The closing stage of NR-717 usually asks how you would know whether the policy worked and whether it would still be working in three years. Both halves are graded. Evaluation means specifying measures, comparison and timing before implementation, and being clear that a policy evaluation is not a controlled trial and should not be written as though it were. Sustainability means naming what keeps a change alive after the attention leaves: funding, ownership, documentation, and the measure that would show decay. Your section may print this as NR 717 or NR717; it is the same course. Chamberlain publishes no syllabi outside Canvas. The placement here is our teaching judgment from the course's catalog arc; your section's rubric decides what your week actually asks.
What NR-717 Week 8 asks for
Eighteen months after a health system adopted a standing-order protocol for a preventive service, an internal audit went looking for the improvement that had been reported at the six-month mark. Completion had risen from 41 percent to 68 percent in the first two quarters, and the project had been closed as a success. The audit found completion back at 47 percent. Nothing had been repealed. The protocol was still in the policy manual, unchanged. What had changed was everything around it: the two nurses who had championed it had moved units, the electronic prompt had been reworded during a system upgrade in a way that moved it below the fold, the monthly report that had carried the measure had been retired when the reporting analyst left, and no committee held the item on a standing agenda. The policy survived. The practice did not.
That audit is the whole argument of this stage. Population health policy work is judged on outcomes, and outcomes are not events that happen at a launch. The written product here is usually an evaluation and sustainability plan attached to the recommendation you have been developing, sometimes combined with a reflective component on your development in policy and population health across the session. Whatever the exact artifact, two things get graded: whether your evaluation design could actually detect the effect you are claiming, and whether your sustainability plan is made of mechanisms rather than intentions.
Doctoral precision matters most here, and it is the place where practice-doctorate writing most often drifts into language it has not earned. Your evaluation of a policy change is a quality and outcomes evaluation, not a research trial. You will usually have no randomization, no control group and no blinding, and secular trends will be running alongside your change. That is normal and acceptable, and the correct response is to design the evaluation so those threats are visible and partially addressed, then write about the results in language that matches the design. Claiming that a policy caused an improvement on the basis of a before-and-after comparison at one site is the single most common overreach in this territory, and it is entirely avoidable.
The same discipline applies to human subjects language. Describing how a quality or policy evaluation would be reviewed under your organization's determination process is appropriate; promising a specific determination outcome is not, and no paper should assert what a review body will decide.
The NR-717 Week 8 method, step by step
Six moves for an evaluation and sustainability plan that survives scrutiny.
-
Write the measure set at three levels
Structure or process measures that show the policy is being followed, an intermediate outcome the policy should move first, and the population outcome you ultimately care about. Specify numerator, denominator, source and frequency for each. Most policy changes cannot show a population outcome shift inside a project window, and saying so is honesty rather than weakness.
-
Add the equity stratification to every measure by design
The gap you documented earlier can widen even while the average improves, which happens when a change is easier to take up for people who already had fewer barriers. Stratify from the first data pull rather than adding it later, and say what result would count as improvement without equity gain.
-
Choose a comparison and name what it cannot rule out
Options include a long baseline series, a comparison unit or site that did not adopt the change, a staggered rollout, or a non-equivalent control measure that should not move if your mechanism is real. Any of these is stronger than a single before-and-after pair. Then write the sentence that names the secular trend or concurrent initiative your design cannot exclude.
-
Set the timing before you see any data
When measurement starts, how long the baseline runs, when the first look happens, and what you will do if the early signal is negative. Deciding these afterwards is how projects end up reporting whichever window flattered the result, and pre-specification is the cheapest credibility you can buy.
-
Build sustainability out of four named mechanisms
A funding source beyond the pilot period, a named accountable owner attached to a role rather than a person, embedding in a system that operates without enthusiasm such as an order set or an automated report, and a recurring review with the measure on a standing agenda. Anything softer than these four is an intention.
-
Write the decay plan
Say which measure would signal that the change is eroding, what threshold triggers attention, and who is responsible for looking. The audit above happened by luck. A plan that specifies the trigger and the owner turns that luck into a mechanism, and it is exactly the addition that separates a doctoral closing paper from a project summary.
A layout and word budget for an evaluation and sustainability plan
Our frame for a closing plan, sized for roughly 1,600 to 1,900 words plus a measure table. If your section pairs this with a reflection on your own development, budget that separately at the assigned length. The outline is ours rather than anything the university issues, and your week's rubric outranks it wherever the two disagree.
| Section | What belongs in it | Word target |
|---|---|---|
| What success would look like | The claim you intend to be able to make afterward, written before the design so the design can be judged against it. | 150 to 190 |
| Measure specification | Process, intermediate and population measures, each with numerator, denominator, source, frequency and stratification. | 350 to 420 |
| Comparison and design | The comparison chosen, why it beats a simple before-and-after, and what it still cannot rule out. | 300 to 360 |
| Timing and decision points | Baseline length, measurement schedule, the first look, and the pre-specified response to an early negative signal. | 200 to 250 |
| Sustainability mechanisms | Funding, accountable role, system embedding and recurring review, each named concretely. | 300 to 360 |
| Decay detection | The signal measure, the threshold, the owner of the watch, and the escalation route. | 180 to 230 |
Evidence craft for policy evaluation writing
Match your verbs to your design, every time. A single-site before-and-after supports was followed by and was associated with. It does not support caused, produced or reduced. This is the most reliable place to gain or lose credibility in a closing paper, and it costs nothing to get right.
Present rates over time rather than two points. Several baseline periods and several follow-up periods let a reader see whether the change was already underway before your policy, which is exactly the question a two-point comparison hides. Where you can only obtain two points, say what a longer series would have shown you that you cannot know.
Count the denominators after the change as carefully as before. Policy changes sometimes alter who is counted, not only what happens to them. If a new protocol changes documentation, coding or eligibility, your denominator may shift and produce an improvement made entirely of measurement. Write the check you would run to rule that out.
Describe review processes without predicting their outcome. It is correct to write that the evaluation would be submitted to the organization's review process for a determination of whether it constitutes human subjects research, and to describe the data protections and de-identification you would apply. It is not correct to state what determination would be returned. Keep that sentence conditional.
Five mistakes that cost points in this week's territory
- Causal language on an uncontrolled design. The fastest credibility loss available at this stage and the easiest to avoid.
- Population outcomes promised inside a short window. Mortality and long-horizon outcomes will not move in a project period, and claiming they will signals a misunderstanding of the measure.
- Unstratified evaluation of an equity intervention. If the whole argument was about a gap, an aggregate result cannot answer whether the gap closed.
- Sustainability written as commitment. Ongoing leadership support and a culture of accountability are not mechanisms and cannot be audited.
- No plan for a negative result. A paper that only describes what happens if it works has not designed an evaluation; it has designed a celebration.
Before you submit
- Measures appear at process, intermediate and population levels with full specifications
- Every measure carries its equity stratification from the first data pull
- The comparison is stronger than a two-point before-and-after, or the limitation is stated
- Verbs match the design throughout, with no causal claims from uncontrolled comparisons
- Timing and the response to an early negative signal are pre-specified
- Sustainability rests on funding, a named role, system embedding and a recurring review
- A decay signal, threshold and owner are named
- Review processes are described without any predicted determination
Closing out NR-717?
Send the rubric and your recommendation out of Canvas. A premium original draft comes back in 24 to 48 hours with measures specified, verbs matched to the design and sustainability written as mechanisms, and revisions run until the grade lands.