Abstract - University of Manchester



Health economic analysis of access site practice in England during changes in practice: Insights from the British Cardiovascular Interventional SocietyRunning title: Economics of access site practice in EnglandMamas A. Mamas BM BChir, DPhil1; Jon Tosh, PhD2; Will Hulme, PhD3; Nicki Hoskins, DPhil2; George Bungey, BSc2; Peter Ludman, MD4; Mark de Belder, MD5; Chun Shing Kwok, MBBS, MSc, BSc1; Nathalie Verin, MSc6; Tim Kinnaird, MD7; Ewan Bennett, MSc, EngD2, Nick Curzen, BM, PhD8; James Nolan, MD1; Evangelos Kontopantelis, PhD.31.Keele Cardiovascular Research Group, Keele University, and Royal Stoke Hospital, University Hospital North Midlands, Stoke-on-Trent, UK.2. DRG Abacus, Bicester, Oxfordshire, UK.3. Health eResearch Centre, Farr Institute for Health Informatics Research, University of Manchester, UK, Faculty of Medical and Human Sciences, University of Manchester, UK.4. Queen Elizabeth Hospital, Birmingham, UK.5. The James Cook University Hospital, Middlesbrough, UK.6. Terumo Europe N.V. Leuven, Belgium.7. University Hospital of Wales, Cardiff, UK.8. University Hospital Southampton and Faculty of Medicine, University of Southampton, UK.Corresponding authorMamas A. Mamas Professor of Cardiology Keele Cardiovascular Research Group Centre for Prognosis Research, Institute for Primary care and Health Sciences,Keele University Stoke-on-Trent, United Kingdom E-mail: mamasmamas1@yahoo.co.ukTwitter handle: @mmamas1973Tel:+44 (0)1782 671654Word count: 4,008Journal subject terms: Percutaneous Coronary Intervention, Revascularization, Cost-Effectiveness, Quality and OutcomesAbstractBackgroundTransradial access (TRA) for percutaneous coronary intervention (PCI) is associated with a reduced risk of mortality compared with transfemoral access (TFA), access site-related bleeding complications and shorter length of stay (LoS). The budget impact from a healthcare system that has largely transitioned to TRA for PCI has not been previously published.Methods and ResultsData from 323,656 patients undergoing PCI between 2010–2014 were obtained from the British Cardiovascular Intervention Society database. Costs for TRA and TFA PCI were estimated based on procedure cost, length of stay, and differences in the rates of complications (major bleeding and vascular complications). In the base case a propensity matched data set between TFA and TRA was used to directly compare the cost per PCI, while in the real-world analysis the full dataset was used. Across all indications and all years, TRA offered an average cost saving of ?250.59 per procedure (22% reduction) versus TFA with the majority of cost saving derived from reduced LOS (?190.43) rather than direct costs of complications (?3.71). In the real-world analysis, adoption of TRA was estimated to have provided cost savings of ?13.31 million across England between 2010–2014; however, if operators in all regions had adopted TRA at the rate of the region with the highest utilization, cost savings of ?33.40 million could have been achieved. ConclusionsThe transition to TRA in England has been associated with significant cost savings across the national healthcare system, in addition to the well-established clinical benefits. Key words: Budget impact, financial implications, transradial access, percutaneous coronary intervention.IntroductionPercutaneous coronary intervention (PCI) is the most common form of revascularization in patients with coronary artery disease in both the elective and emergency setting. Historically, the femoral artery has been the access site of choice. However, both randomized controlled trials and national registries with large unselected patient populations have shown that transradial access (TRA) is associated with a reduced risk of mortality compared with transfemoral access (TFA), as well as lower rates of access site-related bleeding complications and a shorter length of stay (LoS).1-11 TRA is, therefore, increasingly being adopted as the primary access site for PCI across the US and many countries in Europe and Asia.12Despite the evidence and recommendations supporting TRA, there is significant variation in its uptake both between and within countries. In the United Kingdom TRA has become the dominant access site, representing 80.5% of procedures in 2015 with hospital adoption rates varying from 18% to over 95%.13 In the US, studies derived from the National Cardiovascular Data Registry have also demonstrated significant heterogeneity in adoption of TRA, varying from 12% in the Northeast to 38% in the South.14In addition to the clinical benefits of TRA, economic analyses in the US and China have shown healthcare cost savings, primarily due to the lower rate of complications and shorter hospital stay for TRA, although these studies were limited as they were either derived from single centers or excluded patients with acute coronary syndromes where the benefit associated with TRA is arguably at its greatest.14-17 To the best of our knowledge, there has not been a previous national health economic analysis of access site practice in PCI, for an unselected all comer population. Furthermore, no previous studies have considered a European perspective using real-world data, and no studies have examined the impact of the rate of adoption of TRA on national healthcare budgets. We have therefore undertaken a health economic analysis to study differences in healthcare expenditure associated with access site practice in an unselected national cohort of patients undergoing PCI in England during a period of change in access site practice. The objectives of the current study were to estimate (i) the differences in procedure cost, length of stay, complications, and cost of complications between TRA and TFA, and (ii) to estimate the additional costs that could have been saved if TRA uptake has been more uniform across the country.MethodsAn economic model was developed to synthesize clinical outcomes and resource use (such as length of stay and NHS unit costs) for transradial and transfemoral PCI based on data from the British Cardiovascular Intervention Society (BCIS)–National Institute for Clinical Outcomes Research (NICOR) database. The overall approach was based on that of the UK National Clinical Guideline Centre cost analysis that was developed to inform NICE Clinical Guideline for the management of patients presenting with STEMI.18To reflect differences in patient profiles, clinical outcomes, length of stay and costs, the economic model was developed for 3 indications: NSTEMI, STEMI, and stable angina. Clinical dataBCIS collects data on all PCI procedures in the United Kingdom, with data collection coordinated by the National Institute of Cardiovascular Outcomes Research (NICOR) () via the Central Cardiac Audit Database. In 2011, this data set collected information on 99.4% of all PCI procedures performed in National Health Service (NHS) hospitals in England and Wales. The BCIS–NICOR database contains a total of 113 variables, including information on clinical and procedural parameters, as well as patient outcomes. Data on English residents undergoing PCI in the NHS in the UK between 2010–2014 were provided from this database, aggregated by indication (elective/stable, Non-ST elevation myocardial infarction (NSTEMI) /unstable angina, and ST segment elevation myocardial infarction [STEMI]), Strategic Health Authority (SHA; these were part of the structure of the National Health Service in England between 2002 and 2013 and were responsible for enacting the directives and implementing fiscal policy as dictated by the Department of Health at a regional level), and access site.19 A small proportion of procedures (2.9%) were classified as multiple/other, when the access site was not clear or more than one site was used in the same procedure; these data were excluded from this analysis. During PCI it is possible that the primary point of access fails, and that access has to be changed during the procedure (access site crossover). Access site crossover is not captured in the BCIS data, but was included in this analysis as a contributor to procedure costs with access site crossover rates that were obtained from the RIVAL study10 for STEMI and NSTEMI/unstable patients. As crossover rates could not be identified for stable patients from the literature, the pooled crossover rates for STEMI and NSTEMI/unstable patients from the RIVAL study were applied (Supplementary Table 1). Complications that were included in the economic model were major bleeding and vascular complications for rate of event and cost, and major adverse cardiac event (MACE) for rate of event only. MACE was defined as a composite of in-hospital mortality, myocardial infarction, or repeat intervention. Arterial access site complication was defined as any pseudoaneurysm or any access site hemorrhage requiring intervention or delaying discharge. Bleeding was defined as any gastrointestinal bleed, intracerebral bleed, retroperitoneal bleed, or transfusion.1 The 2010–2014 data for England as a whole were used for the base case analysis to compare the cost of transradial and transfemoral PCI.CostsA breakdown of the unit costs used in the economic model and an example flow diagram is provided in the Supplementary Table 2. Procedure and access site crossover costs differ between TRA and TFA, due to different equipment used for each procedural method. It should be noted that the procedure costs are not exhaustive, as they do not include the cost of consumables such as stents and balloons, or costs of using the catheterization lab, as these are likely to be similar for both approaches. However, the cost of a day’s stay in hospital and the costs of treating major bleeding and vascular complications are assumed to be the same regardless of access site, and any differences between the two approaches are due to a different length of stay and different rates of complications associated with each method. The cost per day of a hospital stay is taken from the NHS national schedule of reference costs for 2015–16,20 and is calculated as a weighted average of all the PCI Healthcare Resource Groups (HRGs) (Supplementary Table 2).The cost of major bleeding is assumed to be composed of a blood transfusion, calculated as the cost of a unit of red blood cells21 multiplied by the average number of units used in a transfusion.22 While Clinical Guideline 94 included an extra 6 days in hospital for the management of NSTEMI and unstable angina,23 it was considered that this time was already accounted for in the BCIS–NICOR dataset; this additional time was therefore not included in this model. The cost of vascular complications is assumed to include an ultrasound, a compression device, a thrombin injection, and vascular surgery. The costs of an ultrasound and vascular surgery are taken from the NHS national schedule of reference costs for 2015–16,20 and the cost of a compression device and a thrombin injection are assumed to be the midpoint of the ranges reported in the cost analysis from the National Clinical Guideline Centre for patients presenting with STEMI,18 which are then inflated using Personal Social Services Research Unit (PSSRU) inflation data.24 The percentage utilization of an ultrasound, a compression device, a thrombin injection, and vascular surgery were taken from the cost analysis from the National Clinical Guideline Centre for patients presenting with STEMI.18Base case analysisThe base case analysis directly compares the costs per procedure of transradial and transfemoral PCIs in England from 2010 to 2014 for each indication. In order to account for potential selection bias and confounding in the full BCIS data set, a 1-to-1 logistic regression propensity score matched data set was generated. Matching was exact for procedure year and indication, and additional non-exact matching was performed to adjust for difference in age, sex, history of MI, CABG, PCI, stroke, hypertension, high cholesterol, peripheral vascular disease, valvular heart disease, diabetes, renal dysfunction and cardiogenic shock. Missing values for patient history and comorbidities were imputed using the ‘missing as absent’ heuristic. Operator-modifiable variables were not used to match. The propensity matching algorithm uses “optimal” nearest neighbor matching, with samples chosen to minimize the average absolute different across all matched pairs.Real-world analysisFor the real-world analysis, the full BCIS data set on TRA and TFA PCI procedures for 2010–2014 was used to examine differences in procedure uptake and key cost drivers for transfemoral and transradial PCIs, and the impact of recent trends in procedural access uptake over time for each English region. In order to analyze the effect of changes in procedural uptake over time in each region the total hospital length of stay, complications, and costs based on actual procedure uptake in each year from 2011–2014 were compared with those generated by applying the 2010 procedure uptake for each year. These results were then compared with those obtained by applying to each region the procedural uptake rate for the region with the highest TRA utilization to estimate the impact of more rapid TRA uptake. The baseline characteristics for the full BCIS data set used in this analysis are provided in the Supplementary Table 3.The real-world analysis uses the full BCIS dataset to allow for regional variation in length of stay and complications, and to estimate the actual economic impact of changes in TRA update that occurred from 2010-2014. However, patient characteristics may impact both on the choice to perform a TRA or TFA as well as clinical outcomes, which may confound this analysis. Therefore, an additional analysis was performed using the propensity-matched dataset used in the base case in order to assess the impact of potential bias in the unadjusted dataset on the population level results.Sensitivity analysisComplications in the BCIS data are self-reported; it is therefore likely that the rates of complications are underestimated. A sensitivity analysis was consequently performed on the base case using complications rates from a published meta-analysis of randomized trials by Ferrante et al.25 Pooled rates are available from this study for MACE, major bleeding, and vascular complications. For NSTEMI bleeding rates, the Ferrante et al study considered three studies: RIVAL,10 SAFE-PCI26 and MATRIX4. The RIVAL10 study used a fairly conservative definition of major bleeding, whilst the SAFE-PCI study focused on a female-only population.26MATRIX used the BARC criteria,4 which are the current standard for defining bleeding events in clinical trials. Therefore, the current analysis uses data derived from the MATRIX trial.4 The data used in the sensitivity analysis are available in the Supplementary Table 4.The authors are unable to share data and study materials as this would not be allowed under the data sharing agreements with the National Institute of Cardiovascular Outcomes Research; however, analytic methods can be made available to other researchers for purposes of replicating the procedure.ResultsBase case analysisA total of 248,228 propensity matched procedures between 2010–2014 were included in the base case analysis. The clinical and procedural characteristics stratified by access site are presented in Table 1 with clinical outcome data that the analysis was based on presented in Table 2. An example flow diagram of the economic model calculations used in the base case analysis is provided in Supplementary Figure 1. Between 2010–2014, the total cost per procedure (access site/device, length of stay, and complications) was ?1,120.84 for TFA (range ?997.99–1,191.00) and ?870.25 for TRA (range ?832.43–902.54), with TRA offering a cost saving of ?250.59 (22% reduction) (Table 3). The majority (76%) of the cost saving was attributable to length of stay costs. The greatest cost saving was in the STEMI indication (?348.26), followed by NSTEMI/unstable (?282.21) and elective/stable (?153.88). In the sensitivity analysis on complication costs a similar result was observed, with a cost saving of ?251.84 (22%) per procedure for TRA vs TFA (Table 4). Overall costs for TFA and TRA were both higher than the base case analysis due to the higher absolute complication rates in Ferrante et al25 and the MATRIX4 study relative to the BCIS dataset.Real-world analysisThe percentage of PCI procedures performed by TRA has increased between 2010 and 2014 across all regions in England, but at varying rates, with rates between 20.7% and 79.8% reported in 2010 to 44.5–84.3% in 2014 (Figure 1).Using these uptake rates, the actual uptake of TRA has resulted in a cost saving to the National Health Service (NHS) of ?13.31 million over 2010–2014, compared with the costs that would have been incurred if the percentage of TRA procedures had remained constant at the same level as in 2010 (Table 5). As shown in Figure 1, the percentage uptake of TRA was much lower in some regions than others during 2010–2014. For example, TRA uptake in the South East Coast region increased from 20.7% in 2010 to 44.5% in 2014, compared with 79.8% and 84.3%, respectively, for the North East region. An analysis was therefore performed which assumed that all regions adopted TRA at the same rate as the region with highest utilization of TRA (North East). Increasing the rate of TRA uptake increased the total cost savings versus constant 2010 uptake rates to ?33.40 million, a further increase of ?20.09 million over and above the cost savings actually achieved during this period (Table 5). A summary of the potential cost savings in each region versus those actually achieved over 2010–2014 is presented in Figure 2, with a full breakdown provided in the Supplementary Table 5.In the analysis using the propensity-matched dataset, total cost savings of ?12.39 million were found to have been achieved during 2010–2014 based on actual TRA uptake, which increased to ?30.85 million if uptake rates for all regions were assumed to match that of the region with the highest utilization (Supplementary Table 6). DiscussionThis analysis is the first to consider the budget impact of TRA uptake from a national perspective in an unselected real world cohort from a healthcare system that has transitioned to a predominantly TRA strategy. It is also the first study to examine the impact of the rate of adoption of TRA on national healthcare costs. Our current analysis has found that TRA is 24% less costly on a per procedure basis in England, compared with TFA. Based on current adoption rates for TRA, cost savings of ?13.31?million were estimated between 2010 and 2014 in the real-world analysis; however, if all regions had adopted TRA at the same rate as the region with the highest utilization (North East), a total national cost saving of ?33.40?million could have been achieved during this period. The main component of these savings is the reduction in hospital stay; this not only results in a significant cost saving, but will also increase the efficiency of the system by freeing up hospital beds for other patients. Finally, the greatest cost saving was found in the highest risk acute cases (STEMI cases), where the clinical benefit of TRA has been shown to be greatest.2-5 This real world national economic data therefore adds to the already compelling clinical evidence supporting TRA for PCI. These findings should encourage decision-makers to enhance adoption of TRA both for the benefit of patients and the healthcare economy.While the current study is the first to consider the regional and national budget impact of TRA adoption, the cost per procedure for TRA vs TFA has been considered in seven previous studies, five from a US perspective,16,17,27-29 and one each from China15 and Poland.30 In line with the current analysis, these studies have consistently reported cost savings associated with TRA compared with TFA. From a US perspective, two single-center studies by Muthusany et al27 and Roussanov et al28 considered 2,972 and 181 patients, respectively. Muthusany et al reported significant unadjusted procedural and post-procedural cost savings of $1,019 and $957 for TRA vs TFA, respectively (p<0.05). After adjustment for patient characteristics these savings remained significant for procedural costs, but not for post-procedural costs.27 Roussanov et al also reported significantly lower costs for TRA ($369.5±74.6) vs TFA with ($446.9±60.2) or without ($553.4±81.0) a closure device (p<0.001 for both).28 A multicenter national study of patients undergoing elective PCI from the NCDR dataset by Amin et al reported similar results (n=7,121), with total cost savings of $830 (95% CI $296-1,364; p<0.001) for TRA vs TFA, which increased as the risk of bleeding increased.16 In a database analysis by Safley et al, 609 TRA patients were matched with 60,900 TFA patients. Total adjusted costs were significantly lower for TRA vs TFA by $553 (95% CI $45–1,060; p=0.033). Specifically, while day of procedure costs were similar between TRA and TFA, costs after this point through to post-discharge were lower for TRA.29A previous study undertaken in 5 US centers has suggested that cost benefit associated with TRA may relate to bleeding risk, with cost savings calculated in low risk: $642 (95%CI: $43 to $1,236; p = 0.035); moderate risk: $706 (95% CI: $104 to $1,308; p = 0.029); and high risk: $1,621 (95% CI: $271 to $2,971, p = 0.039).31 Our analysis is in agreement with this, with the greatest saving observed in the higher bleeding risk STEMI population (?348.26) and the lowest gains achieved in the elective/stable group (?153.88), which is in line with our previous work that suggests that the greatest clinical benefit associated with TRA is in those patients at highest baseline bleeding risk.2 In addition to these studies, a network meta-analysis by Safley et al included 14 trials and a stochastic simulation model of per-case costs from a US perspective, and reported a cost saving of $275 (95% CI ?$374 to ?$183) per patient from the hospital perspective with TRA vs TFA.29From a non-US perspective, Chen Jin et al conducted a single-center study of 5,306 patients in China which reported a lower total adjusted cost for TRA, compared with TFA (adjusted difference ?8,081 [$1,283]).15 Similarly, Koltowski et al conducted an economic analysis of the OCEAN RACE single-center trial in Poland (n=103), which reported a lower cost of therapeutic success for TRA vs TFA (€3,060 vs €3,374, respectively) and lower costs related to absence from work.30While these studies have reported cost savings with TRA in comparison to TFA, only 1 study including 103 patients from Poland has considered a European perspective and none have used a national all comer dataset that does not exclude acute coronary syndrome (ACS) cases or other high risk non elective cases16 where the clinical benefit derived from TRA is greatest.2 Furthermore, none have considered the budget impact of TRA uptake at a regional or national level. The current analysis uses a robust and transparent economic model based on a large real-world dataset. This dataset includes the largest patient population modeled so far for TRA, and is the first dataset used where TRA has grown to be the most commonly used access site nationally. As with all budget impact economic models, there are a number of limitations with the current analysis; with a number of cost assumptions made. Where possible, a conservative approach has been taken with estimates around costings and therefore are more likely to result in an underestimation of the cost benefits for the TRA vs TFA approach rather than an overestimation. Firstly, differences in pharmacotherapy costs associated with each procedure were not considered due to a lack of available data. It is possible that clinicians using TRA may treat more aggressively with pharmacological therapy, due to a lower risk of bleeding compared with TFA, although drugs costs represent only a small component of total healthcare costs. Similarly, the number and types of stents used were not costed into the analysis as the cost of a stent will vary at the individual center level and these data are not available in the BCIS dataset. Nevertheless, it is unlikely that access site choice will influence stent choice or number of stents used per case. Finally access site crossover is not captured in the BCIS data, but hence we used access site crossover rates that were obtained from the RIVAL study10 for STEMI and NSTEMI/unstable patients. Crossover rates could not be identified for stable patients from the literature, hence pooled crossover rates for ACS patients from the RIVAL study were applied. This is likely to be an overestimate of actual cross over rates observed in stable patients clinically, leading to a more conservative cost estimate than would be the actual case in this group of patients. Secondly, it is likely that the length of stay for NSTEMI/unstable patients has not been fully captured in the BCIS data. This is because in the UK patients with NSTEMI admitted to hospitals without PCI facilities are often transferred as day cases to a PCI center before being discharged back to the center of admission for continuation of care. The BCIS data only captures the length of stay for patients in the PCI centre which may lead to underestimation of total healthcare costs for both access sites, particularly for NSTEMI cases. Our length of stay for NSTEMI cases was shorter than other international registries, for example the median LOS from the US ACTION Registry-GWTG report was 3 days (IQR 2-5 days).32 Nevertheless, patients admitted to non-PCI centers treated as “treat and return cases” that sustain peri-procedural complications will remain in the PCI center so would expect that differences related to LOS occurring as a consequence of vascular complications would still be captured in the data analyzed.Thirdly, because the data come from a national registry, the baseline characteristics of the patients by access site are different, with patients undergoing PCI through the transfemoral route older, with more co-morbidity and tend to be more hemodynamically unstable as 4.0% of TFA patients were admitted with cardiogenic shock and 1.3% required an IABP compared to 2.9% and 1.3% rates respectively in the TRA cohort. This may have contributed to the greater complication rates and extended length of stay in hospital observed in TFA patients particularly in the “real world analysis”; however, we attempted to reduce differences between the cohorts in an analysis using the propensity-matched dataset which found that the cost savings with TRA remained consistent. Nevertheless, there may still be residual confounding even in the propensity score matched cohorts. For example, the femoral cohorts may be frailer and have a greater prevalence of adverse unmeasured characteristics that contribute to the increased LOS still observed after propensity score matching. Similarly, the BCIS registry only provides operator identifiable information for 2013 onwards, so we were unable to assess the experience of operators in each region who are undertaking TRA, which could potentially impact on complications rates, cross-over rates and subsequent costs.Fourthly, costs for MACE were not included in the analysis. MACE is a composite measure of several major adverse cardiac events, and for this reason it was not possible to identify an appropriate and accurate cost of MACE as a complication for patients who have undergone a PCI. However, as MACE complications are infrequent in general, and are less frequent in patients who received TRA than TFA,25 this is a conservative assumption which may underestimate the cost saving with TRA.Furthermore, differences in the number and type of catheters between TRA and TFA were not considered. This approach is generally consistent with previous studies15-17,27,29 with only two of the published analyses considering differences in catheter use.28,30 Roussanov et al reported that fewer catheters were required, and catheter costs were lower, in the TRA group vs TFA.28Koltowski et al did not consider catheter costs in isolation, but grouped these with other costs for single-use materials, and reported that costs for these materials were higher for TRA vs TFA.30 However, both studies concluded that TRA offered overall economic advantages over TFA. The exclusion of differences in number and type of catheters is therefore likely to have a minimal impact on the results of this study.Mortality has not been included as a health economic outcome in this analysis. Mortality is often not included in such health economic analyze, as it does not incur a direct cost on the health service and is difficult to include in budget impact analyses. In addition, our previous analysis has already estimated the number of lives that may have potentially been saved in the United Kingdom due to an increased TRA uptake (450 between 2005 and 2012), and the number that could have been saved if TRA adoption were uniform nationally (264 over the same time period) using this dataset.11 Finally, gains may be clustered by hospital and our analysis, although it accurately provides information on overall costs gains, cannot explore potential heterogeneity. This is due to the great variability in TRA usage across centers, which would make within-center matching inefficient. This might be particularly relevant in situations where there are institutional differences in hospital policies regarding LOS in non-elective cases, particularly between predominantly radial and femoral centers.This analysis was conducted using UK data due to the availability of both clinical and cost data in the UK. However, the cost-saving findings should be generalizable to other countries, due to standardized clinical practice and similar financial considerations, even though the exact cost-saving figures may not be. Studies in the US, China, and Poland have consistently reported savings for TRA vs TFA;15-17,27-30 it is therefore highly likely that applying these cost savings to regional and national uptake rates for TRA would demonstrate that cost savings could potentially be achieved in these countries. Nevertheless, it must be borne in mind that the absolute savings calculated in this analysis may only be applicable to the United Kingdom, as other healthcare systems are financed and structured differently to that in the UK, and the case mix undertaken in countries at the earlier stages of their radial learning curve may be towards lower risk cases, where the magnitude of costs savings is less. Finally, it should be noted that TRA and TFA for stable patients are both typically performed as day case procedures in the UK; in countries which routinely have a longer hospital stay for the TFA procedure than for TFA, the cost savings are therefore likely to increase substantially. ConclusionThis analysis of PCI performed in England demonstrates that the clinical advantages of TRA are accompanied with significant cost savings, and shows that savings of ?13.31 million have already been achieved during 2010–2014 based on current adoption rates. However, further cost savings can be achieved if adoption of TRA is increased, with a national cost saving of ?33.40 million estimated if all regions in England adopted TRA at the same rate as the region with highest utilization. The results of the current study are highly relevant for payers, administrators, and policy-makers who face ever more constrained healthcare budgets, demonstrating that substantial cost savings can be achieved while improving clinical outcomes. The results of this study should therefore encourage a reduction in the variability in uptake between difference English regions and between countries in Europe.AcknowledgementsThe authors would like to thank Chris Livings and Lydia Crowe for the development of the original economic model, which has been further adapted and updated for the purpose of the analysis reported in this manuscript. They would also like to thank Giorgia Magnatti for her assistance in preparation of the manuscript.Funding sources: The economic analysis was funded by an unrestricted educational grant from Terumo?.Disclosures: NV is an employee of Terumo?. All of the other authors have no conflicts of interest to declare. ReferencesRatib K, Mamas MA, Anderson SG, Bhatia G, Routledge H, De Belder M, Ludman PF, Fraser D, Nolan J. Access site practice and procedural outcomes in relation to clinical presentation in 439,947 patients undergoing percutaneous coronary intervention in the United Kingdom. JACC Cardiovasc Interv. 2015;8:20-9.Mamas MA, Anderson SG, Carr M, Ratib K, Buchan I, Sirker A, Fraser DG, Hildick-Smith D, de Belder M, Ludman PF, Nolan J. Baseline bleeding risk and arterial access site practice in relation to procedural outcomes after percutaneous coronary intervention. J Am Coll Cardiol. 2014;64:1554-64.Romagnoli E, Biondi-Zoccai G, Sciahbasi A, Politi L, Rigattieri S, Pendenza G, Summaria F, Patrizi R, Borghi A, Di Russo C, Moretti C, Agostoni P, Loschiavo P, Lioy E, Sheiban I, Sangiorgi G. Radial versus femoral randomized investigation in ST-segment elevation acute coronary syndrome: the RIFLE-STEACS (Radial Versus Femoral Randomized Investigation in ST-Elevation Acute Coronary Syndrome) study. J Am Coll Cardiol. 2012;60:2481-9.Valgimigli M, Gagnor A, Calabro P, Frigoli E, Leonardi S, Zaro T, Rubartelli P, Briguori C, Ando G, Repetto A, Limbruno U, Cortese B, Sganzerla P, Lupi A, Galli M, Colangelo S, Ierna S, Ausiello A, Presbitero P, Sardella G, Varbella F, Esposito G, Santarelli A, Tresoldi S, Nazzaro M, Zingarelli A, de Cesare N, Rigattieri S, Tosi P, Palmieri C, Brugaletta S, Rao SV, HegD, Rothenbuhler M, Vranckx P, Juni P. Radial versus femoral access in patients with acute coronary syndromes undergoing invasive management: a randomised multicentre trial. Lancet. 2015;385:2465-76.Bernat I, Horak D, Stasek J, Mates M, Pesek J, Ostadal P, Hrabos V, Dusek J, Koza J, Sembera Z, Brtko M, Aschermann O, Smid M, Polansky P, Al Mawiri A, Vojacek J, Bis J, Costerousse O, Bertrand OF, Rokyta R. ST-segment elevation myocardial infarction treated by radial or femoral approach in a multicenter randomized clinical trial: the STEMI-RADIAL trial. J Am Coll Cardiol. 2014;63:964-72.Rao SV, Ou F-S, Wang TY, Roe MT, Brindis R, Rumsfeld JS, Peterson ED. Trends in the prevalence and outcomes of radial and femoral approaches to percutaneous coronary intervention: a report from the National Cardiovascular Data Registry. JACC Cardiovasc Interv. 2008;1:379-386.Feldman DN, Swaminathan RV, Kaltenbach LA, Baklanov DV, Kim LK, Wong SC, Minutello RM, Messenger JC, Moussa I, Garratt KN, Piana RN, Hillegass WB, Cohen MG, Gilchrist IC, Rao SV. Adoption of radial access and comparison of outcomes to femoral access in percutaneous coronary intervention: an updated report from the national cardiovascular data registry (2007-2012). Circulation. 2013;127:2295-306.Komócsi A, Aradi D, Kehl D, Ungi I, Thury A, Pintér T, Di Nicolantonio J, Tornyos A, Vorobcsuk A. Meta-analysis of randomized trials on access site selection for percutaneous coronary intervention in ST-segment elevation myocardial infarction. Arch Med Sci. 2014;10:203-212.Bundhoo S, Nallur-Shivu G, Ossei-Gerning N, Zaman A, Kinnaird TD, Anderson RA. Switching from transfemoral to transradial access for PCI: a single-center learning curve over 5 years. J Invasive Cardiol. 2014;26:535-41.Jolly SS, Yusuf S, Cairns J, Niemela K, Xavier D, Widimsky P, Budaj A, Niemela M, Valentin V, Lewis BS, Avezum A, Steg PG, Rao SV, Gao P, Afzal R, Joyner CD, Chrolavicius S, Mehta SR. Radial versus femoral access for coronary angiography and intervention in patients with acute coronary syndromes (RIVAL): a randomised, parallel group, multicentre trial. Lancet. 2011;377:1409-20.Mamas MA, Nolan J, de Belder MA, Zaman A, Kinnaird T, Curzen N, Kwok CS, Buchan I, Ludman P, Kontopantelis E. Changes in Arterial Access Site and Association With Mortality in the United Kingdom: Observations From a National Percutaneous Coronary Intervention Database. Circulation. 2016;133:1655-67.Bertrand OF, Rao SV, Pancholy S, Jolly SS, Rodes-Cabau J, Larose E, Costerousse O, Hamon M, Mann T. Transradial approach for coronary angiography and interventions: results of the first international transradial practice survey. JACC Cardiovasc Interv. 2010;3:1022-31.British Cardiovascular Intervention Society. (accessed March 2018).Bradley SM, Rao SV, Curtis JP, Parzynski CS, Messenger JC, Daugherty SL, Rumsfeld JS, Gurm HS. Change in Hospital-Level Use of Transradial Percutaneous Coronary Intervention and Periprocedural Outcomes Insights from the National Cardiovascular Data Registry. Circ Cardiovasc Qual Outcomes. 2014;7:550-559.Jin C, Li W, Qiao SB, Yang JG, Wang Y, He PY, Tang XR, Dong QT, Li XD, Yan HB, Wu YJ, Chen JL, Gao RL, Yuan JQ, Dou KF, Xu B, Zhao W, Zhang X, Xian Y, Yang YJ. Costs and Benefits Associated With Transradial Versus Transfemoral Percutaneous Coronary Intervention in China. J Am Heart Assoc. 2016;5:e002684.Amin AP, Patterson M, House JA, Giersiefen H, Spertus JA, Baklanov DV, Chhatriwalla AK, Safley DM, Cohen DJ, Rao SV, Marso SP. Costs Associated With Access Site and Same-Day Discharge Among Medicare Beneficiaries Undergoing Percutaneous Coronary Intervention: An Evaluation of the Current Percutaneous Coronary Intervention Care Pathways in the United States. JACC Cardiovasc Interv. 2017;10:342-351.Mitchell MD, Hong JA, Lee BY, Umscheid CA, Bartsch SM, Don CW. Systematic Review and Cost-Benefit Analysis of Radial Artery Access for Coronary Angiography and Intervention. Circ Cardiovasc Qual Outcomes. 2012;5:454-462.National Clinical Guideline Centre. Myocardial infarction with ST-segment elevation. Appendices. 2013. Available from Accessed March 2018.BCIS. Complications rates and mean length of stay by indication for transradial and transfemoral PCI between 2010 and 2014, stratified by Strategic Health Authority. Data on file.Department of Health. National schedule of reference costs 2015-16. Available at . Accessed March 2018.NHS Blood and Transplant. Price list 2015/16. Standard red cells. Available from: . Accessed March 2018.Varney SJ, Guest JF. The annual cost of blood transfusions in the UK. Transfus Med. 2003;13:205-18.National Clinical Guideline Centre. Unstable angina and NSTEMI: The early management of unstable angina and non-ST-segment-elevation myocardial infarction. Clinical guideline 94. 2010. Available at . Accessed March 2018.PSSRU. Unit Costs of Health and Social Care 2016. Available from . Accessed March 2018.Ferrante G, Rao SV, Juni P, Da Costa BR, Reimers B, Condorelli G, Anzuini A, Jolly SS, Bertrand OF, Krucoff MW, Windecker S, Valgimigli M. Radial Versus Femoral Access for Coronary Interventions Across the Entire Spectrum of Patients With Coronary Artery Disease: A Meta-Analysis of Randomized Trials. JACC Cardiovasc Interv. 2016;9:1419-34.Rao SV, Hess CN, Barham B. A registry-based randomized trial comparing radial and femoral approaches in women undergoing percutaneous coronary intervention: the SAFE-PCI for Women (Study of Access Site for Enhancement of PCI for Women) trial. J Am Coll Cardiol Intv 2014;7:857–67.Muthusamy P, Wohns D. Radial Versus Femoral Access in PCI - An economic and quality analysis. Cardiac Interventions Today. 2013;58–65.Roussanov O, Wilson SJ, Henley K, Estacio G, Hill J, Dogan B, Henley WF and Jarmukli N. Cost-effectiveness of the radial versus femoral artery approach to diagnostic cardiac catheterization. J Invasive Cardiol. 2007;19:349-53.Safley DM, Amin AP, House JA, Baklanov D, Mills R, Giersiefen H, Bremer A, Marso SP. Comparison of costs between transradial and transfemoral percutaneous coronary intervention: a cohort analysis from the Premier research database. Am Heart J. 2013;165:303-9.e2.Koltowski L, Filipiak KJ, Kochman J, Pietrasik A, Huczek Z, Balsam P, Lewandowski A, Chojnacka K, Opolski G, Wrona W. Cost-effectiveness of radial vs. femoral approach in primary percutaneous coronary intervention in STEMI - Randomized, control trial. Hellenic J Cardiol. 2016;57:198-202.Amin AP, House JA, Safley DM, Chhatriwalla AK, Giersiefen H, Bremer A, Hamon M, Baklanov DV, Aluko A, Wohns D, Mathias DW, Applegate RA, Cohen DJ, Marso SP. Costs of transradial percutaneous coronary intervention. JACC Cardiovasc Interv. 2013;6:827-34.Vavalle JP, Lopes RD, Chen AY, Newby LK, Wang TY, Shah BR, et al. Hospital length of stay in patients with non-ST-segment elevation myocardial infarction. Am J Med. 2012;125:1085-94.List of Tables and FiguresTable 1: Propensity-matched baseline patient and procedures characteristics, by access site, 2010-2014Table 2: Propensity-matched clinical data, per indication, per yearTable 3: PSM economic model results - costs per procedure for All EnglandTable 4: PSM economic model sensitivity analysis results, total costs per procedure, per indication, per year Table 5: Real-world analysis – estimated total complications, inpatient data, and costs due to increased radial uptake, per SHA, 2010-2014Figure 1: Percentage of PCI procedures that are performed using transradial access by regionFigure 2: Actual savings (millions) based on uptake of TRA (black text) vs possible savings (blue text) or cost increases (red text) if uptake rates had matched those of the region with the highest utilization (North East)Table 1: Propensity-matched baseline patient and procedures characteristics, by access site, 2010-2014Access site: N (%)Variable% non missingAll procedures N=248,228Femoral onlyN=124,114Radial onlyN=124,114Age100Mean (SD)65.5 (12.2)65.6 (12.1)65.4 (12.2)<50, n (%)26,568 (10.7)13,191 (10.6)13,377 (10.8)50-59, n (%)51,819 (20.9)25,729 (20.7)26,090 (21.0)60-69, n (%)71,715 (28.9)35,859 (28.9)35,856 (28.9)70-79, n (%)64,758 (26.1)32,420 (26.1)32,338 (26.1)80+, n (%)33,246 (13.4)16,831 (13.6)16,415 (13.2)Sex, n (%)100Male178,258 (71.8)89,568 (72.2)88,690 (71.5)Female69,970 (28.2)34,546 (27.8)35,424 (28.5)Ethnicity, n (%)83.6Asian14,921 (7.2)8,628 (8.6)6,293 (5.9)Black1,858 (0.9)1,024 (1.0)834 (0.8)White177,414 (85.5)82,311 (81.9)95,103 (88.8)Other13,401 (6.5)8,588 (8.5)4,813 (4.5)Smoking, n (%)89.7Never82,994 (37.3)40,948 (37.7)42,046 (36.9)Former87,321 (39.2)43,085 (39.7)44,236 (38.8)Current52,431 (23.5)24,623 (22.7)27,808 (24.4)Dialysis, n (%)93.52,503 (1.1)1,670 (1.5)833 (0.7)Diabetes, n (%)95.749,580 (20.9)25,228 (21.5)24,352 (20.2)PCI, n (%)97.662,369 (25.7)31,593 (26.3)30,776 (25.2)CABG, n (%)73.620,040 (11.0)12,544 (13.4)7,496 (8.4)MI, n (%)92.56,5987 (28.7)33,799 (30.2)32,188 (27.4)CVA, n (%)95.71,0518 (4.4)5,206 (4.4)5,312 (4.4)HC, n (%)95.7132,024 (55.6)65,735 (55.7)66,289 (55.5)Hypertension, n (%)95.7134,019 (56.4)67,122 (56.9)66,897 (56.0)PVD, n (%)95.71,2608 (5.3)6,480 (5.5)6,128 (5.1)VHD, n (%)95.74,030 (1.7)2,031 (1.7)1,999 (1.7)LVEF, n (%)49.8Good (> 50%)87,220 (70.5)41,342 (68.9)45,878 (72.1)Fair (30-50%)28,870 (23.4)14,621 (24.4)14,249 (22.4)Poor (< 30%)7,544 (6.1)4,050 (6.7)3,494 (5.5)Indication, n (%)100Elective94,856 (38.2)47,428 (38.2)47,428 (38.2)UA/nSTEMI87,904 (35.4)43,952 (35.4)43,952 (35.4)STEMI65,468 (26.4)32,734 (26.4)32,734 (26.4)Shock, n (%)1007,038 (2.8)4,722 (3.8)2,316 (1.9)Ventilated, n (%)88.34,325 (2.0)3,228 (3.0)1,097 (1.0)IABP, n (%)95.74,268 (1.8)3,288 (2.8)980 (0.8)CP support, n (%)95.7331 (0.1)230 (0.2)101 (0.1)Inotropes, n (%)95.72,248 (0.9)1,541 (1.3)707 (0.6)Left-main PCI, n (%)98.910,654 (4.3)6,092 (5.0)4,562 (3.7)Multi-vessel PCI, n (%)98.933,520 (13.6)16,787 (13.7)16,733 (13.6)Stent type, n (%)96.4No stents 18,569 (7.8)9,631 (8.2)8,938 (7.4)BMS only37,853 (15.8)18,085 (15.3)19,768 (16.3)DES only177,146 (74.0)87,715 (74.3)89,431 (73.8)BMS and DES5,706 (2.4)2,651 (2.2)3,055 (2.5)Anti-platelets, n (%)77.6Clopidogrel165,011 (85.7)85,343 (89.5)79,668 (82.0)Prasugrel13,332 (6.9)5,215 (5.5)8,117 (8.4)Ticagrelor13,631 (7.1)4,588 (4.8)9,043 (9.3)Ticlopidine531 (1.3)256 (0.3)275 (0.3)Warfarin, n (%)94.72,707 (1.2)1,119 (0.9)1,588 (1.4)Bivalirudin, n (%)94.711,203 (4.8)4,231 (3.6)6,972 (6.0)GP IIb/IIIa, n (%)95.449,916 (21.1)2,3861 (20.4)2,6055 (21.8)CP, cardio-pulmonary support; CVA, cerebrovascular accident; GP, glycoprotein; HC, hypercholesterolemia; IABP, intra-aortic balloon pump; LVEF, left-ventricular ejection fraction; MI, myocardial infarction; PVD, peripheral vascular disease; VHD, valvular heart disease.Table 2: Propensity-matched clinical data, per indication, per yearIndicationNumber of procedures?Rate of bleeding?Rate of vascular complications?Rate of MACE?Mean length of stay, days?TRATFATRATFATRATFATRATFATRATFA2010STEMI7,1117,1110.0080.0170.0050.0120.0250.0393.3623.735NSTEMI/unstable11,40311,4030.0040.0120.0030.0100.0100.0151.7092.151Elective/stable10,22210,2220.0030.0100.0030.0100.0120.0070.9811.0342011STEMI8,5258,5250.0080.0140.0030.0090.0260.0593.7534.465NSTEMI/unstable10,41410,4140.0050.0150.0050.0140.0090.0201.8992.482Elective/stable11,92911,9290.0020.0110.0030.0110.0070.0080.9621.1832012STEMI7,3827,3820.0060.0160.0030.0080.0380.0843.7384.697NSTEMI/unstable9,1479,1470.0060.0130.0050.0120.0100.0241.8612.564Elective/stable10,75210,7520.0030.0090.0020.0090.0070.0070.9361.1682013STEMI6,0046,0040.0060.0170.0020.0100.0520.1003.7394.740NSTEMI/unstable7,8907,8900.0040.0160.0030.0160.0140.0291.9422.651Elective/stable8,7288,7280.0030.0130.0040.0140.0060.0050.7681.2482014STEMI3,7123,7120.0090.0160.0040.0070.0590.1084.0505.009NSTEMI/unstable5,0985,0980.0030.0150.0090.0130.0120.0311.8052.467Elective/stable5,7975,7970.0040.0150.0080.0170.0070.0060.5870.968Abbreviations: MACE, major adverse cardiac events; NSTEMI, non-ST segment elevation myocardial infarction; STEMI, ST segment elevation myocardial infarction; TFA, transfemoral access; TRA, transradial access.?Source: BCIS data 2014. Table 3: PSM economic model results - costs per procedure for All EnglandCost categorySTEMINSTEMI/unstableElective/stableAll indicationsTRATFADifferenceTRATFADifferenceTRATFADifferenceTRATFADifferenceVascular access site management (all years)Standard Procedure?135.07?198.47-?63.41?135.07?198.47-?63.41?135.07?198.47-?63.41?135.07?198.47-?63.41Access site crossover?5.89?0.87?5.02?9.25?1.17?8.08?8.34?1.09?7.25?8.02?1.06?6.96Hospital LoS2010?1,243.13?1,380.94-?137.82?631.99?713.57-?163.16?382.45?362.79-?19.66?687.46?793.30-?105.842011?1,387.64?1,650.66-?263.02?702.07?917.51-?215.45?355.78?437.37-?81.59?757.58?934.44-?176.852012?1,382.11?1,736.43-?354.32?688.18?947.95-?259.76?345.96?431.73-?85.77?741.08?957.85-?216.782013?1,382.53?1,752.61-?370.08?717.99?980.22-?262.23?283.98?461.58-?177.60?726.91?985.12-?258.202014?1,497.56?1,851.80-?354.24?667.41?912.18-?244.77?217.11?358.04-?140.92?699.67?931.04-?231.38All years?1,366.53?1,652.92-?286.39?679.83?902.74-?222.91?324.90?419.01-?94.11?725.31?915.75-?190.43Complications2010?3.10?6.69-?3.58?1.58?5.00-?3.43?1.16?4.23-?3.07?1.81?5.15-?3.342011?2.92?5.57-?2.65?2.10?6.32-?4.22?1.06?4.79-?3.73?1.93?5.52-?3.602012?2.15?6.03-?3.88?2.31?5.50-?3.18?1.22?3.98-?2.77?1.84?5.04-?3.212013?2.02?6.61-?4.59?1.66?6.85-?5.20?1.24?5.73-?4.48?1.59?6.35-?4.762014?3.15?5.83-?2.68?1.94?6.20-?4.26?1.95?6.52-?4.57?2.25?6.23-?3.98All years?2.65?6.14-?3.49?1.91?5.89-?3.98?1.26?4.87-?3.61?1.86?5.57-?3.71Total costs2010?1,387.19?1,586.97-?199.78?777.88?999.80-?221.92?507.36?586.24-?78.88?832.43?997.99-?165.562011?1,531.52?1,855.57-?324.05?848.48?1,123.48-?274.99?500.24?641.72-?141.47?902.54?1,139.49-?236.942012?1,525.22?1,941.80-?416.58?834.82?1,153.09-?318.27?490.58?635.27-?144.70?885.96?1,162.43-?276.462013?1,525.51?1,958.56-?433.05?863.96?1,186.72-?322.76?428.62?666.86-?238.24?871.58?1,191.00-?319.422014?1,641.67?2,056.97-?415.30?813.66?1,118.02-?304.36?362.46?564.11-?201.65?845.02?1,136.81-?291.79All years?1,510.13?1,858.40-?348.26?826.06?1,108.27-?282.21?469.56?623.44-?153.88?870.25?1,120.84-?250.59Abbreviations: LoS, length of stay; NSTEMI, non-ST segment elevation myocardial infarction; STEMI, ST segment elevation myocardial infarction; TFA, transfemoral access; TRA, transradial access.Table 4: PSM economic model sensitivity analysis results, total costs per procedure, per indication, per yearYearSTEMINSTEMI/unstableElective/stableAll indicationsTRATFADifferenceTRATFADifferenceTRATFADifferenceTRATFADifference2010?1,389.20?1,590.86-?201.65?780.77?1,001.30-?220.54?507.77?590.91-?83.14?834.22?1,001.21-?166.992011?1,533.72?1,860.57-?326.85?850.84?1,123.66-?272.82?500.76?645.83-?145.07?904.15?1,142.52-?238.372012?1,528.19?1,946.34-?418.15?836.96?1,154.10-?317.14?490.94?640.19-?149.25?887.63?1,165.93-?278.302013?1,528.61?1,962.53-?433.92?866.76?1,186.37-?319.61?428.96?670.04-?241.08?873.51?1,193.16-?319.652014?1,643.64?2,061.71-?418.07e?1,118.33-?302.14?362.09?566.50-?204.40?846.25?1,139.06-?292.81All years?1,512.61?1,862.83-?350.23?828.61?1,108.89-?280.28?469.88?627.47-?157.59?871.93?1,123.77-?251.84Abbreviations: LoS, length of stay; NSTEMI, non-ST segment elevation myocardial infarction; STEMI, ST segment elevation myocardial infarction; TFA, transfemoral access; TRA, transradial access.Table 5: Real-world analysis – estimated total complications, inpatient data, and costs due to increased radial uptake, per SHA, 2010-2014YearTransradial uptake remains constant at 2010 levelsResults (A)Transradial uptake based on actual rates between 2010–2014Results (B)Transradial uptake based on region with highest utilization (North East) between 2010-2014Results (C)Events/costs avoided based on actual transradial uptake for 2010-2014 (% change)(B – A)Additional events/costs avoided with uptake based on region with highest utilization (% change)(C– B)East MidlandsInpatient days75,46971,52366,072-3,946 (-5.2%)-5,452 (-7.6%)Major bleeds199162119-36 (-18.1%)-44 (-27.2%)Vascular complications238216184-22 (-9.2%)-32 (-14.8%)MACEs826708563-118 (-14.3%)-145 (-20.5%)Total costs?33,262,476?31,509,176?29,064,881-?1,753,300 (-5.3%)-?2,444,295 (-7.8%)East of EnglandInpatient days78,67471,49667,781-7,177 (-9.1%)-3,715 (-5.2%)Major bleeds340260186-79 (-23.2%)-74 (-28.5%)Vascular complications284221161-63 (-22.2%)-60 (-27.1%)MACEs696551451-145 (-20.8%)-100 (-18.1%)Total costs?35,481,947?32,343,655?30,494,472-?3,138,292 (-8.8%)-?1,849,183 (-5.7%)LondonInpatient days104,490102,27490,665-2,216 (-2.1%)-11,609 (-11.4%)Major bleeds382355216-28 (-7.3%)-139 (-39.2%)Vascular complications451416267-35 (-7.8%)-149 (-35.8%)MACEs1047965619-82 (-7.8%)-347 (-36.0%)Total costs?46,401,691?45,387,362?40,055,898-?1,014,329 (-2.2%)-?5,331,464 (-11.7%)North EastInpatient days40,83140,138N/A-693 (-1.7%)N/AMajor bleeds236226N/A-10 (-4.2%)N/AVascular complications164158N/A-6 (-3.7%)N/AMACEs575533N/A-42 (-7.3%)N/ATotal costs?18,429,298?18,138,237N/A-?291,061 (-1.6%)N/ANorth WestInpatient days83,72880,78879,476-2,940 (-3.5%)-1,312 (-1.6%)Major bleeds178157140-21 (-11.8%)-16 (-10.2%)Vascular complications11910795-12 (-10.1%)-12 (-11.2%)MACEs706593531-113 (-16.0%)-62 (-10.5%)Total costs?36,709,232?35,418,879?34,736,145-?1,290,353 (-3.5%)-?682,734 (-1.9%)South CentralInpatient days52,28650,54648,004-1,740 (-3.3%)-2,542 (-5.0%)Major bleeds305246161-59 (-19.3%)-85 (-34.6%)Vascular complications15412474-30 (-19.5%)-50 (-40.3%)MACEs536465374-71 (-13.2%)-91 (-19.6%)Total costs?23,661,967.96?22,766,399?21,368,548-?895,568 (-3.8%)-?1,397,851 (-6.1%)South East CoastInpatient days60,02859,48159,645-548 (-0.9%)164 (0.3%)Major bleeds214204164-10 (-4.7%)-41 (-20.1%)Vascular complications198189148-9 (-4.5%)-41 (-21.7%)MACEs618570439-48 (-7.8%)-131 (-23.0%)Total costs?27,562,286?27,165,266?26,440,634-?397,021 (-1.4%)-?724,632 (-2.7%)South WestInpatient days76,92274,63370,823-2,289 (-3.0%)-3,810 (-5.1%)Major bleeds349278164-71 (-20.3%)-114 (-41.0%)Vascular complications322256165-66 (-20.5%)-91 (-35.5%)MACEs745649491-96 (-12.9%)-158 (-24.3%)Total costs?35,148,670?33,991,417?32,064,673-?1,157,253 (-3.3%)-?1,926,744 (-5.7%)West MidlandsInpatient days84,45881,53779,731-2,920 (-3.5%)-1,806 (-2.2%)Major bleeds405328277-78 (-19.3%)-51 (-15.5%)Vascular complications242196169-46 (-19.0%)-27 (-13.8%)MACEs1048868769-180 (-17.2%)-99 (-11.4%)Total costs?36,918,857?35,582,292?34,752,712-?1,336,565 (-3.6%)-?829,580 (-2.3%)Yorkshire and the HumberInpatient days63,88358,00652,157-5,877 (-9.2%)-5,849 (-10.1%)Major bleeds469334231-135 (-28.8%)-103 (-30.8%)Vascular complications291216140-75 (-25.8%)-76 (-35.2%)MACEs954746615-208 (-21.8%)-131 (-17.6%)Total costs?28,871,345?26,347,346?23,785,984-?2,523,999 (-8.7%)-?2,561,362 (-9.7%)All EnglandInpatient days719,923690,675648,451-29,248 (-4.1%)-42,224 (-6.1%)Major bleeds297625371822-440 (-14.8%)-715 (-28.2%)Vascular complications245420991515-355 (-14.5%)-584 (-27.8%)MACEs768566365311-1,048 (-13.6%)-1,326 (-20.0%)Total costs?322,045,448?308,739,349?288,644,612?294,757,559-?13,306,098 (-4.1%)-?20,094,737 (-6.5%)Abbreviations: MACEs, major adverse cardiac events; N/A, not applicable. Figure 1: Percentage of PCI procedures that are performed using transradial access by region?Calculated as the percentage of radial and femoral procedures. Some procedures were classified as ‘multiple/other’; when the access site was not clear or more than one site was used in the same procedure these data were excluded from the analysis.Figure 2: Actual savings (millions) based on uptake of TRA (black text) vs possible savings (blue text) or cost increases (red text) if uptake rates had matched those of the region with the highest utilization (North East)Supplemental materialSupplementary Table SEQ Table \* ARABIC 1: Crossover rates per indicationIndicationRate of access site crossover§TRATFASTEMI5.34%1.60%NSTEMI/unstable8.39%2.15%Elective/stable7.56%1.99%Abbreviations: NSTEMI, non-ST segment elevation myocardial infarction; STEMI, ST segment elevation myocardial infarction; TFA, transfemoral access; TRA, transradial access. §Source: Jolly et al1Supplementary Table SEQ Table \* ARABIC 2: Costs included in the modelProcedure costsVascular access site management TransradialTransfemoralStandard procedure? ?135.07?198.47Access site crossover§?110.32?54.47Hospital stayCost per dayNon-elective inpatients (long stay) excess bed days2??369.73Cost of complicationsMajor bleedingUnit costAdditional unitsTotalBlood transfusion3,4??329.001?329.00Total?2,547.36Vascular complicationsUnit cost% of events??Total costUltrasound2,5?74.8550%?37.42Compression device5,6§§?56.8612.5%?7.11Thrombin injection5,6§§?258.4411.25%?29.07Vascular surgery2,5?1,594.571.25%?19.93Total?93.54?These costs are composed of an introducer kit, a diagnostic guidewire, a guiding catheter, and a closure device (used in 70% of TRA and 80% of TFA procedures). §These costs are made up of an introducer kit (used in 70% of procedures), a diagnostic guidewire (used in 30% of procedures), a guiding catheter (used in 15% of procedures), and a closure device (used in 70% of TRA and 80% of TFA procedures). ?Weighted average of currency codes: EY40A – Complex percutaneous transluminal coronary angioplasty with complication and comorbidity (CC) score 12+; EY40B – Complex percutaneous transluminal coronary angioplasty with CC score 8–11; EY40C – Complex percutaneous transluminal coronary angioplasty with CC score 4–7; EY40D – Complex percutaneous transluminal coronary angioplasty with CC score 0–3; EY41A – Standard percutaneous transluminal coronary angioplasty with CC score 12+; EY41B – Standard percutaneous transluminal coronary angioplasty with CC score 8–11; EY41C – Standard percutaneous transluminal coronary angioplasty with CC score 4–7; and EY41D – Standard percutaneous transluminal angioplasty with CC score 0–3. ?Calculated as the cost of a unit of red blood cells3 multiplied by the average number of red blood cells used in a transfusion4. ??Percentage utilization obtained from the National Clinical Guideline Centre cost analysis,5 which states that “It is assumed that 50% of patients with a hematoma will received an ultrasound to check for pseudoaneurysm (false aneurysm). It is assumed that 50% of those given an ultrasound will be diagnosed with pseudoaneurysm. It is assumed that 50% of those diagnosed with pseudoaneurysm will be treated with compression, requiring an external compression device; that 45% will be treated by injection with thrombin; that 5% would require surgery.” §§Costs are assumed to be the midpoint of the range reported in the cost analysis from the National Clinical Guideline Centre for patient presenting with STEMI5, and inflated using PSSRU inflation data6.Supplementary Table 3: Baseline patient and procedures characteristics, by access site, 2010-2014 (full BCIS data set)Access site: N (%)Variable% non missingAll procedures N=323,656Femoral onlyN=130,894Radial onlyN=183,360Multiple/otherN=9,402Age99.9Mean (SD)64.9 (12.1)65.7 (12.1)64.3 (12.0)66.0 (12.6)<50, n (%)36,505 (11.3)13,629 (10.4)21,835 (11.9)1,041 (11.1)50-59, n (%)71,001 (21.9)26,724 (20.4)42,384 (23.1)1,893 (20.1)60-69, n (%)95,210 (29.4)37,756 (28.9)54,905 (30.0)2,549 (27.1)70-79, n (%)81,457 (25.2)34,785 (26.6)44,219 (24.1)2,453 (26.1)80+, n (%)39,305 (12.2)17,912 (13.7)19,930 (10.9)1,463 (15.6)Sex, n (%)99.9Male240,091 (74.3)93,754 (71.7)139,854 (76.4)6,483 (69.0)Female 83,085 (25.7)36,967 (28.3)43,208 (23.6)2,910 (31.0)Ethnicity, n (%)83.2Asian 18,477 ( 6.9)9,103 (8.6)8,934 (5.7)440 (5.7)Black 2,259 (0.8)1,064 (1.0)1,157 (0.7)38 (0.5)White23,2301 (86.2)86,996 (82.2)13,8577 (88.9)6,728 (87.7)Other 16,329 (6.1)8,728 (8.2)7,137 (4.6)464 (6.0)Smoking, n (%)90.0Never106,820 (36.7)43,316 (37.9)60,369 (35.8)3,135 (37.4)Former112,578 (38.7)45,538 (39.8)63,736 (37.8)3,304 (39.4)Current 71,810 (24.7)25,478 (22.3)44,387 (26.3)1,945 (23.2)Dialysis, n (%)93.5 2,783 (0.9)1,880 (1.6)837 (0.5)66 (0.8)Diabetes, n (%)95.7 62,621 (20.2)26,716 (21.6)34,023 (19.2)1,882 (20.8)PCI, n (%)97.8 74,872 (23.7)34,327 (27.1)38,238 (21.2)2,307 (25.0)CABG, n (%)75.8 23,140 (9.4)14,912 (15.1)7,514 (5.4)714 (9.8)MI, n (%)92.8 80,432 (26.8)36,319 (30.7)41,325 (23.9)2,788 (30.7)CVA, n (%)95.5 13,177 (4.3)5,435 (4.4)7,338 (4.2)404 (4.5)HC, n (%)95.5170,910 (55.3)69,961 (56.2)96,021 (54.6)4,928 (55.5)Hypertension, n (%)95.5171,065 (55.3)70,846 (56.9)95,113 (54.1)5,106 (57.5)PVD, n (%)95.5 15,587 (5.0)6,779 (5.4)8,155 (4.6)653 (7.4)VHD, n (%)95.5 4,776 (1.5)2,129 (1.7)2,473 (1.4)174 (2.0)LVEF, n (%)49.3Good (> 50%)112,706 (70.6)43,263 (68.7)66,397 (72.1)3,046 (65.7)Fair (30-50%) 37,497 (23.5)15,446 (24.5)20,897 (22.7)1,154 (24.9)Poor (< 30%) 9,425 (5.9)4,257 (6.8)4,734 (5.1)434 (9.4)Indication, n (%)100Elective116,940 (36.1)52,117 (39.8)61,144 (33.3)3,679 (39.1)UA/nSTEMI117,601 (36.3)44,489 (34.0)69,981 (38.2)3,131 (33.3)STEMI 89,115 (27.5)34,288 (26.2)52,235 (28.5)2,592 (27.6)Shock, n (%)100 8,156 (2.5)5,282 (4.0)2,318 (1.3)556 (5.9)Ventilated, n (%)87.8 5,110 (1.8)3,421 (3.1)1,470 (0.9)219 (2.6)IABP, n (%)95.7 5,413 (1.7)3,596 (2.9)1,188 (0.7)629 (6.9)CP support, n (%)95.7 388 (0.1)248 (0.2)116 (0.1)24 (0.3)Inotropes, n (%)95.7 2,770 (0.9)1685 (1.4)853 (0.5)232 (2.6)Left-main PCI, n (%)99.0 12,595 (3.9)6468 (5.0)5542 (3.1)585 (6.3)Multi-vessel PCI, n (%)99.0 42,399 (13.2)17712 (13.7)23283 (12.8)1404 (15.1)Stent type, n (%)96.4No stents 23,215 (7.4)10,254 (8.2)11,939 (6.7)1,022 (11.2)BMS only 46,482 (14.9)19,345 (15.6)25,770 (14.4)1,367 (15.0)DES only234,853 (75.3)91,927 (73.9)136,431 (76.5)6,495 (71.2)BMS and DES 7,398 (2.4)2,872 (2.3)4,288 (2.4)238 (2.6)Anti-platelets, n (%)78.0Clopidogrel211,791 (83.9)90,273 (89.9)115,162 (79.8)6,356 (80.9)Prasugrel 19,492 (7.7)5,259 (5.2)13,504 (9.4)729 (9.3)Ticagrelor 20,607 (8.2)4,600 (4.6)15,254 (10.6)753 (9.6)Ticlopidine 666 (0.3)264 (0.3)387 (0.3)15 (0.2)Warfarin, n (%)95.2 3,429 (1.1)1,196 (1.0)2,146 (1.2)87 (0.9)Bivalirudin, n (%)95.2 14,702 (4.8)4,313 (3.4)9,896 (5.7)493 (5.4)GP IIb/IIIa, n (%)95.6 65,873 (21.3)25,073 (20.3)39,049 (22.1)1,751 (19.0)CP, cardio-pulmonary support; CVA, cerebrovascular accident; GP, glycoprotein; HC, hypercholesterolemia; IABP, intra-aortic balloon pump; LVEF, left-ventricular ejection fraction; MI, myocardial infarction; PVD, peripheral vascular disease; VHD, valvular heart disease.Supplementary Table SEQ Table \* ARABIC 4: Complication data used in PSM economic model sensitivity analysesIndicationRate of bleeding?Rate of vascular complications?Rate of MACE?TRATFATRATFATRATFASTEMI0.0150.0290.003?0.010?0.0520.063NSTEMI/unstable0.0130.0170.003?0.010?0.0690.077Elective/stable0.0040.0210.0020.0220.0280.041Abbreviations: MACE, major adverse cardiac events; NSTEMI, non-ST segment elevation myocardial infarction; STEMI, ST segment elevation myocardial infarction; TFA, transfemoral access; TRA, transradial access. ?Source: Ferrante et al7. ?Vascular complications are reported for patients with acute coronary syndrome (ACS); therefore, the overall ACS rates are assumed to apply to both NSTEMI/unstable patients and STEMI patients.Supplementary Table SEQ Table \* ARABIC 5: Real-world economic model analysis – estimated total costs due to increased radial uptake, per SHA, per yearYearTransradial uptake remains constant at 2010 levelsResults (A)Transradial uptake based on actual rates between 2010–2014Results (B)Transradial uptake based on region with highest utilization (North East) between 2010-2014Results (C)Events/costs avoided based on actual transradial uptake for 2010-2014 (% change)(B – A)Additional events/costs avoided with uptake based on region with highest utilization (% change)(C– B)East Midlands2010?5,255,456 ?5,255,456?4,806,272N/A-?449,184 (-8.5%)2011?7,017,347?6,904,789?6,175,120-?112,558 (-1.6%)-?729,669 (-10.6%)2012?7,457,874?6,916,766?6,201,824-?541,107 (-7.3%)-?714,942 (-10.3%)2013?6,785,678?6,297,560?6,025,669-?488,118 (-7.2%)-?271,891 (-4.3%)2014?6,746,121?6,134,604?5,855,995-?611,517 (-9.1%)-?278,609 (-4.5%)Total?33,262,476?31,509,176?29,064,881-?1,753,300 (-5.3%)-?2,444,295 (-7.8%)East of England2010?6,347,575?6,347,575?5,967,869N/A-?379,706 (-6.0%)2011?6,593,125?6,414,821?5,954,841-?178,304 (-2.7%)-?459,979 (-7.2%)2012?7,866,737?7,119,207?6,578,024-?747,530 (-9.5%)-?541,184 (-7.6%)2013?7,946,993?7,334,377?7,128,319-?612,616 (-7.7%)-?206,058 (-2.8%)2014?6,727,517?5,127,675?4,865,419-?1,599,843 (-23.8%)-?262,256 (-5.1%)Total?35,481,947?32,343,655?30,494,472-?3,138,292 (-8.8%)-?1,849,183 (-5.7%)London2010?8,511,811?8,511,811?7,395,345N/A-?1,116,466 (-13.1%)2011?9,794,291?9,702,708?8,104,635-?91,583 (-0.9%)-?1,598,073 (-16.5%)2012?10,669,064?10,530,499?9,781,699-?138,564 (-1.3%)-?748,800 (-7.1%)2013?11,787,252?11,283,478?10,218,266-?503,774 (-4.3%)-?1,065,212 (-9.4%)2014?5,639,273?5,358,866?4,555,953-?280,407 (-5.0%)-?802,913 (-15.0%)Total?46,401,691?45,387,362?40,055,898-?1,014,329 (-2.2%)-?5,331,464 (-11.7%)North East2010?3,806,672?3,806,672N/AN/AN/A2011?3,777,530?3,723,250N/A-?54,280 (-1.4%)N/A2012?4,124,973?4,073,273N/A-?51,700 (-1.3%)N/A2013?3,742,491?3,637,878N/A-?104,613 (-2.8%)N/A2014?2,977,632?2,897,164N/A-?80,469 (-2.7%)N/ATotal?18,429,298?18,138,237N/A-?291,061 (-1.6%)N/ANorth West2010?5,964,772?5,964,772?5,727,654N/A-?237,118 (-4.0%)2011?9,590,853?9,595,843?9,609,858?4,990 (0.1%)?14,015 (0.1%)2012?8,009,591?7,730,174?7,573,204-?279,417 (-3.5%)-?156,971 (-2.0%)2013?7,436,962?6,904,625?6,710,193-?532,337 (-7.2%)-?194,432 (-2.8%)2014?5,707,053?5,223,464?5,115,236-?483,589 (-8.5%)-?108,228 (-2.1%)Total?36,709,232?35,418,879?34,736,145-?1,290,353 (-3.5%)-?682,734 (-1.9%)South Central2010?4,762,450?4,762,450?4,384,073N/A-?378,378 (-7.9%)2011?5,262,408?5,161,673?4,712,675-?100,735 (-1.9%)-?448,998 (-8.7%)2012?5,147,135?5,036,249?4,902,961-?110,886 (-2.2%)-?133,288 (-2.6%)2013?5,474,188?5,073,385?4,803,671-?400,803 (-7.3%)-?269,713 (-5.3%)2014?3,015,787?2,732,642?2,565,168-?283,145 (-9.4%)-?167,474 (-6.1%)Total?23,661,967.96?22,766,399?21,368,548-?895,568 (-3.8%)-?1,397,851 (-6.1%)South East Coast2010?4,932,247?4,932,247?5,460,401N/A?528,154 (10.7%)2011?5,544,375?5,478,646?5,039,577-?65,729 (-1.2%)-?439,069 (-8.0%)2012?6,253,172?6,184,835?5,806,311-?68,336 (-1.1%)-?378,524 (-6.1%)2013?6,475,331?6,325,318?6,076,616-?150,013 (-2.3%)-?248,702 (-3.9%)2014?4,357,163?4,244,220?4,057,728-?112,943 (-2.6%)-?186,491 (-4.4%)Total?27,562,286?27,165,266?26,440,634-?397,021 (-1.4%)-?724,632 (-2.7%)South West2010?6,207,196?6,207,196?5,769,509N/A-?437,687 (-7.1%)2011?7,570,318?7,380,377?6,861,570-?189,942 (-2.5%)-?518,807 (-7.0%)2012?7,816,388?7,577,720?7,193,286-?238,669 (-3.1%)-?384,434 (-5.1%)2013?8,200,906?7,696,929?7,262,225-?503,977 (-6.1%)-?434,703 (-5.6%)2014?5,353,861?5,129,196?4,978,083-?224,666 (-4.2%)-?151,113 (-2.9%)Total?35,148,670?33,991,417?32,064,673-?1,157,253 (-3.3%)-?1,926,744 (-5.7%)West Midlands2010?7,745,869?7,745,869?7,501,102N/A-?244,766 (-3.2%)2011?7,733,908?7,460,522?7,180,534-?273,386 (-3.5%)-?279,988 (-3.8%)2012?8,339,520?8,029,438?7,899,531-?310,081 (-3.7%)-?129,908 (-1.6%)2013?8,262,767?7,856,684?7,713,997-?406,083 (-4.9%)-?142,687 (-1.8%)2014?4,836,794?4,489,779?4,457,548-?347,015 (-7.2%)-?32,232 (-0.7%)Total?36,918,857?35,582,292?34,752,712-?1,336,565 (-3.6%)-?829,580 (-2.3%)Yorkshire and the Humber2010?5,502,647?5,502,647?4,774,332N/A-?728,315 (-13.2%)2011?5,982,504?5,698,796?5,012,881-?283,708 (-4.7%)-?685,915 (-12.0%)2012?6,628,572?5,881,762?5,284,304-?746,810 (-11.3%)-?597,458 (-10.2%)2013?6,561,314?5,695,299?5,321,002-?866,015 (-13.2%)-?374,297 (-6.6%)2014?4,196,308?3,568,843?3,393,466-?627,465 (-15.0%)-?175,377 (-4.9%)Total?28,871,345?26,347,346?23,785,984-?2,523,999 (-8.7%)-?2,561,362 (-9.7%)All England2010?59,057,572?59,057,572?54,844,518N/A-?4,213,053 (-7.1%)2011?68,859,682?67,572,480?62,683,268-?1,287,202 (-1.9%)-?4,889,212 (-7.2%)2012?72,194,708?69,120,681?64,565,679-?3,074,027 (-4.3%)-?4,555,002 (-6.6%)2013?73,076,657?68,124,303?64,155,515-?4,952,353 (-6.8%)-?3,968,788 (-5.8%)2014?48,856,829?44,864,313?42,395,631-?3,992,516 (-8.2%)-?2,468,682 (-5.5%)Total?322,045,448?308,739,349?288,644,612-?13,306,098 (-4.1%)-?20,094,737 (-6.5%)Abbreviations: N/A, not applicable.Supplementary Table SEQ Table \* ARABIC 6: Real-world economic model analysis – estimated total complications, inpatient days, and costs due to increased radial uptake for All England, 2010-2014 (propensity matched inputs)Transradial uptake remains constant at 2010 levelsResults (A)Transradial uptake based on actual rates between 2010–2014Results (B)Transradial uptake based on region with highest utilization (North East) between 2010-2014Results (C)Events/costs avoided based on actual transradial uptake for 2010-2014 (% change)(B – A)Additional events/costs avoided with uptake based on region with highest utilization (% change)(C – B)Inpatient days723,269696,461658,609-26,808 (-3.7%)-37,852 (-5.4%)Major bleeds3,0362,6271,960-409 (-13.5%)-667 (-25.4%)Vascular complications2,5502,2341,681-316 (-12.4%)-553 (-24.7%)MACEs8,0207,2136,202-807 (-10.1%)-1,011 (-14.0%)Total costs?323,311,528?310,921,144?292,461,215-?12,390,384 (-3.8%)-?18,459,929 (-5.9%)Supplementary Table SEQ Table \* ARABIC 7: Real-world economic model analysis – estimated total costs due to increased radial uptake for All England, per year (propensity matched inputs)YearTransradial uptake remains constant at 2010 levelsResults (A)Transradial uptake based on actual rates between 2010–2014Results (B)Transradial uptake based on region with highest utilization (North East) between 2010-2014Results (C)Events/costs avoided based on actual transradial uptake for 2010-2014 (% change)(B – A)Additional events/costs avoided with uptake based on region with highest utilization (% change)(C– B)2010?57,977,765?57,977,765?54,447,867N/A-?3,529,898 (-6.1%)2011?69,288,416?68,059,878?63,394,695-?1,228,538 (-1.8%)-?4,665,183 (-6.9%)2012?72,661,404?69,731,610?65,402,802-?2,929,794 (-4.0%)-?4,328,808 (-6.2%)2013?73,543,692?68,766,298?64,958,187-?4,777,394 (-6.5%)-?3,808,111 (-5.5%)2014?49,840,250?46,385,593?44,257,664-?3,454,658 (-6.9%)-?2,127,929 (-4.6%)Total?323,311,528?310,921,144?292,461,215-?12,390,384 (-3.8%)-?18,459,929 (-5.9%)Abbreviations: N/A, not applicable.Supplementary Figure 1: Example flow diagram of economic model calculations in the basecase analysis (based on 2014 STEMI procedures)Abbreviations: TFA, transfemoral access; TRA, transradial access; STEMI, ST segment elevation myocardial infarction.*Note: costs may not sum to total due to rounding.Supplementary referencesJolly SS, Yusuf S, Cairns J, Niemela K, Xavier D, Widimsky P, Budaj A, Niemela M, Valentin V, Lewis BS, Avezum A, Steg PG, Rao SV, Gao P, Afzal R, Joyner CD, Chrolavicius S, Mehta SR. Radial versus femoral access for coronary angiography and intervention in patients with acute coronary syndromes (RIVAL): a randomised, parallel group, multicentre trial. Lancet. 2011;377:1409-20.Department of Health. National schedule of reference costs 2015-16. Available at . Accessed March 2018.NHS Blood and Transplant. Price list 2015/16. Standard red cells. Available from: . Accessed July 2015.Varney SJ, Guest JF. The annual cost of blood transfusions in the UK. Transfus Med. 2003;13:205-18.National Clinical Guideline Centre. Myocardial infarction with ST-segment elevation. Appendices. 2013. Available from Accessed March 2018.PSSRU. Unit Costs of Health and Social Care 2016. Available from . Accessed March 2018.Ferrante G, Rao SV, Juni P, Da Costa BR, Reimers B, Condorelli G, Anzuini A, Jolly SS, Bertrand OF, Krucoff MW, Windecker S, Valgimigli M. Radial Versus Femoral Access for Coronary Interventions Across the Entire Spectrum of Patients With Coronary Artery Disease: A Meta-Analysis of Randomized Trials. JACC Cardiovasc Interv. 2016;9:1419-34. ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download