Quality Assurance - std-1121-98-reaffirmation0125

DOE-STD-1121-98

11.0 QUALITY ASSURANCE

This section addresses quality assurance in general and independent review of dose assessments and

computer software.

11.1 GENERAL NEEDS

Quality assurance needs for various aspects of internal dosimetry programs are described by the

American National Standards Institute in published and soon-to-be published standards (HPS 1996a &

2000 and ANSI 1996). Berger has given an excellent general overview (Berger 1994). Accreditation

through the U.S. Department of Energy Laboratory Accreditation Program (DOELAP) will include all of

the quality assurance features needed for radiobioassay laboratories (DOE 1998b). The DOELAP

program for radiobioassay laboratories follows many of the precedents set in the field of external

dosimetry (DOE 1998b; McDonald et al. 1992).

11.2 INDEPENDENT REVIEW

When doses are large with respect to the IL and there is controversy over a dose assessment,

independent review is indicated. The experience of one such review is provided by La Bone et al. (La

Bone et al. 1992). Agreement within a factor of two among experienced dose assessors is probably the

best that can be hoped for in difficult cases such as transuranic intakes with subsequent chelation. Easier,

more straightforward cases result in better agreement during intercomparisons (Hui et al. 1994).

11.3 COMPUTER SOFTWARE QUALITY ASSURANCE

Computer software is an important tool in internal dosimetry. The software may include commercial

dosimetry codes, site- or contractor-developed dosimetry codes, calculational algorithms incorporated

into commercial application codes (e.g., spreadsheets), and database application software for

management, manipulation, and reporting of data. Quality assurance activities involve configuration

management, code testing, error correction, and security.

11.3.1 Configuration Management

Dosimetry codes should be subject to configuration management, including records of the version of

the code, the user's manual, instructions for running the code, limitations of the code, hardware

requirements, acceptance testing records, and a copy of the code itself.

11.3.2 Verification and Validation (Acceptance) Testing of Codes

Computer codes should undergo a two-step verification and validation (V&V) process as acceptance

testing before their routine use for dosimetry (ANSI 1987). This process shows that the code produces

valid responses when used to analyze problems within a specific set of parameters and parameter values.

Verification involves determining program requirements, range of program results that may be considered

valid, or criteria to be used in evaluating the validity of results. Validation is the process of testing a

computer program under a specific computing system and evaluating the results to ensure the compliance

with specified requirements. Part of the testing should include running selected "benchmark" cases for

comparison against an independent solution process (e.g., hand calculations, published tabulations of

111