Drivers and recent trends of hospitalisation costs related to acute pulmonary embolism

Venous thromboembolism and particularly acute pulmonary embolism (PE) are a leading cardiovascular cause of death after myocardial infarction and stroke [1] and have a substantial impact on the morbidity and mortality of the population in Europe and globally [2‐4]. In recent years, significant progress was made in the advanced reperfusion treatment of intermediate-risk and high-risk PE, which presents with acute right ventricular (RV) pressure overload and is a potentially life-threatening condition [5‐7]. Data available so far suggest that recently introduced catheter-directed treatment (CDT) options for PE may improve survival and reduce complications and the length of hospital stay [8, 9]. However, healthcare systems in many countries remain reluctant to reimburse the costs of these emerging treatment options. Besides the need for more robust evidence on the clinical benefits of novel approaches [5, 10], there is a lack of systematically collected large-scale data on actual hospitalisation costs and the main cost drivers. In fact, a recent comprehensive analysis that provided an update on the economic burden of cardiovascular diseases across the European Union (EU) focused on coronary heart disease and cerebrovascular disease but did not mention any data related to acute PE [11].

Aiming to contribute to the closure of this gap and to begin to inform decision-making, we analysed, in the present study, the entire German nationwide inpatient population hospitalised for acute PE from 2016 to 2020. We aimed to obtain actual reimbursed hospitalisation costs, comparing them to those of the other two most frequent acute cardiovascular syndromes, myocardial infarction and (ischaemic) stroke. We further aimed to identify patient characteristics, procedures and complications independently affecting the cost of hospital stay and calculate annual trends of hospital reimbursements. In this regard, we placed a focus on patients in the upper categories of clinical PE severity and early risk of death.

The present study investigated the economic burden of acute pulmonary embolism, including all patients hospitalised with PE in in Germany, the country with the largest population in Europe (currently 84 million), over 5 years. Our results can be summarised as follows: (1) actual median hospitalisation costs in Germany were higher than those estimated from European registry data, albeit lower than the costs related to myocardial infarction and (ischaemic) stroke; (2) age, PE severity and comorbidity, as well as in-hospital (particularly bleeding) complications, were identified by multivariate logistic regression as significant cost drivers in the study population; (3) use of CDT doubled over time in the elevated-risk group of patients with “severe” PE (28% of the entire population) and (4) catheter-directed procedures, although costly by themselves as reflected by current reimbursements, were not associated with an overall increase in hospitalisation costs among patients with severe PE and shock. A reduced length of hospital stay paralleled this latter finding.

Very little is known about the actual hospitalisation costs imposed by acute PE on national health systems. A recent cost-of-illness analysis, based on data from 1349 patients diagnosed with PE and included in a prospective European registry, provided a low-end cost estimate of 2328 euros and a high-end estimate of 3533 euros for the hospitalisation related to the index acute event [13]. Earlier models analysing cost sources from several European studies had yielded higher median estimates, between 3891 and 4197 euros [14]. Such approximations may have limitations given the relatively small, heterogeneous patient populations analysed; moreover, several adjustments and assumptions must be made to account for the differences in healthcare systems across Europe [13, 14]. By comparison, in the present study, we analysed actual documented and reimbursed hospitalisation costs in the entire population of a single country. Our results in a population of almost half a million patients in Germany revealed median hospitalisation costs of 3572 euros during the study period, lying at the upper end of the most recent European estimate mentioned above [13]. On the other hand, the cost of illness of acute PE in Germany remained, at least over the years studied, substantially lower than the costs reported for PE-related hospitalisations in the US which reached a median of 10,032 (IQR 4467 to 20,330) US dollars in the years 2016–2018 [15].

Time trend analysis revealed that annual hospitalisation costs for acute PE remained relatively stable during the study period, during which no new major diagnostic or therapeutic procedures for this disease (demanding substantial changes in reimbursement) entered broad clinical practice. Interestingly, a remarkable drop in reimbursed costs occurred in the year 2020. This may be related, at least in part, to the observed increase of in-hospital case fatality of PE associated with COVID-19, particularly in the first year of the pandemic [16‐18], in some cases reducing the duration of hospitalisation. In parallel to a change in the natural history of PE itself, the lockdown measures imposed in Germany and other countries significantly impacted hospitalisations for a broad spectrum of potentially life-threatening diseases, including severe oncological and cardiovascular cases and also emergencies such as acute myocardial infarction [19‐25]. This was due to the combination of reduced hospital capacities in view of the large number of beds being reserved for COVID-19 cases and the patients’ fear of contracting the infection by seeking medical help and being admitted to a hospital. It is thus likely that an undetermined number of patients with severe PE never reached the hospital during that period. Moreover, some cases of severe PE may have gone unnoticed in hospitalised patients, their critical condition being attributed to COVID-19 if the virus test was positive. Finally, the reduction of elective surgical procedures during the pandemic, and consequently of PE cases associated with them, may have contributed to this temporary reduction in reimbursements.

A contemporary topic of debate is to what extent new catheter-directed interventions for dissolving or fragmenting/aspirating pulmonary emboli may impact healthcare costs in Europe and other parts of the world in the coming years. CDT systems are undergoing continuous technical evolution and are increasingly popular among physicians in the US [26], particularly at institutions that have established multidisciplinary pulmonary embolism teams (PERT) [27, 28]. In Europe, the introduction of CDT was initially more hesitant [9], but its use has now begun to increase as well as confirmed by the results of our study. For example, modelling recent CDT trends in the US, we could calculate the expected increase of CDT use in Germany for the future period 2025–2030. Our models predict a CDT penetration ranging from 3.1 to 8.7% by 2030, resulting in an increase of annual costs for PE-related hospitalisations between 15.3 and 49.8 million euros [29].

Observational data suggest that CDT, when used in patients with intermediate-risk or high-risk PE, may reduce the risk of bleeding complications, and it may also be associated with lower in-hospital mortality [8, 9, 30]. Nevertheless, current guidelines demand convincing high-quality data from randomised controlled trials before endorsing CDT as first-line therapy in patients without haemodynamic collapse [5, 6, 10]. In addition, direct costs of catheter systems and procedures for advanced PE therapy need to be determined separately for each country’s hospital reimbursement system and, importantly, be weighed against their benefits in terms of reduced early complications, length of hospital and ICU stay, return to work and productivity and prevention of late sequelae of PE [31]. In agreement with previous reports [8, 9, 30], we observed a significant association between CDT and a reduction of in-hospital mortality and the length of hospital stay. Although CDT use was, as expected, associated with higher hospital costs in the entire group of patients with severe PE, this was no longer the case when the subgroup of high-risk patients with shock was analysed. These findings generate the hypothesis that the cost–benefit ratio of CDT might become increasingly favourable with increasing severity of PE.

Some limitations exist and caution is warranted when attempting to translate associations found in an analysis of nationwide data into possible causal relationships between treatment modalities and hospital outcomes, complications and/or costs. Firstly, neither the physicians’ rationale regarding patient selection for individual therapeutic measures nor the exact timing of complications during the hospital stay, particularly in relation to systemic thrombolysis or advanced interventional procedures, can be retrieved from this type of aggregated observational data. A typical example of this limitation is our finding that patients with PE and shock who underwent thrombolysis appeared to have major bleeding less frequently than those who did not, seemingly contradicting the established bleeding risk of this treatment form [32]. This may be due to the fact that patients with an excessive bleeding risk (which is often the case in this risk category) do not receive this type of treatment in clinical practice [6] but bleed nevertheless, even on heparin anticoagulation alone; besides, patients with shock have a high in-hospital case fatality, and major bleeding occurring immediately before death may have been underreported.

Secondly, due to the nature of an ICD- and OPS-code-based analysis, classification of the severity of PE in the present study may not exactly correspond to the definition proposed by European guidelines [6]. Besides, underreporting of adverse events and/or undercoding of procedures cannot be excluded. However, it is very unlikely that costly complications and treatments were ‘forgotten’, considering the reimbursement efforts of all involved hospitals. Thirdly, we cannot exclude an interdependence of patients’ comorbidities which were included in the progressive adjustment models of our multivariable regression analysis. Finally, no follow-up evaluation after hospital discharge was available. Ongoing randomised trials will help to determine not only the efficacy and safety of CDT techniques compared to medical treatment but also if their implementation on a larger scale and across different healthcare reimbursement systems is cost-effective [33].

In conclusion, the present study provided actual cost-of-illness data for the entire population hospitalised for acute PE in a large European country over a 5-year period. Our results help to identify current and emerging cost drivers. They may inform reimbursement decisions by policymakers and help to guide future health economic analysis of advanced treatment options for patients with intermediate-and high-risk PE.