DIFFERENTIAL DIAGNOSIS VERSUS CAUSATION ASSESSMENT: WHY THEY

In Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S.579; 113 S. Ct. 2786; 125L Ed 2d 469 1993) the U.S. Supreme Court enumerated the factors required to reach a well-founded causal conclusion. In that case, the concept of differential diagnosis was neither part of the decision, nor was it relevant. Rather, Daubert involved causation assessment: the methodology by which an external cause is linked to an internal disorder or disease process.

The distinction between "differential diagnosis" and "causation assessment" is critical, yet commonly misunderstood by the Courts. Differential diagnosis, the "methodology" of physicians, has been improperly broadened to include both: uncovering a diagnosis or clinical disorder and identifying the cause of that disorder. This assessment of the cause of a disorder is pertinent to liability determinations, but it differs substantially from differential diagnosis. Physicians have, in some cases, abetted the Courts’ confusion by asserting that they performed a differential diagnosis, in the standard fashion, to determine causation. They may say: "After performing a standard differential diagnosis, I determined that mold caused the asthma," or, "In the course of standard differential diagnosis, I ruled out everything but the benzene in the well water as a cause of his leukemia." In actuality, in the first instance, the differential diagnostic process resulted only in a diagnosis of asthma. In the second instance, it led to a diagnosis of leukemia. In neither example does the exercise of differential diagnosis establish a cause of those disorders diagnosed by the physician.

This commingling of diagnosis and cause has been adjudicated by various trial and appellate Courts with mixed results, often leading to strange, medically-inaccurate decisions. A number of Courts have, erroneously, accepted the proposition that differential diagnosis leads to a causal assessment and that because differential diagnostic methodology is standard in clinical medicine so, too, is the resulting causation assessment. While the former is true, the latter is not.

The distinction between "differential diagnosis" and "causation assessment" is critical for the Courts to understand. If understood and implemented, it should define how conclusions are properly reached, the necessary factual bases for those conclusions, the recognized and appropriate methodologies for arriving at those conclusions and even, at times, who is qualified to testify.

"Differential diagnosis" and "causation assessment," both seek to uncover causes, but of very different things. An error, common to many decisions, is the Court’s failure to ask, "Cause of what?" Differential diagnosis is the methodology which attempts to identify the internal disease or process which produces or causes the patient’s symptoms or findings. Only rarely does it deal with the external cause of that disorder. That determination requires a separate causation assessment, an activity which utilizes an entirely different set of information.

The cases to be discussed illustrate how the Courts have confused diagnosis and differential diagnosis with cause and causation assessment and why better reasoned, more consistent decisions will emerge when these distinct endeavors are better understood. In Daubert v. Dow Merrill, differential diagnosis was not at issue. That case involved a congenital (birth) defect and its alleged causal connection to a drug, Bendectine, used to control morning sickness. The diagnosis was clear: the birth defect was unequivocal and unchallenged. Because the diagnosis was straightforward, a differential diagnosis was neither necessary, nor a part of the case. A birth defect was the diagnosed cause of the patient’s deformity. At issue, however, was the cause of that defect and the methodology employed by experts to reach their conclusion that Bendectine was the responsible agent, that is the cause of the birth defect, or the reason that it occurred. How the experts decided that Bendectine caused the defect was the fatal methodological flaw of the case and the essence of the Daubert decision which focused attention on the causation methodology utilized.

The Federal Judicial Center’s Reference Manual on Scientific Evidence (1994) defined the term "differential diagnosis" properly as:

In this matter, the diagnosis of RADS (Reactive Airways Dysfunction Syndrome), the claimant’s expert incorporated the cause of the RADS within the diagnosis of the RADS, asserting not only that the patient had RADS, but providing a causal attribution. "RADS due to toluene exposure," was the offered diagnosis. The Court agreed, incorrectly, that the determination of cause was part of that "differential diagnostic" activity. The appellate Court concurred. Troy Mattis, et al., vs. Carlton Electrical Products, et al., N. 98-4091, D.S.D.)

Comment: "Differential diagnosis" is not equivalent to "differential etiology." "Etiology" speaks more to the ultimate cause, that is, the cause of the disease process. Diagnosis speaks to the disease process: what is wrong with the patient. Thus, in this case, the differential diagnosis would establish the existence of RADS (Reactive Airways Dysfunction Syndrome). This diagnosis is the disorder which patient has (not what caused it) and is the end of the differential diagnostic analysis.

Etiology or cause of the RADS is the next question. This involves a new process necessitating answers to other types of questions: whether the agent at issue is capable of producing RADS, i.e., whether it has irritant properties; approximately what level of exposure occurred (dose); what was the timing between symptom onset and exposure (clinical course); were there alternate potential causes of the RADS which have been considered and ruled out? The evaluation of these critical elements is not part of diagnosis, but is central to causation analysis methodology.

In Glastetter v. Novartis Pharmaceuticals, the trial Court dismissed the claimant’s expert testimony that the drug Parlodel caused the claimant’s intracranial hemorrhage (ICH). Affirming that decision, the appellate Court discussed properly certain elements of causation assessment, but they incorporated, improperly, cause in the concept of a differential diagnosis. They said:

Comment: The first sentence of the above statement correct delineates the process of differential diagnosis. The second sentence, including ruling out other causes of the disease, is not. Once again, the Court intermingled "differential diagnosis" with "causation assessment."

The statement by Dr. Linz noted above and cited by the appeals Court actually incorporated both methodologies: differential diagnosis and causation assessment. The end of the differential diagnostic exercise was the diagnosis (carpal tunnel syndrome). Whether or not that diagnosis was accurate could be a matter of dispute between medical experts, but the method of arriving at that diagnosis, history, physical examination, laboratory testing, is not. In contrast to the misunderstanding of the Appellate Court, however, the diagnosis - and only the diagnosis - was the outcome of the differential diagnosis. All of the other components (which the claimant’s expert acknowledged as difficult) were elements of causation analysis. This included the general causation question: Can such activities cause this disorder, CTS? The causation analysis also included the specific causation questions: have other causes been considered and ruled out? Was the work activity sufficient? Was it more significantly intensive than non work-related activities?

The differential diagnostic process, then, is methodologically-designed to uncover a physical disorder or a disease state. It asks, for example, "Why is this person having a headache" Is it a brain tumor, a sinus infection, poor reading glasses, neck strain, or a migraine?" The methodology employed to answer this differential diagnostic question is the standard methodology of medicine: history taking, physical examination and diagnostic studies. Assume it is learned (through CT or MRI scanning) that the patient has a brain tumor. The diagnosis determined by the physician, brain tumor, explains the headaches (likely caused by the tumor), but the cause of the brain tumor has not been elucidated. Assume further that occupational exposures to chemicals at a petrochemical facility are the alleged cause of the tumor. This causal attribution: the central medical/scientific exercise, as well as the proper subject of Daubert challenges in toxic tort claims, requires causation assessment (not differential diagnostic) methodology. Both general causation and specific causation elements need to be satisfied before such an attribution can be established.

Thus, the differential diagnosis came first, the diagnosis was made; now, the ultimate claimed cause of the disease requires a separate step. It is that causal assessment step which is not the standard methodology of clinical medical practice. In actuality, it is generally foreign to medical evaluation and to a physician’s expertise. It is a methodology more common to public health, epidemiology, toxicology and other sciences. Because of those differences and because physicians are not, in general, expert in causation assessment, the validity and reliability of that causation testimony is not established by an expert’s statement, "Differential diagnosis is what I did and it is the standard methodology of medicine."

Why is this important? It is important, frankly, because failing to probe causation methodology undermines the intended purpose of Daubert’s emphasis on accepted and proper methodology. Confusing the methodology of clinical practice with the methodology of causal assessment permits physicians to opine about causal conclusions absent a proper analysis. The clinician can ignore temporal relationships, dose, duration of exposure, alternate causes, latency periods and even general causation (Is the agent at issue known to cause this disease?) and other factors, simply averring to causation, while ignoring contradictory facts.

Differential diagnosis is, in fact, the standard methodology of clinical practice. In general, however, it rarely uncovers the ultimate cause of the disorder; in a few instances it may. For infectious diseases, such as influenza or tuberculosis, the diagnosis of the disease, implies a cause as well: Tuberculosis bacillus, or the influenza virus. Similarly, certain acute poisonings bring diagnosis and cause together. If a child has all of the typical manifestations of aspirin toxicity, was found with an empty bottle of aspirin and has high blood salicylate levels, the diagnosis incorporates the cause: aspirin toxicity. The same is true for other acute poisonings such as a rattlesnake bite.

In most instances in clinical medicine, however, that is not the case. The diagnosis of a brain cancer, an autoimmune disease such as lupus, a neurological disorder such as MS, is the end of the differential diagnostic process, leads to no causal attribution. Moreover, assessment of causal attributions is unrelated to most clinical practice. Oncologists diagnose and treat cancers: they do not commonly sleuth out the potential causes. The same is so for the rheumatologist, who manages patients with scleroderma or lupus and the neurologist who deals with Parkinson’s disease or MS. Consider a more common example: a patient sees a physician complaining of pain in the leg. The differential diagnoses might include: soft tissue trauma, a pulled muscle, a fractured bone, arthritis of the joint, a tumor, an autoimmune process, etc. The differential diagnostic process considers these entities, works the patient up appropriately and concludes with the diagnosis.

If the patient does have a fracture, for example, that diagnosis does not establish the cause of that fracture, merely the cause of the patient’s pain. Whether the patient was in an automobile accident, fell out of a tree, or was hit by a baseball bat is not relevant to the diagnosis. It is, however, the essence of a causal assessment, i.e., the reason that the patient’s leg is broken is what causation is all about. For the broken leg, that causal determination may be easy to make by simply asking the patient, but if that injured person won’t or cannot tell the physician what happened, the cause may remain undetermined. In this case, a differential diagnosis was conducted and completed. Causation assessment was not.

It is far more complicated when exposures, rather than recent trauma, are at issue. A medical evaluation for "trouble breathing" may lead to the diagnosis of asthma. This diagnosis ends the differential diagnostic process. Determining the cause, or even or exacerbator or aggravator, is a new exercise. The physician cannot, for example, identify an indoor mold exposure as the cause, if the asthma predated the exposure, or if the patient has not undergone testing to establish allergy to the molds at issue.