Nursing: Levels of Evidence

When you work with patients, you will have to take lots of different information into account in order to find the most appropriate treatment plan for their individual situation. You will not only need to make sure you are using your own clinical expertise; you will also need to take into account the patient's wishes and values, and as well as locate the best available evidence. Using all three of these things together to inform your care plan is what is known as Evidence-Based Practice.

Let's focus here on the last part. What does it mean to find the "best available" evidence?

Quality of Evidence

When we talk about evidence being the "best available" or being "high-quality", what we mean is that we are looking for information that is more likely to reliably predict health outcomes. Evidence quality is frequently represented as a hierarchy or a pyramid. Evidence and sources at the top of the pyramid are typically of higher quality than the evidence below, but are also more scarce.

As you can see in the figure above, systematic reviews and meta-analyses make up the top of the evidence pyramid. These are labeled in the figure as filtered information, which is another way of saying they are secondary research.

The next levels of the pyramid are (in order from higher quality to lower quality) randomized controlled trials, cohort studies, case-controlled studies, and case series and reports. These are labeled in the firgure as unfiltered information, which is another way of saying they are primary research.

The bottom of the pyramid in the figure is background information and expert opinion. 

You may now be asking, "What do these categories mean? Why are some lower quality than others?" Let's go over what each of these evidence types actually are, including how the different studies are designed.

Systematic Reviews and Meta-analyses

Systematic reviews are a specific type of secondary research, where researchers will scour the literature for any and all evidence they can find about a particular topic or therapy and evaluate them in context with each other to make practice recommendations. They are often combined with meta-analyses, which is a statistical analysis of quantitative research. 

Randomized Controlled Trials

Randomized controlled trials are a type of study that investigates the effectiveness of a particular intervention, where research participants are randomly assigned to be in either the experimental group or the control group. These are considered to have less of a risk of bias than other types of prospective research because neither the participants nor the researchers know to which group each person is assigned (this is called study blinding). 

Example: A study examining the effectiveness of a new blood pressure medication, where one group (experimental) gets the medication and another group (control) gets a placebo. 

Cohort Studies

Cohort studies are observational (instead of experimental) studies of groups of participants through some defined time period. They can be either prospective, where participants are followed through the life of the study and new data is collected, or retrospective, which look at past histories and/or data of participants. Participants in cohort studies are divided into groups that share a common trait, disease, exposure, etc., and are examined to see if they develop a specific outcome.

Example: A study examining incidence of cancers among New Yorkers who assisted with emergency response efforts during the 9/11 attacks and those who did not assist.

Case-controlled Studies

Case-controlled studies are also observational studies that can only be retrospective. Participants are chosen because they either have an outcome of interest (cases) or do not have this outcome (controls). The medical histories of both groups are compared to look for exposures that may have led to the outcome of interest.

Example: A study examining the histories of participants with and without pancreatic cancer to see if any possible exposures or contributors can be determined.

Case Studies and Reports

Case studies and case reports are observational studies that provide detailed information about an individual patient (case). These studies include as much information about the conditions, symptoms, outcomes, treatment, etc. as possible. There is no comparison or experiment in case studies, but they can be good sources of information for rare diseases or atypical treatments.

Background Information and Expert Opinion

This last group of background information and expert opinion is not actually considered part of the scientific evidence. It can be useful for providing a good starting point for your research but should not be incorporated into your clinical practice unless you find real evidence to back it up. 

It's important to note that information from Point-of-Care tools (e.g., UpToDate, LexiDrug, VisualDX) should typically be considered to fall into this category. While these tools purport to contain information to support evidence-based practice, much of the synthesis is biased or limited in some way. These tools should be used as a starting point for learning more, not as true high quality evidence.