| Course | DNP 955 DPI Project: Part I |
|---|---|
| Topic | Topic 7 |
| Paper type | Measurement and analysis plan |
| Length | About 990 words, 3 pages |
| Format | APA 7 student paper |
| School | Grand Canyon University (GCU) |
| Program | DNP |
| Updated | October 2026 |
Free sample paper for DNP 955 Topic 7
Committed Before the Data Exist: The Measurement and Analysis Plan for a Statin Outreach Project, With Its Failure Line and What It Will Not Analyze
[Author Name]
College of Nursing and Health Care Professions, Grand Canyon University
DNP-955: DPI Project: Part I
Topic 7 Assignment
[Instructor Name]
July 16, 2029
The organization, setting and figures below are a composite written as a model document. No real employer, client, colleague or patient is described.
Committed Before the Data Exist: The Measurement and Analysis Plan for a Statin Outreach Project, With Its Failure Line and What It Will Not Analyze
The Measures
Table 1
Measures, Sources, Denominators, Intervals and Extractors
| Measure | Source | Denominator | Interval | Extracted by |
|---|---|---|---|---|
| Primary: statin order or fill in the prior 12 months | Network registry, current definition | Cohort patients still attributed on the measurement date | Monthly, March 2029 to February 2030 | Informatics analyst |
| New statin starts | Order and pharmacy records | Cohort patients with no statin on the first day of each month | Monthly | Informatics analyst |
| Outreach reach | Care manager call log | Patients assigned for outreach that month | Monthly | Lead care manager |
| Pended orders signed within 7 days | Order records | Orders pended | Monthly | Informatics analyst |
| Necessity-concern differential | Structured field for the medicines beliefs questionnaire | Patients reached who completed both calls | At each call | Care managers, compiled by me |
| Balancing: muscle symptom calls or visits within 90 days of a new statin | Encounter and message records | New statin starts | Monthly | Informatics analyst |
The plan specifies every measure before implementation begins in September 2029, so that no choice can be adjusted after the results are seen. The cohort is fixed on August 31, 2029: the 971 patients meeting the project definition from Topic 3 on that date. Table 1 gives each measure's source, denominator, collection interval and the person who extracts it.
Fixing the cohort on one date matters for interpretation. Patients who join the network or turn 40 after August 31 are not added, and patients who have a statin started by another clinician outside the project still count, because the measure is about the population, not about who caused the change. The analyst will send the extract on the fifth working day of each month, and I will check the first two extracts against a manual count of 30 charts before trusting the automated figures.
The Analysis, Specified Now
The primary measure is tracked on a run chart with 12 monthly points, six before and six after implementation, and the baseline median carried forward. The standard run chart rules, six consecutive points above the median for a shift and five consecutive increases for a trend, decide whether the chart shows a signal (Perla et al., 2011). The run chart is the main analysis because it shows when change occurred and whether it was already under way before September.
One statistical comparison is specified, chosen from the measurement level and design. The primary measure is a yes-or-no outcome recorded for the same patients at two points, August 31, 2029, and February 28, 2030, so the observations are paired, and McNemar's test for paired proportions is used rather than a chi-square test for independent groups. Alongside its p value, the plan reports the difference in proportions with a 95 percent confidence interval, because the size of the change matters more to the network than whether it crosses a threshold (Sullivan & Feinn, 2012). New starts per month are reported as counts on a run chart without a test. For the mechanism measure, the change in the necessity-concern differential between the two calls is compared between patients who started a statin and those who declined, using the Wilcoxon rank-sum test because the differential is a bounded score likely to be skewed, and reported as a median difference with its interval.
Missing Data, as Rules
Missing data are handled by rules stated now. Patients who leave the network's attribution before February 28 are excluded from the paired comparison and counted; if more than 10 percent of the cohort leaves, the comparison is repeated on the remaining patients and both results are reported. For the mechanism measure, a patient's scores are used only if both calls produced a complete questionnaire. If fewer than 70 percent of patients reached have both, the mechanism analysis is reported descriptively, as medians by group without a test, and the proportion missing is stated with it. No values are imputed, because the questionnaire data would be too sparse to impute sensibly and imputation would imply more certainty than the project has.
What Would Count as Failure
The plan states what result would count as failure, before anyone knows the result. The project will be judged to have failed on its primary aim if the proportion in the cohort with a statin order or fill rises by less than 5 percentage points between August 31 and February 28, or if the run chart shows no shift or trend. It will be judged to have failed on feasibility if care managers reach fewer than half of assigned patients across the six months. It will be judged to have caused harm if muscle symptom contacts among new starts exceed 15 percent within 90 days, which would suggest that patients were started without adequate discussion of what to expect. The 5-point threshold is not arbitrary: it is roughly the effect of the clinician-directed nudge in the most relevant trial, and a smaller change would not justify the care managers' time (Adusumalli et al., 2023).
What Will Not Be Analyzed
The dataset will permit many more comparisons than the project can honestly run, and each additional comparison raises the chance of a false finding. The plan therefore specifies what will not be analyzed. There will be no comparison by clinic, prescriber or prescriber type, except the two-clinic report if the feasibility rule narrows the outreach. There will be no subgroup analyses by sex, race, ethnicity or insurance within the primary measure, because the cohort is too small to support them with any precision; the numbers in each group will be reported for description only. LDL cholesterol will not be analyzed, because it is measured irregularly in this age group and would be available for too few patients. Adherence beyond the first fill will not be analyzed, because pharmacy data cover only about three in five patients. Any of these might be interesting. None was part of the question, and reporting them would invite readers to treat exploratory patterns as findings.
References
Adusumalli, S., Kanter, G. P., Small, D. S., Asch, D. A., Volpp, K. G., Park, S. H., Gitelman, Y., Do, D., Leri, D., Rhodes, C., VanZandbergen, C., Howell, J. T., Epps, M., Cavella, A. M., Wenger, M., Harrington, T. O., Clark, K., Westover, J. E., Snider, C. K., & Patel, M. S. (2023). Effect of nudges to clinicians, patients, or both to increase statin prescribing: A cluster randomized clinical trial. JAMA Cardiology, 8(1), 23-30. https://doi.org/10.1001/jamacardio.2022.4373
Perla, R. J., Provost, L. P., & Murray, S. K. (2011). The run chart: A simple analytical tool for learning from variation in healthcare processes. BMJ Quality & Safety, 20(1), 46-51. https://doi.org/10.1136/bmjqs.2009.037895
Sullivan, G. M., & Feinn, R. (2012). Using effect size, or why the P value is not enough. Journal of Graduate Medical Education, 4(3), 279-282. https://doi.org/10.4300/JGME-D-12-00156.1
DNP 955 Topic 7 instructions, in plain terms
Topic 7 in DNP-955 asks students to plan measurement and analysis for their project. Strong plans specify every measure and its definition before implementation, fix the cohort on a stated date and describe the analysis precisely enough that no choice can be adjusted after the results are seen. Students should write missing data rules with thresholds, set a failure line before seeing results and name analyses they will not run, since a dataset always permits more comparisons than a project can honestly make. The plan should match the design, for example using run chart rules for monthly data from a single site, and report effect sizes as well as significance.
How the DNP-955 Topic 7 example is put together
The finished example commits in advance. Each measure carries its source, its denominator, its collection interval and the person who extracts it. The analysis is then specified before any data exists: which comparison, which test, chosen from the measurement level, and what will be reported alongside significance. Missing data is addressed with a rule rather than an intention, stating the threshold above which a case is excluded and how that exclusion will be reported. The plan states what result would count as failure, which most drafts avoid. It also specifies what will not be analyzed, since a dataset always permits more comparisons than the project can honestly run. Every specification in the plan is written in a form somebody else could execute without asking questions.
DNP 955 Topic 7 rubric: what Excellent looks like
Faculty reward a plan committed before data exist. An Excellent plan fixes the cohort of 971 on August 31, 2029 and specifies a 12-point run chart with Perla and colleagues' rules. Credit goes to missing data rules, such as excluding patients who leave attribution and flagging attrition above 10 percent. The rubric values a failure line stated in advance and effect sizes after Sullivan and Feinn. Strong plans compare results with Adusumalli and colleagues' trial effect. Top marks go to naming subgroup analyses that will not be run.
Common DNP 955 Topic 7 mistakes, and how to avoid them
Analysis plans written as intentions rather than specifications are the standard version, and they leave every decision to be made once the results are visible. A second failure is omitting a missing data rule, which means exclusions get decided case by case in a way nobody could reproduce. Marks also go for plans that name a test without justifying it from the measurement level. Plans that do not close off unplanned comparisons invite analysis continuing until something reaches a threshold. Measures carrying no extraction owner run straight back into the availability problem the earlier courses spent their length on. Plans with no failure criterion make a negative result impossible to declare. Intervals described loosely leave a collection schedule nobody can build a roster around.
Get a DNP-955 Topic 7 example written to your instructions
Send the DNP-955 Topic 7 instructions and the rubric your classroom posts, with your measures and design. We write a custom example to those criteria, specifying the analysis before collection, with a missing data rule, ruled out comparisons and a failure criterion, in 24 to 48 hours. The first is free.
More DNP 955 papers and DNP samples
- DNP 955 Topic 1: Practice Problem Revisit
- DNP 955 Topic 2: Project Decision Record
- DNP 955 Topic 3: Clinical Question and Population Tightening
- DNP 955 Topic 4: Literature Arranged by Argument
- DNP 955 Topic 5: Framework and Design Defense
- DNP 955 Topic 6: Site Approval DQ Pair
- DNP 955 Topic 8: Proposal Assembly and Consistency Pass
- DNP 960A Topic 7: Interim Data Review
- DNP 840 Topic 7: Ethical Leadership Case
- DNP 965A Topic 7: Dissemination Plan
- DNP 838A Topic 7: Curriculum Change Proposal
DNP 955 Topic 7 questions, answered
Why specify the analysis before collecting anything?
Because decisions made after seeing data are made differently, without anyone intending it. Choosing a test, a subgroup or a threshold once the figures are visible tends to produce the more favorable option. Committing in advance is what lets you report a result as a genuine finding rather than as the best of several attempts.
What should a missing data rule say?
The threshold above which a case is excluded, what happens below it, and how exclusions get reported. Deciding case by case during analysis produces a process nobody could repeat and a sample nobody can characterize. A stated rule takes one sentence and makes the eventual sample defensible.
Why name analyses I will not run?
Because a dataset always permits more comparisons than the project planned, and running them until one reaches a threshold is a well documented way to reach a false conclusion. Naming the comparisons you are closing off, and sticking to it, is the strongest evidence a reviewer has that the primary result means what you say it does.
Where can I find a free DNP 955 Topic 7 sample paper?
The whole statin measurement plan is on this page: the measures, the analysis specified now, missing data rules, what would count as failure and what will not be analyzed. Paste in your DNP 955 Topic 7 prompt and a free plan follows.
Why name analyses that will not be run in DNP 955 Topic 7?
Because each extra comparison raises the chance of a false finding. Listing them in advance shows that any later result was not found by searching.