« Previous | Contents | Next »
Listen
5. Investigation of the circumstances of the incident
5.1 Summary of the initial investigation
5.1 Following receipt by the Inspector of an initial incident report from the BOC, a meeting to discuss the circumstances of this incident was held at the BOC on 10 th February 2006. Participants included senior staff from the BOC, staff from the SEHD (including the Inspector) and an adviser from the Radiation Protection Division of the Health Protection Agency.
5.2 A draft report on the incident by Dr Martin, under the title " Preliminary Report of Investigations on Incident Involving Delivery of a Higher Dose than Intended to the Brain During Treatment of the CNS at the Beatson Oncology Centre, Western Infirmary During January 2006" was made available at this meeting.
5.3 A copy of the note of this meeting, which was made by Marianne Cook of the Scottish Executive Health Department, is appended here as Annex 4.
5.4 The main findings of this initial meeting can be summarized as follows:
(i) The principal cause of this incident was identified as a single erroneous entry for the 'Output' on Medulla Planning Form FM.14.014 for Miss Norris, dated 16 th December 2005.
(ii) The entry was made by Planner B who omitted to carry out the normalization procedure required when entering the 'Output' in monitor units ( MU) per hundred centiGrays. A figure of 91 was therefore written on the planning form instead of the correct figure of 54.
(iii) Checking by more senior colleagues failed to identify the error.
(iv) The error was repeated by Planner B subsequently in planning a similar treatment for a new patient but in this case the error was identified by another senior planner (Senior Planner D). It was this discovery that led to investigation of previous treatment plans and to identification of the error in the plan for Miss Norris. No other patients at the BOC were affected.
(v) The training records for Planner B indicate no record of competence as either "training", "competent", or "authorised to train" for the 'Spine/Medulla/ CNS' (whole CNS) procedure in question.
(vi) Planner B had prepared a similar plan, under supervision, in November of 2005, wherein the prescribed radiation dose was input to the Eclipse module and normalization of the output was therefore applicable. Planner B was unaware of the need for normalization and did not apply it. In this case, however, the daily radiation dose was 1 Gray (100 centiGrays) so the normalization procedure, had it been applied, would not have changed the values of the MU in its transference from the Eclipse Treatment Plan Report to the Medulla Planning Form. The correct value for the 'Output' was therefore entered on the Medulla Planning Form.
(vii) There is no evidence of the employer's written procedures for 'Medulla Planning' (WI.14.01.01) having been updated annually as required by their own quality assurance procedures. The available procedure thus did not reflect fully the current practice at BOC and contained no specific instruction regarding the normalization procedure referred to above.
(viii) The course of radiotherapy for Miss Norris was terminated immediately on discovery of the error and the patient and her family were informed. At this point she had received 19 treatments each of 2.92 Grays, giving a total of 55.5 Grays. The prescribed dose was 35 Grays in 20 equal fractions of 1.75 Grays.
(ix) The incident was reported promptly to the Scottish Executive Health Department in accordance with the requirements of Regulation 4(5) of the IR(ME) Regulations 2000.
(x) A number of immediate changes were implemented at the BOC to minimize the possibility of any recurrence of an incident of this type.
(xi) The Inspector was satisfied that the BOC staff co-operated fully with the inspection team and that all documents requested were made available.
5.5 This initial investigation, together with the BOC incident reports established clearly that the cause of the overexposure was the erroneous 'Output' figure that was entered on the Medulla Planning Form. Further investigation therefore sought to establish (a) the circumstances that caused this error to be made, (b) why the error was not detected earlier and (c) what should be done at the BOC and at other radiotherapy centres to minimize the possibility of recurrence of a similar incident.
5.6 These further investigations included a review of the BOC's written procedures and individual interviews with relevant staff from the BOC and from GGHB. Sections 5.2 and 5.3 of this report, which consider issues (a) and (b) in Paragraph 5.5, include information obtained from individual staff interviews. Annex 5 summarizes the information obtained in these interviews.
5.2 Why was the wrong 'Output' figure entered on the Medulla Planning Form?
5.2.1 The effect of changes to treatment planning procedures in May 2005
5.7 Prior to the introduction of Varis 7 in May 2005, the actual prescribed treatment dose was not entered into Eclipse. Therefore, all BOC treatment plans computed by Eclipse were for a 'standardized' dose of 100 centiGrays per treatment fraction and the MU figure in the Treatment Plan Report that was printed by Eclipse was always in units of MU per 100 centiGrays.
5.8 For most treatment procedures (in general the simpler procedures) the treatment plan provided to the radiographer was simply the hardcopy 'Treatment Plan Report' from Eclipse. For some procedures, however, including the whole CNS procedure, the data from this printout was transcribed to a separate planning form (such as FM.14.014 shown in Annex 2). In either case, since the daily dose fraction that was presented to the radiographer was in units of MU per 100 centiGrays, the actual number of MU for each treatment fraction was always calculated by the radiographer in accordance with the procedure described in Section 4.3 of this report.
5.9 With the introduction of Varis 7 came the opportunity for electronic transfer of the treatment delivery parameters computed by the Eclipse software module to the RTChart module.
5.10 To utilise this capability it was necessary for the data input to Eclipse to include the prescribed total radiation dose and the number of fractions. Hence, the treatment plans computed by Eclipse using this input data were no longer for a 'standardized' dose of 100 centiGrays.
5.11 As indicated in Paragraph 4.10, the change to electronic transfer of the calculated data for treatment delivery from Eclipse to RTChart was not made for all planning procedures. However, the capability of transferring data on the prescribed dose from RTChart into the Eclipse module was also adopted at the BOC for some (but not all) of those plans where the data continued to be transferred manually from the Eclipse Treatment Plan Report back to RTChart.
5.12 In planning the head fields for Miss Norris, the data that was input to Eclipse from RTChart did include the prescribed total dose and the number of fractions. For the spine fields it did not. Hence, for the head field treatment plan, the MU figure that was printed in the Eclipse Treatment Plan Report was no longer in units of MU per 100 centiGrays but was now in units of MU per treatment fraction (in this case MU per 167 centiGrays). Therefore, this figure could no longer be transcribed directly to the 'Output' box on the Medulla Planning Form but now had to be normalized back to MU per 100 centiGray, as described in Paragraph 4.20 of this report.
5.13 Prior to the change in May 2005 therefore, it had never been necessary for treatment planners to normalize the MU figure that emerged from Eclipse. BOC work instruction number WI.14.01.01 for 'Medulla Planning' reflected this in that there was no reference to this requirement. (The version of WI.14.01.01 shown in Annex 1 which is dated 11 th August 1998 was the latest version available at December 2005.)
5.14 More specifically, in planning the head fields for Miss Norris, the input to the Eclipse planning module from RTChart was chosen to include the consultant oncologist's prescribed dose of 35 Grays but in 21 fractions rather than the 20 fractions prescribed (see Paragraph 4.16). Eclipse then calculated that her daily treatment required a Linac setting of 91 MU for a dose of 167 centiGrays and printed this number on the Treatment Plan Report. The treatment planner should then have divided the '91' by 1.67 to give a figure of 54 MU per 100 centiGrays for the 'Output'. Instead, the treatment planner omitted this normalization procedure and transcribed the figure of 91 directly from the Eclipse Treatment Plan Report to the Medulla Planning Form.
5.15 Had the potential for error that resulted from the change to inclusion of the prescribed dose in the Eclipse input data been identified, then a number of different alternatives could have been implemented:
(i) The former practice of computing all treatment plans for a dose of 100 centiGrays per treatment fraction could have been retained for all of those plans where the data in the Eclipse Treatment Plan Report was to be transferred manually back to RTChart.
(ii) Medulla Planning Form FM.14.014 could have been changed to require an entry in MU per treatment fraction instead of in MU per 100 centiGrays and BOC Work Instruction WI13.26.06 for radiography staff ( Annex 3 to this report) changed accordingly.
(iii) BOC work instruction number WI 14.01.01 for 'Medulla Planning' could have been amended to include instruction on the need for normalization of the MU output figure from Eclipse and appropriate training given.
(iv) Data for the head fields (planning for which is less complex than for the spine fields) could have been transferred electronically from Eclipse to RTChart
5.16 Had alternative (ii) in Paragraph 5.15 been adopted, then the new practice of entering the prescribed dose into Eclipse might have been advantageous in removing the need for radiographers to 'scale up' the 'output' figure. However, if there remained some planning procedures for which the prescribed dose was not fed into Eclipse then the potential for confusion and for critical error would be considerable. Additional potential for confusion would arise from the inclusion of two different conventions (for the head fields and the spine fields) for reporting the 'output' within the same CNS plan.
5.17 In summary, the introduction of the new procedure involving electronic transfer of data on the prescribed radiation dose from RTChart to Eclipse was wholly appropriate for and compatible with those plans that involved electronic transfer of the treatment delivery data calculated by the Eclipse module back to RTChart. However, for plans that continued to employ manual transfer of the treatment delivery data from Eclipse to RTChart, this change was not essential but, if adopted (for example, for consistency with plans involving electronic transfer of treatment delivery data) should have been accompanied by a comprehensive evaluation of consequences and by changes to the procedures and documents affected, as well as by any necessary re-training.
5.2.2 The role of Planner B in planning this treatment
5.18 Paragraph 11(1) of the IR(ME) Regulations requires that: 'no practitioner or operator shall carry out a medical exposure or any practical aspect without having been adequately trained', with the proviso that (Section 11(3)) 'Nothing in paragraph (1) above shall prevent a person from participating in practical aspects of the procedure as part of practical training if this is done under the supervision of a person who himself is adequately trained'.
5.19 Paragraph 11(4) of the IR(ME) Regulations requires that: 'The employer shall keep and have available for inspection by the appropriate authority an up-to-date record of all practitioners and operators engaged by him to carry out medical exposures or any practical aspect of such exposure or, where the employer is concurrently practitioner or operator, of his own training, showing the date or dates on which training qualifying as adequate training was completed and the nature of the training'.
5.20 In summary, therefore, the IR(ME) Regulations require that any person who participates in practical aspects of the exposure procedure, such as treatment planning, must either be appropriately trained or must do so under the supervision of another person (an 'operator') who is properly trained. The policy at the BOC was that employer's training records should define the training status of the employees involved.
5.21 The Eclipse Treatment Plan Report for the three planned exposure fields for Miss Norris has the following information:
Head.
Plan created: Thursday December 15, 2005, 3:25:38 PM by [Principal Planner A]
Printed 19/Dec/2005 10:41 AM by [Planner B]
Plan Last Modified Monday December 19, 2005 10:40:32 AM by [Planner B]
Upper spine.
Plan created: Friday December 16, 2005, 10:55:51 by [Planner B]
Printed 21/Dec/2005 3:15 PM by [Principal Planner A] .
Plan Last Modified Friday December 16, 2005 12:53:25 PM by [Planner B]
Lower spine
Plan created: Friday December 16, 2005, 11:32:09 AM by [Planner B] .
Printed 19/Dec/2005 3:16 PM by [ by [Principal Planner A] .
Plan Last Modified Monday December 19, 2005 10:40:32 AM by [Planner B]
The data from each of these three forms was transcribed to the Medulla Planning Form by Planner B. Further investigation of relative roles indicates that the bulk of treatment planning for Miss Norris was carried out by Planner B.
5.22 In the period between May and December 2005, two whole CNS procedures were planned at the BOC, one in August and one in November. The treatment plan for Miss Norris was therefore the third plan of this type that had been completed following the upgrade to the Varis 7 computer system. However, for the first of these three plans, the prescribed radiation dose was not included in the data input to the Eclipse planning computer. For the second of these plans, in November 2005, the prescribed radiation dose of 3600 centiGrays in 36 fractions was input to Eclipse but because the prescribed radiation dose per fraction was 100 centiGrays, the calculated number of monitor units per fraction was precisely the number per 100 centiGrays. Hence the normalization process, whilst applicable, would not have produced any change in the values transcribed to the Medulla Planning Form.
5.23 The plan for Miss Norris was therefore the first of its type at the BOC where the calculated number of MU output from the planning computer was anything other than the number per 100 centiGrays. It was, in fact, the first of any type at the BOC, medulla or otherwise, for which there was a need for manual normalization of the number of MU generated in the Eclipse Treatment Plan Reports before transcribing to a paper planning form.
5.24 Planner B did not participate in the planning of the Spine/Medulla/whole CNS procedure in August. For the November procedure he did, under the supervision of Principal Planner A and Senior Planner C.
5.25 Inspection of Planner B's training records on 10 th February 2006 indicated that these were last signed-off on 28 th June 2005 by Principal Planner A. These training records comprise a tabulated list of 'Competences', against each of which are three tick-boxes respectively headed 'L', 'C' and 'T' to indicate that, for each particular competence, the person is deemed to be either Learning, Competent or competent to Train. The competence relevant to the procedure in question is listed as 'Spine/Medulla/whole CNS'. No entries appear in any of the three related boxes, which suggests that, at the time of completion of this record, Planner B had no formal competence in this planning procedure.
5.26 However, at interview ( see Annex 5) Principal Planner A and Planner B both expressed a view that the training and practical experience received by Planner B in November 2005 was such as to place him in the 'Learning' category for planning of the Spine/Medulla/whole CNS procedure. In this regard, the requirements of Paragraph 11(4) of the IR(ME) Regulations for the employer to keep up-to-date records were not being met. (At Interview ( Annex 5) Principal Planner A stated that he tried to review training records for treatment planning at six-monthly intervals.)
5.27 Therefore, it is the view of Principal Planner A and Planner B that when Planner B undertook planning duties for the Spine/Medulla/whole CNS procedure for Miss Norris, he did so as a 'learner' (though not yet recoded as such). However, it is clear that the training that he had received was confined to his supervised participation in the whole CNS procedure that was planned in November for which (see Paragraph 5.22) failure to apply the normalization procedure prior to transcribing the MU data from the Treatment Plan Report to the Medulla Planning Form had no effect on the number transcribed. At interview ( Annex 5) Planner B indicated that no discussion or instruction on the normalization process had been included either in November or subsequently.
5.28 A further critical circumstance was that the written procedures that were available to Planner B, in particular BOC work instruction number WI.14.01.01 for Medulla Planning made no reference to the use of a planning computer or the normalization procedure. The latest available version of this document ( Annex 1 to this report), is dated 11th August 1998. As discussed in sub-section 5.2.1, the need for normalization arose because of the BOC decision taken several years earlier to enter a nominal dose of 100cGy for all computer plans. In 2005 with the introduction of Varis 7 the BOC decided to implement the full functionality of the software using the actual prescribed radiation doses and number of fractions as an input to the planning computer. Hence, the version of WI.14.01.01 available to Planner B in December 2005 did not reflect current practice. Planner B stated at interview that following his supervised participation in planning a previous patient in November 2005, he had made his own notes on how to carry out the planning. He was not aware of the existence of any applicable written work instructions at the BOC.
5.29 Taking these circumstances together, the only indication available to Planner B on the need for normalization appears to have been the label on the relevant box in Medulla Planning Form FM.14.014 ( Annex 2 to this report) requesting that the entry for the 'Output' should be in units of 'MU/100cGy'.
5.2.3 The role of the Principal and Senior Planners in planning this treatment
5.30 Treatment planning for Miss Norris took place between 15 th and 19 th December 2005. Senior Planner C, who was involved in checking her treatment plan was absent from the BOC for the whole of the week beginning 12 th December and was rotated for other duties in the week beginning 19 th December. Senior Planner C therefore had no direct initial role in the creation of the treatment plan.
5.31 On 15 th December 2005, Principal Planner A, using CT images imported into the Eclipse planning computer, set up the positions of treatment fields and shielding blocks for the clinical oncology consultant responsible for Miss Norris to approve.
5.32 The involvement of Planner B in treatment planning began on 16 th December 2005, under the supervision of Principal Planner A. During the interviews that formed part of the incident investigation (between 23 rd February and 8 th March 2006) the nature of this supervision was discussed. Since no record of their respective roles was made at the time, details given were from recollection.
5.33 According to these recollections, Principal Planner A, having set up the positions of treatment fields and shielding blocks, gave Planner B initial direction on the next stages of treatment planning and gave him the opportunity to ask for any necessary clarification. Planner B recalls that at each stage in the treatment planning process he checked with others to ensure that he was following the correct procedure. Principal Planner A was unable to recall what approaches had been made to him by Planner B in this connection. Planner B's recollection was that at no stage in planning did he recognize the need for normalization and therefore did not raise this as an issue.
5.34 It appears therefore, that apart from initial instruction, the process of supervision was largely indirect (or reactive), whereby following initial instruction, Planner B progressed all aspects of the planning process independently but with the opportunity to raise any issues of doubt with appropriate senior colleagues. His senior colleagues therefore were not given the opportunity to comment or advise on the normalization process because Planner B did not raise it with them as he was unaware that this was an issue.
5.3 Why was the wrong 'Output' figure on the Medulla Planning Form not identified in checking?
5.35BOC Quality System Document QS14.13 'Checking and Issuing of Plans' requires that ' Prior to issuing a plan, calculations and plans will be independently checked and initialled by a suitably qualified member of Physics planning Staff'. This requirement is in accordance with The Royal College of Radiologists' Clinical Oncology Information Network . Guidelines for external beam radiotherapy [2], Recommendation 48 of which states that ' The monitor unit calculation must be rigorously and independently checked'.
5.36 The initials on the completed Medulla Planning Form for Miss Norris indicate that the plan was checked by Principal Planner A and by Senior Planner C. The precise role of each in the checking process is a matter of recollection rather than of record and is therefore difficult to assess. However, it is apparent that Principal Planner A played the more significant part in the checking process and, since he was also involved in supervising Planner B during planning, his involvement cannot be regarded as wholly independent. Indeed, the degree of his independence in checking the plan might be said to correlate inversely with the level of his supervision.
5.37 Recollections by those involved indicate that initial checking was by Principal Planner A, who identified a number of errors, particularly in the calculation and drawing of the compensators for the spine fields (see Paragraph 4.14). Principal Planner A corrected these errors and passed the amended plan to Senior Planner C for further checking of the corrected spine fields. Both Principal Planner A and Senior Planner C failed to identify the critical error in the 'Output' entry.
5.38 As discussed in Paragraph 5.23, the plan for Miss Norris was the first whole CNS plan at the BOC where the calculated number of MU in the Eclipse Treatment Plan Report was anything other than the number per 100 centiGrays. Therefore, the checkers had no previous experience of checking a plan where normalization had been applied in calculating the 'Output' entry.
5.39 In his formal incident report, Dr Martin notes that ' The CNS plan is complex and the checking process is not straightforward. The most difficult part involved relative positioning of the fields, and the attention of the physicists carrying out the checks may have been distracted by errors in other parts of the plan.'
5.40 In summary, there was nothing in either of the checkers' previous experience of creating or checking treatment plans to alert them to the fact that a normalization procedure had been needed in the calculation of the 'Output' entry for the head fields in this plan. The attention of Senior Planner C was directed by Principle Planner A to the spine fields where he had already identified and corrected errors. As suggested in Paragraph 5.39 the attention of the checkers might therefore have been focussed on those areas of the plan that they knew from experience to be the most complex, particularly the positioning of the spine fields and aspects of adjustment of the position of the junction between the head and upper spine and between the upper and lower spine fields.
5.41 The question that then arises is why was Senior Planner D able to identify the same error made by Planner B in the subsequent plan (Paragraph 5.4(iv))? In this regard, it should be noted that Planner B was unable to recall who had supervised him in the preparation of this subsequent plan and to what extent. Formal checking was by Senior Planner C (not by Senior Planner D) who signed-off the plan but at this stage had failed to identify the error.
5.42 In principle, the information contained in all Treatment Plan Reports generated by Eclipse would be available for electronic transfer to RTChart. Therefore, for those procedures where manual transfer of data was required, the system used at the BOC was to set the status of the Eclipse Treatment Plan Report to " Rejected". For the 'subsequent plan' referred to in Paragraph 5.4(iv), one of the radiographers in pre-treatment, apparently unaware of this procedure, approached treatment planning staff to ask why the plans in the Varis 7 database had been set to "Rejected" status. Prompted by this enquiry Senior Planner D decided to look at a copy of this patient's planning form to remind himself about the relevant working procedures. It was at this point that he realised that the monitor units calculated by Eclipse had been based on the prescribed dose of 1.67Gy per fraction rather than 1Gy per fraction. Since this did not appear to him to accord with the definition of 'Output' given on the planning form, he asked Principal Planner A to double-check this. Principal Planner A confirmed that the number written on the form by Planner B was incorrect and initiated the immediate review of previous plans of this type which identified the error for Miss Norris.
5.43 The question of whether this error might have remained un-noticed had it not been for the radiographer's enquiry must remain open. However, the seriousness of the consequences for future patients of this having been the case and hence the importance of Senior Planner D's intervention cannot be ignored.
5.4 A missed opportunity to identify the potential for error from another plan
5.44 Between completion of the plan for Miss Norris on 19 th December 2005 and identification of the error by Senior Planner D on 1 st February 2006, a (manual) medulla procedure for another patient was planned by another treatment planner (Senior Planner E) under the supervision of Principal Planner A. The Medulla Planning Form is dated 12 th January 2006. In this case also, the prescribed radiation dose was 35 Grays in 20 equal fractions and this prescribed dose was entered into the Eclipse module of the Varis 7 treatment planning system.
5.45 Normalization of the 'Monitor Units' figure appearing on the Eclipse Treatment Plan Report was therefore necessary in completing the Medulla Planning Form and this was accomplished successfully. Senior Planner E's recollection of how he was able to identify this need indicates that he recognized that the units for MU on the Eclipse Treatment Planning report ( MU for the full 175 centiGray dose fraction) were different from the units for the 'Output' entry on the Medulla Planning Form ('MU/100cGy'). Neither he nor his supervisor for this plan, Principal Planner A, could recollect clearly their respective contributions to identification of this need but Senior Planner E noted that the extent of his experience in the BOC's Treatment Planning Section had allowed considerable familiarity in the manual transfer of information from Eclipse Treatment Plan Reports.
5.46 This presented an opportunity to identify the potential for error. However, this was the first whole CNS procedure that Senior Planner E had planned and he would have been unaware therefore of the change from prior practice. The implications of the change from prior practice also went unnoticed by his supervisor. The opportunity was therefore missed.
5.47 By 12 th January, Miss Norris had received only five of the planned 20 fractions, hence the radiation dose and the related health risk would have been greatly reduced had the error been discovered at this stage in the treatment.
« Previous | Contents | Next »