DNP Project Guide: Structure, Topics, Data, and Ideas

A DNP project is the final scholarly project of a Doctor of Nursing Practice program. Instead of generating new knowledge, as a PhD dissertation does, it takes existing evidence and uses it to improve practice or outcomes in a real clinical or organizational setting, then evaluates whether the change worked. This guide explains how a DNP project differs from a dissertation, how it is usually structured, how to choose a topic, how to plan the data and evaluation, and gives 12 example project ideas.

DNP project vs PhD dissertation

The DNP is a practice doctorate; the PhD in nursing is a research doctorate. The final projects reflect that difference.

Programs differ in what they call the project (DNP project, scholarly project, final project) and in their exact requirements. AACN guidance describes the final DNP project as one that demonstrates the integration of evidence into practice and leads to a tangible, evaluable outcome. Your program handbook is the document that governs your project, so read it before choosing a topic.

Typical structure of a DNP project paper

Most programs use a chapter or section structure close to this. Names vary, but the content is consistent:

1. Introduction and problem statement. The practice problem, local data that show it exists, and why it matters for patients, staff, or cost. End with the project purpose and the PICOT question.
2. Review of the evidence. Search strategy, an evidence table with levels and quality ratings, and a synthesis organized by theme. This is a focused evidence synthesis, not a broad literature review of everything ever written on the topic.
3. Theoretical and EBP frameworks. Usually two: an EBP or implementation model (for example the Iowa Model or the Johns Hopkins model) and a change or nursing theory (for example Lewin's change theory or a relevant middle-range nursing theory). Explain how each guided a specific part of the project.
4. Methods. Setting, population, intervention, timeline, outcome measures and instruments, data collection procedures, data analysis plan, and ethical considerations.
5. Results. Participants, implementation fidelity, and the outcome data with appropriate statistics and charts.
6. Discussion. What the results mean, how they compare with the evidence, limitations, sustainability, and implications for practice, policy, and education.
7. Dissemination. How the results were or will be shared (poster, presentation to leadership, manuscript).

Many programs also expect a needs assessment or gap analysis (sometimes using a SWOT analysis), a budget or cost analysis, and a sustainability plan.

How to choose a DNP project topic

Good DNP topics share four characteristics:

• A real, local problem. You can show it with data from the setting: incident reports, quality dashboards, audit results, or staff surveys.
• Enough existing evidence. There must be published evidence for an intervention. If the evidence does not exist yet, the question is a research question, not a DNP project.
• Feasible in your timeline. Most projects implement and evaluate an intervention over a few weeks to a few months. Choose an outcome that can change in that time.
• A site and a sponsor. You need access to a setting, permission from leadership, and usually a site mentor or preceptor.

A practical way to start: list three problems you see at work, write a PICOT question for each, and run a quick search for each one. Drop any topic with no evidence base or no measurable outcome. Our PICOT question examples post shows how to write the question, and the evidence-based practice nursing guide covers how to search and grade the evidence.

Data and evaluation

The evaluation is where many DNP projects are weakest. Plan it before implementation, not after.

Choosing outcome measures

Use a mix of measures where possible:

• Outcome measures: the clinical result you want to change (fall rate, HbA1c, readmissions, CAUTI rate).
• Process measures: whether the intervention was actually delivered (percentage of patients screened, protocol adherence).
• Balancing measures: unintended effects elsewhere (time added to nurse workload, delays in another process).

When you use a questionnaire (knowledge, confidence, satisfaction), use an existing validated instrument and report its reliability in your sample. Our guide to Cronbach's alpha explains how; the free Cronbach's alpha calculator computes it from item scores.

Common designs

• Pre/post comparison in the same group (staff knowledge before and after education).
• Pre/post comparison in different patient cohorts (fall rate in the three months before versus the three months after a protocol).
• Run charts or statistical process control charts to show change over time and distinguish a real shift from normal variation.

These are not randomized designs, so you cannot rule out other explanations such as seasonal changes, staffing changes, or other initiatives running at the same time. Name them in your limitations. The post on quasi-experimental design explains the main threats to validity and how to reduce them.

Matching the statistical test to the data

For quick checks you can use the free t-test calculator, chi-square calculator, and Mann-Whitney U test calculator. Report effect sizes and confidence intervals, not only p-values, and remember that a small DNP sample often has limited power: a non-significant result does not mean the intervention had no effect. Our power analysis post explains how to think about sample size before you start.

A short worked example

Suppose 24 nurses completed a 20-item sepsis recognition knowledge test before and after an education session. Mean scores rose from 13.1 (SD 2.4) to 16.3 (SD 2.1). Because the same nurses were tested twice, the right test is a paired t-test on the 24 differences. Suppose the mean difference was 3.2 points with a standard deviation of the differences of 2.6. Then:

• Standard error of the mean difference: 2.6 / sqrt(24) = 0.531
• t(23) = 3.2 / 0.531 = 6.03, which gives p well below .001
• Effect size (Cohen's d for paired data, mean difference divided by SD of differences): 3.2 / 2.6 = 1.23
• 95% CI for the mean difference: 3.2 ± 2.069 × 0.531, about 2.10 to 4.30

An APA-style sentence:

In the manuscript, italicize M, SD, t, p, and d as APA 7 requires. Note that knowledge gain is a process measure; whether sepsis recognition on the unit improved is a separate outcome you would need to measure.

12 DNP project ideas

These are starting points. Each would need a local problem, an evidence synthesis, and a feasibility check before it becomes a project.

1. Implementing a nurse-driven urinary catheter removal protocol to reduce CAUTI on a medical unit.
2. Introducing hourly purposeful rounding to reduce inpatient falls in older adults.
3. Standardizing a sepsis screening tool at emergency department triage.
4. A nurse-led telephone follow-up program for patients discharged with heart failure.
5. Routine depression screening with a validated questionnaire in a primary care clinic.
6. A diabetes self-management education program in a community health center.
7. Early mobility protocol for mechanically ventilated ICU patients.
8. Delirium screening and prevention bundle for older adults after surgery.
9. Structured handoff tool (for example SBAR-based) to reduce communication errors at shift change.
10. Pressure injury prevention bundle in a long-term care facility.
11. Evidence-based de-escalation training to reduce restraint use on an inpatient psychiatric unit.
12. Nurse residency or mentoring program to reduce first-year nurse turnover.

Common mistakes

• Choosing a research question instead of a practice question. If you need to discover whether something works, it is research. A DNP project implements something the evidence already supports.
• Scope too large. House-wide change in one semester rarely works. Start with one unit.
• No baseline data. Without a measured baseline, you cannot show change.
• Outcome that cannot move in the time available. Mortality rarely changes in eight weeks; protocol adherence and process measures can.
• Frameworks named but not used. Show how the model shaped each phase.
• Overstating results. Pre/post designs show association, not proof of cause.

Writing up the final paper

Once your data are in, the remaining work is writing the methods, results, and discussion clearly and consistently. If you have your outcome data in a spreadsheet, Data2Paper's Generate Paper can produce a draft paper from a CSV or XLSX file, including the statistical analysis, tables, figures, and references, exported to Word, PDF, or LaTeX. You can see what the output looks like in our sample papers. Every job is generated free so you can preview the first half, and unlocking the full draft is a one-time $19.90. It is an AI draft without expert review: check every number against your own data, follow your program's policy on AI use, and write the interpretation in your own voice.

FAQ

What is a DNP project?

It is the final scholarly project of a Doctor of Nursing Practice program. The student identifies a practice problem, synthesizes the evidence, implements an evidence-based change in a real setting, and evaluates the results.

How is a DNP project different from a dissertation?

A PhD dissertation produces new, generalizable knowledge through original research. A DNP project applies existing evidence to improve practice in a specific setting and evaluates that improvement, usually with quality improvement methods.

Does a DNP project need IRB approval?

It depends on the institution. Many DNP projects are classified as quality improvement and receive a determination that full IRB review is not required, but the IRB or a designated committee usually makes that decision. Never decide on your own.

How long does a DNP project take?

Most programs spread the project over several semesters: topic selection and proposal, implementation, then evaluation and the final paper. The implementation period itself is often a few weeks to a few months.

Do DNP projects require statistics?

Most do, at least descriptive statistics and a simple comparison of outcomes before and after the change. The tests are usually basic (paired t-test, chi-square, Wilcoxon, McNemar) or run charts, but they need to match the data type and design.