Principal‐agent theory‐based cost and reimbursement structures of isavuconazole treatment in German hospitals

Isavuconazole (ISA) is a frequently used antifungal agent for the treatment of invasive fungal diseases (IFDs). However, hospital reimbursement data for ISA is limited.


| INTRODUC TI ON
Invasive fungal diseases (IFDs), such as invasive aspergillosis, are life-threatening infections with high mortality rates, 1 and there has been an increasing incidence in recent years. 2 The primary reasons behind fatal outcomes include the complexity and severity of the patient's primary underlying disease, challenging diagnostic pathways and the potential side effects of currently available antifungal agents (such as nephrotoxicity). 3 Current therapeutic regimens, depending on co-medication, comorbidities and risk factors, and which are recommended by national and international guidelines, include fluconazole, voriconazole, liposomal amphotericin B and/or isavuconazole (ISA). 4,5 ISA is a new triazole with a broad antifungal efficacy and is clinically equivalent to voriconazole, but with superior tolerability in the treatment of aspergillosis. 5,6 ISA is an effective treatment option, particularly when voriconazole is contraindicated for QT prolongation, liver or kidney dysfunction or drug interactions. 7,8 In addition to the clinical burden of IFDs, recently published studies demonstrated high associated treatment costs. [9][10][11] Policymakers, however, are obliged to contain their national healthcare expenditure, while healthcare costs in developed countries have been rising for years due to demographic changes, medical-technical progress and other causes. 12 This raises common concerns that the quality of care may suffer, 13 that is by delay or non-use of highly effective but more expensive treatments. Moreover, little is known about the specific effects of given cost-containment strategies on quality of care. 14 Therefore, it is not clear whether the administration of new antifungal treatments is hampered by higher costs and/or cost-containment strategies. This is even less clear since new, higher-priced treatments seem to incur lower overall healthcare costs due to a reduced length of stay (LOS) in hospital, as an earlier study demonstrated for the usage of ISA. 6 Although this research is more than valuable, it has not clearly shown whether all perspectives of relevant stakeholders were included, or whether it covered only the hospital standpoint. 15,16 This study is focussed on the implementation of ISA reimbursement mechanisms for hospital providers. As innovative antifungals are typically used in the hospital setting, the aim is to investigate the influence of the ISA administration types on the cost-containment strategies and hospital reimbursement in Germany. Since national regulations of healthcare systems are quite complex and differ from each other, the space limitation of an article only allows us to display one system in detail. We chose Germany because it is the largest European market and will provide additional information for other European countries.

| Theoretical framework
The principal-agent theory (PAT) emerged in the 1970s and has previously been used to investigate similar questions in different settings, including health care. 17 The origin of this theory is a matter of debate, but often Jensen and Meckling are cited. 18 PAT played a major role in Bengt Holmström's work and contributed to him winning the Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel (often called the Nobel Prize in Economics) in 2016. PAT is used when economic subjects are limited in their decision-making (eg through asymmetrical information distribution). 19,20 Incomplete information occurs when an individual appraises the actions of others. In a broad definition, there is a principal-agent relationship as soon as the well-being of one depends on the actions of the other. In a narrow definition, there is a client (principal) who, by mutual agreement, entrusts a contractor (agent) with a task for remuneration. Since the two may pursue different goals, conflicts can arise as the agent may take advantage of the information asymmetry to the detriment of the principal.
The theory was extended far beyond economics or institutional studies to the context of information asymmetry, uncertainty and risk. 21,22 The treatment of IFDs is a typical application of PAT. A patient (principal) may rarely decide on her/his treatment due to the lack of information and/or knowledge. Instead, healthcare professionals (agents) will decide on her/his behalf. 23 However, if one of the agents faces disincentives (eg because more effective treatment is more expensive and therefore not reimbursed), there may be reasons to decide on a treatment that is contrary to the principals' interests. To include multiple perspectives and comprehensive information, we considered three relevant healthcare professionals as agents for this PAT analysis: physicians, pharmacists and hospital managers. Based on common hospital workflows and the authors' expertise, assumptions regarding the agents' objectives were made.

| Costs and reimbursement
We analysed costs and reimbursement of ISA for the treatment of IFDs in German hospitals via a cost analysis. 24 The German reimbursement for inpatient treatment (German diagnosis-related group [G-DRG] system) was analysed using publicly available groupers F I G U R E 1 Conceptual basis of PAT applied to the treatment setting. Asymmetric relationship between principal and agent when applied to the treatment setting; relevant principal and agents in this study. PAT, principal-agent theory (software that converts ICD-10 codes into G-DRG codes) and tariffs.
The aim was to identify all G-DRGs relevant to ISA administration due to IFDs.

G-DRGs are coded based on the International statistical
Classification of Diseases (ICD)-10 codes, main and side discharge diagnoses of patients, operation and procedure codes (OPS, www. dimdi.de) 25,26 and current reimbursement handbooks provided by the Institute for the Remuneration System in Hospitals (InEK Siegburg, www.g-drg.de). 27 Cost analysis that followed the hospital standpoint is based on the InEK cost matrix and was conducted by using the G-DRG Cost-tool, available via the G-DRG-Report-Browser (www.medin foweb.de). 24,28 In order to calculate the costs of ISA treatment, patients and G-DRGs were identified by the OPS codes 6-008g and 6-008h (ISA administration parenteral and oral) and based on the G-DRG handbook 2020. 27 Following common base rates in Germany, we assumed that, at national level, the value of a G-DRG point was €3,500. No discounting of costs was performed because cost analysis stretches over a one-year timespan. 24 Cost savings were calculated by multiplying data from the SECURE trial 6 (reduced LOS of 2 days for ISA patients) and the (daily) cost information from the G-DRGs.
We calculated the related G-DRG reimbursements by the G-DRG handbook-determined cost weight of 2020 as an allocation basis of the costs of all regular cases with an average case-cost weight between the lower and upper limit of LOS. To determine G-DRG remuneration, we multiplied the different cost weights per G-DRGs with €3,500 as the value of a G-DRG point and included additional payments for inpatient hospital stay exceeding the G-DRG-specific threshold. To determine the ISA treatment reimbursement, we assumed a 45-day ISA treatment, with patients receiving a loading dose of 3 × 200 mg/day for the first two treatment days, and 1 × 200 mg for each following day for the treatment of invasive aspergillosis. 6 Thereby, the official pricing, specified by Lauer Taxe reference list for pharmaceutical price information (www.lauer-fisch er.de) 29 was considered. By these amounts, we identified the additional payments for innovative treatment procedures (Zusatzentgelte [ZE]) for ISA per inpatient hospital stay.
For the reimbursement procedure, the focus was placed on hospitals that individually negotiated additional payments for ZE and new diagnostic and treatment methods (Neue Untersuchungsund Behandlungsmethoden [NUB]) that are not yet covered by the G-DRG flat tariffs. Therefore, NUBs are usually the preliminary stage before being classified as ZE and eventually being considered in the G-DRG tariff. The aim of additional payments is to include innovative products (such as ISA) into everyday clinical routines. In our study, we referred to the additional payment of ISA as ZE, as ISA was classified as NUB in 2019 30 and was changed by G-DRG handbook 2020 to ZE (ZE2020-166/167: administration of ISA, parenteral/oral 27 ). In the G-DRG system, ZE reimbursement is divided into (a) an individually negotiated reimbursement and (b) a nationwide reimbursement.
As the G-DRG reimbursement calculation is based on the average hospital costs, we used the G-DRG reimbursement as a proxy for a daily full cost in an inpatient setting. 31 To evaluate the influence of cost-containment strategies on the quality of care, the relevant G-DRGs were analysed according to their average reimbursement, as well as their net effect with and without ZE, for ISA administration under consideration of LOS.

| Ethics statement
No ethical approval was required for this study as the underlying data were retrieved from publicly available sources.

F I G U R E 2
Agents and their primary objectives applying PAT. Applying the PAT, different primary objectives of the involved agents are assumed. This further sets the agents' objectives in relation to those of the principal. PAT, principal-agent theory 3 | RE SULTS

| Principal-agent theory
Based on the author's experience, four dimensions of potential (dis) incentives of ISA administration were identified: (a) medical outcomes, (b) treatment process, (c) treatment costs and (d) reimbursement (Table 1). Medical outcomes included the effectiveness of ISA for the treatment of IFDs. Our analysis showed that the incentives of the principal and all agents were aligned. Treatment process focussed on treatment benefits and the reduction of adverse events. We found that principal and agent incentives were either aligned or indifferent (neutral). Reimbursement covered every type of remuneration of hospital care. Our analysis revealed that the incentives of the principal and agent (especially hospital managers) may (partially) diverge.
The same was observed for treatment costs, which considered all costs associated with ISA treatment. Therefore, we have analysed reimbursement and treatment costs in more detail.
Three of these G-DRG codes (E77B, R60C and R60D) were identified for both OPS codes.
In the OPS code 6-008g (parenteral), the overall ZE of €26 265 resulted from a 2-day loading dose of 3 × 200 mg/day followed by a single dose of 200 mg for 43 days for each G-DRG. The reference price for 200 mg parenteral ISA was €536/day. Two main findings resulted (Table 2). Firstly, due to the reduced LOS of 2 days, G-DRG cost savings occurred for each G-DRG, ranging from €963 in R60D to €2828 in A09B. Secondly, the total reimbursement (G-DRG + ZE) varied widely from €40 524 in E77B to €114 251 in A07B.
Considering the OPS code 6-008h (oral), the overall ZE of €5241 resulted from a 2-day loading dose of 3 × 200 mg/day followed by a single dose of 200 mg for 43 days for each G-DRG. The reference price for 200 mg oral ISA was €107/day. Our analysis yielded results similar to parenteral ISA administration (Table 3).
Consequently, G-DRG cost savings were generated for each G-DRG code ranging from €694 in E77D to €2,184 in A04D. The total reimbursement (G-DRG + ZE) varied from €14 547 in E77D to €46 954 in A04D. Small cost savings and low reimbursement were identified in E77D for oral administration. The G-DRG code A04D, however, offered the highest cost savings (€2,184) and the highest total reimbursement (€46 954). To determine an ISA cost-reimbursement comparison, we conducted a sub-analysis for both OPS codes for the overlapping G-DRGs E77B, R60C and R60D. The sub-analysis identified specific cost shares for each G-DRG case referring to material cost (cost category group 4: drugs general and drugs individual) and thus, aimed to indicate the importance of ZE for hospital managers. Based on the InEK cost matrix, we calculated the accrual drug-cost shares within the G-DRGs (general/individual) to be €1070 in E77B, €2455 in R60C and €2821 in R60D Note: Definitions of the G-DRG: A04D, Bone marrow transplantation/stem cell transfusion, allogeneic, with graft-versus-host disease Grade III and IV or except for plasmocytoma, HLA-different or with complex treatment in multi-resistant pathogens; A04E, Bone marrow transplantation/stem cell transfusion, allogeneic, except for plasmocytoma; E77B, Certain other infections and inflammations of the respiratory organs with certain CC or highly complex diagnosis or complex diagnostics for organ transplantation or intensive care complex treatment > 196/-/-effort points; E77D, Certain other infections and inflammations of the respiratory organs, age > 9 y; R60B, Acute myeloid leukaemia with intensive chemotherapy with complicated diagnosis or dialysis or port implantation or intensive care complex treatment > 392/368/-effort points or heaviest CC, age > 15 y; R60C, Acute myeloid leukaemia with intensive chemotherapy, extremely CC or complex diagnostics when leukaemia or with moderate complex chemotherapy with certain complicated factors or extremely difficult CC with complex diagnostics or complex treatment. MDR with dialysis or extremely difficult CC or the heaviest CC; R60D, Acute myeloid leukaemia with intense chemotherapy, without complicated diagnosis, without dialysis, without port implantation, without ICU complex treatment. >392/368/-effort points, without extremely difficult CC without complex diagnosis when leukaemia or with dialysis or extremely severe CC (G-DRG-System, http://g-drg. de, last accessed 2020-03-19).
(Tables S1-S3 incentives. To shift the sole cost risk from hospital providers and to reduce the bureaucratic burden, the current triple-bilateral negotiation processes between (a) hospitals and health insurance providers, (b) hospitals and pharmaceutical companies and (c) health insurance providers with pharmaceutical companies ( Figure 3) need to be reorganised towards a triangular structure that allows for transparent, performance-and outcomes-based cost-coverage, resulting in an adequate reimbursement for healthcare providers.
Merging negotiation processes for the purchase and reimbursement of drugs prevents cost risks for hospitals. Moreover, a trilateral agreement between all stakeholders could lead to efficient, transparent and quality-based cost-reimbursement structures supporting the timely arrival of innovative treatments and drugs into routine care.

| Implications for future research
Our study outlined the relevance of a secured reimbursement of innovative but also expensive treatments, such as ISA, for hospital providers in Germany. Reorganisation of inefficient negotiation processes towards a triangular structure would yield multiple advantages: (a) less bureaucratic burden in general, including more rapid patient access to novel treatments, (b) efficient agreement on payments between all relevant stakeholders, allowing innovative treatments in inpatient care and (c) reduced cost risk for hospital providers.
In other European countries, there are fewer hurdles to overcome to obtain hospital reimbursement, or cost-covering mechanisms for innovative treatments already exist. For example, in Italy, pricing and reimbursement systems of pharmaceutical products are primarily based at a national level, resulting in less bureaucracy for hospital management. 32 In conclusion, our study has shown cost savings of using ISA for hospital providers, and that ZE reimbursement allows for rapid access to novel treatments benefitting patients. With respect to the recommendations of international guidelines, 33,34 we suggest considering cost-covering/reimbursement strategies in European countries as a clinically and economically sound strategy in IFD treatment.