The Battle of Sensitivity and Specificity: Treatment or Prevention?

“The larger threat posed by American medicine is that more and more of us are being drawn into the system not because of an epidemic of disease, but because of an epidemic of diagnoses,” H. Gilbert Welch, Lisa Schwartz and Steven Woloshin wrote in the New York Times on January 2, 2007.

The authors continued, “This epidemic is a threat to your health. It has two distinct sources. One is the medicalization of everyday life. Most of us experience physical or emotional sensations we don’t like, and in the past, this was considered a part of life…The other source is the drive to find disease early. While diagnoses used to be reserved for serious illness, we now diagnose illness in people who have no symptoms at all, those with so-called predisease or those “at risk.””

The last statement raises an interesting question and debate. Is it more efficacious both to an individual’s health as well as more cost effective to identify high risk patients and attempt to prevent disease, or do such actions decrease the specificity of a diagnoses and increase the number of treated false positives; that is a patient who appears to be at risk is treated despite really not having an elevated risk?

Prevention has been touted as being less expensive than treatments. In the United States the majority of health care costs are spent in a patient’s last six months of their life. And with the United States population demographics changing (those over 65 years old will triple in the next 20 years) it would seem to be intuitive to increase early disease detection as well as preventive measures to deflect health care costs.

The definitions of prevention are important here. Primary prevention methods are those used to prevent a disease from ever occurring, such as exercise and eating healthy are preventive against heart disease. Secondary prevention methods are those used in individuals with known risk factors, such as treating blood pressure to reduce stroke risk by up to 40%. Tertiary prevention are essentially treatments, such as someone who has had a myocardial infarction (MI) or heart attack receives a defined series of medications.

Of those three prevention methods, which one sounds like it will decrease the morbidity of a disease and is the cheapest? Obviously primary prevention. However, the cost of primary prevention is a lower sensitivity and more false positives.

Specificity and sensitivity definitions are important as well. Specificity rules in a diagnosis, while sensitivity rules out diseases. For instance, a test called D-Dimer is often ordered to rule out a Pulmonary Embolism (PE), a blood clot in the lung. PE’s often have an elevated D-Dimer, but not all elevated D-Dimers indicate a PE. Therefore, the test is sensitive for PE, but does not diagnosis it. A CT or V/Q Scan confirms a PE diagnosis; they are specific tests.

The authors are concerned today’s medical efforts are leading to more incorrect diagnoses and possible un-needed treatments. The real issue the authors are hinting at is that of specificity v. sensitivity. Has medicine sacrificed specificity for sensitivity?

This over-treatment of which the authors are troubled seems to be most prevalent in psychiatric diseases. Bipolar has become all the rage and everyone under the sun with mood swings now receives this diagnosis as well as the cocktail of pills psychiatrists use to treat it. Restless Leg Syndrome (RLS) has been associated with Parkinson’s Diseases as well as possible malignancies in middle-aged men. Now anyone whose legs move has Restless Leg Syndrome and is treated. But there has not been enough time to see if these preventive methods have decreased the incidence of Parkinson’s Disease or reduced malignancies in middle-aged men.

However, it is certainly advantageous to use a secondary prevention method such as a baby aspirin to prevent stroke after a patient had a transient ischemic attack, TIA or mini-stroke. The authors certainly would not argue this point.

So is medicine correct now to reduce specificity and increase diagnoses in order to potentially avoid future complications of certain syndromes and diseases? It is a very good question. I do not like sacrificing specificity for sensitivity. I prefer to treat people who definitely have a disease, but I am also a big proponent of prevention. So the solution to this is to understand when sensitivity is more advantageous than specificity, such as in the secondary prevention of stroke but not in the treatment of an individual who may or may not have bipolar disorder.

Medicine is moving toward prevention and sensitivity is trumping specificity in diagnoses and treatment/prevention. The authors are correct in showing anxiety about these movements and the potentially unnecessary treatments that result. However, there clearly are advantages to some forms of prevention. The solution to this problem seems to be rather simple: understand when using sensitivities in order to prevent disease outweighs the risks of waiting for specific tests to confirm a diagnosis. It sounds simple, but in reality it really is not. Hopefully, with a better understanding of disease mechanisms and disease course more effective prevention methods will be employed with high sensitivities as well as specificities.