Using the Power of Statistics to Help Patients Detect Misinformation in the Media

Discussion of five key concepts in statistical analysis to help the nurse read and understand research studies, and in turn, educate patients on how to detect misinformation in the media.

Updated:  

Using the Power of Statistics to Help Patients Detect Misinformation in the Media

In today's toxic misinformation environment, the old "studies have shown" moniker cannot be used flippantly. Nurses need to have the ability to sift through flawed research studies and find the ones with true merit. The time has come for us to dust off the cobwebs of our knowledge on statistics and be able to say, "Because science!” with conviction.  

A top priority in nursing is to keep patients informed, and over the last two years, navigating the minefield of misinformation has made this task much more difficult. Therefore, nurses must be well-informed on evidence-based practice and true scholarly research. This article will explain how we as nurses can help our patients, family, and friends detect misinformation in the media. 

The first step in fighting misinformation is to find articles to back up claims. Any person can say anything on social media, so it's important to teach patients to find the research related to any questionable claims. 

Below are 5 concepts the astute nurse must grasp to read and understand research articles. Armed with this knowledge, nurses can help patients and family members alike learn to detect misinformation.

Statistical Significance

Statistical significance means that the probability of the test result happening by chance (pure randomness) is extremely low. Some test results are statistically significant while others are not. Statistical significance is portrayed as a number called a p-value.  

If the applied independent variable caused enough of an observable change, the study is considered statistically significant. The researcher can then confidently 'reject the null hypothesis' and accept the alternative hypothesis2. Don't worry, there's no need to get bogged down in these semantics to understand basic statistical significance. Chances are, if these terms confuse you, they'll confuse your patients, too.

P-value

In research articles, a p-value is used to denote statistical significance. A predefined significance level is a requirement of any reputable research study. This is important because if a certain p-value is reached at the end of the study without having set a predefined significance value beforehand, the study results can be unreliable and unproveable. A reasonable p-value, although not used every time, is less than or equal to 0.05. Important to note, p-values can be skewed by random error and systematic errors1.  

For Example: measuring mid-arm circumference on a small geographical sample size, different people doing the measuring, placement of measurement tape, and how tightly the arm was measured can all contribute to random error1. Systematic errors refer to a broad array of errors, including researcher bias, experience, and incorrect mathematical calculations1.

Confidence Interval

Confidence Intervals are extremely helpful in research studies. They help tell us whether a study outcome has true merit. The confidence interval is a range and is usually portrayed at the 95 percent level in research studies. Think of it this way: if a p-value tells us 'Yes or no', the confidence interval tells us 'How much'2.

Confidence intervals tell the reader, '95% of samples in this study fell within this range'. If a confidence interval falls over the "zero effect" line, the researchers cannot reasonably say that the independent variable had statistically significant effects1. This, in turn, makes the study outcome less useful in clinical practice.

For example, if 95 percent of people with documented HTN had a drop in BP after taking a certain antihypertensive in a research study, then a reasonable conclusion would be that this antihypertensive has a reliable effect (statistically significant effect) in dropping BP and may be clinically useful2.

On the other hand, if only some of the 95 percent of people in the study had a BP drop while others had either no change or an increased BP, the confidence interval then crosses the "zero effect" line1, and the results cannot be considered statistically significant.

Important to note: Strong research studies use both p-values and confidence intervals. Some studies can be statistically significant while having an unreliable confidence interval or vice versa2. These variations can be caused by small sample size or demographics2.

Clinical Significance

Different from statistical significance, clinical significance is not based on one number (e.g., p value= <0.05). Clinical significance is the presumption that the study outcome would be relevant in a clinical setting1. One way to assess possible clinical significance is by looking at the confidence interval. However, clinical significance also requires some common sense.

For instance, using the same BP study from above, let's say that the p-value for the study is <0.05, and the CI does not cross the "zero effect" line, but the mean BP drop is only 2 mmHg. Does a 2 mmHg drop in BP warrant administering this drug instead of another drug?

Clinical significance waxes and wanes based on repeat study results, new research, or findings that refute previous results. Hence, the nuances of science are that research study outcomes can change. This fact, while normal, can be frustrating and can grow public distrust. It's important to teach patients and family members this integral aspect of research.

Get Comfortable Reading Research!

Start getting out there and finding research studies to read. It's a surefire way to get you more comfortable understanding all the terms listed above. A few resources for readers to start delving into research studies are:  

  1. The Epocrates app- It's free on any app downloading service and has study synopses from the New England Journal of Medicine.
  2. UptoDate- If you have access to UptoDate through your workplace, this is a fantastic resource to read new and trending research articles.
  3. Google search- Doing a simple Google search on research topics can yield great articles. Just be careful to use reputable websites such as the NCBI, NIH, or PubMed.

The more you read, the better prepared you will be to explain these differences to your patients and help mitigate any damage misinformation causes.

Happy learning! 

References 

1Statistical significance or clinical significance? A researcher's dilemma for appropriate interpretation of research results

2Confidence Interval or P-Value?

2
Up Vote Up Vote ×
PKN

About PKN, RN

Lacee is a master’s prepared registered nurse with twelve years of nursing experience. She has worked in several clinical settings including acute care, critical care and emergency nursing, drug and alcohol rehab, pediatric home health, hospice case management, and hospice leadership.

1 Article   3 Posts

Share this post

Share on other sites

You forgot one very important thing to look at; the conflict of interests and funding source disclosures. Always know who paid the bill.

1
Up Vote Up Vote ×
Specializes in 25 years ICU 11 years Clinical Research.

This article is fantastic! As a 36 year RN and a CCRC for the last 11 years, I truly appreciate how well written and informative this is.

Up Vote Up Vote ×
Specializes in orthopedic/trauma, Informatics, diabetes.

This is assuming that a nurse can discern the information themselves. Knew a nurse that got into NP school with less than a year of practicing and got her NP in psychiatric nursing in about 18 months. She didn't understand the Microbiology of how pathogens/vaccinations work and listened to certain false news about vaccines. She was anti vax. She didn't know how to disseminate data. 

As a pre-vet Biology major and having a degree in one of the Social Sciences, as well as a MSN in Informatics, I have learned how to break down journal articles and data. Mind blowing what some people think is legitimate "research". 

1
Up Vote Up Vote ×
Specializes in Anesthesia.

I think using research to fight misinformation is great, but many of the problems with misinformation come from cognitive biases/dissonance that no amount of information is likely to breakthrough. 

2
Up Vote Up Vote ×
Specializes in Anesthesia.
On 9/28/2022 at 9:32 AM, mmc51264 said:

This is assuming that a nurse can discern the information themselves. Knew a nurse that got into NP school with less than a year of practicing and got her NP in psychiatric nursing in about 18 months. She didn't understand the Microbiology of how pathogens/vaccinations work and listened to certain false news about vaccines. She was anti vax. She didn't know how to disseminate data. 

As a pre-vet Biology major and having a degree in one of the Social Sciences, as well as a MSN in Informatics, I have learned how to break down journal articles and data. Mind blowing what some people think is legitimate "research". 

It isn’t just nurses as we have physicians that are anti-vax. I think more education reduces the overall rate that people will believe in misinformation, but it doesn’t eliminate it.

Up Vote Up Vote ×
+ Add a Comment