Development and Experimental Evaluation of a Modified Computational Model for Estimating Occupational Whole-Body Radiation Dose in the Conventional Diagnostic Radiography Unit of Federal Medical Center, Jalingo, Taraba State, Nigeria

Abstract

Personal dosimetry using thermoluminescent dosimeters (TLDs) can be constrained in resource-limited settings by equipment, processing, and logistical requirements, motivating the development of complementary computational approaches for occupational radiation-dose estimation. This study developed and experimentally evaluated a calibrated computational model for estimating personnel whole-body radiation dose in a conventional diagnostic radiography unit. The primary model incorporated tube potential (kVp), tube current–time product (mAs), source-to-personnel distance, occupancy factor, and a facility-specific calibration coefficient. We obtained six-month personnel dose measurements using chest-worn TLD badges from 15 radiography personnel at the Federal Medical Centre, Jalingo, Nigeria, and compared them with Monte Carlo estimates based on 10,000 simulations per participant. The TLD-measured doses ranged from 0.96 to 1.78 mSv, with a mean of 1.479 ± 0.216 mSv, whereas the corresponding Monte Carlo estimates ranged from 2.204 to 2.265 mSv, with a mean of 2.234 ± 0.017 mSv. The mean model–measurement difference was 0.755 mSv, indicating systematic overestimation by the computational model. Bland–Altman analysis gave 95% limits of agreement of 0.311 to 1.199 mSv, with 14 of 15 observations (93.3%) falling within these limits. One-at-a-time sensitivity analysis demonstrated that source-to-personnel distance and tube potential had the strongest analytical influence on model output because of their squared dependence in the dose equation, while mAs and occupancy factor exhibited first-order dependence. The findings indicate that the calibrated model reproduced the general magnitude of personnel whole-body dose but systematically overestimated the measured TLD values. The model may therefore be useful as a complementary screening and dose-estimation approach, but further calibration and independent validation using larger datasets and multiple radiography facilities are required before application to individual dose assessment