British Journal of Radiology 74 (2001),209-212 © 2001 The British Institute of Radiology
Clinical governance: risks and quality control in radiotherapy
Report on a meeting organized by the BIR Oncology Committee, held at the British Institute of Radiology, London, on 9 February 2000
S G McNee, BSc, PhD
Department of Clinical Physics and Bio-Engineering and Beatson Oncology Centre, Western Infirmary, Dumbarton Road, Glasgow, G11 6NT UK
 |
Introduction
|
|---|
The extent of the current concern for quality and accuracy in radiotherapy practice was demonstrated by the popularity of this meeting, which attracted a "full house" with over 40 unsuccessful applications for registration. As a consequence, the meeting was repeated with minor programme changes on 22 June 2000. The meeting was split into three sessions covering clinical governance/patient management, implementation of quality systems and practical measures for ensuring quality.
|
Session 1: Clinical governance
|
|---|
Dr Fergus MacBeth of the Clinical Effectiveness Support Unit (Wales) challenged the audience to consider the new demands of healthcare. Clinical governance is not a new concept and has its roots in the Griffiths Report of 1983, which criticized the lack of clinical evaluation and measurement of health outputs. Clinicians have always been accountable for the care of individual patients but now we have to account for the care of groups of patients. Can we answer questions such as "How good is your clinical practice?", "How do you know?", "What are you doing to make it even better?" There are real conflicts between the traditional doctor culture of autonomy and self-regulation vs the organization-based philosophies of corporateness and accountability. Clinical governance is an opportunity to integrate these two cultures, which, if taken forward with staff support, can enhance the delivery of quality healthcare. The audience was given the opportunity to determine a logical solution to a problem of inappropriate prescription and delivery of chemotherapy in a fictitious department. The challenge is to solve these problems as they arise in the real department.
Clinical governance has increased the public's expectations of cure but also their intolerance of failure or complications from treatment. The most common complaints from patients concern delays in treatment, incorrect treatment and morbidity. Patients may be seeking financial compensation, explanations, apologies or recognition. They may wish to prevent similar effects for others, they may have been antagonized by the doctor's attitude or they may be unable to come to terms with incurable illness. The result has been a 10–15% annual increase in the number of medical litigation cases. Professor Stan Dische (Mount Vernon Hospital) considered that there is much about medical shortcomings that can be learned from these cases, but this information is not widely disseminated. Most probably this is because most cases either fail or are settled out of court. In either case, this is out of the public domain. It is also affected by an underlying atmosphere of embarrassment, fear and secrecy. Professor Dische provided a fascinating insight to the recent Breast Radiation Injury Litigation case [1] involving a total of 131 patients. Litigation was a lengthy, detailed and costly process, with many twists and turns. One of the key issues was to define what is acceptable clinical practice at a given time. This is not helped by too great a variation in clinical practice (both of technique and dosage) and insufficient reporting of morbidity. This was certainly the case when the Radiation Action Group Exposure (RAGE) patients received treatment, but it remains a problem to this day. In reality, the costs of litigation are small compared with the total healthcare budget, and most claimants fail to win compensation. The radiotherapy community would benefit greatly from an analysis of litigation cases and open reporting of the key features.
Dr Paul Symonds (University of Leicester) considered the response of a clinician when faced with a radiation incident (i.e. significant overdosage or underdosage of a patient). He sought advice from professional bodies, but this was not readily available. Dr Symonds' message to the audience was therefore based on his personal experience of radiation overdose to a patient. The consultant should take control of the situation and make full disclosure to the patient (in the presence of a senior colleague). An apology can be an expression of regret and sympathy, not necessarily an admission of guilt. Steps need to be taken to provide for the medical management and support of the patient. Avoid a "blame culture" when investigating the incident and instigating preventive procedures. Keep management informed (the last thing they will want are "surprises") and prepare for media interest.
The process was illustrated with an example from Dr Symonds' personal practice in a previous department. This incident was reported with commendable candour. The overt cause of the error was human error when using computer software to calculate treatment monitor units. Confusion between "isocentric" and "fixed SSD" resulted in a 50% overdose to the whole neck, including the spinal cord. The incident was discovered when an independent review of the calculations was instigated after the patient's GP reported severe acute reactions. Closer investigation revealed several contributory factors—lack of staff cover while the consultant was on holiday, compromise on treatment owing to lack of availability of treatment units, weaknesses in computer software design and human failure when conducting checking procedures.
The patient was fully informed of the incident and the probable consequences. The support of the Clinical Director in this action was invaluable. All possible clinical action to minimize radiation effects was taken, including referral for specialist neurological opinion, and an interim payment of compensation was settled at an early date. As a consequence of these actions, the confidence of the patient was largely retained.
|
Session 2: Quality systems
|
|---|
Mr Terry Kehoe, Consultant Physicist (Edinburgh), opened the second session by challenging the audience to consider what is meant by quality and how it can be measured. Have ISO9000 systems facilitated improvement in quality within radiotherapy departments? The response from the audience was uncertain. The natural route followed by departments is to focus on hardware, records, calibration checks, etc. Undoubtedly these are important, and they are easily measured and audited. ISO9000 systems are costly to implement and run, but it is possible to operate such a system without ever really addressing quality issues (i.e. examining the quality of the service or the product). An alternative would be to focus on the patient and on personnel. Have patients benefited from our ISO9000 systems? Have staff observed improvements in quality with ISO9000? How would patients measure the quality of our service? Certainly not by reference to machine checks but more likely by perception of professional competence and efficiency of the process of treatment planning and delivery. Incorporation of such issues into the quality system could be derived from implementation of ISO9004, but this has largely been overlooked. If we want a system that leads to real improvement in quality, the message is to focus more on people and less on hardware and records.
Many of these issues have occupied those currently engaged in revising the ISO9000 Standard. A discussion of the likely changes was presented by Mr Nigel Wright, Healthcare Product Manager for the British Standards Institution. Fundamental to the revised system is that it should tell you how well you are performing against the standards you have set. The system will be more process-based, with a strengthened customer focus. Consequently, the role of top management will be enhanced, with the burden of on-going evaluation of the quality management system and the requirement to ensure continual improvement. This will require establishment of monitoring systems and a need to view the system from the patient's perspective. In the new standard (due for publication in 2000), ISO9000 will describe the fundamental requirements, whilst 9001 and 9004 will contain the requirements and guidelines for performance improvement. This will be followed by an auditing standard (ISO19011). It would be possible to incorporate many of the objectives of clinical governance into the revised ISO9000 standard.
Dr Stuart McNee, Principal Physicist (Glasgow), presented the results of an audit of quality processes within UK radiotherapy centres conducted on behalf of the BIR Oncology Committee. A questionnaire was circulated to all UK centres and produced a response of 70%. Almost all departments had achieved or were in the process of being accredited to ISO9000. There was marked variation in staffing resources allocated to the quality assurance (QA) system, which appeared not to be related to department size. Treatment record and verify systems are not yet universally available and, when they are, the predominant method of data entry is by keyboard. Systems for checking data entry are varied. Use of in vivo dosimetry systems to verify target dose is not commonplace. Significantly more use is made of portal imaging to verify beam alignment but there is still scope for improvement. There is very restricted implementation of medical peer-review of target drawing and dose prescription.
Information was gathered on the incidence of treatment errors and "near misses". Again, the incidence showed marked variation, with a suggested average of 1% of courses of radiotherapy being affected by a treatment error. Few of these would have resulted in significantly compromised treatment, but detection often occurred after several fractions had been delivered. Almost every conceivable type of error has been experienced. These have either a dosimetric or geometric component, which might be more easily detected by routine use of in vivo dosimetry and portal imaging. The cause or errors is frequently attributed to human error, but lack of facilities (e.g. record and verify, data networking) and staff shortages were also important. Many errors had a "technology" component (e.g. inconsistencies in labelling of scales, lack of customization) and we are too dependent on human intervention and "over-riding" in these situations. Overall, the survey demonstrated a strong commitment of staff to quality systems and a belief that quality in radiotherapy had improved as a consequence. Clearly, there is still more progress to be made.
Mr Steve Ebdon-Jackson, from the Department of Health, discussed the change in requirements for reporting errors in radiotherapy treatment following the implementation of the Ionising Radiation (Medical Exposure) Regulations (IR(ME)R). Under POPUMET, reporting was voluntary, but there was a good reporting record within radiotherapy. Typical incidents related to wrong patient identification, wrong site or wrong wedge. The concern of the Department of Health was to ensure that the correct processes were put in place, for example for effective, independent checking. Under IR(ME)R, reporting of doses much greater than intended will be mandatory and guidance will be given on levels of significance. One difference from POPUMET is that clinicians do not, by default, carry full responsibility. Instead, it will be possible to identify an appropriate duty holder. The key objective of the Department of Health will be to educate and improve standards rather than to prosecute.
This prompted an enthusiastic debate on how the radiotherapy community could best disseminate information regarding types and causes of treatment errors. The Inspectorate could share information, but there could be problems preserving anonymity, especially if litigation or prosecution was under consideration. Such constraints might not apply for "near misses" and this might be a process best organized by the professional bodies. It was suggested that creation of a reporting mechanism might be considered by the BIR.
|
Session 3: "Taking action"
|
|---|
Dr Tommy Knöös, Senior Physicist (Lund University Hospital), presented an insight into the verification of delivered dose in Swedish Hospitals. His strategy involves three components: checks on source-to-skin distances (SSDs), independent calculation of monitor units and measurement of entrance doses. All Swedish centres use diodes for in vivo dosimetry—usually just for measurement of entrance doses and generally just on the first treatment. A tolerance of 5% can be achieved with a modest input of effort for diode calibration; 3% tolerance can be achieved but for much greater effort. Error detection rates (at the 5% tolerance) range from 1% to 13% according to department. Errors detected include isocentric/fixed SSD difficulties, wedges (omission or wrong angle) and errors in monitor unit calculation. Use of in vivo dosimetry is long established in Swedish departments and experience shows it to be indispensable as a QA tool. It is thought to result in marginal loss of treatment time but this should be assessed against a background of a much lower patient throughput than is the case for linear accelerators in the UK.
Dr Peter Remeijer (Netherlands Cancer Institute) discussed the value of portal imaging in reducing set-up errors for treatment of common sites. There are 17 sources of geometric errors in the planning and treatment process. Those connected with planning are systematic whilst those connected with treatment delivery are random. Where treatment is with a large number of fractions, the most important point is to reduce systematic errors (the random errors tend to cancel). This is achieved by imaging on successive fractions, performing off-line analysis and making corrections to the isocentre position at the next treatment. Additional imaging for confirmation or further correction is required. On average, images may be acquired on 8 to 10 fractions. Applying this methodology to treatment of the prostate, the maximum deviation in set-up can be reduced from 
12 mm to 5 mm. Adjustment is required for about 50% of patients. This improvement is very significant in relation to the margin typically applied between gross tumour volume and planning target volume for this site. Random errors may be more important where few fractions are delivered. In such cases, e.g. rectum, an automated on-line verification and correction procedure is implemented fraction by fraction, producing improvements similar to that for prostate.
What practical methods can be taken to assure correct radiotherapy dose in the UK? Dr David Thwaites, Consultant Physicist (Edinburgh), stressed the importance of taking a global perspective. It is not simply a matter of assuring dose to the ICRU point but also to critical structures. Tolerances, both dosimetric and geometric, have to be defined. Much effort has been directed at QA of planning computers, dosimetry audit and machine performance. Planning processes must also be incorporated into the system but little effort has been made to standardize medical processes. How do we assure the quality of target drawing, the application of appropriate margins and the verification of set-up? Variations in time–dose–fractionation schedules will change biological doses.
Quality control must be cost effective, reliable and readily usable. In the UK, pressure on patient throughput means that adding even 1 min to each treatment time may not be acceptable. To date, the aims may have been to prevent accidents and to minimize their consequences. Perhaps we should be actively trying to improve precision and outcomes. Quality control should be applied across all processes as well as to individual patient treatments and should incorporate independent checking and audit. If in vivo dosimetry and portal imaging are not to be fully utilized, can we be satisfied with the alternatives? Ensuring consistent dose delivery within acceptable tolerances is not simple and, as treatments become more complex, so QA will become more demanding, both in sophistication and resources.
|
Summary
|
|---|
The professions involved in radiotherapy constantly strive to improve quality and to minimize errors and their effects. We are aware of our shortcomings but sometimes are powerless to take corrective action. We have seen that clinical governance requires some change in the medical culture but much has yet to be achieved in the standardization of medical processes. We would gain by sharing experiences of system failures, whether that be by reporting near misses or by disseminating details of litigation cases. In a specialty that depends on human processes and interactions, it is no surprise to find a measurable incidence of human errors. Sometimes this is also related to technological or design deficiencies and we rely too much on human interventions to overcome these. We would like to take positive steps to improve quality, such as those inherent in the new ISO9000 standards, but we are hindered by a variety of constraints—manpower, old technology, pressure to achieve high patient throughput.
We need to make decisions for the future development of radiotherapy in the UK. We could settle for a system that is characterized by high throughput and high risk. Hopefully, we will receive the necessary encouragement and resources to strive for high standards of treatment and low risk.
Received for publication July 13, 2000.
Revision received September 11, 2000.
Accepted for publication October 16, 2000.
|
Reference
|
|---|
-
Dische S, Joslin C, Miller S, Bell N, Holmes J. The breast radiation injury litigation and the clinical oncologist. Clin Oncol 1998;10:367–71.
Top
Introduction
Session 1: Clinical governance
