Whatever their specialty or practice area, all nurses should strive to become more sophisticated consumers of nursing research by learning how to critically appraise, synthesize, and communicate research findings. Such critical appraisal shows your commitment to evidence-informed practice and empowers you to create a practice culture based on the best available evidence.
To critically appraise nursing research, you must ask focused, meaningful questions to determine the overall integrity and applicability of the research. This article will help you better understand—and undertake—the process of critical appraisal. With practice, even nurses who once were intimidated by the research process should be able to efficiently and effectively determine the clinical relevance of scientific studies.
The problem statement should appear at the beginning of the article and should include enough information for you to determine if the study results can be generalized to a specific patient population. Effective problem statements include independent and dependent variables, population of interest, and key concepts of the study.
The research problem should provide a clear rationale for the study. The problem statement may be in one of two forms:
• a research question that indicates the who, what, when, where, and why of the study
• a “purpose” statement that describes the researcher’s purpose in conducting the study.
Also, the research problem should fill a gap in the current body of nursing research or theory or should pinpoint a single, relevant nursing issue that’s meaningful to nurses and patients.
A poorly worded or inappropriate problem statement can cause flaws in a study’s methods, protocols, samples, and analyses. Be wary of broadly stated or overly generalized problem statements, as well as those whose research questions can’t be answered by the methods proposed.
The literature review is a systematic, critical review of the most important scholarly literature on a given topic. It should:
• highlight critical weaknesses in previous studies
• identify previously studied concepts or variables
• relate the current research project to historical research
• identify the current knowledge deficit about a particular phenomenon and state what more needs to be done to overcome that deficit.
Look for a broad range of references—for example, peer-reviewed journal articles, systematic reviews of relevant research, professional standards, position statements, dissertations, and conference proceedings. In general, references should be no more than 5 years old, unless the research cited is a classic or historically important work.Conceptual framework
Depending on the nature of the study, a conceptual or theoretical framework may be presented near the beginning of the article.
• Theoretical frameworks are narrower in scope and can be tested directly.
• Conceptual frameworks express assumptions and can’t be tested directly.
A well-defined conceptual framework allows the reader to better understand the relationship between major concepts of the study and more fully explicates the relationship between the variables. Additionally, a conceptual framework may help you better understand how the researcher’s hypothesis and research question were developed.
The methods section of the article tells you how the principal investigator went about answering the research question. It includes information about the sample selection, study design, data collection, and data or statistical analysis. This section should also provide sufficient information to permit duplication of the study and should address the protection of human subjects.
Sample selection occurs on the basis of eligibility criteria that the researcher establishes in accordance with the study’s objectives. The sample’s eligibility and exclusion criteria, as well as its demographic composition, should be appropriate to achieve the study’s objectives. The study should have an adequate number of participants and a low dropout rate to protect against compositional and statistical bias and make the study more representative of the population.
The study design should be clearly stated and appropriate for the research question being asked. The most common design associated with quantitative research is experimental design—commonly considered the most rigorous design and, for many researchers, the gold standard. In this design, the researcher controls both the selection of study subjects and introduction of the independent variable.
Also, in experimental design, subjects are randomly assigned to treatment and control groups, thereby reducing study bias because the researcher can’t influence the assignment of subjects. Ideally, groups in an experimental study are similar in all respects, and any differences between them result from the intervention administered by the researcher. Experimental design is commonly used when new drugs or medical products are being studied.
In contrast, the nonexperimental study design is more qualitative and is used when the researcher wants to observe a particular phenomenon but lacks the ability and desire to manipulate the independent variables. The quasi-experimental design is more closely related to experimental design but lacks random assignment of subjects and subsequently may introduce bias.
Data collection procedures should be fully explained in the methods section and should provide a clear understanding of how data were collected and who collected them. Issues such as inter-rater reliability, instrument reliability and validity, and training of data collectors should be addressed. A clear explanation of data collection lends credibility to the study.
A thorough review of the statistical analysis is important. (If you’re uncomfortable or unfamiliar with basic statistics, consult a clinical nurse specialist, nurse researcher, or other advanced practice nurse for assistance.) Study results should be presented in a logical, systematic format. For quantitative studies, both descriptive and inferential statistics should be provided.
• Descriptive statistics, including measures of central tendency (mean, median, and mode) as well as measures of dispersion or variability (variance, standard deviation, and range) provide information about the characteristics of the subjects or phenomenon being studied.
• Inferential statistics allow researchers to make assumptions about the population based on a sample. Through the use of significance tests and other measures, inferential statistics help researchers understand the probability that the results of their study occurred by chance. The level of significance, expressed as a “p” value, represents the probability of obtaining the computed value by chance. If “p” is less than 0.01, the probability of obtaining the computed results by chance is less than 1%. In other words, if the study were repeated 100 times, the difference between groups would be attributed to the study intervention 99 times and to chance only one time.Protection of human subjects
To protect research subjects, the study must be approved by an institutional review board. The researcher also must obtain informed consent from participants after giving them oral and written information on the nature of the study and potential safety risks or conflicts of interest.
Results and discussion of findings
Here the author should succinctly relate any descriptive and inferential statistical findings to the study’s independent and dependent variables (as stated in the original research question or research hypothesis). This section should state clearly whether the data analysis supports, or fails to support, the research hypothesis. You should be able to read the study’s results and determine the relationship between the independent and dependent variables.
Some articles may contain a separate “discussion” section in which the author interprets, analyzes, and summarizes the study’s conclusions and its relevance to the larger theoretical framework. Expect the author to objectively state any limitations or weaknesses in the study’s design, method, sample, or data collection procedures. The discussion also should identify any conceptual or theoretical relationships in need of further investigation.
Dunning M, Abi-Aad G, Gilbert D. Experience, Evidence and Everyday Practice: Creating Systems for Delivering Effective Health Care. London: King’s Fund; 1999.
Greenhalgh T. How to read a paper. The Medline database. BMJ. 1997;315(7101):180-183.
Griffin-Sobel JP. Research in practice: immersing yourself in research. Gastroenterol Nurs. 2003;26(5):219-220.
Hudson-Barr D. From research idea to research study: the how. J Spec Pediatr Nurs. 2005;10(3):147-150.
LoBiondo-Wood J, Haber J. Nursing Research: Methods, Critical Appraisal, and Utilization. St. Louis, Mo: C.V. Mosby; 2002.
Valente S. Critical analysis of research papers. J Nurs Staff Dev. 2003;19(3):130-142.
Kenneth J. Rempher, PhD, RN, MBA, CCRN, APRN, BC, is Director of Professional Nursing Practice at Sinai Hospital of Baltimore in Baltimore, Md. Cory Silkman, MAR, BSN, RN, C, is a Clinical Leader in the Comprehensive Inpatient Rehabilitation Unit at Sinai Hospital of Baltimore.