Unintended overexposure of a patient during radiotherapy treatment at the Edinburgh Cancer Centre, in September 2015

9. Recommendations arising

9.1 Recommendations concerning manual calculations

Regarding the ECC procedures for manual calculations, 'Towards Safer Radiotherapy' [2] recommends that :

'Calculations should be checked by a different entitled operator, preferably using a different method and a separate data set. Reverse checking is an example of the use of a different method.'

Prior to the autumn of 2013, in-house computer programs called 'T' and 'Check' were used as part of the treatment planning calculation and checking process. However, due to some lack of certainty about the applicability of the 'T' program for treatments using the new 'true beam' linear accelerators, a memo was circulated to the radiographers informing them that both the first and second calculations were to be done using the data tables.

Clearly, had a 'different method and a separate data set' been used, then it is highly unlikely that both manual calculations would have given the same wrong answer.

Recommendation 1: Relevant Employer's Written Procedures for manual calculations should be changed to ensure that the first and second calculations are carried out independently using a different method, and entitled operators should be appropriately retrained.

Further to this, the practice at the ECC was to carry out the two manual calculations using plain paper, check that the answers matched, and then transcribe the results of these manual calculations to the Radiotherapy Prescription Sheet ( Figure 3). Since there is no formal procedure for subsequently checking the accuracy of this transcription, this allows that transcription errors might arise.

Recommendation 2: The practice of carrying out manual calculations using plain paper, and then transcribing the results to the Radiotherapy Prescription Sheet should be reviewed. In particular, the results of the second manual calculation should be checked or rechecked against the data that had already been transcribed to the Radiotherapy Prescription Sheet.

As discussed in Sub-section 4.5 of this report, the method of manual calculation for 'parallel opposed pair with 100cm FSD' as currently documented in ECC Employer's Written Procedure EP2\ ECC\3402 'Calculating and Checking Monitor Units of Photon Beam Treatments-Manual Calculations' lacks clarity.

Recommendation 3: Review the description of the current method of manual calculation for parallel opposed pair treatments in ECC Employer's Written Procedure EP2\ ECC\3402 and the way in which this calculation is laid out on the Radiotherapy Prescription Sheet. In particular, consider defining a method of calculation that divides the prescribed dose between the two beams, and modifying the layout of the manual calculation tables in the Radiotherapy Prescription Sheet to offer greater clarity about the entries required. An example of this is shown here in Figure 12. Any resulting changes should be followed by appropriate re-training of operators.

9.2 Recommendations concerning RadCalc calculations

The producers of the RadCalc software, Lifeline Software Inc., have advised that the method currently described for 'lateral parallel opposed pair at 100cm FSD', treatments in ECC Employer's Written Procedure EP\ ECC\3422 'RadCalc Instructions', is not optimal.

Recommendation 4: A review of all Employer's Written Procedures and Protocols relevant to the use of RadCalc should be undertaken, with appropriate input from Lifeline Software Inc., to ensure that the methods used are in keeping with those recommended. This should include reconsideration of the terminology used (such as 'Isocentre_3') and steps to ensure that consistent terminology is used in RadCalc and all related Employer's Written Procedures and Protocols.

From the interviews referred to in Sub-section 5.1 of this report, it is clear that whereas the Radiographers concerned considered themselves capable of operating RadCalc in accordance with the relevant ECC procedures and protocols, their understanding of the electronic calculation process was limited. This, coupled with their lack of confidence in the use of RadCalc for this particular form of treatment, led to failure to notice and take proper account of what should have been clear on-screen indications that, not only was their calculation out of tolerance, but since it was approximately 100% out, they had in all likelihood failed in their manual calculation to divide the prescribed dose of radiation between the two opposing fields.

Recommendation 5: Following a review of all relevant Employer's Written Procedures and Protocols, appropriate retraining of everyone involved in the use of RadCalc should be undertaken, and this should seek to ensure that all users have a clear understanding of the calculation process, and a high level of confidence in its results.

As discussed in Section 6, documented provisions for training in the use of RadCalc differ between the two Divisions involved, and there is no separate competence within the scope of entitlement of either the Radiographers or the Physicists for the use of RadCalc.

Recommendation 6: Training requirements for use of RadCalc should be formally documented within the ECC quality system and should be consistent for all Divisions involved. This should include pro-forma training records with provision for 'sign-off' by appropriately qualified and entitled trainers. In addition to an appropriate number and range of 'practice calculations', consideration should be given to the need for 'real patient' calculations carried out under supervision.

Recommendation 7: Appropriate competences that provide a clear separation between any other aspects of the calculation and checking procedures should be added to the list of authorized competences for operators in the Divisions involved. For example 'Competent to undertake RadCalc calculations for non-isocentric treatments to unequal FSD'.

Recommendation 8: The relevant Employer's Written Procedure should include an instruction that where the difference between the monitor units calculated manually and those calculated by RadCalc is out of tolerance for reasons that are not immediately evident, then the source data, its input to Aria, and the manual calculations must be re-visited.

Towards Safer Radiotherapy [2} recommends when new or changed treatment techniques or processes are to be introduced, a risk assessment should be undertaken and consideration given to additional verification procedures for the initial cohort of patients.

Recommendation 9: Following the introduction of any changes arising from these recommendations or otherwise, a risk assessment should be undertaken and appropriate verification procedures implemented for, at least, the initial cohort of patients.

Recommendation 10: Regarding the RadCalc programme itself, consideration should be given by Lifeline Software Inc. to the mistakes made by these operators and to whether any additional safeguards could be added to prevent or highlight inappropriate data entry or manipulation. For example, whereas at present the alert referred to in Sub-section 4.7.of this report [' The cumulative dose per fraction for the beams associated with this prescription exceeds that prescribed dose per fraction by 100%'] disappears following any change between screens, it might be more appropriate that this should persist.

9.3 Recommendations concerning the recording of training and entitlement of duty holders

This investigation has identified deficiencies in the current provisions for holding of operator training records and inconsistencies between the two Divisions involved. Concerns have also arisen about the linkage between training of operators and definition of their scope of entitlement.

Recommendation 11: Provisions for operator training for Radiographers and Physicists should be reviewed to ensure consistency in the training provided and in the means of recording satisfactory completion of training. Pro-forma training plans that include a number of elements that must be completed should include provision for identification of the person confirming satisfactory completion of that element, and the date.

Recommendation 12: The current instruction for staff in Treatment Planning Section that they need not retain evidence of training should be rescinded, and, in accordance with Regulation 11(4) and related guidance, training records that show clear evidence of 'the nature of the training' should be retained. Where possible, pro-forma documents intended to record evidence of training should be held as quality controlled documents.

Recommendation 13: Each member of staff should have a personal record of their scope of entitlement comprising a list of competences for their Division that has been authorised by the Division Head. The competences among those listed for which the duty holder is entitled should be clearly indicated by means which includes identification of the person conferring the entitlement and the date. This 'sign-off' should also clarify whether the duty holder concerned is also authorized to supervise trainees in this competence, and to provide and 'sign-off' the relevant training.

Recommendation 14: Where possible, there should be a clear linkage between each of these authorized competences and the training required prior to entitlement. For example, with regard to Sub-section 5.1 of this report, the prerequisite for sign-off of an operator competence 'Competent to carry out data entry to Aria for all treatment plans for which the operator is entitled' would be completion of the training defined in ECC controlled document number EP2\ ECC\2037 'Training Plan: Data entry'.

Recommendation 15: Where a particular competence within the scope of entitlement of a duty holder is undertaken infrequently, consideration should be given to defining a separate requirement for maintenance of that competence. For example, where an operator has been deemed 'Competent to carry out manual calculations for all parallel opposed pair treatments to unequal FSD', it might be appropriate to require that evidence should be recorded of at least three such actual or practice calculations having been undertaken in the previous year.

9.4 Recommendations concerning Employer's Written Procedures and Protocols

This investigation has identified inconsistencies between the content of the Employer's Written Procedures and Protocols for the two Divisions involved, and with the content of the ECC's Level 1 and Level 2 Employer's Written Procedures.

Recommendation 16: A joint review of those Employer's Written Procedures and Protocols for treatment planning held by the Oncology Physics and by the Treatment Radiotherapy Divisions should be carried out to ensure that these are consistent in terms of the instructions given and the terminology used, and that, as appropriate, they are fully in keeping with the ECC's Level 1 and Level 2 Employer's Written Procedures. This should include due consideration of the need to avoid duplication of instructions, and to ensure that all such documents are freely available to staff in both Divisions and are compatible with their training and entitlement. For example Employer's Written Procedures EP2\ ECC\0050, EP2\ ECC\2200, EP2\ ECC\2201 and EP2\ ECC\3402 each describe the manual calculation process.

9.5 Further measures to reduce the risk of overexposures

Towards Safer Radiotherapy [2] recommends that all centres should have protocols for in-vivo dosimetry monitoring for most patients at the beginning of treatment. In vivo and transit dosimetry, which can be described as being the use of detectors to measure the amount of radiation delivered, can detect some significant errors, and, if carried out at an early stage in the course of treatment, might allow corrective action to be taken.

Recommendation 17: The use of an in-vivo dosimetry tool should be considered for palliative patients.

While there were clear differences between the circumstances of the incident considered here and those of the Glasgow incident [3], a fundamental similarity is that in both cases the MU used were much higher than the figure normally expected for the treatment in question. This would suggest that if the prescribed MU could have been compared with an appropriate stored list of expected MU values for particular treatments, then the Radiographers would have been alerted to the fact that the MU figure was unusually high at some point prior to delivery of the first fraction of the treatment.

In this regard, Towards Safer Radiotherapy [2] notes that 'For some commonly delivered treatments, ... the range of monitor units per fraction falls within a predictable range for the majority of patients' and recommends that ' Centres should consider which kinds of treatment fit into such categories and draw up lists of ranges of expected monitor units for certain beam configurations to assist staff in establishing familiarity with standard protocols'.

The simplest form of such a check might be comparison of manually calculated MU with locally listed expectation values. However, an additional possibility is that this comparison could be carried out electronically. The optimal point for such an electronic comparison would be immediately prior to delivery of the first fraction of the treatment so that any foregoing errors in the MU calculation whether in the prescription, manual planning, transcription of data, or in electronic planning could be identified.

Recommendation 18: For those treatments for which 'the range of monitor units per fraction falls within a predictable range for the majority of patients' consideration should be given to how the prescribed MU can be compared to appropriate stored data, both manually and electronically, so that significant departures can be flagged ahead of treatment delivery.