Midway through a 192-hour implementation block the question stops being whether the change is a good idea and becomes whether it is actually happening, in the form it was designed, to the people it was designed for. That is fidelity and adoption, and the written work at this stage reports both with counts rather than impressions. This is the stage the extra hours in a 3-credit block exist to fund, because process evaluation takes observation time that a lighter block cannot spare. Your section may print this as NR 707B or NR707B; 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-707B Week 4 asks for
On a medical-surgical unit running a new post-discharge callback, a project can look healthy at a distance and be hollow underneath. The calls are being made. The completion rate looks respectable. Then you sit with the outreach nurse for two afternoons and discover that the structured question set is being compressed into a single "how are you feeling," that the medication reconciliation prompt is skipped whenever the call runs past four minutes, and that the escalation pathway has never once been used because nobody is sure who answers it. The intervention that is running is not the intervention that was designed. Writing that distinction clearly, without either alarm or excuse, is the doctoral task of this stage.
The boundary that governs this manual. The hours you spend observing, the log that records them, encounter counts, site paperwork, and preceptor or mentor evaluations are your own record of work you personally performed, verified by your school and your organization. They are never drafted, reconstructed, estimated, or shaped with anyone's help. What a manual can teach is the written layer: how a fidelity audit is reported, how adoption is distinguished from reach, how a shortfall is described so that it reads as analysis. Every observation that reaches your page is de-identified, at the level of roles, counts, and aggregates.
Four terms carry this stage and they are routinely collapsed into one another. Reach is what proportion of the eligible population the change touched at all. Adoption is what proportion of the intended deliverers actually used it. Fidelity is whether the change, where delivered, was delivered as designed. Dose is how much of it each person received. Keeping them separate is not pedantry; it is the only way to explain a result. A change that reached ninety per cent of admissions but was delivered at half its intended content is a completely different finding from one that reached forty per cent and was delivered perfectly, and the two demand opposite responses.
Expect the deliverable to be a process evaluation or fidelity write-up, often with a small table of counts, sometimes embedded in a mid-block progress document, and frequently accompanied by a posted discussion about what is and is not working. The tone that scores is neither defensive nor apologetic. It is the tone of somebody reporting instrument readings.
The NR-707B Week 4 method, step by step
Six analytic moves for reporting whether the change is really happening.
-
1. Partition reach, adoption, fidelity, and dose
Give each its own number and its own denominator. Reach counts patients or encounters; adoption counts deliverers; fidelity counts components delivered as designed; dose counts intensity. Four numbers, four sentences, no blending.
-
2. Decompose the intervention into checkable components
List the discrete elements the change is made of, then say for each whether it happened. A callback with four required elements produces a component-level table, and that table is what turns "partial adherence" into a finding somebody can act on.
-
3. Verify a sample directly rather than trusting the count
Take a defined sample, state the sampling rule, and check the underlying record or observe the delivery yourself. Report how many you checked and how many matched. Direct verification of twenty cases outranks a system-generated rate over two hundred.
-
4. Attribute each fidelity gap to a cause you can name
Time pressure, unclear ownership, a build that buries the field, a component nobody understood, or a component that genuinely does not fit the workflow. The cause determines the fix, and a gap reported without a cause invites the reader to supply their own.
-
5. Distinguish drift from adaptation
Drift is erosion under pressure. Adaptation is a deliberate local modification that preserves the mechanism. Both change the intervention, but only one is defensible, and the write-up should say which occurred and who decided it.
-
6. Convert findings into a stated mid-block correction
Name the one or two things you are changing in response, when they take effect, and how you will know whether they worked. A mid-block correction dated in the record is exactly the material your week seven interpretation will need.
A layout for a fidelity and adoption report
Our frame for a mid-block process evaluation, sized for roughly 1,200 to 1,500 words plus a component table. It is our own outline rather than anything the university issues, and your week's rubric outranks it wherever they disagree.
| Section | What belongs in it | Word target |
|---|---|---|
| What was supposed to happen | The intervention broken into its required components, stated once, in the form the design specified. | 150 to 190 |
| Reach and adoption | Eligible population touched, deliverers who used the change, each with counts and denominators. | 190 to 230 |
| Component-level fidelity | For each component, delivered as designed in how many of how many verified cases. | 230 to 280 |
| Verification method | Sampling rule, how many records or observations were checked, by whom, and against what standard. | 160 to 200 |
| Causes of the gaps | Each shortfall matched to a named cause, with the observation that supports the attribution. | 230 to 280 |
| Mid-block correction | What changes, when, who owns it, and what would show in the next two weeks if it worked. | 170 to 210 |
Evidence craft for process evaluation writing
Counts with denominators, in every sentence that carries a claim. "Component three was delivered in 11 of the 19 verified encounters" is a finding. "Fidelity to component three was inconsistent" is a summary of a finding you did not write down. On unit-scale work the numbers are small, and small numbers reported honestly are more persuasive than percentages that conceal them.
Say how you know, immediately after you say what happened. Direct observation, chart verification, system report, or staff self-report are four different grades of evidence, and a reader assigns weight based on which one produced the number. Self-reported delivery is usually the most generous and should be labeled as such rather than presented alongside verified counts without comment.
Use a named process-evaluation framework and hold to its terms. Implementation literature provides established vocabularies for reach, adoption, fidelity, and maintenance. Attributing one with a year gives your section a structure a doctoral reader can follow, and it stops you from inventing categories that overlap.
Keep the improvement frame intact. This is a local evaluation of a local process, not a trial of an intervention. Nothing here supports a claim about what would happen at another hospital, and the language should not drift toward efficacy. Where a determination about human subjects oversight applies at your organization, describe the process you followed rather than predicting its outcome.
Report the uncomfortable number first. A fidelity section that opens with the strongest component and buries the weakest reads as advocacy. Opening with the largest gap and then explaining it reads as evaluation, and it is the pattern that consistently scores higher because it signals that you are the project's most rigorous critic.
Five mistakes that cost points in this week's territory
- One adherence number for a multi-component change. An average across components hides exactly the thing the reader needs, which is which component is failing.
- Self-report presented as verification. Staff recollection is useful context and weak evidence, and mixing it silently with checked counts damages the whole section.
- Blaming staff for a design problem. A component that is skipped by everyone is usually a component that does not fit the work, and saying so is stronger analysis than describing non-compliance.
- Adaptation described as failure. A local modification that preserves the mechanism is legitimate, and it belongs in the record as a decision rather than as an apology.
- Findings with no correction attached. Mid-block is the last point where a correction can still influence results, and a report that ends at description wastes it.
Before you submit
- Reach, adoption, fidelity, and dose each appear as separate numbers with their own denominators
- Fidelity is reported at the component level, not as a single average
- The verification method, sample size, and sampling rule are stated
- Every gap is matched to a named cause with an observation behind it
- Drift and deliberate adaptation are distinguished, with the decision-maker's role named
- A dated mid-block correction is written with an owner and a check point
Writing the fidelity report for NR-707B?
Send the rubric and your verification notes out of Canvas. A premium original draft comes back in 24 to 48 hours with component-level counts, stated verification methods, and gaps matched to causes, and revisions run until the grade lands.