Interpretation of biossay data

Draft document: Interpretation of biossay data
Submitted by Luiz Bertelli, Los Alamos National Laboratory
Commenting as an individual

Luiz Bertelli’s Comments on the Supporting Guidance Document Interpretation of Bioassay Data, Draft – 21 February 2007. - MINIMUM DETECTION LIMITS for 239Pu Page 10, paragraph 1: “The 1 mSv criterion is not achievable by individual monitoring in some cases, e.g. exposures to some physico-chemical forms of 239Pu or other actinides when assessed from urine monitoring.” Comment: This assumption is not correct since many laboratories have been adopting techniques like the Thermal Ionization Mass Spectrometry (TIMS), which shows minimum detection limits of 4 µBq per sample. Assuming that the urine monitoring corresponds to inhalation of a 239Pu Type S compound with AMAD = 5 µm with a monitoring period of 360 days, this limit corresponds to a committed effective dose of only about 0.2 mSv, which shows that the 1mSv criterion is achievable. I also recommend inserting this comment in section 4.3 Analysis of Excreta and Other Biological Materials (page 34, last paragraph) as it was done for ICP-MS for analyzing uranium and thorium samples. - THE ASSUMPTION OF A CONSTANT CHRONIC INTAKE AS AN UNBIASED ESTIMATE OF INTAKE Excerpts from the draft document: a) Page 38: Section 5.2 General Principles for the Design of Individual Monitoring Programmes: “Routine monitoring is performed where intakes could arise as a result of an essentially continuous risk of contamination of the workplace from normal operations, or where undetected accidental intakes could occur.” b) Page 42, second paragraph, Section 5.3 Routine Monitoring “The approach adopted in ICRP Publication 78 (ICRP, 1997) defines a simple rule that limits the possible error on the estimate of intake arising from the unknown time of exposure. Monitoring intervals are selected so that any underestimation introduced by the unknown time of intake is no more than a factor of three.” c) Page 42, third paragraph, Section 5.3 Routine Monitoring “In practice the intake is often taken to occur at the mid-point of the sampling interval. The approach was used in ICRP Publication 78 and for consistency is also adopted in the OIR series of publications. It has been reported by Puncher et al that this can lead to a biased estimate of intake (Punchner et al 2006) and the assumption of a "Constant Chronic" method has been proposed which has been shown to give unbiased estimates of a workers intake. However this approach differs very little from the mid-point method (see Annex F). “ d) ANNEX F, Page 5, 5. CONCLUSIONS “Theoretical considerations aside, the choice of an appropriate method of estimating intakes is also dependent on broader considerations such as ease of implementation, and in many situations, the Mid-point method is adequate, considering other uncertainties. Indeed, this is the default method suggested in the main text of this document. However, it is always preferable to avoid bias, and if possible, the Constant Chronic method should be used. If a software program is being used to estimate the intake, then this method can be simply applied by selecting the intake regime to be constant and chronic throughout the monitoring interval.” e) 7.3.2 Time(s) and Pattern of Intake, page 55 “ While the mid point assumption is a pragmatic and simple approach and for consistency with previous advice in publications 54 and 78 is also applied in the OIR series of publications, it does not result in an unbiased estimate of the intake. Appendix F demonstrates that if this assumption is applied regularly, to an individual worker, it is subject to some bias and will tend to overestimate the worker’s real intake. It is always preferable to avoid bias, and it is demonstrated that this can be avoided by assuming a constant chronic intake. If this is of concern and a software program is being used to estimate intake, then the same result can be obtained by setting the intake to be constant and chronic throughout the monitoring interval.” Comment: I am strongly opposed to this proposal. Reasons: According to a), the fact that the risk of contamination can be continuous does not imply that the probability of an intake is high, otherwise the workplace will show unacceptable conditions of radiation protection. Since b) was adopted the frequencies of monitoring of 60 days for HTO and 30 and 60 days for 131I , which are shown in Table A1 of Annex F are unacceptable according to the ICRP78 criterion. In this way the values of 1.79, 1.8, 2.82, 2.85 and 8.28 are misleading and very biased! Moreover, true situations of "Constant Chronic" intake could be found perhaps in environmental exposures but wouldn’t be neither acceptable nor realistic in a workplace. However I support the use of continuous intake scenarios, as done in ICRP-78, as a way to see the trends in situations of long term accumulation of contents in tissues and in fecal and urinary excretion measurements and to use the corresponding graphs together with tables showing values of bioassay contents for short times after the single intake to help to analyze the contribution of short-decay terms when bioassay samples are collected after the weekend. Mathematically speaking, in a real workplace which has good radiological protection practices the probability of intake is as low as one thousandth. This means that even in a workplace which has many radiation workers there wouldn’t be sufficient intake cases to justify this theory and the potential biased cases would be minimal. In addition, higher estimated dose values require a follow-up sample, which will require another intake estimate which can substantially minimize uncertainties on the intake. Moreover, even assuming that the “Constant Intake” approach is correct and that the values for the magnitudes of the bias of intake estimation shown in Table A1 are considered, excluding those associated with the unacceptable frequencies discussed above, the practical question that must be discussed here is what is the true impact of, say 30% bias, on the total measurement uncertainty (IDEAS Types A and B errors) ? Finally, it would be very complicated to communicate to a worker that the method used to calculate his doses assumes a continuous intake in his workplace. The evidences above are strong enough to require that the text “It has been reported by Puncher et al that this can lead to a biased estimate of intake (Punchner et al 2006) and the assumption of a "Constant Chronic" method has been proposed which has been shown to give unbiased estimates of a workers intake. However this approach differs very little from the mid-point method (see Annex F).“ to be TOTALLY EXCLUDED from the third paragraph, Section 5.3 Routine Monitoring, page 42 as well as the whole Appendix F. The same should be done with the paragraph shown in e) (page 55, section 7.3.2 Time(s) and Pattern of Intake) and in the following text extracted from Section Routine monitoring “A recent paper has, however, noted that this can lead to some bias in the estimated dose and in some circumstances a more appropriate assumption would be to assume a constant chronic intake. This is described further in Annex F. “ It is almost certain that they will cause more confusion than elucidation if left in the document. - A PROVISIONAL TABLE IN IDEAS DOCUMENT BECAME ACCEPTED IN THE DRAFT. Section 7.6.5, page 65 The Table 7.12 had provisional status in the IDEAS documents and became accepted in the draft. Why ? - TABLE 7.12: NUMBER AND TYPE OF DATA REQUIRED FOR ASSESSMENT OF DOSE FOR SOME CATEGORIES OF RADIONUCLIDES AND THE RESPECTIVE MONITORING PROCEDURES The proposed numbers of bioassays are not connected with the reality. No one will give four bioassay samples in 10 days for estimating H-3 doses. It is financially unfeasible as well. Experience in practical radiation protection also tells that 3 urine and 3 fecal samples for Pu intake estimates are only requested in extreme accidental situations and that, specifically in the case of alpha-emitters, due to the radiochemical procedures it can take up to two weeks to have a result from a single sample. In this way, not necessarily the more it is requested the more improved will be the intake estimate. It must be pointed out that values recommended by ICRP are supposed to be applied in the many different countries all over the world and that they will become part of documents of national regulatory organisms and that laboratories must comply with the rules stated in those documents. In this way, practical and feasible recommendations are really important. - ANNEX B, PAGE 7 “Alternatively, the best estimate of intake can be determined using appropriate internal dosimetry software. The IMBA Professional software was used here to assess this case. It implements the current ICRP dosimetric and biokinetic models but enables the user to alter parameter values from the ICRP defaults. It uses the maximum likelihood method to fit multiple data and can assess the intake by fitting predicted values to different types of data simultaneously. The intake was estimated by fitting the predicted values to both the whole body data and the urine data simultaneously. This is identical to calculating the intake using the above equations.” ICRP has never recommended, suggested or mentioned any computer software and has never indicated any software as an “official” ICRP software. Quotes like those above must be deleted. - ANNEX G Structured Approach to Advanced Dose Assessment Other comments already sent to ICRP have sufficiently shown that any result can be obtained after approximately Step 5.16, from the simple fact that multiple statistically valid results can be obtained when fitting the data. In my opinion remarks like this must be clearly stated in the text if the IDEAS proposal is adopted in the guidance. - MINOR ISSUES: Page 70: 8.2.2 Stage 1. Level 0, and for higher exposures [Fig. 8.1] The text right after Step 1.1: Identify monitoring value (M) and duration of monitoring interval (T) says in the end “ …. which may add significantly to the dose, or even exceed that from the radionuclide measured (see Chapter 8). “ Since this is chapter 8 the parenthesized remark (see Chapter 8), should be deleted.