|Inspection report details|
|Licence holder:||ANSTO Health Products|
|Location inspected:||Lucas Heights|
|Date of inspection:||2 March 2021|
An inspection was conducted as part of ARPANSA’s augmented inspection process to assess compliance with the Australian Radiation Protection and Nuclear Safety Act 1998 (the Act), the Australian Radiation Protection and Nuclear Safety Regulations 2018 (the Regulations), conditions of facility licence F0262 and Performance Objectives and Criteria (POCs).
The inspection consisted of a review of records, interviews, and physical inspection of the facility.
The inspection was conducted as a result of an event which saw workers exposed to contamination from Lutetium-177 (Lu-177) waste, contained within the bunding designed to hold shielded Ytterbium (Yb) waste, but were not contaminated themselves. Given multiple exposure events at ANSTO in recent years, ARPANSA sought to undertake an augmented inspection to better understand the event itself and the issues involved with consideration of human and organisational factors that ultimately led to an abnormal exposure of staff who work in the facility.
ANSTO’s Health Products facility produces multiple radiopharmaceutical products for the diagnosis and treatment of diseases such as cancer. Lu-177 is just one of these many products. An enriched Yb-176 (>99%) target undergoes irradiation in the OPAL reactor to produce Yb-177. Lu-177 is the decay product of Yb-177. The target is then moved across the site until it finally reaches the production suite where the lutetium is separated from the target material and eventually dispensed into its final product form.
Following an upgrade to the waste system in 2017, waste previously residing within a compartment below the production cell was transferred to the lower ground basement area that was purpose built to house and segregate the waste streams for the facility. Lu-177 waste has its own dedicated bunker area within the basement which is cordoned off from the main area. Generally, the basement is a low occupancy area with only staff having responsibilities for the waste streams required to be present.
Lu-177 waste is collected in 3 different bottles: Bottle 1, Bottle 2 and the Yb-bottle. Each bottle is changed over after a designated number of production runs with the Yb-waste collected and left to decay for its eventual recycling.
The exposure event which took place on 23 December 2020 occurred when a leak was discovered in the lutetium bunker area during maintenance activities (a crystalline substance was captured within the bunding that holds the Yb-bottle). The production operator involved, following radiation monitoring efforts and obtaining more appropriate PPE, then decided to remove the leaked waste material and contain it within a radioactive waste bin. Following this first attempt at removing the leaked material, a Health Physics Surveyor (HPS) attended the scene. A second removal effort was conducted with the waste being handled by both the operator and the HPS.
ANSTO’s investigation report identifies 29 recommended actions. Some of these are solely in relation to the event itself while others have a much greater reach and apply to organisational strategies and functions. However, the progress of the implementation of these recommendations will be monitored by ARPANSA and should be done in a timely manner.
A more general area for improvement (AFI) is described so that recurring themes, as identified through the body of this inspection report as previously identified AFIs, continue to be improved on. A second AFI is aimed at ensuring standard practices continue to be fully implemented.
A non-compliance has also been identified.
It is recognised that all ANSTO staff who participated in the inspection are committed to improving the facility to ensure that such an event is prevented in the future.
Performance reporting & verification
Within less than a day of the event, ANSTO had raised the event through its internal reporting system (GRC), performed initial conservative dose estimates and informed ARPANSA. This event, however, is not unique. Within the 12 months leading up to the event, four leaks (including this event) in the lutetium waste bunker have been identified and all seemingly with different causes. In each one of these events workers used the STAR (Stop, Think, Act, Review) principle as noted in the GRC reports. The employment of this principle is emphasised by ANSTO at an organisational level for all workers, partners and contractors during everyday workplace activities. In addition to this, workers are continuously urged to adopt a questioning attitude. These strategies are aimed at building a strong safety culture and can generally be found throughout the organisation’s literature (i.e. training material, guides, procedures, instructions, etc.).
In the previous 3 events, as indicated by individual reports in GRC, the STAR principle allowed for appropriate decision making to occur, given the nature of the hazard and permitted remediation activities to be undertaken without issue due to the additional controls put in place. In this latest event, the STAR principle was not implemented to its fullest and therefore contributed to workers receiving an abnormal exposure. While it is evident that some form of ‘dynamic risk assessment’ was made by the production operator involved in the activity, in that further monitoring equipment was gathered to help aid decision making and more specific PPE was acquired to further guard against the hazard, other standard ANSTO practices were missing which could have prevented the event.
ANSTO’s investigation report (the report) produced by independent parties discusses communication issues within the facility and describes the ineffectiveness of three-way communication. During the inspection it was identified that communication is not required to be so formal and that communication between the production supervisor, operator and HPS is identified as a contributing human factor to the incident. Rather than the repeat-back process, due diligence reporting is encouraged whereby operators inform staff involved/in the immediate vicinity and then relay the information to the appropriate supervisors. It will then be escalated upwards through management depending on the importance/severity of the event.
The report also considers lessons learnt from previous incidents. Toolbox talks are a recurring tool implemented at the facility to ensure communication of recent incidents, amongst other things, and lessons learnt in safety. The staff in this recent event have also been involved in previous lutetium events and, as the report suggests, were aware of their potential impacts demonstrating that information is disseminated and awareness created. A recommendation of the ANSTO Health Independent Report (following an accident in 2017) relating to human factors was centred on learning not just from ‘failures’ but also when things ‘go right’. It is unclear how this information is used by the operating staff, to demonstrate the lesson has been learnt, in order to ensure further incidents do not occur. However, it is understood that ANSTO is developing a system to account for this.
Inspection, testing and maintenance
A general summation of the process for creation of work orders normally has maintenance works set at a planned frequency. A work order is released to the owner of the work, and a period is set in which the work must be completed. If the activity could not be completed within this time, assessments are done to determine risks involved with having the work go uncompleted at which point maintenance planners escalate the issue to the responsible officer.
The leak at the centre of the most recent event was identified during a preventative maintenance (PM) inspection in the lutetium waste bunker. A PM-work order (PM-WO) was released through ANSTO’s computerised asset management and maintenance system (SAP) for an inspection of the waste system in the basement bunker to check valves, fittings, etc., and determine if there were any leaks, loose connections or other identifiable problems. Further notifications would then be raised with the maintenance team as required.
ARPANSA inspectors sought clarification of the PM-WO in regard to what the actual requirements were for the work to be performed as the document lacked specificity. This work is scheduled quarterly and was to be completed in December. Due to availability of both production operators and maintenance technicians, it was decided that the maintenance task would be performed slightly earlier than originally planned. Inspectors were informed that only a visual check was to be undertaken with any additional requirements necessitating notification to the maintenance team for the generation of further work orders as per the standard process.
The report delved into the lack of detail of the PM-WO and the lack of clarity around its ownership (operations or maintenance) which led to recommendations and actions to remedy such issues. An example of the same PM-WO, updated as a result of this event, was given to inspectors. It now separates the work performed by operations staff and maintenance (the example provided is solely for operations) and pre-requisite details are now listed (additional PPE, radiological monitoring, communication requirements, etc). While this is only a first iteration improvement, the facility intends for other work orders to receive similar updates where required as the actions stemming from the report’s recommendations are implemented and developed over time.
A common tool used where work activities involve a radiological component is the Safe Work Method and Environment Statement (SWMES). This is a job-specific document which details the risks associated with the work and controls put in place to mitigate them as decided by the appropriate stakeholders. Such an assessment was not performed as part of this PM-WO as workers were never intended to have interaction with the system i.e. visual inspection only. Had standard ANSTO practice been implemented, a SWMES or equivalent would have been conducted prior to hands-on activities. This contributes to an AFI.
The maintenance worker involved in the exposure event did not take part in the remediation activities and therefore did not receive a skin dose from the radioactive contamination.
Each leak that has occurred so far has been unique and as such there is no formalised process for verifying how the leak occurred. Currently, the facility is investigating multiple connection fittings that will be tested to potentially stop any future leaks.
ARPANSA inspectors reviewed multiple safety assessments that have been developed to demonstrate the risks involved with Lu-177 production activities as well as the waste transfer process. Inspectors were aware of the original assessment (ANSTO/T/TN/2013-09 rev 0, dated June 2013) used to approve a regulation 51 submission (as it was known at the time) for hot commissioning and operation of the Lu‑177 hot cells. Through a desktop review and subsequent requests for information a more recent revision was provided (ANSTO/T/TN/2013-09 rev 1, dated July 2014). Inspectors primarily focussed on waste and in doing so multiple discrepancies were found; however, they are not limited to these scenarios.
Both assessment reports consider production occurring up to once per week with each run processing a 1 g target. In 2013, typical product activity was 20 GBq with a concentration of 40 GBq/ml. The maximum activity was 150 GBq with a concentration of 75 GBq/ml. Yet, in 2014, activity could then range between 10 – 100 GBq with a maximum concentration of 200 GBq/ml. Each document states that if modifications were to occur that caused the limits to be breached, a revision would be performed. While it is evident that a revision did occur from the increase in the original limits, such a revision to the original assessment (and therefore the safety case) appears to have not been provided to ARPANSA. The process inventories within both documents from a single target are also quite different in terms of product activity and Yb-waste activity yet the initial and receipt target values remain unchanged.
Focussing on just the waste scenarios, no detail is provided as to how many runs are considered to be captured by the Yb-bottle. Judging by values in the process inventory for both assessment revisions, the Yb-waste stream is considered to have the cumulative waste (and therefore considers decay) from 4 production runs. However, what was once considered to only contain 33.8 GBq as per revision 0 now has a value of 167 GBq in revision 1. This is further compounded by the volume of waste in which the radioactivity is contained. The former value is held in 9.6 L of liquid in comparison to the latter which is 4 L. Speaking broadly, the 2014 revision has a greater activity in a smaller volume thus giving a greater concentration of radioactivity that is more than an order of magnitude greater than the initial assessment (41.75 MBq/ml vs 3.4 MBq/ml). To put this in perspective in terms of potential radiological dose, revision 1 states that the skin dose received by staff (if exposed to a droplet of the associated waste liquid) is a factor of ~7.5 times higher than the initial assessment (901 mSv/h vs 118 mSv/h).
Furthermore, an optimisation radiological risk assessment performed in December 2016 lists a maximum activity of 860 GBq in 8.4 L yielding a concentration of 102 MBq/ml. This amount is unclaimed in any risk assessment used as a basis for approval of continued operations.
A third risk assessment is available relating to the Lu-177 waste transfer project categorised by the licence holder under Regulation 52 (now section 64) meaning a change has occurred which is not significant to safety. This assessment was conducted as part of a change that saw the waste stream relocated in the basement from the waste compartment in the production suite upstairs. Considering the scenario for skin exposure due to the operator coming into contact with a droplet from the waste stream the doses received by staff (within five minutes) are the same as for the aforementioned revision 1 of the production assessment. However, the concentration within the waste volume is different and appears to be from one production run, 306 GBq and 84 GBq of Lu-177 and Yb-175 respectively in 840ml, equating to 306 MBq/ml (Lu-177) and 100 MBq/ml (Yb-175). The reason that the doses between the waste transfer assessment match the first revision of the production assessment is due to the assumption that the operator will only come into contact with a droplet (as per the other assessments) but this has been reduced by 90% due to flushing of the waste line. The issue is not whether the assumptions are valid but that another amount of waste has been claimed providing a different activity concentration.
Also, a recent submission was made for activities to be performed with decayed waste. Documentation used to support this submission provided data detailing how many cumulative production runs contributed to a singular waste volume before being replaced and the process restarted. This data shows that there have been multiple occasions where the assumed waste contribution (4 production runs) has been within a lesser volume (by almost 50%) as well as an increase in the number of runs per waste volume (10 production runs contributing to an 8.4 L volume). Unfortunately, neither of these situations has been described within an approved assessment demonstrating operational safety.
Staggered reviews of these assessments have also taken place yet these do not provide assurance as to which document was the correct source at that point in time. A “risk assessment – periodic review report” from October 2017 only considers the revision 0 production assessment even though revision 1 had been issued prior to that date. Further, a review of high risk/hazard tasks was performed in June 2018. This assessment did review revision 1; however, at the time of the review, the number of production runs contributing to a waste volume had already been increased (10 runs/8.4 L) which is not reflected as part of revision 1 and hence was not considered in the review.
The inspectors find the licence holder has failed to comply with subsection 60(1) of the Regulations in that it has not effectively managed nor stayed within the safety case as it relates to Lu-177. Subsection 60(1) requires the licence holder to take all reasonably practicable steps to manage the safety of the facility described in the licence.
All staff involved in the event had completed the requisite training and were deemed competent. Each has completed facility-specific and organisational inductions (inclusive of radiation safety) as well as their own job-specific training and are recognised as suitably qualified and experienced persons (SQEP).
The report identifies that reliance on SQEP meant documentation lacked suitable hazard/safety information. Following the ANSTO Health event in August 2017 and the subsequent recommendations, the inclusion of hazard warnings in procedures has been a recurring issue during ARPANSA inspections of facilities where, due to the requirements of the work being performed, there are manual handling tasks involving unsealed radioactive sources i.e. radiopharmaceutical production facilities. Further, multiple areas for improvement have been identified (e.g. inspection reports R19/09834, R18/07456, and R20/04132) to the licence holder with the aim of addressing the issue. Such warnings in procedures should be implemented across the organisation in a structured and consist manner in order to effectively communicate potential consequences associated with those sources. Examples of such efforts were provided to inspectors for different products (Gentech Mo-99/Tc-99m generators, sodium iodide (I-131) capsules and Lu-177 hot cell and waste compartment entry, process solution (waste) handling procedure and checklist). While the format between the Mo-99 and I-131 differed, the information appears consistent. This could not be said for the Lu-177 documents that either did not include warnings or were generic. These documents also had different PPE requirements with entry procedures indicating that radiation-attenuating gloves are mandatory in comparison to the instructions for the waste system that only required operators to ensure the gloves were available. While inspectors were informed that such PPE was available in the area, it was not a formalised requirement for the work being conducted. The lacking/ambiguous information in Lu-177 procedures is considered to be a contributing human factor as staff rely on these procedures and without the correct information may not appreciate the inherent risk of tasks being performed.
Standard practice throughout the organisation found in documentation at both an organisational and facility specific level (along with the implementation of the STAR principle) is to stop work, alert people in the area (and those responsible), make the area safe, and to not perform recovery activities. Staff are to contact Health Physics for radiological advice if they find themselves in an abnormal situation or are unsure of radiological conditions in the area. In all previous events, it is clear that these practices were applied. In this case, standard practices were not fully implemented which brings into question how effectiveness of training is determined (see R18/07456). This contributes to an AFI.
It is recognised that training and operational experience did have a positive influence in this event. Extra protection measures not listed as mandatory in procedures specific to this activity but found elsewhere were instinctively put in place, which reduced the dose to the operator involved.
As a result of this event, training improvements are currently in development stages for scenario-based response action training which is expected to be rolled out through the entirety of ANSTO Nuclear Medicine (ANM and Health Products).
Dose estimates determined on the day of the incident calculated 3 different dose values (under estimate, best estimate and over estimate) for both of the exposed workers based on a range of assumptions. Best estimates, with conservative assumptions applied, calculated that the production operator and HPS received an extremity dose of 52 mSv (or mGy) and 160 mSv (or mGy) respectively (the annual limit is 500 mSv). Exposure to such doses is not considered to be an accident under ARPANSA’s regulatory framework. However, it appears conservative decision-making was applied in informing the regulator.
The PPE which allowed for dose reduction of the operator was a pair of radiation-attenuation gloves. Experimental studies and computational modelling performed by ANSTO’s Radiation Protection Services (RPS) demonstrate that, in the case of Lu-177, doses are reduced by approximately 90%. At the time of the event, the mandatory requirement for the use of those gloves was not apparent in the procedure nor were they used by HPS. Given the nature of the radioisotope being able to produce a high skin dose compounded by HPS only wearing thin nitrile gloves (considered negligible in terms of their shielding effect), the HPS staff member received a higher dose. Had the operator not been wearing the radiation-attenuating gloves or HPS handled both contaminated samples, it is possible that dose thresholds may have been exceeded.
RPS is currently developing a Health Physics ‘kit’ of sorts for each of its staff to address recommendations made in the report. Aside from other standard PPE and monitoring equipment, this is to include radiation-attenuating gloves.
Current working arrangements, shift patterns and on-site requirements due to work demand were reported to cause fatigue felt by HPS and affected the decision-making process. That, and proceeding while lacking information and not seeking further clarification are considered to be contributing human factors culminating in the handling of radioactive contamination.
During the inspection, it was noted there could be a potential lack of awareness and understanding of the contribution that the beta-emitting isotope has in terms of skin dose versus external exposure as presented through radiation monitoring. While specific details may be unknown, a more general approach to contamination response is communicated to staff through radiation safety training by way of general terminology i.e. ‘it’s dirty, do not touch it, contact Health Physics’. While some scenario training has been conducted in previous years in relation to response to spills, the effectiveness of such training in relation to human factors to prevent potential exposure is an ongoing issue. This is further highlighted by the method of remediation whereby the operator removed the contamination directly rather than creating distance through the use of tongs, tweezers or CV reachers.
During the physical inspection, inspectors were shown that the PPE and radiation monitoring equipment are currently kept in the immediate vicinity. New and improved signage has also been placed in direct view of the area which immediately identifies what to do/who to contact before entering the waste bunker area alongside area safety hazard noticeboards and a flowchart describing the process to be followed in case of a spill/contamination event.
It should also be noted that if conservative assumptions were removed and the parameters surrounding the dose estimation were refined, RPS has indicated that a more realistic dose calculation would be closer to the under estimate value.
The inspection revealed the following non-compliance:
- The licence holder has failed to comply with subsection 60(1) of the Regulations in that it has not effectively managed nor stayed within the safety case as it relates to Lu-177. Compliance with section 60 of the Regulations is a licence condition of F0262. Subsection 30(2) of the Act requires the licence holder to comply with the licence conditions.
The inspection revealed the following areas for improvement:
- Continued implementation of actions to address previously identified AFIs.
- Adherence to standard ANSTO practices.
It is expected that improvement actions will be taken in a timely manner.