Medical testing accuracy varies in properly identifying disease

By Brandon Peters, MD 

Updated on March 16, 2022

 Fact checked by Angela Underwood

In the context of health care and medical research, the terms sensitivity and specificity may be used in reference to the confidence in results and utility of testing for conditions. Learn about these terms and how they are used to select appropriate testing and interpret the results that are obtained.

 RapidEye / E+ / Getty Images

Uses of Medical Tests

As soon as you start telling your healthcare provider the constellation of symptoms that you have, they will begin to formulate a hypothesis of what the cause might be based on their education, prior experience, and skill. The cause may be obvious. However, in some cases, several potential diseases may be suspected. Additional testing may be necessary to sort out the underlying contributors. The selection of these tests may rely on the concepts of sensitivity and specificity.

To make a diagnosis, healthcare providers may perform a complete physical examination, take body fluid samples (such as blood, urine, stool, or even saliva), or perform other medical tests to confirm or reject their initial hypotheses. Useless tests that cannot rule in or out certain diseases should be avoided. Ideally, a test will be chosen that can accurately confirm the diagnosis that is suspected.

Another use of medical testing is in screening tests given to identify diseases that a certain group may be at higher risk of developing.¹ They are not done to diagnose an illness, but to find one that may not yet be producing symptoms. As well, personal risk factors may increase the risk of an unidentified disorder and suggest earlier or more frequent screening. These factors include ethnicity, family history, sex, age, and lifestyle.

Considering the purpose of a test in certain populations requires careful consideration of both sensitivity and specificity. This helps both healthcare providers and patients to make the best decisions about testing and treatment.

Understanding Sensitivity and Specificity

Not every test is useful to diagnose a disease. Unfortunately, modern health care also cannot sustain the costs associated with unlimited testing. A healthcare provider must carefully select the most appropriate test for an individual based on specific risk factors. Choosing the wrong test may be useless, a waste of time and money, or it may even lead to a false positive test, suggesting the presence of a disease that is not actually present. Let’s consider how these characteristics of testing impact the test that is chosen and the interpretation of the results that are obtained.

When medical research develops a new diagnostic test, the scientists try to understand how effective their test is at properly identifying the target disease or condition. Some tests may not find a disease often enough in patients who are really sick. Others may incorrectly suggest the presence of a disease in someone who is actually healthy.

Healthcare professionals take into consideration the strengths and weaknesses of tests. They try to avoid any choices that might lead to the wrong treatment. For example, in diagnosing someone with cancer, it may be important not only to have an image that suggests the presence of the disease, but a tissue sample that helps to identify the characteristics of the tumor so the right chemotherapy may be used.²It would be inappropriate to solely depend on a single test that is not accurate in identifying the presence of cancer, and then start a treatment that may not actually be needed.

In situations where one test is less than certain, multiple tests may be used to increase the confidence of a diagnosis. Two useful measures of a test’s diagnostic strengths are sensitivity and specificity. What do these terms mean?

Sensitivity indicates how likely a test is to detect a condition when it is actually present in a patient.¹ A test with low sensitivity can be thought of as being too cautious in finding a positive result, meaning it will err on the side of failing to identify a disease in a sick person. When a test’s sensitivity is high, it is less likely to give a false negative. In a test with high sensitivity, a positive is positive.

Specificity refers to the ability of a test to rule out the presence of a disease in someone who does not have it.¹ In other words, in a test with high specificity, a negative is negative. A test with low specificity can be thought of as being too eager to find a positive result, even when it is not present, and may give a high number of false positives. This could result in a test saying that a healthy person has a disease, even when it is not actually present. The higher a test’s specificity, the less often it will incorrectly find a result it is not supposed to. 

It may seem logical that both a false negative and false positive should be avoided. If the presence of a disease is missed, treatment may be delayed and real harm may result. If someone is told they have a disease that they do not the psychological and physical toll may be significant. It would be best if a test had both a high sensitivity and a high specificity. Unfortunately, not all tests are perfect. It can be necessary to find a balance that matches the purpose of the testing to the individual being evaluated.

Comparing Tests

The best test (or group of tests) for diagnosing a disease is called the gold standard.¹ This may consist of the most comprehensive and accurate testing or measurements available. When new tests are developed in research, they will be compared to the best available testing currently in use. Before being released for wider use in the medical community, the new test’s sensitivity and specificity are derived by comparing the new test’s results to the gold standard. In some cases, the purpose of the test is to confirm the diagnosis, but some testing is also used more widely to identify people at risk for specific medical conditions.

Screening is when a medical test is given to a large population of patients, with or without current symptoms, who may be at risk for developing a specific disease. Some examples of proposed screening tests for potential medical conditions include but are not limited to the below:¹

• Breast cancer (mammography)

• Prostate cancer (prostate-specific antigen or PSA)

• High cholesterol (cholesterol panel)

• Cervical cancer (pap smear)

Not everyone needs to be screened for colon cancer at a young age, but someone with a specific genetic condition or a strong family history may require the evaluation. It is expensive, and somewhat invasive, to do the testing. The test itself may have certain risks. It is important to strike a balance between selecting the appropriate person to be tested, based on their risk factors and relative likelihood of having the disease, and the utility of the testing available.

Everyone is not tested for every disease. A skilled clinician will understand the pre-test probability of a specific measurement, or the likelihood that a test will have an anticipated result.

Screening for specific diseases is targeted to at-risk people. To find and treat a condition in the highest number of people possible, the costs of the testing must be justified and false positives must be avoided.

Positive and Negative Predictive Value

It is appropriate for healthcare providers to consider the risks of a disease within an untested group through the lens of two additional considerations: PPV and NPV.¹ 

Positive predictive value (PPV) is the number of correct positive results of a test divided by the total number of positive results (including false positives). A PPV of 80% would mean that 8 in 10 positive results would accurately represent the presence of the disease (so-called “true positives”) with the remaining two representing “false positives.”

Negative predictive value (NPV) is the number of correct negative results a test gives divided by the total number of negative results (including false negatives). An NPV of 70% would mean that 7 in 10 negative results would accurately represent the absence of the disease (“true negatives”) and the other three results would represent “false negatives,” meaning the person had the disease but the test missed diagnosing it.

PPV and NPV, combined with the frequency of a disease in the general population, offer predictions about what a broad-scale screening program might look like.