NR-704 · Week 3 of 8 · Surveillance sources and rate construction

NR-704 Week 3 Surveillance and Rate Construction: How to Write It

The short answer

The middle of the first half of this course is where population health stops being conceptual and becomes arithmetic. The written work is measurement reasoning: choosing a surveillance or administrative data source, understanding what that source can and cannot see, constructing a rate correctly, and interpreting a comparison without overreading it. Doctoral graders are watching for whether you know the difference between a count and a rate, between a crude and an adjusted rate, and between a real change in an outcome and a change in how the outcome was captured. Your section may print this as NR 704 or NR704; 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.

NR-704 Week 3 grading scale at Chamberlain, the criterion levels this assessment is scored on, from Chamberlain Tutors
How Chamberlain grades NR-704 Week 3, visualized by Chamberlain Tutors.

What NR-704 Week 3 asks for

A quality director in a long-term care organization brought us a paragraph she was proud of. Her facility's pressure injury count had fallen from 31 to 19 across two quarters, and she had written it up as a 39 percent improvement. Then the census came out. Occupancy had dropped by nearly a fifth over the same period because a wing was closed for renovation, and the number of resident-days at risk had fallen with it. Expressed per 1,000 resident-days, the rate had barely moved. The count told a story of success; the rate told a story of a smaller building. Nothing dishonest had happened. A denominator had simply been left out.

This stage exists so that mistake never reaches your project. The territory is the measurement layer of population health: national and state surveillance systems, publicly reported facility and hospital measures, registries, mandated assessment instruments, claims-derived indicators, and the internal reports your own organization already produces. Each of these sees a different slice of reality through a different aperture, and a doctoral paper is expected to say which aperture it is looking through before it reports a number.

Deliverables at this depth usually involve locating real data on your population, presenting it in a table or figure, and writing an interpretation. The interpretation is what is being scored. A table anyone could copy from a public dashboard demonstrates nothing; a paragraph explaining that the apparent rise in a condition follows a coding change rather than a change in incidence demonstrates the competency the course is built around. Where a discussion runs this week, hold the same standard, because a misreported rate is easy to check and posts do not reopen once submitted in Canvas.

Three distinctions carry most of the marks. Count against rate, already illustrated. Crude against adjusted, which matters enormously when populations differ in age or acuity, since an older facility will look worse on almost any outcome until age is accounted for. And numerator ascertainment against true occurrence, which is the question of whether a change in the measure reflects a change in the world or a change in surveillance intensity. Write those three explicitly and most of the analysis row takes care of itself.

The NR-704 Week 3 method, step by step

Six moves for writing about population data without overreading it.

  1. Reduce the rubric rows to verbs before opening any dataset

    Locate, present, interpret and evaluate each require different amounts of prose. A row asking you to evaluate the quality of a data source is asking about completeness, lag and case definition, not about whether the website was easy to use.

  2. Name the source with its custodian, its year and its case definition

    Who collects it, from whom, under what reporting obligation, and how a case is defined for inclusion. A rate cannot be interpreted without its case definition, and two sources with different definitions will disagree by design.

  3. Build the rate deliberately, one part at a time

    Write the numerator, then the denominator, then the period, then the multiplier. Choose a denominator that reflects exposure rather than headcount where the two differ: resident-days, patient-days or person-years capture time at risk in a way a simple census does not.

  4. Decide whether adjustment is required and say so

    If you are comparing populations that differ in age, acuity or case mix, a crude comparison is misleading and you should either use an adjusted figure the source provides or state clearly that your comparison is crude and in which direction the bias runs.

  5. Interrogate any change before you explain it

    Ask four questions of every rise or fall: did the case definition change, did surveillance intensity change, did the denominator change, and is the movement larger than the ordinary variation between periods. Only then reach for a substantive explanation.

  6. Write the interpretation as a claim with its limits attached

    One sentence saying what the data supports, one saying what it does not, and one saying what additional information would settle the question. That three-part close reads as doctoral judgment and is quick to write once the analysis is honest.

A layout and word budget for a data and measurement paper

The frame our tutors keep beside a surveillance and rates submission, sized for roughly 1,300 to 1,600 words plus a table. It is our own outline rather than anything the university issues, and your week's rubric outranks it wherever the two disagree.

SectionWhat belongs in itWord target
Question and populationThe measurement question stated as a question, with the denominator carried forward from earlier work in one clause.90 to 120
Source descriptionCustodian, collection route, reporting obligation, case definition, reporting lag and known completeness issues.250 to 300
Rate constructionNumerator, denominator, period and multiplier written out, with the reason this denominator represents exposure.200 to 250
PresentationThe table or figure, plus the sentences that tell the reader what to look at first and in which direction the axis runs.160 to 200
InterpretationWhat the pattern supports, the artefactual explanations ruled out, and any stratification that changes the picture.320 to 400
Limits and next dataWhat this source cannot see, and the one additional measure that would most sharpen the conclusion.180 to 220

Evidence craft for population measurement writing

Choose exposure-based denominators wherever time matters. Events per 1,000 resident-days behaves correctly when length of stay varies; events per admission does not. State which you used and why, because the choice changes the number materially and a grader reading a facility-level analysis will look for it.

Treat small numbers with visible caution. Rates built on a handful of events swing wildly for reasons that have nothing to do with care. Where your numerator is small, say so, give the raw counts alongside the rate, and resist writing a trend across three unstable points. Doctoral judgment shows more in restraint here than in analysis.

Distinguish measures of process from measures of outcome in every sentence. The proportion of residents who received an intervention is a process measure. The proportion who experienced the event is an outcome measure. Blending them in one paragraph makes it impossible for a reader to tell whether delivery or effect is being described.

Cite the dataset itself, not only the article that mentioned it. If you pull a figure from a public reporting system, attribute the system, the extraction date and the reporting period. Public data updates on its own schedule, and a number without an extraction date cannot be reproduced by the person grading it.

Five mistakes that cost points in this week's territory

  • Counts presented as performance. Fewer events with a shrinking denominator is not improvement, and this is the error graders in outcomes courses look for first.
  • Crude comparisons between unlike populations. Comparing a high-acuity facility to a state average without acknowledging case mix produces a conclusion the data cannot support.
  • Trends drawn through noise. Three quarters of small numbers is not a trend line, and calling it one invites a direct challenge in feedback.
  • A table with no reading instructions. Dropping a dashboard export into the paper and moving on leaves the interpretation row unearned.
  • Ignoring reporting lag. Presenting the most recent published year as current, when the source runs eighteen months behind, misstates the present state of your population.

Before you submit

  • Every rate shows numerator, denominator, period and multiplier
  • The data source is named with custodian, case definition and extraction date
  • Adjustment is either used or its absence is declared with the direction of bias
  • Artefactual explanations for any change are considered before substantive ones
  • Small numerators are flagged and raw counts appear beside the rates
  • The paper closes with what the source cannot see

Working with population data for NR-704?

Send the rubric and the dataset details out of Canvas. A premium original draft comes back in 24 to 48 hours with rates constructed properly and an interpretation that rules out the artefacts first, and revisions run until the grade lands.

Questions students ask about this stage

Am I allowed to use my employer's internal quality reports as data?
Often yes for coursework, with three conditions that protect both you and the setting. Report only aggregate figures, never anything traceable to an individual resident, patient or staff member. Check whether your organization requires permission before internal performance data appears in an academic paper, because many do and the requirement is easy to satisfy in advance and painful to satisfy afterwards. And describe the report honestly, including how the measure is defined internally, since in-house definitions frequently differ from the public ones and a reader comparing your figure to a national benchmark needs to know that. Where permission is not available, published data on a comparable population supports the same reasoning, and saying why you used it is a legitimate line in the paper rather than an admission of weakness.
What do I do when two credible sources report different rates for the same thing?
Treat the disagreement as the finding and go looking for its cause in the case definitions. Sources diverge for predictable reasons: they define the event differently, they cover different populations, they use different denominators, they capture different time periods, or one relies on voluntary reporting while the other is mandated. Lay the two definitions side by side in your own words, identify which difference is doing the work, and then say which figure is the appropriate one for your question and why. That paragraph demonstrates exactly the competency the stage is testing. What does not work is averaging the two, or picking the one that better supports your argument without mentioning that the other exists, since a grader familiar with the field will know both.
How much statistics do I need for a rate-based paper at doctoral level?
Less computation than you expect and more interpretation than you are used to. You need to construct a rate correctly, understand why adjustment exists and when it is required, read a confidence interval as a statement about precision, and recognize when a difference between two periods is too small relative to ordinary variation to be worth an explanation. You do not need to run inferential tests on a facility's quarterly numbers, and doing so in a paper about a handful of events usually signals a misunderstanding rather than sophistication. The graded skill is measurement judgment: knowing what a number can carry. Where your section does ask for a specific calculation, follow the instructions given there, since requirements vary between sections and the rubric outranks any general guidance.

Keep going

Online now