National health expenditure per capita is associated with CRT implantation practice: findings from the ESC CRT Survey II with 11 088 patients

Abstract

Aims

Cardiac resynchronization therapy (CRT) reduces morbidity and mortality in selected patients with heart failure (HF) and electrical dyssynchrony. This treatment receives class IA recommendations in European Society of Cardiology (ESC) guidelines. However, despite these strong recommendations, CRT implantation practice varies greatly in Europe. The purpose of the sub-analysis of CRT Survey II data was to describe how countries’ health per capita expenditure affects CRT implantation practice.

Methods and results

Between 2015 and 2016, two ESC associations, European Heart Rhythm Association and Heart Failure Association, conducted the CRT Survey II, a survey of CRT implantations in 11 088 patients in 42 ESC member states. We analysed CRT patient selection and guideline adherence in those countries according to high or low health expenditure per capita. There were 21 high health expenditure countries (n = 6844 patients) and 21 (n = 3852) with low health expenditure. The countries with the lowest health expenditure were more likely to implant CRT in patients who had strong guideline recommendations for implantation, younger patients and those recently hospitalized for HF or with symptomatic HF (67% vs. 58%, P < 0.001). The ratio of CRT-Pacemaker (CRT-P) to CRT-Defibrillator (CRT-D) was similar in both spending groups, as was the percentage of CRT implantation in women.

Conclusion

CRT Survey II has demonstrated a non-uniform delivery of healthcare. Countries with low health expenditure per capita appear to be reserving CRT therapy for younger patients, those with class IA indication and patients with more severe symptoms of heart failure.

Introduction

Cardiac resynchronization therapy (CRT) reduces morbidity and mortality in selected patients with heart failure (HF) and electrical dyssynchrony. This treatment receives Class IA recommendations in European Society of Cardiology (ESC) guidelines, including the European Hearth Rhythm Association (EHRA) guidelines on cardiac pacing and CRT and the Heart Failure Association (HFA) Guidelines on Diagnosis and Treatment of Acute and Chronic Heart Failure.^1–4 However, despite these strong recommendations, CRT implantation practice varies greatly in Europe.⁵ Such variation in delivery of medical technology and cardiac implantable devices is consistent with other large studies.^6–8 There are several factors postulated to explain such implantation variation including the high upfront cost of implanting a CRT device within the context of a limited health budget, inadequate reimbursement, limited device implantation training, and unfamiliarity with guidelines.^9–11

In CRT Survey II, we collected data on 11 088 patients implanted with CRT during 2015–16 in 42 countries. These data demonstrate large variations in adherence to the ESC Guidelines recommendations regarding CRT.^3,4,12 We have speculated that this could be related to funding rather than guideline awareness. In this sub-analysis of the CRT Survey data, we describe how countries’ health expenditure per capita affects CRT implantation practice and guidelines adherence.

Methods

Study design

The CRT survey was jointly conducted by two ESC associations, EHRA and HFA following the success of CRT Survey I.¹³ The rational, study design, and main findings of the survey have been published previously.^14,15 The national cardiology societies in 47 ESC member countries were identified using the EHRA White Book 2014 and were requested to participate in the CRT Survey II.¹⁶ There were 42 countries that participated with 288 centres and 11 088 patients. Only de novo CRT implantations or upgrades from a previous pacemaker or implantable cardioverter defibrillator (ICD) were included in the survey.

Data collection and analyses

The Institut für Herzinfarktforschung Ludwigshafen (IHF) together with the CRT Survey II Scientific Committee revised and modified the EBogen electronic case report form (eCRF) from the first ESC CRT Survey and created the statistical analysis plan.^13,14 The IHF was responsible for data collection, monitoring and verification. Stavanger University Hospital was responsible for the logistics of running the survey.

Survey population

Implanting centres were requested to complete an eCRF for all consecutive patients implanted with a CRT device, both de novo procedures and upgrades. All patients were eligible for inclusion whether or not the procedure was successful (intention to treat). The eCRF included information regarding patient characteristics, indications for CRT, implant procedures and short-term outcomes including adverse events during hospitalization.⁵ Information following discharge from hospital was not collected. The eCRF was designed to be completely anonymized to facilitate recruitment and comply with Institutional Review Board regulations. Furthermore, each implanting centre was requested to complete a one-time site questionnaire. This questionnaire described hospital type, size, population served, infrastructure, facilities, and implantation routines for the CRT device programme.

Statistical analysis

All data analyses were performed by IHF using SAS^®, release 9.3 (SAS Institute Inc., Cary, NC, USA) on a Microsoft^® Windows^® 7 Enterprise platform. All percentages were relative to the total number of patients with available information. P-values were calculated using the Mann–Whitney–Wilcoxon test when comparing medians and chi square-test when comparing percentages. Missing values were not inputted. Using the World Health Organization (WHO) global health data, the 42 countries were classified into two groups by their level of health expenditure per capita expressed in Purchasing Power Parities as percentage of gross domestic product (GDP): greater or lower than the median.¹⁷ We then compared these two groups, looking specifically at patient demographic pertaining to ESC guidelines on CRT during the survey period in the different health spending categories.^3,4 Both these guidelines were updated in 2021. However, the indications for CRT that were strengthened from 2013 to 2016 have remained as strong in 2021.^1,2

Results

Health expenditure per capita

From the 42 countries included in CRT Survey II, 21 countries (6844 patients) were included where health expenditure per capita was high and 21 countries, (3852 patients) with low health expenditure per capita (Table 1). For most of these countries, the percentage spent on health mirrors their countries GDP level.¹⁸

Implantation practice

CRT Implantations according to health expenditure per capita is shown in Table 2 and Figure 1. The countries with the lowest health expenditure, compared to those with the highest, implanted more CRT in patients that had strong guidelines recommendations for implantation such as symptomatic HF— New York Heart Association (NYHA) Classes II–IV (99% vs. 95%, P < 0.001), LVEF ≤35% (89% vs. 86%, P < 0.001), sinus rhythm (72% vs. 68%, P < 0.001), QRS duration ≥150 ms (73% vs. 66%, P < 0.001) and LBBB (75% vs. 72%, P < 0.001). Patients with all the above characteristics have a recommendation class I, evidence level A indication for CRT in both the ESC CRT guidelines from the HFA (2016) and EHRA (2013).^3,4 These recommendations have stayed as strong in the 2021 guidelines from both HFA and EHRA.²

Overall, 67% of patients were implanted with indications with class IA recommendation class I, in the lowest health spending per capita countries vs. 58% in the highest spending group.

For CRT indications with different recommendation strengths in the ESC EHRA 2013 and ESC HFA 2016 guidelines such as upgrades from a previous device and pacemaker indication with expected RV pacing dependence, these weaker indications were less frequent in countries with the lowest expenditure on health (20% vs. 24%, P < 0.001 and 20% vs. 24%, P < 0.001 respectively). Furthermore, countries with lower health expenditure compared with those with higher health expenditure implanted a greater percentage of patients symptomatic of HF, with 53% vs. 43%, P < 0.001 having been hospitalized for HF during the past year and 66% vs. 55%, P < 0.001 being in NYHA Class III or IV. These countries also implanted less CRT in elderly patients compared with the group with the higher health expenditure per capita (median age 67 years vs. 71 years, P < 0.001 with 24% vs. 37%, P < 0.001 of patients ≥75 years of age). The percentage of females implanted with a device on the other hand was similar in the low and high health expenditure countries (24% vs. 25%, P = 0.315).

CRT-P vs. CRT-D

The ratio of CRT-Pacemaker (CRT-P) to CRT-Defibrillator (CRT-D) implants was similar in both spending groups with CRT-P percentages being 31% in the lower health spending group and 30% in the highest spending group.

Hospital stay and remote monitoring

Furthermore, patients in high health expenditure countries had a 1 day shorter hospital stay [4 days (low) vs. 3 days (high)]. These patients were also more frequently discharged with remote device follow-up monitoring [13% (low) vs. 40% (high), P < 0.001].

Discussion

Factors influencing health resources utilization

This sub-analysis of data from CRT Survey II has shown that countries with the lowest percentage of their budget allocated to health are much more likely to implant CRT in patients with class I A indications for CRT, in younger patients and in patients with more severe HF symptoms. However, fewer patients with NYHA II HF were implanted with CRT in low health spending countries compared with countries with high health spending. For CRT indications with different recommendation strengths in the 2013 ESC EHRA and 2016 ESC HFA guidelines, such as upgrades from a previous device and pacemaker indication with expected RV pacing dependence, these indications were less frequent in countries with the lowest expenditure on health. This suggests that the high guideline adherence may not be related to compliance but rather to the limited resources to explore indications with less strong guideline recommendations. Our data also suggest that countries with lower health expenditure are limiting CRT devices not only to those with the strongest recommendation class but also to the patients with most symptoms of HF.

Furthermore, the countries with the lowest health expenditure were also less likely to implant CRT in elderly patients, thereby perhaps actively prioritizing these devices to younger patients. On the other hand, the ratio of CRT-P to CRT-D implants was similar in the spending groups. The reason for this is not apparent but one could speculate that once selected for CRT in certain countries, the physician is free to decide on CRT-D or P without justifying the extra cost for an ICD. Furthermore, countries that spend less on health, implant more patients with ischaemic HF than those countries that spend more (49% vs. 42%, P < 0.001) which may have favoured the CRT-D choice. Therefore, a further hypothesis could be that low health expenditure countries prioritized care for ischaemic heart disease and thus a CRT-D may be chosen for primary prevention in patients with HF. Moreover, adherence to guidelines indicated that HF medication was significantly stronger in low than in high health expenditure countries, which is in keeping with guidelines underlining that patients eligible for CRT must be on guidelines’ indicated medication. This difference in HF medication rates also demonstrates that the same economic constraints that apply for devices do not apply for these medications.

Guideline adherence

When discussing guideline adherence, it is helpful to consider two separate categories of apparent non-adherence. One of these is non-optimal patient care, such as implanting patients with CRT under guideline recommendation level III, where there is evidence that the treatment is not useful, and in some cases may be harmful.³ Fortunately, CRT Survey II found little evidence of such practice in Europe.

The other type of non-adherence is exploring indications, where one believes the intervention may likely be beneficial to the patient despite lacking a strong evidence base. These include IIa and IIb guideline recommendations levels such as CRT in patients with permanent atrial fibrillation. It appears that such indications, including implanting CRT in patients ≥75 years, are more common in countries that dedicate more of their national budget to health. Therefore, guideline adherence is not necessarily related to how well the physician knows and respects the guidelines but rather how much economic leeway and reimbursement is available for CRT implantations in patients with recommendation class II level indications.

Unwarranted variation

CRT Survey II has demonstrated unwarranted variation in CRT delivery. The term ‘unwarranted variation’ has been coined to describe variations in medical practice between geographical regions or provider groups (hospitals of physicians) that cannot be explained by the patient's morbidity, risk factors or preferences.^19,20 Such ununiformed delivery of healthcare demonstrated in CRT Survey II is unfortunately not unique; previous studies have shown that patients’ morbidity is not the main determinant of healthcare utilization and expenditure; rather, it is geography.^19,21,22 Global comparison analysis of HF treatment in the ASCEND-HF, ASTRONAUT, and PARADIGM-HF trials all showed large differences in percentage of HF patients that were implanted with CRT and ICD/CRT-D between continents.^23–25 The sub-analysis of the ASTRONAUT trial on acute HF treatment found that rates of CRT were 11–12% in North America and Western Europe but only 3–5% in the other regions.²⁵ Furthermore, evidence suggests that the problem of unequal health delivery is not only related to differences in individual countries but that there also exists uneven healthcare delivery between centres and physicians for similar medical conditions.¹⁹ In Europe, a collaboration has been initiated to understand and address these variations in medical practice—the European Collaboration for Healthcare Optimization.²²

An important explanation for the large variations in CRT delivery demonstrated in CRT Survey II and worldwide is that these devices are not reimbursed in all countries or only reimbursed within limited indications. These variations in reimbursement regimes contribute to the variations in implementation of CRT. HF treatment with CRT requires a large infrastructure to select, implant, and follow-up CRT devices and thus implantation rates per capita are usually higher in well-resourced healthcare systems.²⁴

Another explanation for unwarranted variation could be related to the level of education and individual understanding of health failure in the population. A study looking at variations in cardiac implantable electronic devices (including CRTs but also pacemakers and ICDs in EU countries) found that level of tertiary education among the regional labour force was positively associated with implantation rates.²⁶ The Swedish HF registry also found similar results regarding higher education and CRT.²⁷

Limitations

CRT Survey II contains data on 11 088 patients implanted with a CRT. We have estimated that about 11% of patients implanted with CRT devices in participating countries during the recruitment period were enrolled in the survey. However, recruitment rates between countries varied considerably. As participation in the survey was voluntary, we are not able to adequately assess selection bias. Investigators were encouraged to include patients consecutively, but this cannot be verified. It is possible that implanters were less likely to enter data on patients that were implanted unsuccessfully or outside guideline recommendations. Furthermore, as in all non-mandatory registries or surveys we are only able to comment on the patients who received a CRT device; we have no data on patients with a recommendation class I indication who did not receive a CRT device. Finally, the completeness of documentations in countries with low health expenditure per capita is moderately higher than those countries with the highest expenditure.

Conclusions

The ESC CRT Survey II has demonstrated a non-uniform delivery of healthcare. Countries with low health expenditure per capita appear to be reserving CRT therapy for younger patients, those with guideline recommendation level I indications and patients with more severe symptoms of HF. The findings of this survey demonstrate substantial unwarranted variation with regard to CRT implantation practice, which deserves further inquiry.

Acknowledgements

Thanks to Tessa Baak for her dedication and organizational skills. We express our gratitude to our sponsors and industry representatives Joel Courville, Thomas Herrmann, Dave Hollants, Marijke Laarakker, Art Pilmeyer, Nico Uwents, and Alphons Vincent. Thank you to the national coordinators and the implanters for providing us with the data.

Funding

The ESC CRT Survey II was supported by the European Heart Rhythm Association, the Heart Failure Association, with financial support from Biotronik, Boston Scientific, Medtronic, Sorin, St. Jude, Abbott, Bayer, Bristol Myers Squibb, and Servier.

Conflict of interest

CRT Survey II complies with the Declaration of Helsinki. The data are completely anonymized, and the locally appointed ethics committees have confirmed that informed consent from the subjects was not required. C.L. reports receiving research support from the Swedish Heart Lung Foundation, Swedish Royal Society of Science, Stockholm County Council, consulting fees from AstraZeneca, Roche Diagnostics, speaker honoraria from Novartis, Astra, Bayer, Vifor Pharma and Medtronic, and Impulse Dynamics and serves on advisory boards for AstraZeneca. S.T. was a Medtronic Inc. employee at the time of authoring. Z.I. is an employee of Abbott, Zaventem, Belgium. C.N., N.B., and K.D. have no conflict of interest.

Data availability

The data underlying this article are available in the article.

References