38 CFR § 1.17 - Evaluation of studies relating to health effects of radiation exposure.
(a) From time to time, the Secretary shall publish evaluations of scientific or medical studies relating to the adverse health effects of exposure to ionizing radiation in the “Notices” section of the Federal Register.
(b) Factors to be considered in evaluating scientific studies include:
(1) Whether the study's findings are statistically significant and replicable.
(2) Whether the study and its findings have withstood peer review.
(3) Whether the study methodology has been sufficiently described to permit replication of the study.
(4) Whether the study's findings are applicable to the veteran population of interest.
(5) The views of the appropriate panel of the Scientific Council of the Veterans' Advisory Committee on Environmental Hazards.
(c) When the Secretary determines, based on the evaluation of scientific or medical studies and after receiving the advice of the Veterans' Advisory Committee on Environmental Hazards and applying the reasonable doubt doctrine as set forth in paragraph (d)(1) of this section, that a significant statistical association exists between any disease and exposure to ionizing radiation, § 3.311 of this chapter shall be amended to provide guidelines for the establishment of service connection.
(1) For purposes of paragraph (c) of this section a significant statistical association shall be deemed to exist when the relative weights of valid positive and negative studies permit the conclusion that it is at least as likely as not that the purported relationship between exposure to ionizing radiation and a specific adverse health effect exists.
(2) For purposes of this paragraph a valid study is one which:
(i) Has adequately described the study design and methods of data collection, verification and analysis;
(ii) Is reasonably free of biases, such as selection, observation and participation biases; however, if biases exist, the investigator has acknowledged them and so stated the study's conclusions that the biases do not intrude upon those conclusions; and
(iii) Has satisfactorily accounted for known confounding factors.
(3) For purposes of this paragraph a valid positive study is one which satisfies the criteria in paragraph (d)(2) of this section and whose findings are statistically significant at a probability level of .05 or less with proper accounting for multiple comparisons and subgroup analyses.
(4) For purposes of this paragraph a valid negative study is one which satisfies the criteria in paragraph (d)(2) of this section and has sufficient statistical power to detect an association between exposure to ionizing radiation and a specific adverse health effect if such an association were to exist.
(e) For purposes of assessing the relative weights of valid positive and negative studies, other studies affecting epidemiological assessments including case series, correlational studies and studies with insufficient statistical power as well as key mechanistic and animal studies which are found to have particular relevance to an effect on human organ systems may also be considered.
(f) Notwithstanding the provisions of paragraph (d) of this section, a significant statistical association may be deemed to exist between exposure to ionizing radiation and a specific disease if, in the Secretary's judgment, scientific and medical evidence on the whole supports such a decision.