The single most consequential classification a Doctor of Nursing Practice student makes is whether their project is quality improvement, research, or evidence-based practice. The label is not a formality. It determines whether the project requires review by an Institutional Review Board, and getting it wrong in either direction carries a real cost. A project that is quality improvement in substance but treated as research invites unnecessary regulatory burden and delay. A project that is research in substance but conducted as quality improvement, without the IRB review that research requires, is a compliance failure that can block publication and stall a degree. The distinction is not intuitive, because the three categories overlap in their methods and share a commitment to improving care. What separates them is intent, and federal regulation defines that intent precisely. Our DNP project help begins with getting this classification right, because everything downstream depends on it.
This guide explains how the Common Rule defines research, why the phrase 'generalizable knowledge' is the pivot on which the whole distinction turns, how quality improvement and evidence-based practice differ from research and from each other, and what design choices push a project across the line into research.
Quick Answer:
Under the federal Common Rule (45 CFR 46.102), research is a systematic investigation designed to develop or contribute to generalizable knowledge. Quality improvement asks a local question: how do we improve this process in this setting? It applies existing knowledge to close a performance gap; it is usually not research and usually does not require IRB review. Evidence-based practice translates established evidence into practice for eligible patients. Research that seeks to produce new, generalizable knowledge requires IRB review. The determinant is intent: if the project is designed to generate findings meant to apply beyond the local setting, it is research, regardless of its methods. Because design choices that create generalizable intent, such as randomization, hypothesis testing, or an intent to publish findings in a generalizable way, can push a project across the line, the classification should be confirmed in writing with the IRB before the project begins.
Why This Distinction Exists, and Why It Is Hard
The three categories are difficult to separate precisely because they look alike in practice. All three collect data. All three may measure outcomes before and after a change. All three aim to improve patient care. A quality improvement project and a research study can use the same statistical test on the same kind of data and reach superficially similar conclusions. The difference is not in the tools, but in the purpose the project was designed to serve, and the purpose is not always visible in the methods section.
This is why the distinction cannot be settled by looking at whether a project is rigorous, or whether it collects a lot of data, or whether it uses inferential statistics. Rigorous quality improvement is still quality improvement. The question is narrower and more specific: was the project designed to answer a local operational question, or to produce knowledge intended to generalize beyond the setting in which it was conducted? Federal regulation calls into question the entire test.
The Three Categories, Defined
Setting out each category by its defining purpose, rather than by its methods, is the only way to reliably separate them.
Research is defined by the Common Rule as a systematic investigation, including research development, testing, and evaluation, designed to develop or contribute to generalizable knowledge. Two elements must both be present: a systematic investigation and the intent to produce generalizable knowledge. The full regulatory definition is set out at 45 CFR 46.102. When both elements are present, and the project involves human subjects, IRB review is required.
Quality improvement applies existing knowledge to improve a process, outcome, or system within a specific local setting. The question is operational and local: how do we improve this outcome, in this unit, with these patients? It uses rapid cycles of change, often the Plan-Do-Study-Act model, and existing data, to close a gap between current and desired performance. Because its intent is to improve local care rather than to generate generalizable knowledge, quality improvement is usually not human subjects research and usually does not require IRB review, though many institutions require a formal non-research determination to confirm this.
Evidence-based practice is the translation of the best available evidence into clinical practice for all eligible patients. It does not seek to generate new knowledge, generalizable or local; it applies existing knowledge, integrating research evidence with clinical expertise and patient values. An EBP project takes an established finding and implements it, which is why selecting the right implementation framework matters, a choice our guide on the Iowa Model, Johns Hopkins, and JBI frameworks addresses directly.
Table 1: Quality Improvement vs Research vs Evidence-Based Practice
Dimension | Quality Improvement | Research | Evidence-Based Practice |
|---|---|---|---|
Purpose | Improve a local process or outcome | Produce new generalizable knowledge | Apply existing evidence to practice |
Core question | How do we improve this, here? | What is true, beyond this setting? | How do we apply what is known? |
Knowledge produced | Local | Generalizable | None new; applies existing |
Typical IRB pathway | Non-research determination | Full IRB review required | Usually non-research |
Reporting guideline | SQUIRE 2.0 | CONSORT, STROBE, PRISMA, etc. | Framework-dependent (Iowa, JHNEBP, JBI) |
Generalizable Knowledge: The Phrase the Whole Distinction Turns On
If a single phrase decides whether a DNP project needs IRB review, it is generalizable knowledge. Everything in the classification reduces to whether the project intends to produce it.
Knowledge is generalizable when it is intended to be applied beyond the specific setting in which it was produced, to inform practice elsewhere, to contribute to a scientific field, or to be presented as a finding that others can rely on in different contexts. Quality improvement produces local knowledge: this change improved this outcome in this unit. That knowledge is valuable, but it is not designed to establish a generalizable truth. Research produces generalizable knowledge by design: this intervention produces this effect, in a way intended to hold beyond the study site.
The intent is what matters, and it is assessed at the design stage, not after the results are in. A project designed from the outset to test a hypothesis and contribute to the answer to the wider field is research, even if it is conducted in a single unit. A project designed to solve a local problem is a quality improvement, even if its results are interesting to others. The reason intent is assessed prospectively is that the regulatory obligation, IRB review to protect human subjects, must be satisfied before the project begins, not judged retrospectively once data have been collected.
What Pushes a Project Into Research
Certain design choices signal an intent to generate generalizable knowledge, and their presence can move a project from quality improvement into research, even when the student intended quality improvement. Recognizing them lets you classify honestly at the design stage.
Randomizing participants to test which of two approaches works better is a hallmark of research, because randomization exists to produce a generalizable causal comparison rather than to improve a local process. Testing a formal hypothesis, particularly one framed as a contribution to knowledge rather than a local performance target, signals research intent. Designing the project from the outset with the primary aim of publishing the findings as generalizable knowledge, rather than disseminating a local improvement, points toward research; note that publishing a quality improvement project is entirely permissible and common, and does not by itself make it research, but designing it in order to produce a generalizable finding does. Withholding an established effective treatment from a comparison group to measure a difference is a research design choice. Enrolling participants and collecting data primarily to answer a research question, rather than using data generated by routine care and operations, also points toward research.
The presence of one or more of these features does not automatically make a project research, but it raises the question, and it is exactly the kind of judgment an IRB exists to make. This is why the determination lies with the IRB, not the student. Our DNP project proposal service frames the project so that its classification is defensible before it reaches the board.
Table 2: Design Choices That Can Push a Project Into Research
Design Choice | Why It Signals Research Intent |
|---|---|
Randomizing participants | Exists to produce a generalizable causal comparison, not to improve a local process |
Testing a formal hypothesis | Frames the aim as a contribution to knowledge rather than a local performance target |
Designing primarily to publish a generalizable finding | Intent to generalize, not merely to disseminate a local improvement |
Withholding effective treatment from a comparison group | A research design choice, not a care-improvement one |
Enrolling participants to answer a research question | Data collected for knowledge, not generated by routine care and operations |
The AACN Framing: The DNP Project Is Practice-Focused
The classification is reinforced by how the profession defines the DNP project itself. The American Association of Colleges of Nursing, in The Essentials: Core Competencies for Professional Nursing Education (2021), frames the DNP scholarly project as practice-focused: its purpose is to translate, apply, and evaluate evidence to improve patient or practice outcomes, not to generate original, generalizable knowledge in the way a PhD dissertation does. The distinction between the practice doctorate and the research doctorate is precisely this. The PhD produces new knowledge for the discipline; the DNP applies and translates knowledge to improve care.
This matters for classification because it means the default expectation for a DNP project is that it is quality improvement or evidence-based practice, not research. A DNP project designed as original research sits in tension with the practice-focused purpose of the degree, and a student whose project has drifted into research intent should pause and reconsider whether the design still fits the doctorate. AACN is also clear that a stand-alone literature review, integrative review, or systematic review is not by itself sufficient to constitute a DNP project; the project must apply evidence to practice. Establishing the practice gap that the project will address is the proper starting point, which our guide on identifying a practice gap for your DNP project walks through.
The Consequences of Getting It Wrong
The two directions of error carry different but real costs, and naming them makes the stakes concrete.
Classifying a project as quality improvement when it is, in substance, research means conducting human subjects research without IRB approval. This is a serious compliance and ethics problem. It can render the work ineligible for publication because journals require documentation of IRB review or exemption for research, and it can require the student to halt, seek retrospective review that may not be grantable, and, in some cases, redo the work. Retrospective IRB approval is generally not available because the entire purpose of prospective review is to protect participants before they are exposed to risk, so a misclassification discovered late can be difficult to remedy.
Classifying a project as research when it is genuinely quality improvement imposes the opposite costs: unnecessary IRB review, added delay, and administrative burdens the project did not require. This is less dangerous than the first error but still wasteful, and it can push a time-limited DNP project past its deadlines.
Both errors are avoided the same way: by determining the classification honestly at the design stage, against the generalizable-knowledge test, and by obtaining the IRB's determination in writing before the project begins. Because the IRB, not the student or the faculty advisor, makes the binding determination, the safe course is always to submit the project for a determination rather than to assume one.
Not sure whether your DNP project needs IRB review? |
|---|
Send us your project aims and design. A nurse methodologist will assess it against the generalizable-knowledge test, tell you whether it reads as QI, research, or EBP, and flag any design choice that would push it toward research before your IRB sees it. Get your project classified and receive an itemized quote within 2 to 4 business hours, no obligation. |
How the Determination Is Actually Made
Because the classification carries such weight, it helps to know how institutions make it in practice, so that a student can anticipate the questions and prepare a defensible answer. Most institutions do not leave the judgment to intuition. They use a structured self-determination tool, a short checklist that walks the project through the criteria that separate research from quality improvement, and many require this tool to be completed and, where it indicates research or is ambiguous, submitted to the IRB for a formal determination.
These checklists ask a consistent set of questions, each of which maps to the generalizable-knowledge test. Is the primary intent to improve a local process or to generate new knowledge? Will the findings be applied only to the local setting, or are they designed to inform practice elsewhere? Does the project involve randomization, a control group, or the withholding of an established treatment? Is there a hypothesis framed as a contribution to knowledge? Are participants being enrolled and data collected specifically to answer a research question, or is the project using data generated by routine care? A pattern of answers pointing toward local improvement indicates quality improvement; a pattern pointing toward generalizable intent indicates research. A widely used formal version of this logic is the instrument developed by Ogrinc and colleagues in 2013 to differentiate clinical research from quality improvement, which many IRBs and investigators use to structure exactly this decision.
The practical value of completing such a tool early is twofold. First, it forces the design questions to the surface before the project is built, when they can still be answered deliberately rather than discovered late. Second, it produces a documented rationale for the classification, which is what an IRB, a faculty committee, and later a journal editor will want to see. A student who can show a completed determination tool and a written IRB determination has resolved the question as the system expects. Framing the project and its aims so that this determination is clear and defensible is a core part of our DNP project proposal service.
Reporting the Project: SQUIRE 2.0
Once a project is correctly classified as a quality improvement project, it should be reported using the standard designed for it. SQUIRE 2.0, the Standards for Quality Improvement Reporting Excellence, is the reporting guideline for quality improvement work, published by Ogrinc and colleagues in 2016. It is the QI counterpart to what PRISMA is for systematic reviews or CONSORT is for trials, and using it signals that the project was conceived as quality improvement and reported to the appropriate standard.
SQUIRE 2.0 emphasizes several elements that distinguish a strong QI report: the use of theory or a conceptual framework in planning, implementing, and evaluating the intervention; a clear account of the local context, because context is central to whether a local improvement will transfer; and a careful study of the intervention itself, including how it evolved over successive cycles. Reporting a DNP quality improvement project to SQUIRE 2.0, available through the SQUIRE statement, both strengthens the write-up and reinforces the project's identity as quality improvement rather than research. The data analysis these reports rest on is handled by our statistical analysis service, and the full write-up by our capstone writing service.
Frequently Asked Questions
Does my DNP project need IRB review?
It depends on whether the project is research under the Common Rule, meaning a systematic investigation designed to develop or contribute to generalizable knowledge. If your project applies existing knowledge to improve a local process, it is usually quality improvement and usually does not require full IRB review, though many institutions require a formal non-research determination. If it is designed to produce findings that can be generalized beyond your setting, it is research and requires IRB review. The IRB, not you, makes the binding determination, so submit it for a determination before beginning.
What is the difference between quality improvement and research?
The difference is intent, not method. Quality improvement applies existing knowledge to improve a process in a specific local setting, asking how to improve this outcome here. Research is designed to produce new, generalizable knowledge intended to apply beyond the study setting. The same statistical methods can appear in both; what separates them is whether the project was designed to answer a local operational question or to contribute generalizable knowledge to a field.
What does generalizable knowledge mean?
Generalizable knowledge is knowledge intended to be applied beyond the specific setting where it was produced, to inform practice elsewhere or contribute to a scientific field. Quality improvement produces local knowledge about a specific setting; research produces knowledge designed to generalize. The intent is assessed at the design stage, because the obligation to obtain IRB review must be met before the project begins, not judged after results are collected.
Can a quality improvement project be published?
Yes. Publishing a quality improvement project is common and permissible, and it does not by itself make the project research. What would make it research is designing the project from the outset with the primary aim of producing a generalizable finding, rather than disseminating a local improvement. Quality improvement projects are appropriately reported in accordance with the SQUIRE 2.0 guidelines.
Is a DNP project supposed to be original research?
Generally no. The AACN Essentials (2021) frame the DNP project as practice-focused: its purpose is to translate, apply, and evaluate evidence to improve patient or practice outcomes, rather than to generate original, generalizable knowledge as a PhD dissertation does. A DNP project that has drifted toward original research intent sits in tension with the practice doctorate, and the student should reconsider whether the design still aligns with the degree.
What design choices turn a QI project into research?
Randomizing participants to compare approaches, testing a formal hypothesis framed as a contribution to knowledge, designing the project primarily to publish a generalizable finding, withholding an established effective treatment from a comparison group, or enrolling participants and collecting data primarily to answer a research question rather than using routine-care data. Any of these can push a project into research, which is why an IRB determination should be sought when they are present.
Can I get IRB approval after I have already collected data?
Usually not. Retrospective IRB approval is generally unavailable because the purpose of prospective review is to protect participants before they are exposed to any risk. This is why a project that is research in substance but conducted as quality improvement without review is difficult to remedy once discovered. Obtain the IRB's written determination before the project begins.
You May Also Find Useful
Classifying With Confidence
The DNP students who avoid a late, costly reclassification are the ones who ask the generalizable-knowledge question honestly before they design the study, not after. If the aim is to improve care in your setting, you are almost certainly doing quality improvement or evidence-based practice, and you should build the project on that footing, report it to SQUIRE 2.0, and confirm the non-research determination with your IRB in writing. If the aim is to produce a finding meant to generalize, you are doing research, and you need IRB review before you start. The classification is a design decision, and deliberately making it keeps the project on schedule and publishable.
Bring us the project before you finalize the design, and a nurse methodologist will classify it against the Common Rule test and flag anything that would push it toward research. Send your project for a classification check, and you will have an itemized quote within 2 to 4 business hours, with no obligation.
