 Assessment 3 sample paper?
This page holds the full AI tool governance plan, with decision rights, training, data ownership, misuse, monitoring, communication and review. A custom plan for your technology is available free.