Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Cost-effectiveness of ultrasound-guided foam sclerotherapy, endovenous laser ablation or surgery as treatment for primary varicose veins from the randomized CLASS trial

Cost-effectiveness of ultrasound-guided foam sclerotherapy, endovenous laser ablation or surgery... Abstract Background The treatment of patients with varicose veins constitutes a considerable workload and financial burden to the National Health Service. This study aimed to assess the cost-effectiveness of ultrasound-guided foam sclerotherapy (UGFS) and endovenous laser ablation (EVLA) compared with conventional surgery as treatment for primary varicose veins. Methods Participant cost and utility data were collected alongside the UK CLASS multicentre randomized clinical trial, which compared EVLA, surgery and UGFS. Regression methods were used to estimate the effects of the alternative treatments on costs to the health service and quality-adjusted life-years (QALYs) at 6 months. A Markov model, incorporating available evidence on clinical recurrence rates, was developed to extrapolate the trial data over a 5-year time horizon. Results Compared with surgery at 6 months, UGFS and EVLA reduced mean costs to the health service by £655 and £160 respectively. When additional overhead costs associated with theatre use were included, these cost savings increased to £902 and £392 respectively. UGFS produced 0·005 fewer QALYs, whereas EVLA produced 0·011 additional QALYs. Extrapolating to 5 years, EVLA was associated with increased costs and QALYs compared with UGFS (costing £3640 per QALY gained), and generated a cost saving (£206–439) and QALY gain (0·078) compared with surgery. Applying a ceiling willingness-to-pay ratio of £20 000 per QALY gained, EVLA had the highest probability (78·7 per cent) of being cost-effective. Conclusion The results suggest, for patients considered eligible for all three treatment options, that EVLA has the highest probability of being cost-effective at accepted thresholds of willingness to pay per QALY. Introduction The treatment of patients with varicose veins results in a considerable workload and financial burden to the National Health Service (NHS)1. Ultrasound-guided foam sclerotherapy (UGFS) and endovenous laser ablation (EVLA) are now commonly used alternatives to surgery. Despite their widespread adoption, uncertainty exists regarding their cost-effectiveness and about which method offers the best value for money. To date there have been three published cost–utility analyses2–4 comparing these treatment modalities from a UK NHS perspective, based on estimated treatment costs and outcomes, follow-up costs and 5-year recurrence rates. These have revealed differing results. One study2 found EVLA and radiofrequency ablation (RFA) to be the most cost-effective, another3 reported day-case EVLA to be cost-effective, whereas a third study4 found UGFS to be the most cost-effective. CLASS (Comparison of LAser, Surgery and foam Sclerotherapy) is a multicentre randomized clinical trial5 that compared the clinical and cost-effectiveness of three treatments: UGFS, EVLA and surgery. Participant-level cost and effect data were collected at 6 weeks and 6 months. A Markov decision model was developed to extrapolate the trial cost-effectiveness data over a 5-year time horizon. Results of these analyses are presented here. Methods A total of 798 participants were recruited from 11 centres in the UK between November 2008 and October 2012. All three treatment options were available in eight centres, whereas three centres offered only UGFS and surgery. Participants were randomized between the treatments with even allocation. To conduct a valid three-way comparison on costs and outcomes, as required for cost-effectiveness analysis, this article reports on the eight centres at which all three treatment options were available. The allocation used a minimization algorithm that included centre, age (less 50 years, 50 years or older), sex, great saphenous vein (GSV) or small saphenous vein (SSV) reflux, and unilateral or bilateral veins. Inclusion criteria were5: age over 18 years; primary unilateral or bilateral symptomatic varicose veins (Clinical Etiologic Anatomic Pathophysiologic (CEAP) class C2 or above); GSV and/or SSV involvement; and reflux duration more 1 s on duplex ultrasound imaging. Exclusion criteria comprised: current deep vein thrombosis; acute superficial vein thrombosis; a GSV or SSV diameter smaller than 3 mm or greater than 15 mm; tortuous veins considered unsuitable for EVLA; and contraindications to UGFS, or to general/regional anaesthesia that would be required for surgery. Further details of the trial have been published elsewhere5. Treatments Surgery was performed under general anaesthetic. It consisted of proximal GSV/SSV ligation and stripping (all GSV) and concurrent phlebectomies. EVLA of truncal veins was performed under local anaesthetic in either an operating theatre (9·5 per cent of patients) or treatment room (90·5 per cent). UGFS was administered to residual varicosities at 6-week follow-up if required, with the exception of one centre that performed concurrent phlebectomies. Foam was produced using the Tessari technique6 at a ratio of 0·5 ml sodium tetradecyl sulphate (STS) to 1·5 ml air (3 per cent STS for GSV/SSV truncal veins, 1 per cent for varicosities, maximum 12 ml foam per session). Resource use and costs Patient-level resource use data were collected for each treatment modality and entered on case report forms. Details of the following resource items were collected: location and time of treatment and recovery, and grade of surgeon and anaesthetist. A separate survey of participating centres collected information on nursing and allied health professionals present for each procedure, and equipment and consumables used. All resource inputs were costed in 2010–2011 unit prices (Table 1). The time of entering and time of leaving the operating theatre or treatment room was used to measure the total time requirement of staff present for the procedure. Staffing information and treatment/recovery duration times were combined with national unit cost data (Personal Social Services Research Unit, 2011)7 to estimate the total cost of staff time. Although the staff cost multipliers included an allocation for overheads and use of hospital space, these may not adequately capture the additional overhead costs associated with procedures done in theatre under general anaesthetic. Therefore, a secondary analysis was conducted whereby an estimate of theatre overhead cost (£218 per hour)9 was applied to procedures carried out in this location. Capital equipment costs were amortized over the expected useful lifespan of the item, and allocated on a per-patient basis using estimated annual clinical throughput obtained from sites. The majority of centres in CLASS reported receiving the laser generator for EVLA on loan, but paying for laser fibres. Therefore, the generator costs were excluded from the base-case analysis, but were included in a sensitivity analysis. A list of consumables required for each procedure was established and costed using unit prices obtained from participating centres. Table 1 Unit costs applied in the analysis Input variables . Unit cost (£) . Source . Nursing staff (cost per h) Band 2 20 PSSRU, 20117 Band 3 24 PSSRU, 20117 Band 4 30 PSSRU, 20117 Band 5 82 PSSRU, 20117 Band 6 107 PSSRU, 20117 Band 7 129 PSSRU, 20117 Band 8b (consultant nurse) 147 PSSRU, 20117 Medical staff (cost per h) Foundation Year 1 33 PSSRU, 20117 Foundation Year 2 core trainee, CCT 42 PSSRU, 20117 Specialty trainee 59 PSSRU, 20117 Staff grade 95 PSSRU, 20117 Associate specialist 131 PSSRU, 20117 Consultant medical 136 PSSRU, 20117 Equipment (cost per use) Ultrasound machine 8·78 Participant centres ECG, pulse oximeter, blood pressure monitor 4·15 Participant centres Laser generator* 34·67 Participant centres Consumables Surgery 159·56 Participant centres UGFS 26·23 Participant centres EVLA 321·06 Participant centres Follow-up costs GP visit 36 PSSRU, 20117 Practice nurse visit 12 PSSRU, 20117 Vascular surgery follow-up outpatient 123 Department of Health, 20118 appointment Duplex imaging 53 Department of Health, 20118 Input variables . Unit cost (£) . Source . Nursing staff (cost per h) Band 2 20 PSSRU, 20117 Band 3 24 PSSRU, 20117 Band 4 30 PSSRU, 20117 Band 5 82 PSSRU, 20117 Band 6 107 PSSRU, 20117 Band 7 129 PSSRU, 20117 Band 8b (consultant nurse) 147 PSSRU, 20117 Medical staff (cost per h) Foundation Year 1 33 PSSRU, 20117 Foundation Year 2 core trainee, CCT 42 PSSRU, 20117 Specialty trainee 59 PSSRU, 20117 Staff grade 95 PSSRU, 20117 Associate specialist 131 PSSRU, 20117 Consultant medical 136 PSSRU, 20117 Equipment (cost per use) Ultrasound machine 8·78 Participant centres ECG, pulse oximeter, blood pressure monitor 4·15 Participant centres Laser generator* 34·67 Participant centres Consumables Surgery 159·56 Participant centres UGFS 26·23 Participant centres EVLA 321·06 Participant centres Follow-up costs GP visit 36 PSSRU, 20117 Practice nurse visit 12 PSSRU, 20117 Vascular surgery follow-up outpatient 123 Department of Health, 20118 appointment Duplex imaging 53 Department of Health, 20118 * The laser generator cost was included only in a sensitivity analysis (see text). PSSRU, Personal Social Services Research Unit; CCT, Certificate of Completion of Training; ECG, electrocardiograph; UGFS, ultrasound-guided foam sclerotherapy; EVLA, endovenous laser ablation; GP, general practitioner. Open in new tab Table 1 Unit costs applied in the analysis Input variables . Unit cost (£) . Source . Nursing staff (cost per h) Band 2 20 PSSRU, 20117 Band 3 24 PSSRU, 20117 Band 4 30 PSSRU, 20117 Band 5 82 PSSRU, 20117 Band 6 107 PSSRU, 20117 Band 7 129 PSSRU, 20117 Band 8b (consultant nurse) 147 PSSRU, 20117 Medical staff (cost per h) Foundation Year 1 33 PSSRU, 20117 Foundation Year 2 core trainee, CCT 42 PSSRU, 20117 Specialty trainee 59 PSSRU, 20117 Staff grade 95 PSSRU, 20117 Associate specialist 131 PSSRU, 20117 Consultant medical 136 PSSRU, 20117 Equipment (cost per use) Ultrasound machine 8·78 Participant centres ECG, pulse oximeter, blood pressure monitor 4·15 Participant centres Laser generator* 34·67 Participant centres Consumables Surgery 159·56 Participant centres UGFS 26·23 Participant centres EVLA 321·06 Participant centres Follow-up costs GP visit 36 PSSRU, 20117 Practice nurse visit 12 PSSRU, 20117 Vascular surgery follow-up outpatient 123 Department of Health, 20118 appointment Duplex imaging 53 Department of Health, 20118 Input variables . Unit cost (£) . Source . Nursing staff (cost per h) Band 2 20 PSSRU, 20117 Band 3 24 PSSRU, 20117 Band 4 30 PSSRU, 20117 Band 5 82 PSSRU, 20117 Band 6 107 PSSRU, 20117 Band 7 129 PSSRU, 20117 Band 8b (consultant nurse) 147 PSSRU, 20117 Medical staff (cost per h) Foundation Year 1 33 PSSRU, 20117 Foundation Year 2 core trainee, CCT 42 PSSRU, 20117 Specialty trainee 59 PSSRU, 20117 Staff grade 95 PSSRU, 20117 Associate specialist 131 PSSRU, 20117 Consultant medical 136 PSSRU, 20117 Equipment (cost per use) Ultrasound machine 8·78 Participant centres ECG, pulse oximeter, blood pressure monitor 4·15 Participant centres Laser generator* 34·67 Participant centres Consumables Surgery 159·56 Participant centres UGFS 26·23 Participant centres EVLA 321·06 Participant centres Follow-up costs GP visit 36 PSSRU, 20117 Practice nurse visit 12 PSSRU, 20117 Vascular surgery follow-up outpatient 123 Department of Health, 20118 appointment Duplex imaging 53 Department of Health, 20118 * The laser generator cost was included only in a sensitivity analysis (see text). PSSRU, Personal Social Services Research Unit; CCT, Certificate of Completion of Training; ECG, electrocardiograph; UGFS, ultrasound-guided foam sclerotherapy; EVLA, endovenous laser ablation; GP, general practitioner. Open in new tab Details of secondary healthcare services use after treatment (hospital admissions and outpatient attendances) were collected at 6 weeks and 6 months. Use of primary care services was collected from a 6-month patient-completed questionnaire. All CLASS participants were invited to attend a 6-week and 6-month assessment appointment, but only the 6-week assessment following UGFS or EVLA (to assess the need for further foam treatment) was considered potentially consistent with routine practice. Therefore, the cost of an outpatient appointment plus ultrasound scan was incorporated for patients attending this appointment following UGFS or EVLA. For patients in the EVLA or UGFS groups who underwent additional foam treatment at 6 weeks, the cost of the treatment, rather than the cost of follow-up assessment, was applied. The impact of applying the cost of an assessment visit before receiving top-up foam treatment was also assessed, as was the impact of removing all post-treatment assessment costs. National unit cost data were used to cost secondary and primary care contacts (Table 1)7,8. The total cost to the health service (to 6 months) was calculated for each participant by summing their treatment and follow-up costs. Health outcomes The EQ-5D™ 3 level (EQ-5D-3 L) (EuroQol Group, Rotterdam, The Netherlands) was completed by participants at baseline, 6 weeks and 6 months. A preference-based utility score was derived for each participant's response to the EQ-5D™ using the UK population time trade-off tariff10. These scores represent the desirability of the different heath states defined by the EQ-5D™ on a scale where 0 represents death and 1 denotes full health. Quality-adjusted life-years (QALYs) can be derived from responses to the EQ-5D™ by multiplying the time (in years) spent in different states of health by the utility score associated with each state. The QALYs accrued to each participant at 6 months were calculated assuming that the change in utility between measures at adjacent time points followed a straight line. Analysis of participant-level cost and outcome data Data were summarized and analysed by intention to treat using Stata™ version 12.111. The mean incremental costs and QALYs associated with EVLA and UGFS versus surgery were estimated using generalized linear regression, with adjustment for minimization variables and baseline values, as appropriate. The method of recycled predictions12 was used to recover the estimated mean cost and 6-month utility estimates for surgery, and the mean utility and cost increments associated with UGFS and EVLA were used to recover the expected utilities and costs for patients receiving these treatments. This analysis was also repeated using a multiple imputation data set generated using chained equations13. Modelling of longer-term cost-effectiveness The cost and effect estimates from the trial data were used to populate the Markov model, developed in TreeAge Pro14 to extrapolate cost-effectiveness over a 5-year time horizon. This is the duration to which the CLASS study follow-up will ultimately be extended, providing a validity check on initial model-based estimates of cost-effectiveness. The model was constructed to simulate transitions between discrete health states over 6 months (Markov cycle), for a cohort of patients with mean age and sex matching that of trial participants (Fig. 1). For the first 6-month cycle, the model was populated using the estimated mean cost and utility data obtained from the analysis of individual patient data. Beyond 6 months, monthly probabilities of clinical recurrence, derived from a network meta-analysis undertaken to inform the recent National Institute for Health and Care Excellence (NICE) guideline2 (Table 2), were incorporated into the model. Clinical recurrence was assumed to result in a drop in health state utility to the mean baseline value observed across the treatment allocation groups in CLASS. Following clinical recurrence, patients could either present for further treatment (surgery, UGFS or EVLA), and transit to the ‘post-treatment of recurrent varicose veins’ state, or remain in the ‘clinical recurrence’ state. In line with the modelling undertaken to inform the NICE guideline, it was assumed that 75 per cent of patients experiencing a clinical recurrence would receive further treatment (42 per cent UGFS, 12 per cent surgery, 46 per cent EVLA), and before this treatment they would consult their general practitioner twice and attend a vascular surgery outpatient appointment2. The initial treatment modality was assumed not to influence the method of treatment for clinical recurrence. For patients experiencing a second recurrence, a simplifying assumption was made that these patients would not proceed to further treatment. Modelled future costs and QALYs were discounted at the recommended rate of 3·5 per cent per annum15. Fig. 1 Open in new tabDownload slide Diagram of model structure. p, Probability; #, complement of p; ACM, all-cause mortality; Tx, treatment; pTreat_recurrence, probability of proceeding for further treatment following a recurrence Table 2 Post-treatment recurrence risks applied in the analysis . Monthly probability . Estimated s.e.m. . Distribution . Source . Surgery 0·008818  0·00306  Beta NICE, 20132 UGFS 0·0115    0·009929 Beta NICE, 20132 EVLA 0·006532  0·003448 Beta NICE, 20132 . Monthly probability . Estimated s.e.m. . Distribution . Source . Surgery 0·008818  0·00306  Beta NICE, 20132 UGFS 0·0115    0·009929 Beta NICE, 20132 EVLA 0·006532  0·003448 Beta NICE, 20132 Monthly probabilities were transformed into 6-month probabilities. NICE, National Institute for Health and Care Excellence; UGFS, ultrasound-guided foam sclerotherapy; EVLA, endovenous laser ablation. Open in new tab Table 2 Post-treatment recurrence risks applied in the analysis . Monthly probability . Estimated s.e.m. . Distribution . Source . Surgery 0·008818  0·00306  Beta NICE, 20132 UGFS 0·0115    0·009929 Beta NICE, 20132 EVLA 0·006532  0·003448 Beta NICE, 20132 . Monthly probability . Estimated s.e.m. . Distribution . Source . Surgery 0·008818  0·00306  Beta NICE, 20132 UGFS 0·0115    0·009929 Beta NICE, 20132 EVLA 0·006532  0·003448 Beta NICE, 20132 Monthly probabilities were transformed into 6-month probabilities. NICE, National Institute for Health and Care Excellence; UGFS, ultrasound-guided foam sclerotherapy; EVLA, endovenous laser ablation. Open in new tab Model-based analysis To characterize uncertainty, each model input was assigned a distribution appropriate to the nature of the variable, and the model was analysed probabilistically. For treatment cost and utility parameters, distributions were derived from the analysis of patient-level data. Beta distributions were applied to the probabilities of clinical recurrence based on reported mean values and standard errors2. The model was estimated 10 000 times, each time sampling a value for each parameter from its assigned distribution. The mean costs and effects from these 10 000 iterations are reported for each treatment option. To help identify the optimal approach to treatment, the net monetary benefit (NMB) framework was used, where the NMB for a given strategy is equal to the accrued QALYs multiplied by the ceiling ratio (Rc) of willingness to pay (WTP) per QALY, minus the strategy costs. NMB=QALYs×Rc–costs The value of £20 000, which is typically used by NICE to inform judgements on cost-effectiveness, was placed on Rc15. The probability of each strategy generating the greatest NMB at this value of Rc is reported. Extensive deterministic sensitivity analysis was undertaken to assess the robustness of the findings to various parameter and structural assumptions applied in the base-case analysis. This included basing model input parameters on the analysis of multiple imputed patient data, to assess the sensitivity of findings to missing data. To assess the sensitivity of findings to model structuring assumptions, an alternative model structure was developed using pretreatment and post-treatment health states defined by clinical severity, as assessed by the Venous Clinical Severity Score (VCSS)16. In this model changes in quality of life were driven solely by transitions across four clinical severity states defined using the VCSS (0, 1–3, 4–6, more than 6). The patient data were then used to estimate the effects of the alternative treatments on transitions between these states from baseline to 6 months after treatment. The defined states were found to correlate reasonably with participants' EQ-5D™ scores, which reflect self-reported general health status and were used for estimating QALYs within this model. Results The mean estimates of resource use, costs and EQ-5D™ values are summarized in Table S1 (supporting information). The cumulative mean time spent in treatment was considerably lower for participants randomized to UGFS than it was for EVLA or surgery. Based on estimates of treatment duration, associated staff costs, consumable and equipment costs, the estimated total treatment costs were £245, £737 and £916 for participants randomized to UGFS, EVLA and surgery respectively. Recovery time costs were highest for participants randomized to surgery. The cost of consumables was considerably higher in the EVLA group, on account of the high cost of the laser fibre (unit cost £256). Total NHS costs at 6 months, incorporating routine follow-up (following EVLA and UGFS) and further unplanned use of health services, were £453, £951 and £1113 for participants randomized to UGFS, EVLA and surgery respectively. When an additional cost of theatre use was included to reflect higher overhead costs, the corresponding costs increased to £465, £975 and £1367 respectively. The regression-based estimates of the mean differences in cost and utility outcomes between treatment modalities, which were applied in the Markov model, are summarized in Table 3. Using multiple imputation data, the mean cost savings associated with UGFS and EVLA compared with surgery decreased somewhat (Table 3). Table 3 Estimated EQ-5D™ utility values and 6-month costs (including initial treatment) applied in the Markov model . Complete case data . Multiple imputation data . Distributional form . EQ-5D ™ at baseline (across all treatment groups) QALYs at 6 months 0·790(0·009) – Beta Surgery 0·432(0·006)* 0·435(0·006)* Beta UGFS (increment)‡ −0·005(0·010)† −0·011(0·009)† Normal EVLA (increment)‡ 0·011(0·006)† 0·007(0·007)† Normal EQ-5D™ at 6 months Surgery 0·884(0·010)* 0·888(0·011)* Beta UGFS (increment)‡ −0·009(0·018)† −0·020(0·023)† Normal EVLA (increment)‡ 0·016(0·012)† 0·011(0·016)† Normal Costs at 6 months (£) Surgery 1110(80)* 1047(80)* Gamma UGFS (increment)‡ −655(85)† −579(75)† Normal EVLA (increment)‡ −160(83)† −98(79)† Normal . Complete case data . Multiple imputation data . Distributional form . EQ-5D ™ at baseline (across all treatment groups) QALYs at 6 months 0·790(0·009) – Beta Surgery 0·432(0·006)* 0·435(0·006)* Beta UGFS (increment)‡ −0·005(0·010)† −0·011(0·009)† Normal EVLA (increment)‡ 0·011(0·006)† 0·007(0·007)† Normal EQ-5D™ at 6 months Surgery 0·884(0·010)* 0·888(0·011)* Beta UGFS (increment)‡ −0·009(0·018)† −0·020(0·023)† Normal EVLA (increment)‡ 0·016(0·012)† 0·011(0·016)† Normal Costs at 6 months (£) Surgery 1110(80)* 1047(80)* Gamma UGFS (increment)‡ −655(85)† −579(75)† Normal EVLA (increment)‡ −160(83)† −98(79)† Normal Values are mean(s.e.); * cluster bootstrapped standard error; † cluster robust standard error. ‡ Estimated 6-month utility and cost increments associated with ultrasound-guided foam sclerotherapy (UGFS) and endovenous laser ablation (EVLA) compared with surgery. Cost and utility estimates were derived using linear regression adjusting for baseline EQ-5D™ score and minimization variables used in the randomization process (age (less than 50 years, 50 years or older), sex, great saphenous vein or small saphenous vein reflux, and unilateral or bilateral veins). Open in new tab Table 3 Estimated EQ-5D™ utility values and 6-month costs (including initial treatment) applied in the Markov model . Complete case data . Multiple imputation data . Distributional form . EQ-5D ™ at baseline (across all treatment groups) QALYs at 6 months 0·790(0·009) – Beta Surgery 0·432(0·006)* 0·435(0·006)* Beta UGFS (increment)‡ −0·005(0·010)† −0·011(0·009)† Normal EVLA (increment)‡ 0·011(0·006)† 0·007(0·007)† Normal EQ-5D™ at 6 months Surgery 0·884(0·010)* 0·888(0·011)* Beta UGFS (increment)‡ −0·009(0·018)† −0·020(0·023)† Normal EVLA (increment)‡ 0·016(0·012)† 0·011(0·016)† Normal Costs at 6 months (£) Surgery 1110(80)* 1047(80)* Gamma UGFS (increment)‡ −655(85)† −579(75)† Normal EVLA (increment)‡ −160(83)† −98(79)† Normal . Complete case data . Multiple imputation data . Distributional form . EQ-5D ™ at baseline (across all treatment groups) QALYs at 6 months 0·790(0·009) – Beta Surgery 0·432(0·006)* 0·435(0·006)* Beta UGFS (increment)‡ −0·005(0·010)† −0·011(0·009)† Normal EVLA (increment)‡ 0·011(0·006)† 0·007(0·007)† Normal EQ-5D™ at 6 months Surgery 0·884(0·010)* 0·888(0·011)* Beta UGFS (increment)‡ −0·009(0·018)† −0·020(0·023)† Normal EVLA (increment)‡ 0·016(0·012)† 0·011(0·016)† Normal Costs at 6 months (£) Surgery 1110(80)* 1047(80)* Gamma UGFS (increment)‡ −655(85)† −579(75)† Normal EVLA (increment)‡ −160(83)† −98(79)† Normal Values are mean(s.e.); * cluster bootstrapped standard error; † cluster robust standard error. ‡ Estimated 6-month utility and cost increments associated with ultrasound-guided foam sclerotherapy (UGFS) and endovenous laser ablation (EVLA) compared with surgery. Cost and utility estimates were derived using linear regression adjusting for baseline EQ-5D™ score and minimization variables used in the randomization process (age (less than 50 years, 50 years or older), sex, great saphenous vein or small saphenous vein reflux, and unilateral or bilateral veins). Open in new tab Model-based incremental cost-effectiveness analysis The model-based estimates of mean costs and QALYs at 5 years are presented in Table 4. EVLA was associated with increased costs and QALYs in comparison with UGFS, with an incremental cost per QALY gained of £3640. Surgery was associated with increased costs compared with EVLA (£206), and on average produced slightly fewer QALYs (−0·078). This was driven by the slightly lower number of QALYs observed for surgery at 6 months compared with EVLA, and also the slightly higher clinical recurrence rate applied to surgery compared with EVLA. Applying a ceiling WTP ratio of £20 000 per QALY gained, EVLA had the highest probability (78·7 per cent) of being cost-effective. Table 4 Model-based incremental cost-effectiveness analysis (over a 5-year time horizon) . Cost (£) . Incremental cost (£) . QALYs . Incremental QALYs . Incremental cost per QALY (£) . Probability cost-effective at Rc £20 000 . UGFS 664 – 4·000 – – 0·168 EVLA 1095 431 4·119 0·118 3640 0·787 Surgery 1300 206 4·040 −0·078 Dominated 0·045 . Cost (£) . Incremental cost (£) . QALYs . Incremental QALYs . Incremental cost per QALY (£) . Probability cost-effective at Rc £20 000 . UGFS 664 – 4·000 – – 0·168 EVLA 1095 431 4·119 0·118 3640 0·787 Surgery 1300 206 4·040 −0·078 Dominated 0·045 QALY, quality-adjusted life-year; Rc, ceiling ratio of willingness to pay per QALY gained; UGFS, ultrasound-guided foam sclerotherapy; EVLA, endovenous laser ablation. Open in new tab Table 4 Model-based incremental cost-effectiveness analysis (over a 5-year time horizon) . Cost (£) . Incremental cost (£) . QALYs . Incremental QALYs . Incremental cost per QALY (£) . Probability cost-effective at Rc £20 000 . UGFS 664 – 4·000 – – 0·168 EVLA 1095 431 4·119 0·118 3640 0·787 Surgery 1300 206 4·040 −0·078 Dominated 0·045 . Cost (£) . Incremental cost (£) . QALYs . Incremental QALYs . Incremental cost per QALY (£) . Probability cost-effective at Rc £20 000 . UGFS 664 – 4·000 – – 0·168 EVLA 1095 431 4·119 0·118 3640 0·787 Surgery 1300 206 4·040 −0·078 Dominated 0·045 QALY, quality-adjusted life-year; Rc, ceiling ratio of willingness to pay per QALY gained; UGFS, ultrasound-guided foam sclerotherapy; EVLA, endovenous laser ablation. Open in new tab The proportion of probabilistic iterations favouring each of the alternative strategies by increasing levels of WTP per QALY gained is shown in Fig. 2. This figure shows that, as WTP increases beyond £30 000, EVLA has approximately an 80 per cent chance of being the optimal strategy on grounds of cost-effectiveness. Fig. 2 Open in new tabDownload slide Cost-effectiveness acceptability curve based on 5-year time horizon. UGFS, ultrasound-guided foam sclerotherapy; EVLA, endovenous laser ablation; QALY, quality-adjusted life-year Sensitivity analysis The results of various analyses assessing the sensitivity of results to uncertainty surrounding several key input parameters and modelling assumptions are shown in Table S2 (supporting information). In general, the findings were robust to changes, including the cost of follow-up assessment in patients requiring further foam sclerotherapy treatment following EVLA and UGFS, the use of multiple imputation for missing data, and the inclusion of the annuitized cost of a laser generator. The application of additional theatre overhead costs, and the removal of all follow-up clinical assessment costs, increased the estimated cost savings associated with EVLA compared with surgery to £439 and £332 respectively. It was also noted that in some centres EVLA was carried out with a greater number of staff present than for UGFS. When all EVLA procedures were costed using the same nursing staffing profile as for UGFS, the cost saving with EVLA compared with surgery increased to £368. When applying the assumption of equivalent QALYs between the strategies to 6 months, and equivalent EQ-5D™ values at 6 months (rather than the treatment-specific mean estimates), UGFS had the highest probability of being cost-effective at 5 years (scenario 13; Table S2, supporting information). However, the base-case ordering of treatment modalities was restored when the time horizon for this analysis was extended to 10 years (scenario 14; Table S2, supporting information). The analysis of the alternative model structure, based on clinical severity states defined by the VCSS, predicted higher QALYs with EVLA than UGFS, and slightly higher QALYs with surgery than EVLA (scenario 22; Table S2, supporting information). However, the probability of surgery being cost-effective remained lower than that for EVLA at the Rc of £20 000 per QALY. Discussion The results of this study suggest that, over a 5-year time horizon, EVLA is likely to be the preferred option on grounds of cost-effectiveness, costing £3640 per QALY gained in comparison with UGFS, and costing less and producing slightly more QALYs than surgery. EVLA had approximately an 80 per cent chance of being cost-effective at a ceiling WTP ratio of £20 000 per QALY. These findings were generally found to be robust to uncertainty surrounding various model inputs and assumptions, including multiple imputation of missing data, and alternative model structuring assumptions. The base-case estimated reduction in cost (£206) associated with EVLA compared with surgery is likely to be a conservative estimate that does not fully capture the increased overhead costs associated with day-case admission and use of theatre for surgery. When a separate estimate of theatre overhead costs (per hour) was factored into the analysis, the cost saving associated with EVLA increased to £439. This additional overhead cost was omitted from the base-case analysis because it was available for only one recruiting centre and it may also double-count some of the overhead costs included in the staff time unit cost multipliers. However, even when this cost was excluded, sensitivity analysis indicated that the cost savings associated with EVLA would increase to £368 if performed using a similar staff profile to UGFS, and that further savings could also be realized if patients with a good outcome following initial EVLA treatment were not followed up at 6 weeks in routine practice. Strengths of the study include that the model was populated, where possible, using estimates of the mean differences in costs and effects derived from the analysis of prospectively collected individual patient data on resource use. Therefore, the results are internally valid and generalizable across settings in the UK. In the absence of data on long-term clinical recurrence for the CLASS cohort, the risk of clinical recurrence was modelled using data derived from an existing network meta-analysis of trials reporting clinical recurrence of varicose veins as an outcome2. Although this provides the best current source of evidence on recurrence, the quality of included trials varied and follow-up generally extended to only 12 or 24 months. Furthermore, clinical recurrence was not always well defined and varied between studies. Uncertainty remains regarding the applicability of these recurrence rates to patients in the CLASS trial (and more generally), and also regarding the risks of recurrence beyond 24 months. This underlines the importance of collecting further data on clinical recurrence and its impact on generic health-related quality of life via the extended follow-up of CLASS participants, which is currently in progress. In general, the approach used in this paper is consistent with that used in previous economic modelling studies of treatments for varicose veins. The structure and assumptions applied in the primary analysis are in line with those used in the model developed to inform the recent NICE clinical guideline2, which also found that endothermal treatment (EVLA or RFA) had the highest probability of being cost-effective under most scenarios assessed. However, the cost and utility inputs applied in the present model were based on prospectively collected data from randomized patients. Consequently, based on a bottom-up costing approach, a somewhat narrower difference in cost was estimated between EVLA and surgery compared with the estimates used in the NICE guideline model. This was due to differences in estimated mean treatment durations, and the reported staffing profiles for the different procedures, compared with those used in the NICE model. The present estimates are more likely to be representative of current practice because they were derived using time and staffing data collected prospectively alongside the CLASS trial. Furthermore, rather than assuming equivalent utility values following the different treatment modalities at 6 months (as was done in the NICE guideline model), the estimated 6-month mean incremental differences between UGFS and surgery, and between EVLA and surgery, were applied in the present study. Gohel and colleagues3 similarly developed a Markov model to assess the cost-effectiveness of conservative management, UGFS, EVLA, RFA and surgery. Based on their model, they estimated that EVLA carried out under local anaesthetic would have the highest chance of being cost-effective at the Rc of £20 000 per QALY. In contrast to the findings here, they reported that surgery (on a day-case basis) had the next highest probability of being cost-effective, followed by RFA and UGFS3. The present findings appear inconsistent with those of a recently conducted modelling study which concluded that UGFS would be more cost-effective than surgery, EVLA and RFA, based on it having the lowest cost while generating similar clinical outcomes. In this modelling study, Carroll and colleagues4 used data from two single-centre UK trials17,18 to assess the relative cost differences between UGFS, RFA, EVLA and surgery; the cost of EVLA was estimated to be 2·02 times that of surgery. However, this multiplier was derived from cost data collected for 88 patients randomized between surgery and RFA17, and in this particular study RFA was performed under general anaesthesia and took 30 min longer than surgery. The staff and theatre costs were thus higher for RFA than for surgery. By contrast, in CLASS, all EVLA procedures were performed under local anaesthetic; the majority (90·5 per cent) were carried out in a clinic treatment room rather than an operating theatre and the measured procedure time was lower for EVLA than for surgery. This information was collected prospectively for patients recruited to CLASS from multiple treatment centres, and so should be generalizable to standard NHS practice. This explains the lower estimate for the cost of EVLA compared with surgery, and also helps explain why EVLA was more cost-effective than UGFS. The present modelling study suggests that, for patients with primary varicose veins in whom all three of the investigated treatment modalities offer a clinically viable option, EVLA performed under local anaesthetic has the highest probability of being cost-effective at accepted thresholds of WTP per QALY. This finding is consistent with the results of recent modelling undertaken to inform the NICE clinical guideline on the management of varicose veins. In the interests of cost-effectiveness, EVLA should be performed where possible in a clinic treatment room and the levels of nursing support required should be considered carefully. Acknowledgements The authors thank the following contributors to the CLASS project for ongoing advice and support for the trial: J. Cruden, T. Davidson, J. Francis, A. McDonald, G. McPherson, S. Wileman, I. Chetter, J. J. Earnshaw, T. Lees, J. Scott, S. Baker, G. MacLennan, M. Prior and D. Bolsover. In addition, they thank J. Wallace, E. Wilson, S. Hardy, K. Hargreaves, D. Rittoo, S. Baker, L. Park, D. Parker, I. Chetter, J. Hatfield, J. Scott, N. Dewhirst, T. Lees, N. Parr, D. Dodd, J. Sorrell, K. R. Makhdoomi, B. Rorison, I. Nyamekye and P. Matheson for information on resource use and costs for treatment modalities at participating centres. This work was carried out as part of a project funded by the National Institute for Health Research (NIHR) Health Technology Assessment (HTA) Programme (project number 06/45/02). The views and opinions expressed herein are those of the authors and do not necessarily reflect those of the NIHR HTA Programme, NIHR, NHS or Department of Health. Disclosure The authors declare no conflict of interest. References 1 Health and Social Care Information Centre . Hospital Episode Statistics . http://www.hesonline.nhs.uk/Ease/servlet/ContentServer?site=1937&categoryID=192 [accessed 1 August 2013 ]. 2 National Institute for Care and Health Excellence . Varicose Veins in the Legs (CG168) . http://guidance.nice.org.uk/CG168 [accessed 1 August 2013 ]. 3 Gohel MS , Epstein DM, Davies AH. Cost-effectiveness of traditional and endovenous treatments for varicose veins . Br J Surg 2010 ; 97 : 1815 – 1823 . Google Scholar Crossref Search ADS PubMed WorldCat 4 Carroll C , Hummel S, Leaviss J, Ren S, Stevens JW, Everson-Hock E et al. Clinical effectiveness and cost-effectiveness of minimally invasive techniques to manage varicose veins: a systematic review and economic evaluation . Health Technol Assess 2013 ; 17 : i–xvi , 1 – 141 . Google Scholar OpenURL Placeholder Text WorldCat 5 National Institute for Health Research . The CLASS Trial: Comparison of LAser, Surgery and Foam Sclerotherapy. Research Protocol . http://www.nets.nihr.ac.uk/__data/assets/pdf_file/0004/51448/PRO-06-45-02.pdf [accessed 1 January 2014 ]. 6 Breu FX , Guggenbichler S. European Consensus Meeting on Foam Sclerotherapy, April, 4–6, 2003, Tegernsee, Germany . Dermatol Surg 2004 ; 30 : 709 – 717 . Google Scholar Crossref Search ADS PubMed WorldCat 7 Personal Social Services Research Unit (PSSRU) . Unit Costs of Health and Social Care . http://www.pssru.ac.uk/pdf/uc/uc2011/uc2011.pdf [accessed 1 October 2012 ]. 8 GOV.UK . 2010–11 Reference Costs Publication . https://www.gov.uk/government/publications/2010-11-reference-costs-publication [accessed 1 October 2012 ]. 9 Information Services Division (ISD) Scotland . Theatres. Costs – Detailed Tables. R141X: Average Theatre Tunning Costs, by Board; SFR 5.10: Source Data for Theatre Services . http://www.isdscotland.org/Health-Topics/Finance/Costs/Detailed-Tables/Theatres.asp [accessed 1 October 2012 ]. 10 Dolan P , Gudex C, Kind P, Williams A. The time trade-off method: results from a general population study . Health Economics 1996 ; 5 : 141 – 154 . Google Scholar Crossref Search ADS PubMed WorldCat 11 StataCorp . Stata Multiple-Imputation Reference Manual Release 12 . StataCorp LP : College Station , 2012 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 12 Glick H , Doshi J, Sonnad S, Polsky D. Economic Evaluation in Clinical Trials (1st edn). Oxford University Press : New York , 2007 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 13 Royston P Flexible alternatives to the Cox model, and more . Stata J 2001 ; 1 : 1 – 28 . Google Scholar Crossref Search ADS WorldCat 14 Software TreeAge . TreeAge Pro 2013, R2.0 Released . http://www.treeage.com/news/treeage-pro-2013-r2-0-released/ [accessed 1 August 2013 ]. 15 National Institute for Health and Care Excellence . Guide to the Methods of Technology Appraisal 2013 . http://www.nice.org.uk/media/D45/1E/GuideToMethodsTechnologyAppraisal2013.pdf [accessed 1 June 2013 ]. 16 Vasquez MA , Rabe E, McLafferty RB, Shortell CK, Marston WA, Gillespie D et al. ; American Venous Forum Ad Hoc Outcomes Working Group. Revision of the venous clinical severity score: venous outcomes consensus statement: special communication of the American Venous Forum Ad Hoc Outcomes Working Group . J Vasc Surg 2010 ; 52 : 1387 – 1396 . Google Scholar Crossref Search ADS PubMed WorldCat 17 Brittenden J , Cotton SC, Elders A, Tassie E, Scotland G, Ramsay CR et al. Randomised controlled trial comparing foam sclerotherapy, alone or in combination with endovenous laser therapy, with conventional surgery as a treatment for varicose veins . Health Technol Assess 2014 ; (in press). Google Scholar OpenURL Placeholder Text WorldCat 18 Subramonia S , Lees T. Radiofrequency ablation vs conventional surgery for varicose veins – a comparison of treatment costs in a randomised trial . Eur J Vasc Endovasc Surg 2010 ; 39 : 104 – 111 . Google Scholar Crossref Search ADS PubMed WorldCat © 2014 BJS Society Ltd. Published by John Wiley & Sons Ltd This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) © 2014 BJS Society Ltd. Published by John Wiley & Sons Ltd http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png British Journal of Surgery Oxford University Press

Cost-effectiveness of ultrasound-guided foam sclerotherapy, endovenous laser ablation or surgery as treatment for primary varicose veins from the randomized CLASS trial

Download PDF
10 pages

Loading next page...
 
/lp/oxford-university-press/cost-effectiveness-of-ultrasound-guided-foam-sclerotherapy-endovenous-B1S8doGff5

References (33)

Publisher
Oxford University Press
Copyright
Copyright © 2022 BJS Society Ltd.
ISSN
0007-1323
eISSN
1365-2168
DOI
10.1002/bjs.9595
Publisher site
See Article on Publisher Site

Abstract

Abstract Background The treatment of patients with varicose veins constitutes a considerable workload and financial burden to the National Health Service. This study aimed to assess the cost-effectiveness of ultrasound-guided foam sclerotherapy (UGFS) and endovenous laser ablation (EVLA) compared with conventional surgery as treatment for primary varicose veins. Methods Participant cost and utility data were collected alongside the UK CLASS multicentre randomized clinical trial, which compared EVLA, surgery and UGFS. Regression methods were used to estimate the effects of the alternative treatments on costs to the health service and quality-adjusted life-years (QALYs) at 6 months. A Markov model, incorporating available evidence on clinical recurrence rates, was developed to extrapolate the trial data over a 5-year time horizon. Results Compared with surgery at 6 months, UGFS and EVLA reduced mean costs to the health service by £655 and £160 respectively. When additional overhead costs associated with theatre use were included, these cost savings increased to £902 and £392 respectively. UGFS produced 0·005 fewer QALYs, whereas EVLA produced 0·011 additional QALYs. Extrapolating to 5 years, EVLA was associated with increased costs and QALYs compared with UGFS (costing £3640 per QALY gained), and generated a cost saving (£206–439) and QALY gain (0·078) compared with surgery. Applying a ceiling willingness-to-pay ratio of £20 000 per QALY gained, EVLA had the highest probability (78·7 per cent) of being cost-effective. Conclusion The results suggest, for patients considered eligible for all three treatment options, that EVLA has the highest probability of being cost-effective at accepted thresholds of willingness to pay per QALY. Introduction The treatment of patients with varicose veins results in a considerable workload and financial burden to the National Health Service (NHS)1. Ultrasound-guided foam sclerotherapy (UGFS) and endovenous laser ablation (EVLA) are now commonly used alternatives to surgery. Despite their widespread adoption, uncertainty exists regarding their cost-effectiveness and about which method offers the best value for money. To date there have been three published cost–utility analyses2–4 comparing these treatment modalities from a UK NHS perspective, based on estimated treatment costs and outcomes, follow-up costs and 5-year recurrence rates. These have revealed differing results. One study2 found EVLA and radiofrequency ablation (RFA) to be the most cost-effective, another3 reported day-case EVLA to be cost-effective, whereas a third study4 found UGFS to be the most cost-effective. CLASS (Comparison of LAser, Surgery and foam Sclerotherapy) is a multicentre randomized clinical trial5 that compared the clinical and cost-effectiveness of three treatments: UGFS, EVLA and surgery. Participant-level cost and effect data were collected at 6 weeks and 6 months. A Markov decision model was developed to extrapolate the trial cost-effectiveness data over a 5-year time horizon. Results of these analyses are presented here. Methods A total of 798 participants were recruited from 11 centres in the UK between November 2008 and October 2012. All three treatment options were available in eight centres, whereas three centres offered only UGFS and surgery. Participants were randomized between the treatments with even allocation. To conduct a valid three-way comparison on costs and outcomes, as required for cost-effectiveness analysis, this article reports on the eight centres at which all three treatment options were available. The allocation used a minimization algorithm that included centre, age (less 50 years, 50 years or older), sex, great saphenous vein (GSV) or small saphenous vein (SSV) reflux, and unilateral or bilateral veins. Inclusion criteria were5: age over 18 years; primary unilateral or bilateral symptomatic varicose veins (Clinical Etiologic Anatomic Pathophysiologic (CEAP) class C2 or above); GSV and/or SSV involvement; and reflux duration more 1 s on duplex ultrasound imaging. Exclusion criteria comprised: current deep vein thrombosis; acute superficial vein thrombosis; a GSV or SSV diameter smaller than 3 mm or greater than 15 mm; tortuous veins considered unsuitable for EVLA; and contraindications to UGFS, or to general/regional anaesthesia that would be required for surgery. Further details of the trial have been published elsewhere5. Treatments Surgery was performed under general anaesthetic. It consisted of proximal GSV/SSV ligation and stripping (all GSV) and concurrent phlebectomies. EVLA of truncal veins was performed under local anaesthetic in either an operating theatre (9·5 per cent of patients) or treatment room (90·5 per cent). UGFS was administered to residual varicosities at 6-week follow-up if required, with the exception of one centre that performed concurrent phlebectomies. Foam was produced using the Tessari technique6 at a ratio of 0·5 ml sodium tetradecyl sulphate (STS) to 1·5 ml air (3 per cent STS for GSV/SSV truncal veins, 1 per cent for varicosities, maximum 12 ml foam per session). Resource use and costs Patient-level resource use data were collected for each treatment modality and entered on case report forms. Details of the following resource items were collected: location and time of treatment and recovery, and grade of surgeon and anaesthetist. A separate survey of participating centres collected information on nursing and allied health professionals present for each procedure, and equipment and consumables used. All resource inputs were costed in 2010–2011 unit prices (Table 1). The time of entering and time of leaving the operating theatre or treatment room was used to measure the total time requirement of staff present for the procedure. Staffing information and treatment/recovery duration times were combined with national unit cost data (Personal Social Services Research Unit, 2011)7 to estimate the total cost of staff time. Although the staff cost multipliers included an allocation for overheads and use of hospital space, these may not adequately capture the additional overhead costs associated with procedures done in theatre under general anaesthetic. Therefore, a secondary analysis was conducted whereby an estimate of theatre overhead cost (£218 per hour)9 was applied to procedures carried out in this location. Capital equipment costs were amortized over the expected useful lifespan of the item, and allocated on a per-patient basis using estimated annual clinical throughput obtained from sites. The majority of centres in CLASS reported receiving the laser generator for EVLA on loan, but paying for laser fibres. Therefore, the generator costs were excluded from the base-case analysis, but were included in a sensitivity analysis. A list of consumables required for each procedure was established and costed using unit prices obtained from participating centres. Table 1 Unit costs applied in the analysis Input variables . Unit cost (£) . Source . Nursing staff (cost per h) Band 2 20 PSSRU, 20117 Band 3 24 PSSRU, 20117 Band 4 30 PSSRU, 20117 Band 5 82 PSSRU, 20117 Band 6 107 PSSRU, 20117 Band 7 129 PSSRU, 20117 Band 8b (consultant nurse) 147 PSSRU, 20117 Medical staff (cost per h) Foundation Year 1 33 PSSRU, 20117 Foundation Year 2 core trainee, CCT 42 PSSRU, 20117 Specialty trainee 59 PSSRU, 20117 Staff grade 95 PSSRU, 20117 Associate specialist 131 PSSRU, 20117 Consultant medical 136 PSSRU, 20117 Equipment (cost per use) Ultrasound machine 8·78 Participant centres ECG, pulse oximeter, blood pressure monitor 4·15 Participant centres Laser generator* 34·67 Participant centres Consumables Surgery 159·56 Participant centres UGFS 26·23 Participant centres EVLA 321·06 Participant centres Follow-up costs GP visit 36 PSSRU, 20117 Practice nurse visit 12 PSSRU, 20117 Vascular surgery follow-up outpatient 123 Department of Health, 20118 appointment Duplex imaging 53 Department of Health, 20118 Input variables . Unit cost (£) . Source . Nursing staff (cost per h) Band 2 20 PSSRU, 20117 Band 3 24 PSSRU, 20117 Band 4 30 PSSRU, 20117 Band 5 82 PSSRU, 20117 Band 6 107 PSSRU, 20117 Band 7 129 PSSRU, 20117 Band 8b (consultant nurse) 147 PSSRU, 20117 Medical staff (cost per h) Foundation Year 1 33 PSSRU, 20117 Foundation Year 2 core trainee, CCT 42 PSSRU, 20117 Specialty trainee 59 PSSRU, 20117 Staff grade 95 PSSRU, 20117 Associate specialist 131 PSSRU, 20117 Consultant medical 136 PSSRU, 20117 Equipment (cost per use) Ultrasound machine 8·78 Participant centres ECG, pulse oximeter, blood pressure monitor 4·15 Participant centres Laser generator* 34·67 Participant centres Consumables Surgery 159·56 Participant centres UGFS 26·23 Participant centres EVLA 321·06 Participant centres Follow-up costs GP visit 36 PSSRU, 20117 Practice nurse visit 12 PSSRU, 20117 Vascular surgery follow-up outpatient 123 Department of Health, 20118 appointment Duplex imaging 53 Department of Health, 20118 * The laser generator cost was included only in a sensitivity analysis (see text). PSSRU, Personal Social Services Research Unit; CCT, Certificate of Completion of Training; ECG, electrocardiograph; UGFS, ultrasound-guided foam sclerotherapy; EVLA, endovenous laser ablation; GP, general practitioner. Open in new tab Table 1 Unit costs applied in the analysis Input variables . Unit cost (£) . Source . Nursing staff (cost per h) Band 2 20 PSSRU, 20117 Band 3 24 PSSRU, 20117 Band 4 30 PSSRU, 20117 Band 5 82 PSSRU, 20117 Band 6 107 PSSRU, 20117 Band 7 129 PSSRU, 20117 Band 8b (consultant nurse) 147 PSSRU, 20117 Medical staff (cost per h) Foundation Year 1 33 PSSRU, 20117 Foundation Year 2 core trainee, CCT 42 PSSRU, 20117 Specialty trainee 59 PSSRU, 20117 Staff grade 95 PSSRU, 20117 Associate specialist 131 PSSRU, 20117 Consultant medical 136 PSSRU, 20117 Equipment (cost per use) Ultrasound machine 8·78 Participant centres ECG, pulse oximeter, blood pressure monitor 4·15 Participant centres Laser generator* 34·67 Participant centres Consumables Surgery 159·56 Participant centres UGFS 26·23 Participant centres EVLA 321·06 Participant centres Follow-up costs GP visit 36 PSSRU, 20117 Practice nurse visit 12 PSSRU, 20117 Vascular surgery follow-up outpatient 123 Department of Health, 20118 appointment Duplex imaging 53 Department of Health, 20118 Input variables . Unit cost (£) . Source . Nursing staff (cost per h) Band 2 20 PSSRU, 20117 Band 3 24 PSSRU, 20117 Band 4 30 PSSRU, 20117 Band 5 82 PSSRU, 20117 Band 6 107 PSSRU, 20117 Band 7 129 PSSRU, 20117 Band 8b (consultant nurse) 147 PSSRU, 20117 Medical staff (cost per h) Foundation Year 1 33 PSSRU, 20117 Foundation Year 2 core trainee, CCT 42 PSSRU, 20117 Specialty trainee 59 PSSRU, 20117 Staff grade 95 PSSRU, 20117 Associate specialist 131 PSSRU, 20117 Consultant medical 136 PSSRU, 20117 Equipment (cost per use) Ultrasound machine 8·78 Participant centres ECG, pulse oximeter, blood pressure monitor 4·15 Participant centres Laser generator* 34·67 Participant centres Consumables Surgery 159·56 Participant centres UGFS 26·23 Participant centres EVLA 321·06 Participant centres Follow-up costs GP visit 36 PSSRU, 20117 Practice nurse visit 12 PSSRU, 20117 Vascular surgery follow-up outpatient 123 Department of Health, 20118 appointment Duplex imaging 53 Department of Health, 20118 * The laser generator cost was included only in a sensitivity analysis (see text). PSSRU, Personal Social Services Research Unit; CCT, Certificate of Completion of Training; ECG, electrocardiograph; UGFS, ultrasound-guided foam sclerotherapy; EVLA, endovenous laser ablation; GP, general practitioner. Open in new tab Details of secondary healthcare services use after treatment (hospital admissions and outpatient attendances) were collected at 6 weeks and 6 months. Use of primary care services was collected from a 6-month patient-completed questionnaire. All CLASS participants were invited to attend a 6-week and 6-month assessment appointment, but only the 6-week assessment following UGFS or EVLA (to assess the need for further foam treatment) was considered potentially consistent with routine practice. Therefore, the cost of an outpatient appointment plus ultrasound scan was incorporated for patients attending this appointment following UGFS or EVLA. For patients in the EVLA or UGFS groups who underwent additional foam treatment at 6 weeks, the cost of the treatment, rather than the cost of follow-up assessment, was applied. The impact of applying the cost of an assessment visit before receiving top-up foam treatment was also assessed, as was the impact of removing all post-treatment assessment costs. National unit cost data were used to cost secondary and primary care contacts (Table 1)7,8. The total cost to the health service (to 6 months) was calculated for each participant by summing their treatment and follow-up costs. Health outcomes The EQ-5D™ 3 level (EQ-5D-3 L) (EuroQol Group, Rotterdam, The Netherlands) was completed by participants at baseline, 6 weeks and 6 months. A preference-based utility score was derived for each participant's response to the EQ-5D™ using the UK population time trade-off tariff10. These scores represent the desirability of the different heath states defined by the EQ-5D™ on a scale where 0 represents death and 1 denotes full health. Quality-adjusted life-years (QALYs) can be derived from responses to the EQ-5D™ by multiplying the time (in years) spent in different states of health by the utility score associated with each state. The QALYs accrued to each participant at 6 months were calculated assuming that the change in utility between measures at adjacent time points followed a straight line. Analysis of participant-level cost and outcome data Data were summarized and analysed by intention to treat using Stata™ version 12.111. The mean incremental costs and QALYs associated with EVLA and UGFS versus surgery were estimated using generalized linear regression, with adjustment for minimization variables and baseline values, as appropriate. The method of recycled predictions12 was used to recover the estimated mean cost and 6-month utility estimates for surgery, and the mean utility and cost increments associated with UGFS and EVLA were used to recover the expected utilities and costs for patients receiving these treatments. This analysis was also repeated using a multiple imputation data set generated using chained equations13. Modelling of longer-term cost-effectiveness The cost and effect estimates from the trial data were used to populate the Markov model, developed in TreeAge Pro14 to extrapolate cost-effectiveness over a 5-year time horizon. This is the duration to which the CLASS study follow-up will ultimately be extended, providing a validity check on initial model-based estimates of cost-effectiveness. The model was constructed to simulate transitions between discrete health states over 6 months (Markov cycle), for a cohort of patients with mean age and sex matching that of trial participants (Fig. 1). For the first 6-month cycle, the model was populated using the estimated mean cost and utility data obtained from the analysis of individual patient data. Beyond 6 months, monthly probabilities of clinical recurrence, derived from a network meta-analysis undertaken to inform the recent National Institute for Health and Care Excellence (NICE) guideline2 (Table 2), were incorporated into the model. Clinical recurrence was assumed to result in a drop in health state utility to the mean baseline value observed across the treatment allocation groups in CLASS. Following clinical recurrence, patients could either present for further treatment (surgery, UGFS or EVLA), and transit to the ‘post-treatment of recurrent varicose veins’ state, or remain in the ‘clinical recurrence’ state. In line with the modelling undertaken to inform the NICE guideline, it was assumed that 75 per cent of patients experiencing a clinical recurrence would receive further treatment (42 per cent UGFS, 12 per cent surgery, 46 per cent EVLA), and before this treatment they would consult their general practitioner twice and attend a vascular surgery outpatient appointment2. The initial treatment modality was assumed not to influence the method of treatment for clinical recurrence. For patients experiencing a second recurrence, a simplifying assumption was made that these patients would not proceed to further treatment. Modelled future costs and QALYs were discounted at the recommended rate of 3·5 per cent per annum15. Fig. 1 Open in new tabDownload slide Diagram of model structure. p, Probability; #, complement of p; ACM, all-cause mortality; Tx, treatment; pTreat_recurrence, probability of proceeding for further treatment following a recurrence Table 2 Post-treatment recurrence risks applied in the analysis . Monthly probability . Estimated s.e.m. . Distribution . Source . Surgery 0·008818  0·00306  Beta NICE, 20132 UGFS 0·0115    0·009929 Beta NICE, 20132 EVLA 0·006532  0·003448 Beta NICE, 20132 . Monthly probability . Estimated s.e.m. . Distribution . Source . Surgery 0·008818  0·00306  Beta NICE, 20132 UGFS 0·0115    0·009929 Beta NICE, 20132 EVLA 0·006532  0·003448 Beta NICE, 20132 Monthly probabilities were transformed into 6-month probabilities. NICE, National Institute for Health and Care Excellence; UGFS, ultrasound-guided foam sclerotherapy; EVLA, endovenous laser ablation. Open in new tab Table 2 Post-treatment recurrence risks applied in the analysis . Monthly probability . Estimated s.e.m. . Distribution . Source . Surgery 0·008818  0·00306  Beta NICE, 20132 UGFS 0·0115    0·009929 Beta NICE, 20132 EVLA 0·006532  0·003448 Beta NICE, 20132 . Monthly probability . Estimated s.e.m. . Distribution . Source . Surgery 0·008818  0·00306  Beta NICE, 20132 UGFS 0·0115    0·009929 Beta NICE, 20132 EVLA 0·006532  0·003448 Beta NICE, 20132 Monthly probabilities were transformed into 6-month probabilities. NICE, National Institute for Health and Care Excellence; UGFS, ultrasound-guided foam sclerotherapy; EVLA, endovenous laser ablation. Open in new tab Model-based analysis To characterize uncertainty, each model input was assigned a distribution appropriate to the nature of the variable, and the model was analysed probabilistically. For treatment cost and utility parameters, distributions were derived from the analysis of patient-level data. Beta distributions were applied to the probabilities of clinical recurrence based on reported mean values and standard errors2. The model was estimated 10 000 times, each time sampling a value for each parameter from its assigned distribution. The mean costs and effects from these 10 000 iterations are reported for each treatment option. To help identify the optimal approach to treatment, the net monetary benefit (NMB) framework was used, where the NMB for a given strategy is equal to the accrued QALYs multiplied by the ceiling ratio (Rc) of willingness to pay (WTP) per QALY, minus the strategy costs. NMB=QALYs×Rc–costs The value of £20 000, which is typically used by NICE to inform judgements on cost-effectiveness, was placed on Rc15. The probability of each strategy generating the greatest NMB at this value of Rc is reported. Extensive deterministic sensitivity analysis was undertaken to assess the robustness of the findings to various parameter and structural assumptions applied in the base-case analysis. This included basing model input parameters on the analysis of multiple imputed patient data, to assess the sensitivity of findings to missing data. To assess the sensitivity of findings to model structuring assumptions, an alternative model structure was developed using pretreatment and post-treatment health states defined by clinical severity, as assessed by the Venous Clinical Severity Score (VCSS)16. In this model changes in quality of life were driven solely by transitions across four clinical severity states defined using the VCSS (0, 1–3, 4–6, more than 6). The patient data were then used to estimate the effects of the alternative treatments on transitions between these states from baseline to 6 months after treatment. The defined states were found to correlate reasonably with participants' EQ-5D™ scores, which reflect self-reported general health status and were used for estimating QALYs within this model. Results The mean estimates of resource use, costs and EQ-5D™ values are summarized in Table S1 (supporting information). The cumulative mean time spent in treatment was considerably lower for participants randomized to UGFS than it was for EVLA or surgery. Based on estimates of treatment duration, associated staff costs, consumable and equipment costs, the estimated total treatment costs were £245, £737 and £916 for participants randomized to UGFS, EVLA and surgery respectively. Recovery time costs were highest for participants randomized to surgery. The cost of consumables was considerably higher in the EVLA group, on account of the high cost of the laser fibre (unit cost £256). Total NHS costs at 6 months, incorporating routine follow-up (following EVLA and UGFS) and further unplanned use of health services, were £453, £951 and £1113 for participants randomized to UGFS, EVLA and surgery respectively. When an additional cost of theatre use was included to reflect higher overhead costs, the corresponding costs increased to £465, £975 and £1367 respectively. The regression-based estimates of the mean differences in cost and utility outcomes between treatment modalities, which were applied in the Markov model, are summarized in Table 3. Using multiple imputation data, the mean cost savings associated with UGFS and EVLA compared with surgery decreased somewhat (Table 3). Table 3 Estimated EQ-5D™ utility values and 6-month costs (including initial treatment) applied in the Markov model . Complete case data . Multiple imputation data . Distributional form . EQ-5D ™ at baseline (across all treatment groups) QALYs at 6 months 0·790(0·009) – Beta Surgery 0·432(0·006)* 0·435(0·006)* Beta UGFS (increment)‡ −0·005(0·010)† −0·011(0·009)† Normal EVLA (increment)‡ 0·011(0·006)† 0·007(0·007)† Normal EQ-5D™ at 6 months Surgery 0·884(0·010)* 0·888(0·011)* Beta UGFS (increment)‡ −0·009(0·018)† −0·020(0·023)† Normal EVLA (increment)‡ 0·016(0·012)† 0·011(0·016)† Normal Costs at 6 months (£) Surgery 1110(80)* 1047(80)* Gamma UGFS (increment)‡ −655(85)† −579(75)† Normal EVLA (increment)‡ −160(83)† −98(79)† Normal . Complete case data . Multiple imputation data . Distributional form . EQ-5D ™ at baseline (across all treatment groups) QALYs at 6 months 0·790(0·009) – Beta Surgery 0·432(0·006)* 0·435(0·006)* Beta UGFS (increment)‡ −0·005(0·010)† −0·011(0·009)† Normal EVLA (increment)‡ 0·011(0·006)† 0·007(0·007)† Normal EQ-5D™ at 6 months Surgery 0·884(0·010)* 0·888(0·011)* Beta UGFS (increment)‡ −0·009(0·018)† −0·020(0·023)† Normal EVLA (increment)‡ 0·016(0·012)† 0·011(0·016)† Normal Costs at 6 months (£) Surgery 1110(80)* 1047(80)* Gamma UGFS (increment)‡ −655(85)† −579(75)† Normal EVLA (increment)‡ −160(83)† −98(79)† Normal Values are mean(s.e.); * cluster bootstrapped standard error; † cluster robust standard error. ‡ Estimated 6-month utility and cost increments associated with ultrasound-guided foam sclerotherapy (UGFS) and endovenous laser ablation (EVLA) compared with surgery. Cost and utility estimates were derived using linear regression adjusting for baseline EQ-5D™ score and minimization variables used in the randomization process (age (less than 50 years, 50 years or older), sex, great saphenous vein or small saphenous vein reflux, and unilateral or bilateral veins). Open in new tab Table 3 Estimated EQ-5D™ utility values and 6-month costs (including initial treatment) applied in the Markov model . Complete case data . Multiple imputation data . Distributional form . EQ-5D ™ at baseline (across all treatment groups) QALYs at 6 months 0·790(0·009) – Beta Surgery 0·432(0·006)* 0·435(0·006)* Beta UGFS (increment)‡ −0·005(0·010)† −0·011(0·009)† Normal EVLA (increment)‡ 0·011(0·006)† 0·007(0·007)† Normal EQ-5D™ at 6 months Surgery 0·884(0·010)* 0·888(0·011)* Beta UGFS (increment)‡ −0·009(0·018)† −0·020(0·023)† Normal EVLA (increment)‡ 0·016(0·012)† 0·011(0·016)† Normal Costs at 6 months (£) Surgery 1110(80)* 1047(80)* Gamma UGFS (increment)‡ −655(85)† −579(75)† Normal EVLA (increment)‡ −160(83)† −98(79)† Normal . Complete case data . Multiple imputation data . Distributional form . EQ-5D ™ at baseline (across all treatment groups) QALYs at 6 months 0·790(0·009) – Beta Surgery 0·432(0·006)* 0·435(0·006)* Beta UGFS (increment)‡ −0·005(0·010)† −0·011(0·009)† Normal EVLA (increment)‡ 0·011(0·006)† 0·007(0·007)† Normal EQ-5D™ at 6 months Surgery 0·884(0·010)* 0·888(0·011)* Beta UGFS (increment)‡ −0·009(0·018)† −0·020(0·023)† Normal EVLA (increment)‡ 0·016(0·012)† 0·011(0·016)† Normal Costs at 6 months (£) Surgery 1110(80)* 1047(80)* Gamma UGFS (increment)‡ −655(85)† −579(75)† Normal EVLA (increment)‡ −160(83)† −98(79)† Normal Values are mean(s.e.); * cluster bootstrapped standard error; † cluster robust standard error. ‡ Estimated 6-month utility and cost increments associated with ultrasound-guided foam sclerotherapy (UGFS) and endovenous laser ablation (EVLA) compared with surgery. Cost and utility estimates were derived using linear regression adjusting for baseline EQ-5D™ score and minimization variables used in the randomization process (age (less than 50 years, 50 years or older), sex, great saphenous vein or small saphenous vein reflux, and unilateral or bilateral veins). Open in new tab Model-based incremental cost-effectiveness analysis The model-based estimates of mean costs and QALYs at 5 years are presented in Table 4. EVLA was associated with increased costs and QALYs in comparison with UGFS, with an incremental cost per QALY gained of £3640. Surgery was associated with increased costs compared with EVLA (£206), and on average produced slightly fewer QALYs (−0·078). This was driven by the slightly lower number of QALYs observed for surgery at 6 months compared with EVLA, and also the slightly higher clinical recurrence rate applied to surgery compared with EVLA. Applying a ceiling WTP ratio of £20 000 per QALY gained, EVLA had the highest probability (78·7 per cent) of being cost-effective. Table 4 Model-based incremental cost-effectiveness analysis (over a 5-year time horizon) . Cost (£) . Incremental cost (£) . QALYs . Incremental QALYs . Incremental cost per QALY (£) . Probability cost-effective at Rc £20 000 . UGFS 664 – 4·000 – – 0·168 EVLA 1095 431 4·119 0·118 3640 0·787 Surgery 1300 206 4·040 −0·078 Dominated 0·045 . Cost (£) . Incremental cost (£) . QALYs . Incremental QALYs . Incremental cost per QALY (£) . Probability cost-effective at Rc £20 000 . UGFS 664 – 4·000 – – 0·168 EVLA 1095 431 4·119 0·118 3640 0·787 Surgery 1300 206 4·040 −0·078 Dominated 0·045 QALY, quality-adjusted life-year; Rc, ceiling ratio of willingness to pay per QALY gained; UGFS, ultrasound-guided foam sclerotherapy; EVLA, endovenous laser ablation. Open in new tab Table 4 Model-based incremental cost-effectiveness analysis (over a 5-year time horizon) . Cost (£) . Incremental cost (£) . QALYs . Incremental QALYs . Incremental cost per QALY (£) . Probability cost-effective at Rc £20 000 . UGFS 664 – 4·000 – – 0·168 EVLA 1095 431 4·119 0·118 3640 0·787 Surgery 1300 206 4·040 −0·078 Dominated 0·045 . Cost (£) . Incremental cost (£) . QALYs . Incremental QALYs . Incremental cost per QALY (£) . Probability cost-effective at Rc £20 000 . UGFS 664 – 4·000 – – 0·168 EVLA 1095 431 4·119 0·118 3640 0·787 Surgery 1300 206 4·040 −0·078 Dominated 0·045 QALY, quality-adjusted life-year; Rc, ceiling ratio of willingness to pay per QALY gained; UGFS, ultrasound-guided foam sclerotherapy; EVLA, endovenous laser ablation. Open in new tab The proportion of probabilistic iterations favouring each of the alternative strategies by increasing levels of WTP per QALY gained is shown in Fig. 2. This figure shows that, as WTP increases beyond £30 000, EVLA has approximately an 80 per cent chance of being the optimal strategy on grounds of cost-effectiveness. Fig. 2 Open in new tabDownload slide Cost-effectiveness acceptability curve based on 5-year time horizon. UGFS, ultrasound-guided foam sclerotherapy; EVLA, endovenous laser ablation; QALY, quality-adjusted life-year Sensitivity analysis The results of various analyses assessing the sensitivity of results to uncertainty surrounding several key input parameters and modelling assumptions are shown in Table S2 (supporting information). In general, the findings were robust to changes, including the cost of follow-up assessment in patients requiring further foam sclerotherapy treatment following EVLA and UGFS, the use of multiple imputation for missing data, and the inclusion of the annuitized cost of a laser generator. The application of additional theatre overhead costs, and the removal of all follow-up clinical assessment costs, increased the estimated cost savings associated with EVLA compared with surgery to £439 and £332 respectively. It was also noted that in some centres EVLA was carried out with a greater number of staff present than for UGFS. When all EVLA procedures were costed using the same nursing staffing profile as for UGFS, the cost saving with EVLA compared with surgery increased to £368. When applying the assumption of equivalent QALYs between the strategies to 6 months, and equivalent EQ-5D™ values at 6 months (rather than the treatment-specific mean estimates), UGFS had the highest probability of being cost-effective at 5 years (scenario 13; Table S2, supporting information). However, the base-case ordering of treatment modalities was restored when the time horizon for this analysis was extended to 10 years (scenario 14; Table S2, supporting information). The analysis of the alternative model structure, based on clinical severity states defined by the VCSS, predicted higher QALYs with EVLA than UGFS, and slightly higher QALYs with surgery than EVLA (scenario 22; Table S2, supporting information). However, the probability of surgery being cost-effective remained lower than that for EVLA at the Rc of £20 000 per QALY. Discussion The results of this study suggest that, over a 5-year time horizon, EVLA is likely to be the preferred option on grounds of cost-effectiveness, costing £3640 per QALY gained in comparison with UGFS, and costing less and producing slightly more QALYs than surgery. EVLA had approximately an 80 per cent chance of being cost-effective at a ceiling WTP ratio of £20 000 per QALY. These findings were generally found to be robust to uncertainty surrounding various model inputs and assumptions, including multiple imputation of missing data, and alternative model structuring assumptions. The base-case estimated reduction in cost (£206) associated with EVLA compared with surgery is likely to be a conservative estimate that does not fully capture the increased overhead costs associated with day-case admission and use of theatre for surgery. When a separate estimate of theatre overhead costs (per hour) was factored into the analysis, the cost saving associated with EVLA increased to £439. This additional overhead cost was omitted from the base-case analysis because it was available for only one recruiting centre and it may also double-count some of the overhead costs included in the staff time unit cost multipliers. However, even when this cost was excluded, sensitivity analysis indicated that the cost savings associated with EVLA would increase to £368 if performed using a similar staff profile to UGFS, and that further savings could also be realized if patients with a good outcome following initial EVLA treatment were not followed up at 6 weeks in routine practice. Strengths of the study include that the model was populated, where possible, using estimates of the mean differences in costs and effects derived from the analysis of prospectively collected individual patient data on resource use. Therefore, the results are internally valid and generalizable across settings in the UK. In the absence of data on long-term clinical recurrence for the CLASS cohort, the risk of clinical recurrence was modelled using data derived from an existing network meta-analysis of trials reporting clinical recurrence of varicose veins as an outcome2. Although this provides the best current source of evidence on recurrence, the quality of included trials varied and follow-up generally extended to only 12 or 24 months. Furthermore, clinical recurrence was not always well defined and varied between studies. Uncertainty remains regarding the applicability of these recurrence rates to patients in the CLASS trial (and more generally), and also regarding the risks of recurrence beyond 24 months. This underlines the importance of collecting further data on clinical recurrence and its impact on generic health-related quality of life via the extended follow-up of CLASS participants, which is currently in progress. In general, the approach used in this paper is consistent with that used in previous economic modelling studies of treatments for varicose veins. The structure and assumptions applied in the primary analysis are in line with those used in the model developed to inform the recent NICE clinical guideline2, which also found that endothermal treatment (EVLA or RFA) had the highest probability of being cost-effective under most scenarios assessed. However, the cost and utility inputs applied in the present model were based on prospectively collected data from randomized patients. Consequently, based on a bottom-up costing approach, a somewhat narrower difference in cost was estimated between EVLA and surgery compared with the estimates used in the NICE guideline model. This was due to differences in estimated mean treatment durations, and the reported staffing profiles for the different procedures, compared with those used in the NICE model. The present estimates are more likely to be representative of current practice because they were derived using time and staffing data collected prospectively alongside the CLASS trial. Furthermore, rather than assuming equivalent utility values following the different treatment modalities at 6 months (as was done in the NICE guideline model), the estimated 6-month mean incremental differences between UGFS and surgery, and between EVLA and surgery, were applied in the present study. Gohel and colleagues3 similarly developed a Markov model to assess the cost-effectiveness of conservative management, UGFS, EVLA, RFA and surgery. Based on their model, they estimated that EVLA carried out under local anaesthetic would have the highest chance of being cost-effective at the Rc of £20 000 per QALY. In contrast to the findings here, they reported that surgery (on a day-case basis) had the next highest probability of being cost-effective, followed by RFA and UGFS3. The present findings appear inconsistent with those of a recently conducted modelling study which concluded that UGFS would be more cost-effective than surgery, EVLA and RFA, based on it having the lowest cost while generating similar clinical outcomes. In this modelling study, Carroll and colleagues4 used data from two single-centre UK trials17,18 to assess the relative cost differences between UGFS, RFA, EVLA and surgery; the cost of EVLA was estimated to be 2·02 times that of surgery. However, this multiplier was derived from cost data collected for 88 patients randomized between surgery and RFA17, and in this particular study RFA was performed under general anaesthesia and took 30 min longer than surgery. The staff and theatre costs were thus higher for RFA than for surgery. By contrast, in CLASS, all EVLA procedures were performed under local anaesthetic; the majority (90·5 per cent) were carried out in a clinic treatment room rather than an operating theatre and the measured procedure time was lower for EVLA than for surgery. This information was collected prospectively for patients recruited to CLASS from multiple treatment centres, and so should be generalizable to standard NHS practice. This explains the lower estimate for the cost of EVLA compared with surgery, and also helps explain why EVLA was more cost-effective than UGFS. The present modelling study suggests that, for patients with primary varicose veins in whom all three of the investigated treatment modalities offer a clinically viable option, EVLA performed under local anaesthetic has the highest probability of being cost-effective at accepted thresholds of WTP per QALY. This finding is consistent with the results of recent modelling undertaken to inform the NICE clinical guideline on the management of varicose veins. In the interests of cost-effectiveness, EVLA should be performed where possible in a clinic treatment room and the levels of nursing support required should be considered carefully. Acknowledgements The authors thank the following contributors to the CLASS project for ongoing advice and support for the trial: J. Cruden, T. Davidson, J. Francis, A. McDonald, G. McPherson, S. Wileman, I. Chetter, J. J. Earnshaw, T. Lees, J. Scott, S. Baker, G. MacLennan, M. Prior and D. Bolsover. In addition, they thank J. Wallace, E. Wilson, S. Hardy, K. Hargreaves, D. Rittoo, S. Baker, L. Park, D. Parker, I. Chetter, J. Hatfield, J. Scott, N. Dewhirst, T. Lees, N. Parr, D. Dodd, J. Sorrell, K. R. Makhdoomi, B. Rorison, I. Nyamekye and P. Matheson for information on resource use and costs for treatment modalities at participating centres. This work was carried out as part of a project funded by the National Institute for Health Research (NIHR) Health Technology Assessment (HTA) Programme (project number 06/45/02). The views and opinions expressed herein are those of the authors and do not necessarily reflect those of the NIHR HTA Programme, NIHR, NHS or Department of Health. Disclosure The authors declare no conflict of interest. References 1 Health and Social Care Information Centre . Hospital Episode Statistics . http://www.hesonline.nhs.uk/Ease/servlet/ContentServer?site=1937&categoryID=192 [accessed 1 August 2013 ]. 2 National Institute for Care and Health Excellence . Varicose Veins in the Legs (CG168) . http://guidance.nice.org.uk/CG168 [accessed 1 August 2013 ]. 3 Gohel MS , Epstein DM, Davies AH. Cost-effectiveness of traditional and endovenous treatments for varicose veins . Br J Surg 2010 ; 97 : 1815 – 1823 . Google Scholar Crossref Search ADS PubMed WorldCat 4 Carroll C , Hummel S, Leaviss J, Ren S, Stevens JW, Everson-Hock E et al. Clinical effectiveness and cost-effectiveness of minimally invasive techniques to manage varicose veins: a systematic review and economic evaluation . Health Technol Assess 2013 ; 17 : i–xvi , 1 – 141 . Google Scholar OpenURL Placeholder Text WorldCat 5 National Institute for Health Research . The CLASS Trial: Comparison of LAser, Surgery and Foam Sclerotherapy. Research Protocol . http://www.nets.nihr.ac.uk/__data/assets/pdf_file/0004/51448/PRO-06-45-02.pdf [accessed 1 January 2014 ]. 6 Breu FX , Guggenbichler S. European Consensus Meeting on Foam Sclerotherapy, April, 4–6, 2003, Tegernsee, Germany . Dermatol Surg 2004 ; 30 : 709 – 717 . Google Scholar Crossref Search ADS PubMed WorldCat 7 Personal Social Services Research Unit (PSSRU) . Unit Costs of Health and Social Care . http://www.pssru.ac.uk/pdf/uc/uc2011/uc2011.pdf [accessed 1 October 2012 ]. 8 GOV.UK . 2010–11 Reference Costs Publication . https://www.gov.uk/government/publications/2010-11-reference-costs-publication [accessed 1 October 2012 ]. 9 Information Services Division (ISD) Scotland . Theatres. Costs – Detailed Tables. R141X: Average Theatre Tunning Costs, by Board; SFR 5.10: Source Data for Theatre Services . http://www.isdscotland.org/Health-Topics/Finance/Costs/Detailed-Tables/Theatres.asp [accessed 1 October 2012 ]. 10 Dolan P , Gudex C, Kind P, Williams A. The time trade-off method: results from a general population study . Health Economics 1996 ; 5 : 141 – 154 . Google Scholar Crossref Search ADS PubMed WorldCat 11 StataCorp . Stata Multiple-Imputation Reference Manual Release 12 . StataCorp LP : College Station , 2012 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 12 Glick H , Doshi J, Sonnad S, Polsky D. Economic Evaluation in Clinical Trials (1st edn). Oxford University Press : New York , 2007 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 13 Royston P Flexible alternatives to the Cox model, and more . Stata J 2001 ; 1 : 1 – 28 . Google Scholar Crossref Search ADS WorldCat 14 Software TreeAge . TreeAge Pro 2013, R2.0 Released . http://www.treeage.com/news/treeage-pro-2013-r2-0-released/ [accessed 1 August 2013 ]. 15 National Institute for Health and Care Excellence . Guide to the Methods of Technology Appraisal 2013 . http://www.nice.org.uk/media/D45/1E/GuideToMethodsTechnologyAppraisal2013.pdf [accessed 1 June 2013 ]. 16 Vasquez MA , Rabe E, McLafferty RB, Shortell CK, Marston WA, Gillespie D et al. ; American Venous Forum Ad Hoc Outcomes Working Group. Revision of the venous clinical severity score: venous outcomes consensus statement: special communication of the American Venous Forum Ad Hoc Outcomes Working Group . J Vasc Surg 2010 ; 52 : 1387 – 1396 . Google Scholar Crossref Search ADS PubMed WorldCat 17 Brittenden J , Cotton SC, Elders A, Tassie E, Scotland G, Ramsay CR et al. Randomised controlled trial comparing foam sclerotherapy, alone or in combination with endovenous laser therapy, with conventional surgery as a treatment for varicose veins . Health Technol Assess 2014 ; (in press). Google Scholar OpenURL Placeholder Text WorldCat 18 Subramonia S , Lees T. Radiofrequency ablation vs conventional surgery for varicose veins – a comparison of treatment costs in a randomised trial . Eur J Vasc Endovasc Surg 2010 ; 39 : 104 – 111 . Google Scholar Crossref Search ADS PubMed WorldCat © 2014 BJS Society Ltd. Published by John Wiley & Sons Ltd This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) © 2014 BJS Society Ltd. Published by John Wiley & Sons Ltd

Journal

British Journal of Surgery – Oxford University Press

Published: Oct 2, 2014

Keywords: sclerotherapy; ultrasonography; cost effectiveness; quality-adjusted life years; surgical procedures, operative; varicosity; surgery specialty; laser ablation; class trial; national health service (uk)

There are no references for this article.