Purpose: Routine real time in vivo dosimetry (IVD) is performed in HDR prostate brachytherapy to independently verify dose delivery. This study investigates impact of position uncertainty on error detection thresholds for IVD.Methods: IVD is implemented using a microMOSFET placed centrally in the prostate using an additional needle. 144 IVD measurements were made for 15 Gy or 19 Gy single fraction treatments. Needle insertion and treatment planning used real-time trans-rectal ultrasound. Source-MOSFET position thresholds of ±1, ±2 and ±3 mm were used to calculate per-needle and total plan error detection thresholds for the measured dose using an uncertainty analysis based on the treatment plan data.Results: The median dose difference from 144 total plan measurements was -5.2% (range +7.4% to -17.3%). 3 plans measured outside the total plan error detection threshold for position threshold ±1 mm, no plans measured outside the total plan error detection threshold for larger position thresholds. For 2233 individual needle measurements, for position thresholds of ±1 mm, ±2mm and ±3 mm the number of needles outside the per-needle error detection threshold was 103, 25 and 10 respectively and the number of treatments that would have required interruption based on these thresholds for real-time IVD was 66, 16 and 8 respectively.Conclusion: IVD in HDR prostate brachytherapy using a microMOSFET provides a high level of confidence that we are correctly delivering the planned dose to our patients. A ±2-3 mm position threshold gives an appropriate balance between error detection and avoiding unnecessary treatment interruptions.