NR-709A Week 5 Fidelity and What Was Delivered: How to Write It

What NR-709A Week 5 asks for

How much can one intervention vary while still being called the same intervention? A community health worker program delivering a post-discharge check-in in patients' homes had four workers, a one-page script and a shared expectation of a thirty-minute visit. In practice one worker completed the script in eleven minutes because her caseload had grown, one had extended into an hour and was doing medication reconciliation that nobody had asked for, one had translated the script into the language most of her patients spoke and had changed several questions in doing so, and one had stopped using the script at all because she found it made people uncomfortable. Every one of those adaptations was reasonable. Together they mean that the program's outcome data describes four different interventions averaged together.

Doctoral fidelity writing is not about catching anyone out. It is about being able to say what the results describe. The standard components are dose, meaning how much was delivered; reach, meaning what proportion of the eligible population received it; adherence, meaning how closely delivery matched the design; and adaptation, meaning what was deliberately changed and why. Adaptations in particular deserve respect rather than apology, because frontline modification is usually how a change becomes deliverable at all, and documenting it is what makes the change reproducible elsewhere.

Expect a deliverable that reports implementation alongside outcome, or a section that does so within a growing report. The writing should be structural rather than personal throughout: describe roles, workflows and conditions, not individuals. In a document that will be read at your site, that discipline is both an ethical obligation and the thing that keeps the analysis usable.

The boundary that governs this manual. Clinical hours, hour logs, encounter counts, preceptor evaluations, signatures and site documentation are yours and your site's, never drafted, reconstructed or estimated with outside help, and no one else performs, observes or documents clinical activity on your behalf. The supportable layer is the writing: describing implementation accurately, structuring a fidelity account, and analyzing work you genuinely did. Every patient and staff detail is de-identified before it reaches a page, and the setting is described by type and volume rather than by name.

The NR-709A Week 5 method, step by step

Seven moves for writing a fidelity account that changes how the results are read.

1. Write the intervention as designed, in components

   Break it into the discrete things that were supposed to happen: a screening step, a documentation action, a referral, a follow-up contact, a training session. Fidelity can only be assessed component by component, never as a whole.

2. Report reach as a proportion with its denominator

   How many of the eligible people received the component, out of how many eligible. Reach is where most improvement projects lose their effect, and a project reaching a third of its population cannot produce a population-level result.

3. Quantify dose where the component has an amount

   Number of contacts, minutes, sessions attended, messages sent. Where dose varied widely, report the distribution rather than the mean, because an average of two contacts can mean everyone got two or half got none.

4. Assess consistency across the people and settings delivering it

   Compare delivery across roles, sites, shifts or clinic sessions using whatever evidence exists. Report variation structurally, by role or by session, never by naming individuals.

5. Build an adaptation log with dates and reasons

   What was changed, when, at whose initiative, and what problem it solved. Then classify each adaptation as one that preserved the core function or one that altered it, because those have different implications for your conclusion.

6. Identify the conditions that made delivery possible or impossible

   Staffing levels, room availability, competing initiatives, a champion's presence, a workflow that placed the step where it fit. These are the variables that determine whether your change survives, and this stage is where you observe them.

7. State explicitly what the fidelity picture means for interpreting the outcome

   One paragraph: given this reach, this dose and this consistency, here is what the outcome data can and cannot tell us. That paragraph is often the most valuable in the entire report.

A layout and word budget for a fidelity and implementation account

Our frame for an implementation section, sized for roughly 1,200 to 1,500 words. It is our own outline rather than anything the university issues, and your week's rubric outranks it wherever the two disagree.

Evidence craft for implementation reporting

Use a named implementation framework and apply its categories. Published frameworks exist for describing implementation outcomes and contextual determinants, and adopting one gives your section a structure a reader recognizes. Cite it with author and year, and use its terms as defined rather than loosely.

Report fidelity from records where you can, and label observation as observation. Documentation-based measures are checkable; your impressions from being present are valuable and are a different class of evidence. Mark which is which, and where you rely on what you saw, say how often you were there and in what capacity.

Distinguish adaptation from drift in your language. A deliberate modification made for a stated reason and communicated to the team is an adaptation. An unnoticed decay in how consistently something is done is drift. Both belong in the paper and they carry opposite implications for whether the change can be maintained.

Never attribute variation to individual quality. Write that delivery was lower during sessions where a role was covered by float staff, or that a component requiring an unassigned handoff was completed inconsistently. Statements about how conscientious particular people are have no place in a doctoral report and destroy the working relationships the project needs.

Handle staff information as carefully as patient information. In a small clinic, delivery data broken down by role can identify one person. Aggregate, suppress small cells, and describe by function rather than by title where a title is held by a single individual. De-identification obligations run in every direction here, not only toward patients.

Five mistakes that cost points in this week's territory

• Fidelity asserted rather than measured. The intervention was implemented as planned, with no reach or dose figures behind it, is an assumption wearing the clothes of a finding.
• Reach omitted entirely. Without it, a modest outcome change cannot be distinguished from a change delivered to a small fraction of the population.
• Adaptations treated as failures. Frontline modification is usually how a change survives, and framing it apologetically hides the most transferable knowledge in the project.
• Individual attribution. Naming or implicating staff is an ethical problem and converts a system analysis into a personnel document.
• No link drawn to the outcome. A fidelity section that does not say what the implementation picture means for the results has left the analytic work undone.

Before you submit

• The intervention is broken into named components with responsible roles
• Reach is reported per component with numerator, denominator and period
• Dose is reported as a distribution where amount varies
• Delivery variation is described by role, site or session, never by person
• Every adaptation carries a date, a reason and a judgment about core function
• Adaptation and drift are distinguished explicitly
• Enabling and constraining conditions are described as system properties
• A paragraph states what this fidelity picture means for the outcome
• Staff-level data is aggregated so no individual is identifiable

Questions students ask about this stage

Keep going