This work evaluated the relationship between the utilization of 25 technologies for different anatomical systems in German inpatient care and available clinical evidence, as well as changes in funding and safety information.
The number of included publications per technology ranged from two to 498, with case-series and other non-comparative designs constituting the bulk of the evidence body. Although this reflects the interest of clinicians in sharing their experience with a certain technology in a real-world setting, such studies do not provide an adequate foundation to conclude on a technology’s comparative value. For seven of the 25 technologies, a maximum of one publication from the two highest evidence levels (systematic reviews of RCTs or individual RCTs) was identified despite high utilization numbers. In fact, for most included technologies there were few RCTs, predominantly with high RoB. A high volume of publications does not necessarily predicate the robustness of the evidence body on the benefit of a technology. Similarly, the availability of multiple systematic reviews of comparative studies (LoE 1a/2a) for the same technology does not necessarily result in information gain. For example, several systematic reviews/meta-analyses on stent retrievers for mechanical thrombectomy in acute stroke were identified in this work, published within two years and combining the same six RCTs. What is more, findings in systematic reviews are only as robust as the studies included in the review allow.
The trend towards higher LoE over time was not always observable. For several technologies, a robust body of evidence (e.g. at least one RCT with low risk of bias) did not materialise for several years or even until the end of the observation period. Across technologies, the number of years and the number of patients treated before adequate scientific evidence became available varied. The type of FDA approval, which is tied to post-market evidence generation requirements, and could thus have influenced the number and type of available studies (36, 37), was not investigated further. However, it seems like the lack of mandatory benefit assessment before reimbursement potentially exposed patients to undue harm (including lack of benefit) and the health system to inefficient spending. Other European countries, such as France (38) and the UK (39), have established pathways of comparative effectiveness assessment of new technologies prior to reimbursement of innovative technologies. An evaluation pathway was established for certain high-risk innovative technologies in German inpatient care in 2016 (§137h SGB V), mandating hospitals seeking to negotiate extrabudgetary payments for the first time to provide the FJC with available evidence on effectiveness. This did not apply to any of the technologies in the sample.
A relationship between evidence and utilization could be shown for some technologies, but not all. Up to a point, this was to be expected, as there are many factors that influence the adoption of new technologies in healthcare organisations that could not be accounted for in the regression model. The available number of alternative technologies, operator experience, user-friendly operation, organizational culture, individual beliefs and preferences of operators and patient demand (40, 41) can play a role in the diffusion process. It is also thinkable that high initial acquisition costs lead to the continued use of technology despite the availability of better alternatives. Finally, even if a technology is reimbursable and has a positive evidence profile, insurers may not be willing to agree to extra-budgetary or supplementary payments that cover all costs.
The relationship between utilization and funding, as well as the relationship between utilization and safety notifications were explored in a qualitative manner, without clear results. However, this does not fully exclude the possibility that such relationships existed for any of the technologies. This study was not designed to predict what utilization might have looked like if funding had not changed over time from the most insecure (extra-budgetary payments) to more secure types of funding (e.g. adequate depiction in a DRG). For at least one technology, safety notices and restrictions could have strengthened the effect of clinical evidence, but this relationship also proved impossible to evaluate fully with this study design.
To ensure quality of care, it is important that reimbursed technologies are safe and effective. The first step towards achieving this is to have regulatory processes in place that only allow such medical devices to enter the market. The changes introduced by the MDR, which took effect in May 2021 (5, 8), could reduce the number of years of utilization without robust evidence in German inpatient care and elsewhere in Europe (and the time lag between CE-certification and FDA approval (42)) but this will also depend on how the regulation is implemented.
The second step is to assess the comparative effectiveness or even cost-effectiveness of new technologies prior to reimbursement. Every health system needs to balance timely access with certainty on the safety and effectiveness of an innovation, while distributing limited resources wisely. For new technologies that show promise but are not supported by adequate evidence yet, coverage with evidence development (CED) may provide a solution. CED is used by several European countries, such as Belgium, England, France, the Netherlands, Spain, and Switzerland (43), and has also been introduced in Germany (44). It is important to design such programs carefully, particularly as public institutions are often not experienced in planning and conducting clinical trials, which can lead to delays and – depending on program set-up – challenges with access or inefficiencies (43). Further, CED has the potential to support innovation from small and medium size manufacturers of medical devices, who may not be able to afford large clinical trials otherwise.