Four Elements in Statistics and Their Importance in Ethics of Medical Care
You may wonder why any patient should concern themselves with the statistical terms of specificity, sensitivity, positive predictive value and negative predictive value. I would say to become educated in these four statistical elements can be more meaningful and useful to the patient than the education provided by direct to consumer drug advertising. Drug companies might disagree with me but then educating the public in order to think about, even by those who don't need it, and perhaps get prescribed their new drug is the drug company's primary goal. Educating the public about these statistical elements has no associated conflict of interest and is of importance to all the public.
Back to the statistical elements: Patients are frequently being ordered and submitting to tests and procedures which are designed to help make a diagnosis or provide some scientific basis for appropriate therapy. And this is where the need for not only understanding but finding the data to support these tests is essential for the ordering physician but also should be essential for the patient on the receiving end to be aware. After all, tests and procedures are often expensive but also may take up the patient’s time. Tests and procedures may be invasive to the patient and there occasionally may be a unwanted reaction. Tests and procedures may unexpectedly turn up results that suggest some abnormality not initially considered but later after more tests, more cost, more risk that abnormality was a mistaken impression,not related to the disorder under investigation and was never present in the first place. And finally, with regard to that disorder under investigation, the test or procedure may provide the wrong answer. The disease will be determined as confirmed as present but it really is not present but is a false positive. The disease can also be confirmed as absent but it really is not absent but just missed by the testing and becomes a false negative.
Before the doctor ordering and the patient accepting most tests, both have to understand the significance of the test in the four different statistical elements:
Sensitivity-how sensitive is the test to correctly detect the disease if the disease is actually present?
Specificity-how specific is the test to correctly determine that the disease is absent when it is actually absent.
Positive predictive value- Considering how common the disease is in the population, what are the chances that a positive test result will be correct for a specific patient.
Negative predictive value-Considering how common the disease is in the population, what are the chances that a negative test result will be correct for a specific patient.
What does this all have to do with medical ethics? These days, patients are endowed with the ethical principle of autonomy. They should have the opportunity through information provided by the physician and elsewhere to make their own final medical decisions including the acceptance of tests or procedures for diagnosis. Yes, the physician should be aware of the statistical value of the tests or procedures, he or she advises and should provide some education to the patient regarding the basis and value of such tests. Of course, it is not expected that the patient will have the same education and background experience of the doctor to evaluate any particular test. On the other hand, patients should be aware of the factors that the physician should be considering before giving advice which should include the statistical value of the test (as delineated above) which is part of the benefit but also the risks and burdens involved for the patient such as including side-effects and financial costs and the errors of the test. If a patient is educated about these four statistical factors, the patient can then have some understanding as to what to ask the doctor about the tests if the benefit and risks are still not fully understood by the patient in making their autonomous decision. It’s all about ethics and there is even more.
Beyond autonomy involved in decisions about tests is justice and the use of scarce resources and one of the scarce resources is the cost of medical care in the United States as an example. The statistical elements I have described play an even more important role in this regard when considering screening tests and exams where there are large numbers of the public being screened who do not have the disease being screened or when the significance of the disease with regard to life span or symptoms is trivial in certain segments of the population to be tested. Is there real benefit to the population for screening? Or do the statistical elements and the nature of the disease make for waste of health care dollars, time, physical risks and discovery of disorders which are not present but require additional testing to find that they are false positives?
All patients who are interested in making truly informed consent or dissent, should go to the British Medical Journal Sept. 27 2003 and read the article “Understanding sensitivity and specificity with the right side of the brain” by Tze-Wey Loong where with graphics, Loong helps the reader understand all four statistical elements.
I hope I have shown the importance of some knowledge of simple statistical analysis for all patients and suggest, with that knowledge, patients should consider challenging their physicians if the there is some question about the advised tests or procedures. If you disagree, write your comments here. ..Maurice.
Graphic: Louis Pasteur (1822-1895), Print Collection, Reynolds Historical Library.