Skip to Main Content
It looks like you're using Internet Explorer 11 or older. This website works best with modern browsers such as the latest versions of Chrome, Firefox, Safari, and Edge. If you continue with this browser, you may see unexpected results.

Statistical Literacy: Epidemiology

Key concepts in epidemiology for APRNs

Gold Standard versus Point of Care Tests for Diseases

Let’s begin by saying there are gold standard tests. These tests are accurate in determining if someone has a disease or not. Examples of gold standard tests include biopsy or culture. These take time to analyze and may be painful for the patient.  So, people developed tests that are quicker and/or less invasive. Some of them are called Point of Care (POC) tests, rapid tests, etc. Some are not rapid, but do serve the purpose of being less invasive. For example, think of a simple urinary tract infection (UTI): in clinic, the patient provides a urine sample to test. Cultures typically take three days, agonizing and dangerous for the patient. Alternately, a urine dip takes less than one minute.  

There is a trade-off for using POC tests: although they are quicker and less invasive, POC tests are typically not as accurate at determining disease status as the gold standard. There are more false positives and false negatives than in the gold standard test. In fact, the manufacturer chooses whether the test errs on the side of false positives or false negatives, which depends on the consequences of being wrong.  

In short, when we consider how the selected POC test compares to the gold standard test or, ideally, the "truth" about disease status, you must understand specificity and sensitivity. As a provider, you must know how to use these tests as tools. 

Sensitivity and Specificity

Sensitivity refers to true positives: how often you are able to get a positive test when disease is present.  

We are talking about positive people. So the test either correctly identifies them as positive (true positive) or incorrectly identifies them as negative (false negative). In other words, tests with a high sensitivity have few false negatives.  
How to Interpret:  If 100 patients known to have a disease were tested, and 60 correctly test positive, then the test has 60% sensitivity.  

Specificity refers to true negatives: how often you are able to get a negative test when disease is absent.  

We are talking about negative people. So the test either correctly identifies them as negative (true negative) or incorrectly identifies them as positive (false positive). In other words, tests with a high specificity have few false positives. 
How to Interpret:  If 100 patients known to have no disease were tested, and 60 correctly test negative, then the test has 60% specificity.


SPIN and SNOUT
Some like to use the SPIN and SNOUT method to apply sensitivity and sensitivity concepts for clinical application. One hint: if you use SPIN (specific tests rule in disease) and SNOUT (sensitive tests rule out disease), put in an extra letter. SPIN could be written more accurately as SPPIN for specific tests that are positive help rule in disease. SNOUT could be written as SNNOUT or sensitive tests that are negative help rule out disease.

Use in Patient Care:  The sensitivity and specificity of a test are listed in the pamphlet that comes with the test. You can also find this information online. Knowing these values is the first step.

EXAMPLE

Let’s use a Rapid Strep (POC) test as an example:

  • The sensitivity of the test in your clinic is 70%.
  • The specificity of these tests is 95%.

In plain language:

  • Sensitivity means the ability to detect disease in those who have disease. In positive people, these tests would correctly identify disease in 70% of the cases (true positive). On the other hand, we would not identify disease in 30% of the positive people (false negative).
  • Specificity means the ability to detect no disease in those who do not have disease.  In negative people, these tests would correctly identify no disease in 95% of the cases (true negative). On the other hand, we would identify disease in 5% of the negative people (false positive).

Likelihood Ratios

Likelihood ratios (LR) are another format used to assess the value of performing a diagnostic test. A positive likelihood ratio will tell us how likely the test will be positive in a patient with the disease. A negative likelihood ratio will tell us how likely the test will be negative in a patient without the disease. Likelihood ratios use both sensitivity and specificity to assess a given test.
How to Interpret:

  • For a positive likelihood ratio (numbers typically greater than 1), the larger the number the more likely the patient with a positive test has the disease. The threshold for a positive LR is typically 3.0 or more as it increases the probability that there is disease.
  • For a negative likelihood ratio (numbers should be less than 1), the smaller the number the more likely the patient with a negative test is without the disease.The threshold for a negative LR  is typically 0.3 or less as it increases the probability that there is not disease.
  • Likelihood ratios are actually discussing odds. Therefore, if you see an LR of 1.0 it means no difference in odds of having a disease based on the diagnosis from the test. In other words, the test result does not tell us about the likelihood of disease presence or absence.

Reliability and Validity

Validity refers to if the test measures what it is supposed to measure. If the patient has a blood sugar of 60 will the glucose monitor read 60?  In terms of surveys, is the survey discovering who is undernourished?
 
Reliability refers to the consistency of the test. Will the test yield the same results if it were applied to the same person over and over? In terms of surveys, if the person was surveyed more than once, would the results be the same?

EXAMPLES
  • Reliability: the repeatability or reproducibility of measurements
    • Patient Care: If one nurse obtains a blood pressure reading and then another nurse repeats the measurement, the results of their measurements should be almost identical.    
    • Research: Can another researcher reproduce what was done in a study because there was enough detail provided that the research can be replicated? If so, and the results are the same, then the measurements are reliable. Is there evidence of interrater reliability in the research report? If so, then there is reliability for that aspect of the research.
  • Validity: accuracy of the measurement
    • External: Does the method used to answer the research question make sense? In other words, does it address the phenomena of the question? This is known as being valid on its face, or prima facie validity.
    • Internal: the measurement truly measures the variable of interest.

Use in Patient Care:  Each nurse takes the patient's blood pressure using the correct size cuff and using the correct technique. The result is that measurements obtained are reliable (repeatable) and valid (accurate).

Predictive Values

How Predictive Values are Calculated
The predictive values are calculated using the sensitivity and specificity of the test, and the prevalence of the disease in the population.

Positive Predictive Value (PPV) is used to indicate the likelihood of disease in a patient when the test result is positive. This answers the question “If I have a positive test, how likely is it that I have a disease?”
The positive predictive value of a test would be higher if the prevalence of disease was higher, and would be lower if the prevalence was lower.  In other words, if the disease is very common, a person with a positive test result is more likely to actually have the disease than if a person has a positive test in a population where the disease is rare.
How to Interpret:  For PPV, the higher the value, the more likely the patient has the disease when they have a positive test.  The scale used for this is 0-100%.

Negative Predictive Value (NPV)  is used to indicate the likelihood that the patient does not have the disease when the test result is negative. This answers the question “If I have a negative test, how likely is it that I do not have a disease?”
The negative predictive value of a test would be lower if the prevalence of disease was higher, and would be higher if the prevalence was lower. In other words, if the disease is very common, a person with a negative test result is more likely to actually have the disease than if a person has a negative test in a population where the disease is rare.
How to Interpret:  For NPV, as for PPV, the higher the value the more likely the patient does not have the disease when they have a negative test. The scale used for this is 0-100%

Use in Patient Care:  If the PPV of a screening test is 0.12, or 12% this means that among those who had a positive screening test, the probability of disease was 12%. That is not very high!  We would be giving a lot of people a scare if we were not careful with how we shared this positive result. A positive in this case means further evaluation is needed. If we wanted to increase this PPV we would only test in a population where the prevalence of disease is higher.  
If the NPV of a screening test is .99 or 99% this means that among those who had a negative screening test the probability of not having the disease is 99%. That is reassuring. If we tested in a population with a high prevalence of disease, the NPV would decrease (opposite of what the PPV does).