The Department of Labor (DOL) is responsible for determining the probability that a cancer included in a claim under EEOICPA (The Act)
was caused by workers' exposure to radiation during nuclear weapons production. DOL will determine the "probability of
causation" for each claim for which NIOSH is required to complete a dose reconstruction. Generally, with some exceptions,
these claims are for workers who are not a member of the Special Exposure Cohort.
To answer questions about the probability of causation, NIOSH developed a list of Frequently Asked Questions (FAQs) below.
Links to other program FAQs are located on the "Find It!" navigation box under "On this page . . . "
Select the question you are interested in below by clicking its link. You will be taken to the answer located on this page. Links throughout the FAQs will guide you to further information.
Top of Page
Probability of Causation
 What is "probability of causation"?
Under EEOICPA (The Act), probability of causation is a measure of how likely it is that an energy employee's cancer was "at
least as likely as not" caused by occupational exposure to ionizing radiation.
Top of Page
 What does dose reconstruction have to do with the probability of causation?
The Department of Labor (DOL) will use the energy employee's personal characteristics, employment, medical information, and dose reconstruction
results developed by NIOSH to determine the probability of causation.
Top of Page
 What is the Probability of Causation rule?
This rule provides the guidelines that DOL uses to determine the degree of likelihood that a cancer included in a case
was "at least as likely as not" caused by the energy employee's exposure to radiation during nuclear weapons
production.
Top of Page
 Where can I find a copy of the Probability of Causation rule?
A copy of the Probability of Causation rule can be found on our Web site on The Act page.
Top of Page
 How is probability of causation calculated under EEOICPA (The Act)?
NIOSH developed a computer program that is used by DOL to determine the probability of causation. The computer program,
known as NIOSHIREP, uses cancer risk models developed primarily by the National Cancer Institute (NCI) to estimate the
probability of causation under The Act.
The risk models take into account the energy employee's cancer type, year of birth, year of cancer diagnosis, exposure
information (years of exposure, radiation type, and dose). Smoking history is taken into account for lung cancer cases.
Ethnicity is taken into account for skin cancer cases. Under Subtitle B of The Act, exposures to other occupational,
environmental, or dietary carcinogens are not currently taken into consideration for the probability of causation
calculation.
Top of Page
 What is the significance of the probability of causation calculation?
Probability of causation is expressed as a percentage between 0 and 100. A value of 100% means that it is certain that the
radiation dose was the cause of the cancer. A value of 0% means that it is certain that the radiation dose was not the cause
of the cancer. As defined in The Act, probability of causation values greater than or equal to 50% at the upper 99th
percentile credibility limit mean that it is "at least as likely as not" that the radiation dose caused the
energy employee's cancer. Therefore, if the probability of causation for your case is less than 50%, you will not be
entitled to compensation. If your probability of causation is between 50% or greater, DOL will recommend that you receive
compensation.
Top of Page
 Does NIOSH determine/calculate the percent that determines the compensability of the
claim?
No, NIOSH reconstructs a worker's radiation dose and determines a final estimate of the worker's exposure to radiation.
The results from a worker's dose reconstruction are sent to DOL. DOL determines the probability of causation  the
likelihood that the worker's cancer was related to radiation exposure received while employed at a covered facility.
Top of Page
 What is uncertainty?
A term used to describe the lack of precision and accuracy of a given estimate, the extent of which depends upon the amount and quality of the evidence or data available.
Top of Page
 What are some examples of uncertainty?
Some statistical issues associated with epidemiological studies (e.g., possible measurement error, possible effects of confounding variables, etc) are:
 Amount of radiation exposure
 Doseresponse assumptions
 Transfer of "excess risk" from Japanese and other exposed populations to EEOICPA claimant population
Top of Page
 What is the effect of uncertainty on the probability of causation?
 Uncertainty is built into nearly every part of the process of estimating causation
 Uncertainty distributions are used in dose reconstruction, cancer risk models, etc.
 "Credibility limits" are placed around the central estimate of PC
 Upper 99th percentile determines final PC result and claim outcome
 An accounting of uncertainty is necessary because the upper 99% credibility limit of PC is used in adjudicating claims for compensation
 Uncertainty is often a major contributor to PC
Top of Page
 What is an uncertainty distribution?
A range of discrete or continuous values arrayed to encompass, with high confidence, the true but unknown values of a given quantity or parameter.
Top of Page
 What is the Upper 99th percent credibility limit?
The Upper 99th percent credibility limit is the 99th percentile of the range of values in an uncertainty distribution. The uncertainty distribution of the probability of causation (PC) is first estimated, and the upper 99th percentile of this distribution is compared to the decision criterion (a PC of 50%) to determine eligibility for compensation.
Top of Page
 What factors could influence the probability of causation result?
 Dose reconstruction efficiency process
 Erring on the side of the claimant
 During case development by DOL
 Selection of IREP cancer model: policy often requires running 2 or more models and using the highest PC
 Use of NIOSHIREP "combined" lung cancer risk model
 Handling of "unknown primary" cancers
 Handling of claims with more than one primary cancer
 Special "multiple primary" equation increases the likelihood of compensation.
 NIOSHIREP cancer risk models
 Personal characteristics: gender, age at exposure, age at diagnosis, latency, smoking history, ethnicity
 Reconstructed radiation dose: organ dose, radiation type, radiation energy, exposure rate
 Uncertainty: in cancer risk models, in dose
 Policy: multiple primary cancers, unknown primary cancer, alternate cancers
Top of Page
 How does NIOSH calculate the probability of causation when a claim has more than one primary cancer?
 NIOSHIREP includes a procedure to calculate an upper 99% credibility limit of PC when a claim for compensation involves more than one primary cancer.
 In all these cases, NIOSHIREP is run for each primary cancer separately, and the upper 99% credibility limit of probability of causation (PC) for all cancers combined is computed based on the upper 99% credibility limits of all the cancers involved, using the Multiple Cancer Equation:
PC_{total} = 1  [(1  PC_{1})x(1  PC_{2})x...x(1  PC_{n})]
Top of Page
 NIOSH completed my dose reconstruction and DOL calculated the probability
of causation for my claim to be less than 50%. As a result, DOL's recommended decision was to deny compensation
for my claim. Since that time, I have had two additional cancers. I reported the two additional cancers to DOL and
my case was returned to NIOSH for a new dose reconstruction. NIOSH completed the new dose reconstruction using
my two additional cancers and when DOL determined my probability of causation, it was again less than 50%. In
fact, once the two new cancers were added to my dose reconstruction, the probability of causation percentage
calculated by DOL was lower than the value of my first dose reconstruction. How is this possible? Shouldn't the
addition of two more cancers increase the probability of causation value?
NIOSH understands this concern. It may seem obvious that when additional cancers are added to a dose reconstruction
it would increase the probability that radiation caused the multiple cancers. However, in some cases, the NIOSH
dose reconstruction will result in a lower probability of causation when additional cancers are added. An
explanation for how and why this can happen is as follows:
Accurately estimating the exposure a worker received while working at a covered facility is a time consuming process.
In order to complete dose reconstructions in as timely and efficient a manner as possible, NIOSH may make
assumptions on dose estimates that are favorable to the claimant in order to simplify the dose reconstruction
process.
For instance, instead of completing a dose reconstruction which precisely estimates the worker's exposure for
cases that would most likely result in a probability of causation well below the 50% necessary for the claim to
be compensable by DOL, NIOSH will significantly overestimate the exposure based on the highest levels of exposure
observed or possible for the facility. If the claim will not be compensable even when using these significantly
overestimated exposure estimates, then no further refinement to the dose reconstruction is required.
This manner of dose reconstruction is called an "efficiency measure." It allows NIOSH to issue a
timely dose reconstruction where attempts to refine the exposure estimate would not result in a compensable claim
(i.e., a full dose reconstruction would in all likelihood produce a much lower probability of causation than the
overestimated exposure values used).
Anytime a new cancer(s) is reported to NIOSH by DOL, NIOSH is required to rework the dose reconstruction to include
the new cancer(s). If the new cancer(s) creates a potentially compensable case, NIOSH must refine the dose
reconstruction using more probable and precise exposure estimates and not the significant overestimates as mentioned
above. These new, more accurate exposure estimates will be lower than the original estimate because we no longer
use the assumptions which overestimated the radiation dose for the cancer found in the initial dose reconstruction.
The revised radiation dose estimate may result in a probability of causation for the new cancer(s) that is lower than
the noncompensable estimate that was obtained for the first cancer using the overestimated radiation exposure. That
is why the addition of a new cancer(s) may result in a probability of causation that is lower than the probability
found for the single original cancer. However, this does not always occur when a new cancer(s) is added to a claim and
NIOSH has to rework the dose reconstruction. It is possible for additional cancer(s) to produce a full dose reconstruction
that will be determined to be compensable by DOL.
The use of "efficiency measures" and procedures for dose reconstructions that NIOSH follows under The Act
are described in detail in Final Rule: Methods for Radiation Dose Reconstruction
under EEOICPA42 CFR 82 [ PDF 129 KB (23 pages)].
Top of Page
 Who do I contact if I have additional questions about probability of causation?
If you have additional questions about probability of causation, please contact DCAS by email at dcas@cdc.gov or ocas@cdc.gov or 5135336800 (tollfree at 18772227570).
Top of Page
