Early in my career as a medical officer with NIOSH, I became acutely aware that the occupational link for illnesses was frequently being missed. I commented in a letter published in the New England Journal of Medicine in 1980 that a high percentage of pathologists missed a straightforward case of asbestosis on autopsy. Since then, occupational safety and health professionals have made great strides in identifying work-related health effects. But the question remains, are we doing a good job of determining when something is caused by work?
I am sure we are all frustrated by some clinicians' matter-of-fact opinion "well, he works with chemicals; therefore his asthma (or you fill in the condition) must be the result of his job." Another good old standby is "I can't find any other reason; therefore, it must be work!"
Let's discount the obvious motivators for making an arbitrary work-relationship:
Higher medical fee schedule
Worker's compensation indemnity payments
"Keeping the patient happy"
What unfortunately remains is that few clinicians employ a formal scientifically based methodology to make a determination of causation. A true "weight of the evidence" decision based on sound science requires criteria weighting on either a broad substance exposure effect/ causal association or an individual worker basis. For example, did exposure to triethyl doorknob cause my permanent hair loss (a concern near and dear to my head!)? In order to begin to address a causal relationship in this instance, one must answer the following questions:
What is permanent hair loss?
Did the exposure occur before or after the primary hair loss?
How much exposure must occur over what timeframe for the effect to occur?
Did the exposure simply aggravate an underlying condition?
Has the effect been documented in humans?
Armchair scientists will use deductive reasoning like Sherlock Holmes in causal determination: "If you rule out all probabilities, the remaining possibility, no matter how unlikely is the culprit!" The fatal flaw in this approach is that, in medicine, we simply don't know the cause of the disease in many instances. For example, most causes of interstitial fibrosis are unknown or idiopathic. Simply because a worker happened to work in a building containing asbestos doesn't mean his interstitial fibrosis is asbestosis, the interstitial fibrotic disease produced by high-level, long-term asbestos exposure.
On a global basis, inappropriate deductive reasoning can also lead one astray. Consider this statement: "Everyone who ate a pickle in 1869 is dead today. Therefore pickles cause death."
Statistically speaking, there is a 100% association in the pickle example, but, epidemiologically, it is clearly without scientific foundation. It is simply not biologically plausible. But as some might point out, "somebody could choke to death on a pickle"! Possibility does not equate to probability, much less certainty.
