3 Evidence and interpretation
Assessing the risk of transmission of CJD via interventional procedures was the first step in the guidance development process. The risk assessment was informed by a number of evidence sources.
A risk assessment by the Department of Health Economics, Statistics and Operational Research (ESOR) Division.
A systematic review of issues such as the prevalence of CJD in the general population and efficacy of decontamination arrangements (see the Review Group's final report).
A formal elicitation process using leading experts in CJD and decontamination and frontline practitioners (see the Review Group's final report).
Empirical data from experimental studies coordinated by the Department of Health Research and Development Programme.
Hospital Episodes Statistics (HES) were used to derive the number of operations per specialty per year and the rates of repeat operations.
The NHS Purchasing and Supplies Agency (PASA) database provided information on costs of reusable instruments.
Submissions by manufacturers on decontamination arrangements, tracking systems and costs of single-use instruments (see appendix E).
Additional information provided during the public consultation period.
Overall, there was a lack of data on many of the key parameters such as the prevalence of vCJD infectivity in the UK population and the effectiveness of current decontamination practices against human prions, resulting in a high degree of uncertainty (that is, wide reference ranges in the model outputs). Probabilistic sensitivity analysis was used to quantify the uncertainty surrounding these key parameters. The risk assessment took account of the dynamics of CJD transmission via interventional procedures. Finally, the assessment also considered the costs and cost effectiveness of various methods of reducing the risk of transmission.
Three scenarios based on the inputs from the sources listed above were constructed: pessimistic, median and optimistic. The model outputs for each scenario were presented to the Committee for all medium- and high-risk operations as defined in the guidance remit (over 2 million operations per year in England and Wales). Initial model runs showed that, even when pessimistic assumptions about a high prevalence of CJD and a low effectiveness of decontamination were made, the number of cases of CJD resulting from medium-risk operations was still less than 25 cases from a total of approximately 2 million medium-risk operations undertaken in England each year. Therefore, the modelling suggested that medium- and low-risk interventions were unlikely to result in a self-sustaining epidemic and that changes in practice for these interventions would not be cost effective. As a result, further modelling focused on high-risk procedures only.
A number of potentially more effective decontamination methods are currently in development. These include decontamination solutions or changes in decontamination practice, such as keeping instruments wet immediately after use until they are cleaned. Preliminary commercial and academic in-confidence submissions on the efficacy of decontamination solutions and their likely costs were reviewed by the Committee and used to model their effects on transmission of CJD via interventional procedures. Based on the submissions considered, the modelling assumed that these new decontamination agents would be introduced universally throughout the NHS within the next 5 years. The model also assumed that these agents would reduce the average infectivity of all tissues by 5 log, which was at the lower end of the efficacy ranges submitted to the Institute. The safety and efficacy of these agents against human prions require rigorous evaluation. SEAC has described the principles which should govern this work.
Despite attempts to obtain data on instrument swapping and the use of supplementary instruments, very limited information about these practices exists. However, after considering the evidence available, the Committee assumed that the probability of at least one instrument, on average, being swapped during an operation or the subsequent decontamination process was 50%. Where a supplementary instrument was used, it was assumed that there was a 50% probability that it would become part of the main set with a similar instrument from the original set becoming the supplementary instrument (see the Review Group's final report).
For high-risk procedures, the model showed that keeping reusable instruments together within sets significantly reduced the number of cases of CJD transmission. Threshold analysis showed that, at an assumed willingness to pay £30,000 per quality-adjusted life year (QALY), it would be, on average, cost-effective for the average neurosurgical (brain surgery) unit to spend up to £5.9 million over a 5-year period in purchasing additional supplementary instruments and/or introducing instrument tracking methods, where appropriate, in order to eliminate instrument migration. For posterior eye surgery it would be cost effective to spend up to £10.9 million per surgical unit over a 5-year period (see the Review Group's final report).
For medium-risk procedures, there was no evidence that keeping reusable instruments together in sets reduced CJD transmission because of the lower levels of infectivity. However, the Committee considered that keeping sets together in medium- and low-risk procedures was desirable for other reasons (see 3.2.6).
Systematic review of the literature identified a number of studies on the complication rates associated with single-use compared with reusable instruments in surgery, but these were related almost entirely to adenotonsillar surgery. The studies took the form of three prospective audits, six retrospective audits and a survey (see the Review Group's final report for details). Among these studies, the Welsh National Tonsillectomy Audit was considered to provide the highest quality evidence. It was therefore considered inappropriate to weaken the findings of this audit by combining these data with data from other studies in a meta-analysis.
The Welsh National Tonsillectomy Audit indicated that, compared with reusable instruments, single-use instruments were initially associated with a more than 100% increase in complication rates. Primary haemorrhage resulting in return to theatre doubled from a baseline rate of 0.6% with reusable instruments to 1.2% when single-use instruments were introduced. The rates returned to baseline levels (that is, equivalent to reusable instrument complication rates) when appropriate procurement and surveillance mechanisms were introduced, which ensured that single-use instruments met detailed specifications. The investigators also conducted two separate and detailed audits of suppliers of single-use tonsillectomy instruments in the UK, which showed considerable variation in the quality and consistency of the single-use instruments available. Careful specification and quality control of instruments were therefore emphasised as key requirements to ensure patient safety.
Scenarios were presented which assumed higher complication rates for single-use instruments in high-risk specialties (see the Review Group's final report).
The model suggested that for every case of transmission of CJD via an interventional procedure, the average number of QALYs lost, based on the age distribution of the patients undergoing a procedure, was in the order of 17 for neurosurgery and 11 for posterior eye surgery.
The direct cost to the NHS of treating patients with CJD (approximately £40,000 per patient) is based on estimates provided by the Department of Health (see the Review Group's final report). Many other costs are borne by the public sector, individuals and society in general, but these have not been included in the cost-effectiveness modelling, in accordance with the NICE reference case (see the glossary, appendix F).
Modelling was performed using the assumption that single-use instruments would cost the same as their reusable counterparts. This assumption was made because there are almost no single-use instruments on the market for high-risk operations, meaning that, in most cases, reusable instruments would need to be used and disposed of after first use. In practice, the costs of single-use instruments may be lower than this.
Lists of the instruments used in all relevant procedures were prepared with the help of healthcare professionals and Committee members. For high-risk procedures, specific lists were compiled of those instruments that normally come into contact with high-risk tissues – these are the instruments that would need to be replaced with single-use alternatives. However, for many such instruments single-use versions are not currently available.
The costs associated with procurement and surveillance protocols for single-use instruments were also accounted for. These were based on the Welsh experience with single-use instruments for tonsillectomy. The costs of disposing of single-use instruments were assumed to be similar to the costs of autoclaving reusable instruments.
3.1.7 Cost effectiveness of preventing transmission of CJD through high-risk procedures using single-use instruments
The results of the model strongly suggested that preventing instrument migration for high-risk procedures was both effective and cost effective in reducing the risk of CJD transmission. Modelling of further measures to minimise the risk of CJD transmission was therefore performed assuming that reusable instruments stay together within sets.
Specifically, modelling was carried out to assess the effectiveness and cost-effectiveness of introducing single-use instruments for patients who have had previous high-risk procedures. Patients who have undergone high-risk procedures may have an increased risk of having contracted CJD and could potentially transmit the disease via subsequent high-risk procedures. Using single-use instruments on these patients would prevent onward transmission and reduce the overall number of infections via interventional procedures.
The estimated average mean incremental cost per QALY of introducing single-use instruments for patients who have previously undergone high-risk surgery was approximately £100,000 for intradural brain surgery and £45,000 for surgery on the retina and optic nerve. For surgery on the spine, the average cost per QALY was around £400,000. The estimated average incremental cost per QALY of introducing single-use instruments for all patients undergoing high-risk operations was in excess of £50,000.
For neuroendoscopy, the average incremental cost per QALY of introducing single-use accessories was around £16,000 for rigid and £22,000 for flexible neuroendoscopes. The average incremental cost per QALY of introducing single-use neuroendoscopes was £126,000 for rigid and over £1 million for flexible neuroendoscopes. These figures assume that rigid reusable neuroendoscopes are autoclaved. For the above calculations, the full price of reusable surgical instruments, neuroendoscopes and accessories was used, as there are no single-use equivalents currently available, with the exception of single-use rigid neuroendoscopes (see the Review Group's final report).
The confidence intervals around these means were very wide, with median cost-effectiveness values being generally higher than the means, as the distributions were highly skewed. Cost-effectiveness acceptability curves were used to depict the uncertainty around the mean (see the Review Group's final report).
3.2.1 The Committee reviewed the available evidence on the risk of transmission of CJD via interventional procedures and on the safety and clinical effectiveness of a range of interventions aimed at reducing this risk. Prevention of a self-sustaining epidemic was considered particularly important. The work of the Committee was hindered by a lack of firm evidence and considerable uncertainty surrounding many key aspects of CJD, such as the prevalence of vCJD infectivity in the UK population and the efficacy of current decontamination practices against human prions. These uncertainties caused considerable problems for the Committee in its attempt to identify appropriate risk-reducing and cost-effective interventions which could be rapidly introduced in the NHS.
3.2.2 In its deliberations, the Committee was mindful of the need to ensure that its advice took account of the efficient use of NHS resources. At the same time, it was particularly concerned about the need to avoid even a low risk of a self-sustaining CJD epidemic occurring through surgical transmission. The Committee considered three main measures as possible means of reducing the risk of transmission of CJD.
The introduction of more effective decontamination practices.
The prevention of instrument migration between sets.
The introduction of single-use instruments.
The Committee took account of evidence relating to the safety, efficacy, costs and practicalities of implementing these different interventions for various types of procedures and patient groups, with a view to reducing the risk of transmission of CJD via interventional procedures.
3.2.3 The Committee considered in-confidence information about agents that are being developed to decontaminate instruments from human prions. The manufacturers' submissions, together with information provided by research groups and the Department of Health Research and Development Programme, satisfied the Committee that effective and practical decontamination methods are likely to be available in the NHS within the next 5 years, and that, once available, they could prevent surgical transmission of CJD via most types of instruments. The Committee therefore decided to limit the time horizon for modelling to 5 years. However, the data were considered insufficient to support any recommendations about the use of these decontaminants. Further assessments of efficacy against human prions, safety, and costs of these decontaminants were regarded by the Committee as urgently needed to support future recommendations to the NHS.
3.2.4 Initial outputs from the model showed that it is not cost effective to change to single-use instruments for medium-risk procedures, and, by implication, for low-risk procedures. Medium-risk procedures include tonsillectomy, for which single-use instruments are readily available, in contrast to many other procedures. These instruments are relatively inexpensive and have been the subject of a number of studies. More detailed modelling indicated that for tonsillectomy, the mean cost of changing to single-use instruments was over £500,000 per QALY using median scenarios, and over £50,000 per QALY for pessimistic scenarios.
Model outputs indicated that continuing with reusable instruments for medium-risk procedures (including tonsillectomy) would not confer a risk of a self-sustaining CJD epidemic. In view of these results, it was decided that the model should concentrate on high-risk procedures only, namely intradural neurosurgery, neuroendoscopy and posterior eye surgery (see appendix C for a full list of high-risk procedures). Further modelling demonstrated that any change in practice for surgery on the spine, which is considered to be 100 times less infectious than the brain and posterior eye, would not be cost-effective. Therefore, the Committee focused on intradural brain surgery, neuroendoscopy and posterior eye surgery.
Approximately 56,000 high-risk and 2 million medium-risk procedures are performed each year in England and Wales.
3.2.5 The Committee was particularly concerned about the possible risks associated with instruments migrating from one set to another and the use of supplementary instruments. Modelling showed that migration of instruments between sets increased the risk of transmission of CJD in high-risk procedures and that preventing instrument migration through tracking and abolishing the use of supplementary instruments was cost effective in these procedures. The Committee considered it most important that systems are put in place to ensure no instruments are swapped between sets in high-risk procedures and that the effectiveness of these systems is demonstrated through regular audit. The Committee noted that this practice is line with the guidance set out in Health Service Circular 2000/032.
3.2.6 For medium- and low-risk procedures, instrument swapping had no significant effect on the overall risk of onward transmission of CJD. However, the Committee noted that there may be other reasons for keeping instruments together in sets in these procedures, as suggested in Health Service Circular 2000/032, such as to allow identification of patients who may have been exposed to the risk of any infectious disease at the time of an interventional procedure.
3.2.7 The Committee was aware of advice by professional organisations, such as the British Society for Gastroenterology, recommending replacement of reusable endoscope accessories with single-use items. Although this is not a cost-effective measure to prevent transmission of CJD (except in the case of accessories for neuroendoscopy), single-use instruments (including accessories) currently represent the only means of eliminating all risk of transferring infectious diseases from one patient to another. This provides justification for continuing the carefully monitored development and introduction of single-use instruments.
3.2.8 The Committee considered neuroendoscopy and decided that reusable rigid neuroendoscopes that can be autoclaved should be used whenever possible and should be thoroughly cleaned and autoclaved after each use. Only when the use of rigid neuroendoscopes is impractical should reusable flexible neuroendoscopes be used.
3.2.9 The Committee considered children born since 1 January 1997, who are unlikely to have been exposed to BSE or CJD via diet or blood transfusion, respectively. Therefore, the prevalence of CJD in this population is close to zero. The Committee decided that new reusable surgical instruments for high-risk procedures and new neuroendoscopes should be purchased and used solely on those children born after 1 January 1997 who have not previously undergone high-risk procedures. These instruments and neuroendoscopes should not be used on children who have been identified as being at risk of any form of CJD, including inherited CJD.
3.2.10 The Committee discussed the performance of single-use compared with reusable instruments, with a focus on patient safety. There were repeated concerns that complication rates might be increased if single-use instruments were of inadequate quality. The published literature, and in particular the Welsh National Tonsillectomy Audit, demonstrated that this could occur, but also showed that single-use instruments can be of equivalent quality to reusable instruments, provided that appropriate quality-control mechanisms are in place. The Committee therefore accepted the premise that the quality and performance of single-use instruments could be equivalent to those of reusable instruments, provided appropriate procurement, quality control and audit mechanisms are in place. The modelling was undertaken assuming equivalent performance and taking account of the costs of an adequate quality-control system.
3.2.11 The Committee noted that single-use instruments are used regularly in some high-risk procedures, but that no single-use instruments exist to replace the reusable ones for other high-risk procedures. Based on the Welsh experience with tonsillectomy instruments, there would be a significant time lag (18 to 24 months) before single-use instruments could be introduced with mechanisms to guarantee appropriate levels of consistency and quality.
The Committee considered the scenario of introducing single-use instruments for high-risk procedures, either for all patients or only for those who have had previous high-risk surgery. However, this was not cost effective, except in the case of single-use accessories for neuroendoscopy for all patients. Taking account of both costs and the practical challenges of introducing single-use instruments, the Committee decided against recommending their use, with the exception of single-use accessories for neuroendoscopy.
3.2.12 Throughout its deliberations the Committee was mindful of the potential implications of a self-sustaining CJD epidemic. This was particularly important given the episodic nature of the possible transmission of CJD via high-risk operations. The model demonstrated that, unless instruments are kept with their sets, a single high-risk procedure in an infectious patient could result in a large number of cases of onward transmission. For example, the model suggested that if instrument swapping and use of supplementary instruments continued at assumed current levels, there was a 10% chance that approximately 700 new fatal CJD infections could occur as a result of transmission by interventional procedures every year in England and Wales. This could result in an epidemic occurring within the 5-year timeframe considered by the Committee. The likelihood of an epidemic would be significantly reduced if instruments were kept with their sets. Therefore, the Committee felt it was essential that instrument swapping should not occur and that the effectiveness of any measures taken to ensure instruments are kept together is clearly demonstrated through audit.
 Assessing the risk of vCJD transmission via surgery: an interim review (June 2005).