Reading and Interpreting the Scientific Evidence

Article Outline

• Levels of Evidence
• Interpreting a Research Report
• Conclusions
• Disclosures
• References
• Biography
• Copyright

Key words: experiment, interpreting research, levels of evidence, randomized controlled trial, scientific evidence

Nurses often need to find scientific evidence to help them determine the best practice available in a particular situation or in a periodic review of policies and procedures. Conversely, a nurse may read an article and may assume that it warrants consideration for implementation in the practice setting. It is common for people to find the research literature confusing. A previous column (To Engage in Evidence Based Practice, You Must First Find the Evidence) addressed the issue of locating relevant scientific evidence; this column is intended to demystify research articles.

Back to Article Outline

Levels of Evidence 

The U.S. Preventive Services Task Force of the Agency for Healthcare Research and Quality (2008) has identified three levels of evidence. Evidence is considered good if it includes consistent results from well-designed, well-conducted studies in representative populations that directly assess effects on health outcomes. Evidence is considered fair if it is sufficient to determine effects on health outcomes but the strength of the evidence is limited by the number, quality, or consistency of the individual studies; generalizability to routine practice; or indirect nature of the evidence on health outcomes. Evidence is regarded as poor if it is insufficient to assess the effects on health outcomes because of the limited number or power of studies, important flaws in design or conduct, gaps in the chain of evidence, or lack of information on important health outcomes. The U.S. Preventive Services Task Force uses this evaluation of the evidence to issue grades ranging from a grade of A, in which a practice is recommended on the basis of available evidence, to a grade of D, in which a practice is not recommended, or a grade of I, in which case the evidence is insufficient to make a determination. Although this is not the only model for examining the relevant science, it is a credible one from the federal agency charged with fostering safer health care. A continuing problem in health care is that till recently, insufficient evidence has been generated for our practice in most cases, and therefore we are forced to rely on the expertise of seasoned clinicians. If you are able to locate research relevant to your clinical practice question, the next challenge will be to interpret it.

Back to Article Outline

Interpreting a Research Report 

Research articles are written in a fairly standard format. The abstract provides an overview of the purpose of the study, methods, and results. This is probably sufficient information to indicate whether the article is relevant to your clinical question or not.

The introduction of the article, which may not be labeled, provides further detail on the purpose and focus of the study and the background information regarding as to why the authors thought it was a valuable study to conduct. The review of previously published data will be found in the introductory section, and this should provide the logical argument for both the purpose of the study and the way the authors decided to design it. You need to determine whether the review of previously published data included current information and whether it was sufficient to justify the study. It is often useful to further read the discussion section of the article in which the authors summarize what they found and the limitations or flaws from which the study may suffer. This serves as a good preparation before reading the Methods and Results sections, which together form the heart of the article.

The methods section should describe the type of study that was conducted. The “gold standard” for clinical evidence is regarded to be the randomized controlled clinical trial, or experiment. This means that the participants were randomly assigned to either the experimental treatment group or to a control group. Ideally, people responsible for implementing the intervention being studied will not know which participant is being assigned to which group, so that they cannot skew the outcomes, thereby making it a double-blinded study.

It is often not possible to require people to be in one group or the other; therefore, you will often find quasi-experiments reported in the published data. In a quasi-experimental study, some participants undergo the experimental treatment and others are placed in a comparison group, but they will not have been randomized to one group or the other. It could have happened perhaps because either the participants volunteered to take part after being informed about the study, or the data were collected long after the intervention was completed to compare outcomes. In other studies, individuals may act as their own controls; they are evaluated before and after the experimental intervention. It is possible that they might receive a standard treatment in addition to the experimental treatment; in this case, the order can be randomized, thereby making it an experimental design. In cases when there are no comparisons being made, the nature of the study is descriptive.

Descriptive studies can be qualitative or quantitative in nature. Qualitative studies compile participants’ responses to open-ended questions on a topic of interest, whereas quantitative studies involve the analysis of mathematical data such as, for example, scores on survey instruments, laboratory values, length of stay, or cost of care. We are best able to determine cause and effect between an intervention and an outcome with an experiment; descriptive research is useful in providing information related to the topic of interest, including issues related to patients’ satisfaction with their care, but it is hard to determine a cause and effect relationship between an intervention and outcomes.

As you read through the Methods section of the article, you will want to determine the following: whether or not subjects’ rights were protected through institutional review board approval and consent to participate; procedure of recruitment of subjects and how the number of participants was determined; what procedures were used to ensure control of the study; what measurements were taken, and whether or not the authors addressed the quality of those measures, in terms of their reliability and validity.

As you read through the results of the study, you will want to read information related to a description of the participants in the study so that you are able to consider the relevance of this study to your own clinical setting. You will also note whether or not the analysis of data was appropriate. If it is a quantitative study, you may be able to determine whether the statistic was appropriate by referring to a statistical handbook such as the one by Munro (2005), or by way of a consultation with a Master’s degree-prepared nurse in your institution or a statistician working with your quality assurance department. Of particular interest is whether or not the experimental intervention yielded a statistically significant difference in outcomes from the comparison treatment. In addition, you should read to determine whether or not the findings that were reported addressed the research question that was asked. The conclusion of the paper summarizes what the authors believed that their contribution was, as well as implications for health care or health policy.

You may want to invest in a research textbook to provide assistance as you develop your skill in reading scientific articles. Many such textbooks are commercially available, and there are some very good examples that have withstood the test of time such as Burns & Grove (2009) or Polit & Beck (2007). You may find it useful to engage in dissecting relevant research articles with others with whom you work and ideally with someone who is accustomed to reading them. You will find that the more you read, the easier it will be to understand. Research articles often are unable to address with relevance in your clinical setting with your particular patient population and resources; that requires staff deliberation. In addition, staff nurses need to consider the clinical significance of research findings that may be statistically significant; for instance, would the improvement in outcomes be worth the resources required to bring them about? Although they cannot provide definitive answers to your practice issues, scientific publications do provide the information that you need to make an informed decision on whether or not practice changes are worthy of consideration.

Back to Article Outline

Conclusions 

Nurses who seek to answer a clinical question or solve a clinical problem have access to a vast array of print and electronic resources, and the U.S. Preventive Task Force of the Agency for Healthcare Research and Quality (2008) rating system helps with the evaluation of existing evidence. In cases where a recent compilation and assessment of evidence is unavailable, a review of previously published research is essential. The step-by-step approach to reading a scientific article described previously should ease the challenge of reading and interpreting research. Do not be afraid to ask others in your organization for help in understanding the information. Remember that evidence-based practice requires a team effort.

Back to Article Outline

Disclosures 

The author reports no real or perceived vested interests that relate to this article (including relationships with pharmaceutical companies, biomedical device manufacturers, grantors, or other entities whose products or services are related to topics covered in this manuscript) that could be construed as a conflict of interest.

Back to Article Outline

References 

1. Burns N, Grove SK. The practice of nursing research: Appraisal, synthesis, and generation of evidence. 6th ed.. St. Louis, MO: Elsevier; 2009;
   • View In Article
2. Munro BH. Statistical methods for health care research. 5th ed.. Philadelphia, PA: Lippincott, Williams & Wilkins; 2005;
   • View In Article
3. Polit DF, Beck CT. Nursing research: Generating and assessing evidence for nursing practice. Philadelphia, PA: Lippincott Williams & Wilkins; 2007;
   • View In Article
4. U.S. Preventive Services Task Force of the Agency for Healthcare Research and Quality . Grade definitions after May 2007. Retrieved from http://www.ahrq.gov/clinic/uspstf/grades.htm2008;
   • View In Article