Errors in clinical laboratory test selection and result interpretation: commonly unrecognized mistakes as a cause of poor patient outcome

In April of 1984, I was a clinical pathology resident carrying a beeper for the coagulation laboratory at Barnes Hospital in St. Louis, Missouri. I received a page from a highly regarded first-year resident in internal medicine about a patient who had a prolonged PTT for which he did not know the cause. I came to meet him for a face-to-face consultation about what to do next. At that time, I possessed the belief, like virtually everyone else in medicine, that doctors who treat patients know which laboratory tests to select and what the test results mean for every test which is ordered. What I learned from this resident is that in 1984, both residents and attending physicians were frequently guessing about which tests were necessary and what the test results meant. Considering how much the laboratory test menu has expanded in size, complexity and cost since 1984, imagine how much worse the problem is now. The resident informed me that if it was not heparin prolonging the PTT, he had no idea what it might be. He said that the response of most physicians to a prolonged PTT was to simply transfuse the patient with fresh frozen plasma until the PTT normalized. In 1984, one out of 20 bags of fresh frozen plasma was contaminated with hepatitis C or HIV. The internal medicine resident had no idea that removing fresh frozen plasma units out of the freezer at that time was equivalent to playing Russian roulette for the patients. I am sure there are thousands of individuals who have either died or received a liver transplant because they received fresh frozen plasma unnecessarily. Today, patients may attribute their need for a liver transplant from an infection they received after transfusion of a blood product, without ever inquiring whether the blood transfusion was necessary in the first place. Devastating errors associated with mistakes in clinical laboratory testing and result interpretation have been unrecognized for decades, and that problem is much worse today than it was in 1984. Only in the past few years has there been even a modest recognition that picking the wrong laboratory tests causes more than a minor inconvenience. The challenge of ordering the correct tests and interpreting the results correctly has become so great with the large test menu of tests that there is an acceptance of these errors as a part of medical practice today, no matter what the consequences are to the patient [1]. Most physicians have not been educated to confidently know simple things – like when the sensitivity of a test is 80%, the laboratory test in use will miss 20% of patients who have the disease for which they are testing.

A surgeon who removes the healthy kidney and allows the diseased kidney to remain makes a medical error that is recognized in minutes. Medication overdoses with highly toxic effects are noted immediately. On the other hand, consider the patient who appears in the emergency room short of breath, and has a pulmonary embolism requiring anticoagulation. Not uncommonly at all, the emergency room doctor mistakenly concludes that the dypneic patient has asthma and sends the patient home with a nebulizer. If this patient dies in the parking lot walking to his car as the pulmonary embolism expands because he was misdiagnosed and not immediately anticoagulated, few physicians recognize this as a medical error, despite the devastating consequences, attributed to the emergency room doctor who failed to select a D-dimer test to demonstrate that the patient was suffering from a pulmonary embolism. In 2013, risk management experts and fellow physicians rarely recognize this error, and even more surprising, the doctor who failed to order the correct test to diagnose the thrombosis may not even recognize that the failure to order a laboratory test could have prevented the patient’s death.

Over my career, I have created patient specific, expert driven interpretations for more than 50,000 cases that were evaluated with special coagulation tests [2]. From personal experience, I have observed that a high percentage of very knowledgeable physicians fail to order the correct tests and frequently misinterpret the test results. It extends beyond coagulation to autoimmunity, endocrinology, microbiology, and many, many other areas. With the explosion of genetic information, doctors are still asking me “what is an allele?”

It is not their fault. We are still teaching medical students what a myocardial infarction looks like through a microscope, without providing information about the changing thresholds for the troponin test which is central to the diagnosis of cardiac ischemia in a patient with chest pain. In a survey of medical schools that is currently in progress, we have learned that graduating medical students receive approximately 9 h of education over their entire 4 years regarding clinical laboratory test selection and result interpretation. Currently, they receive between 60 to 300 h of anatomic pathology training in medical school, but when they begin to practice medicine, there is an anatomic pathologist with the expertise to provide microscopic diagnoses. This is quite different for laboratory medicine. Selecting appropriate laboratory tests is an activity which physicians perform every day, and for which there is no specialist easily available in most institutions to guide them. We learned in another survey recently completed that most physicians “do not know what they don’t know.” Surprisingly, most physicians are confident evaluating disorders with laboratory tests that they have used since they were in training, even though many of these tests have been supplanted by more accurate and more efficient assays.

It is the problem of doctors not knowing what they don’t know that has led to a failure to recognize what may possibly be the most frequent cause of medical error – greater in number than errors in surgery almost certainly – because almost every patient is evaluated with laboratory tests. The presence of a prolonged PTT, a positive antinuclear antibody test, and an infectious agent which is not eliminated with apparently appropriate antibiotic treatment are all common encountered in clinical medicine [3]. If there is a serious medical error associated with even 1% of these 3 scenarios because the doctor has not ordered the correct tests or misinterpreted the test results, the number of preventable poor patient outcomes from laboratory associated medical errors is enormous.

At our institution (Vanderbilt University Medical Center), we have teams of physicians and residents who routinely review the medical record and laboratory test results to provide a patient specific narrative with the differential diagnosis, without a request from another physician to do so. This activity is similar to what occurs in anatomic pathology without a microscope. This occurs for every case for patients being evaluated for complex coagulation disorders, microbiology evaluations when the patient outcome is less than expected, all hematologic malignancies, all problems in transfusion medicine, CNS malignancies for which a genetic profile is informative, and endocrine associated hypertension. In short order, we will be doing the same thing for breast cancer with the radiologist, pathologist, and molecular genetic specialists meeting daily, and other major cancers where the testing is complex or the choice of chemotherapy requires a genetic evaluation to determine if the treatment will help or harm the patient. The teams are called “diagnostic management teams.” In 2005, the Institute for Quality in Laboratory Medicine at the Centers for Disease Control and Prevention recognized the efforts of what is now called the diagnostic management team in coagulation, which began at the Massachusetts General Hospital in the mid-1990s, as a novel and essential practice for accurate and rapidly generated diagnoses involving clinical laboratory tests [4]. Now that this diagnostic management team approach continues to be expanded through the whole of diagnostic medicine at Vanderbilt, and we are producing studies for publication to show the clinical and financial benefits of this approach, diagnostic management teams will hopefully be instituted, at least initially, in other major medical centers in the USA. The goal is to generate a new standard of care and greatly reduce the under-recognized medical errors associated with laboratory test selection and result interpretation.