Historical Dose Reconstruction for U.S. Radiologic Technologists

During 2008, we completed a significant methodological enhancement to an historical dose reconstruction for a cohort of U.S. radiologic technologists (Simon et al, in preparation). The previous version of the dosimetry system, described in 2006 (Simon et al), used physical and statistical models, literature-reported exposure measurements for the years before 1960, and archival cohort badge dose readings for the period from 1960 to 1984. Annual estimates of badge and organ doses were generated for 90,000 cohort members -- based on their employment practices reported in a baseline survey conducted in the mid-1980s. The 2009 enhanced system uses newly obtained cohort badge dose and work history data, and improved shielding and organ dose factor models to provide annual estimates of the personal dose equivalent (badge dose) and radiation absorbed doses through 2006 to twelve organs and tissues for 110,000 cohort members. Key features of this 2009 version include:

• Dosimetry Models based on differential availability of dose data in two time periods:
  • Pre-1960 (divided further into pre-1940, 1940-1949, and 1950-1959)
  • 1960-2006
• Available Badge Readings and Survey Data:
  • 1,231,961 annual cohort badge doses from the 1977-2006 computerized records, and 2,100 from the 1959-1972 microfilm reels, of the nation's largest commercial dosimetry provider
  • 16,000 annual cohort badge doses for 1950-2004 from the Army, Navy and Air Force dose registries
  • 629 annual cohort badge doses for 1960-1976 from a major U.S. hospital
  • Literature readings and other reports on badge dose and other measurements before 1960, occupational radiation regulations, changes in protective measures, and technical procedures and specifications
  • Extensive individualized work history data from cohort members in three comprehensive surveys
• Badge Dose Estimation:
  • Estimated annual and cumulative doses are currently available for 110,000 cohort members who completed the first or second cohort surveys
  • All annual badge doses account for the fraction-of-year an individual worked during each year, 1901-2006
  • Pre-1960 doses were based on annual “standard” probability density functions (PDFs) derived from a synthesis of data from 11 publications (Simon et al 2006)
  • For 1960 onward, standard PDFs were developed for each year based on the 1.25 million cohort badge dose readings and a Tobit regression model.
  • Combination facility (hospital, physician office, combination, other) and workplace (civilian, military) effects were described by time-constant multiplicative factors estimated from the data
  • “Standard” densities (the “average” PDF for a given year and facility type) were partitioned into high, low, middle PDFs and assigned to cohort members based on work history (for third survey responders).
  • New third order autoregressive correlation structure was introduced in badge dose simulation sampling
  • 1940-1949 badge PDFs are now not constant, rather 100 simulations of 1940-1949 literature data were used
  • Dose realizations were based on conditional distribution when badge dose reading was below the minimal detectable limit (MDL). The MDL changes by time period.
  • Bias adjustment (shared error) was incorporated in geometric mean of badge dose for years 1916-1976
• Protection/Shielding:
  • Used apron categorical percentage of use -- for 4 radiological procedure types (from third survey), by decade
  • Total apron protection weighted by procedure-specific weights to obtain overall yearly protection percentage
  • Probabilistic apron thickness (0.25mm and 0.5 mm)
  • Probabilistic transmission factors
  • Used beta densities to calibrate first and third survey apron protection responses from categorical to numeric values
  • Badge readings were treated as “outside an apron”
• Organ Dose Factors (used to modify the badge dose to obtain specific organ doses):
  • Dose factors for each organ were not treated as constant, but as mean estimates with uncertainty
  • Each organ dose factor was specific for: time-period; probabilistic values of peak X-ray energy (kV ranges from 70-90); ratio of tissue or organ dose to air kerma free-in-air; and ratio of personal dose equivalent or badge dose to air kerma free-in-air
  • Each organ dose factor was further refined for badge location under versus outside an apron
• Organ Dose Estimation:
  • Badge doses (estimated or actual individual badge readings) were used together with protection/shielding data and organ dose factors to estimate organ doses.
  • Estimated annual and cumulative organ doses for 12 organs and tissues (breast, lung, thyroid, brain, colon, ovary, testis, bone marrow, skin of head and neck, skin of trunk and extremities, heart, lens of eye) are currently available for 110,000 cohort members who completed the first or second cohort surveys.

References:

Simon SL, Weinstock RM, Doody MM, Neton J, Wenzl T, Stewart PA, Mohan AK, Yoder RC, Hauptmann M, Freedman DM, Cardarelli J, Feng HA, Bouville A, Linet M. Estimating historical radiation doses to a cohort of U.S. radiologic technologists. Radiat Res 2006;166:174-92.

Simon SL. Organ-specific external dose coefficients and protective apron transmission factors for an historical dose reconstruction for medical personnel (in preparation).

Simon SL, Weinstock RM, Doody MM, Preston DL, Sigurdson AJ, Kwon D, Alexander BH, R Yoder RC, Bhatti P, Kim KP, Miller JS, Linet MS, et al. Radiation organ doses received by US radiologic technologists: Estimation methods and findings (in preparation).