The Report on Carcinogens
What Went Wrong and What Can Be Done to Fix It
Full Document Available in PDF
There is nothing wrong in principle with publishing periodic reports identifying substances that pose carcinogenic risks to humans. Cancer remains a serious disease even though advances in diagnosis and treatment have rendered most types much less often fatal than they were when President Richard Nixon declared the “war on cancer” in 1971. But it would be a mistake to continue basing these reports on scientific knowledge and primitive technology dating from the 1960s.
The National Toxicology Program’s (NTP) Report on Carcinogens (RoC) is one such periodic report. The NTP has interpreted its statutory charge in a way that never was consistent with the law authorizing its preparation, resulting in Reports that never could live up to Congress’ original intent. Though the law requires the NTP to estimate the number of Americans actually exposed, and to list substances only if a significant number of Americans are exposed to them, the NTP functionally ignores exposure. The law also requires the NTP to estimate the reduction in cancer incidence resulting from regulatory standards, but it does not perform that required task, either.
Problems with the RoC begin with the NTP’s listing criteria. A careful review of the text shows that they are mere tautologies. For example, a substance is deemed to be a known carcinogen if the NTP decides that the evidence from human studies is sufficient. The minimum threshold for designation as a known carcinogen is unknown to the public because the NTP never says what is required for evidence to be sufficient. Thus, a substance is a known human carcinogen if the NTP says the evidence is sufficient. Conversely, the evidence is sufficient if the NTP says the substance is a known carcinogen. Similarly circular logic pervades the definition of a reasonably anticipated human carcinogen.
Worse, the NTP appears to be institutionally incapable of incorporating decades of advancements in scientific knowledge into its listing decisions, and there is no transparent way to scientifically rebut or reverse a listing decision once it has been made.
Both of the statutory categories for assignment (known and reasonably anticipated) imply that a causal relationship has been demonstrated with near certainty in the first case, and with an unspecified but lesser confidence level in the second. But the NTP’s listing criteria do not require any demonstration of causality. Rather, the NTP assumes that causality is demonstrated when it decides to list. This is clear from the grammatical structure of the criteria, which treats causality as a merely parenthetical element.
This enables the NTP to reserve to itself the discretion to consider whatever information it wants, to exclude whatever information it wants, and to evaluate that information in accordance with whatever ad hoc criteria it wants to apply. The NTP does not constrain itself to scientific information, either. By withholding B. 2 Belzer: The Report on Carcinogens from the public the weight of evidence scheme, the NTP preserves the policy discretion to give any weight it wants to policy goals and objectives, and to keep those weights hidden from public view.
This paper concludes with specific recommendations for statutory reforms that would improve the scientific quality of listing decisions and the practical utility of the RoC for screening-level risk-benefit decision-making. Each recommendation would help restore science to its intended role and end the NTP’s science charade.