Possible Health Technology Assessment Pathways

6.1 Introduction

Two intersections exist between in silico methodologies and Health Technology Assessment (HTA). The most obvious is when in silico methods are used as predictive Software as Medical Devices (SaMD), e.g., as clinical decision support systems. In such cases, HTA is used, like any other medical technology, to evaluate if its adoption is cost-effective and clinically appropriate (criteria under which a certain intervention is properly prescribed to a patient; according to the Italian Medicine Agency (AIFA), appropriateness is defined as “adequacy of the actions adopted to manage a disease, concerning both the patient’s needs and the correct use of resources” (Garattini & Padula, 2017)).

A second intersection is when the use of In Silico Trials in the regulatory qualification of a medical product impacts the HTA assessment of that new product. Using In Silico Trials to replace, reduce or refine human experimentation could improve the ability to detect change (which would turn into a more sensitive assessment of differences in efficacy/performance). It could also provide an efficacy/performance assessment closer to real-world effectiveness, as virtual patients may make it easier to explore the efficacy of under-represented sub-groups in clinical trials. In addition, in silico methodologies could reduce or replace Phase 4 trials, produce early estimates of quantities of interest for the HTA assessment of medicinal products, and support the so-called early discourse on HTA. This second perspective is the focus of this chapter.

6.2 Assessing in Silico Methodologies for HTA

Modelling and simulation is a constantly expanding field, in terms of both methodology and application sectors. As the models evolve in complexity and increased uptake, it becomes essential to clarify the most appropriate tools for evaluating in silico methodologies, especially those that can contribute to HTA.

Developers often make very strong claims with poor reporting and/or weak VVUQ evidence supporting those models. These tools still have a long way to go in terms of implementation and public adoption, as well as rigour in their use, which can be inconsistent and unbalanced at the moment (Musuamba et al., 2021). This chapter aims to provide input on the scientific evaluation of in silico methodologies of health interventions (drugs and other technologies) from the HTA point of view and the role that such technologies can play in HTA.

6.3 Introduction to Health Technology Assessment (HTA)

HTA is a multidisciplinary process that uses explicit methods to determine the value of health technology at different points in its lifecycle. The purpose is to inform decision-making to promote an equitable, efficient, high-quality health system.¹ In many countries, it is now common to perform this systematic and multidimensional evaluation of health technologies aimed at informing coverage, reimbursement, or pricing decisions within public healthcare systems.

The process is formal, systematic, and transparent, using state-of-the-art methods to consider the best available evidence. The dimensions of value for a health technology may be assessed by examining the intended and unintended consequences of using a health technology compared to existing alternatives. These dimensions often include clinical effectiveness, safety, costs and economic implications, ethical, social, cultural, and legal issues, organisational and environmental aspects, and wider implications for the patient, relatives, caregivers, and the population. The overall value may vary depending on the perspective taken, stakeholders, and decision context.

HTA can be applied at different points in the lifecycle of health technology, i.e., pre-market, during market approval, post-market, and through the disinvestment of health technology. The approach and methods used at each phase will differ and depend on the available evidence (whether primary or secondary data) and the decision about the technology.

Whilst licensing approval is mainly focused on the technical and safety profile of the medical device, HTA bodies have different interests and, therefore, different evidence requirements. Normally, the requirements aim to inform policymakers (and decision-makers in general) of the rationale allocation of resources within finite budgets to the funding (or using) of healthcare interventions. For this reason, data required for market access might go beyond those used or developed for licensing, particularly in medical devices, where regulatory requirements have historically been low.

This additional evidence generation could also be worthwhile from the manufacturer's perspective. With prepaid financing mechanisms for health systems through general taxation or private/social insurance, third-party payers’ coverage strongly influences market prospects for medical technology companies. For example, granting a CE mark does not imply that the product will be available to patients everywhere in the EU. If the HTA assessment leads to declined public reimbursement in a particular country, the vast majority of patients cannot afford the product in that country.

It is important to mention that “health technology” is a broad concept. The accepted international definition of a health technology is an intervention developed to prevent, diagnose, or treat medical conditions; promote health; provide rehabilitation; or organise healthcare delivery. The intervention can be a diagnostic test, device, medicine, vaccine, procedure, program, or system.² As we explained in Chap. 1, in silico technologies can be used as medical devices, as they are used in the diagnostic, prognostic, or therapeutic process. In addition, they can be used to evaluate the safety, efficacy/performance, prescriptive appropriateness, and cost-effectiveness (HTA) of a new medical product, whether a medical device or a drug. This chapter will mainly focus on this second use, touching on the first in the final section of future challenges.

It is also worth mentioning that the modelling and simulation methods are frequently used to evaluate different types of (implemented) medical interventions, often in the context of HTA. These studies have mainly been used to supplement systematic reviews in an effort to increase the usefulness of the evidence summary. Uncertainty about the optimal choice among available interventions for important patient-relevant outcomes may persist even after synthesising the best available evidence. Indeed, decision-makers are increasingly interested in complementing the results of systematic reviews of empirical evidence with information from modelling and simulation studies. That is, integrating empirical evidence on benefits and harms, values (preferences), and/or resource utilisation while accounting for all relevant sources of uncertainty (Dahabreh et al., 2008, 2017). Some of the most frequent applications of this type of modelling and simulation are used:

• to synthesise data from disparate sources (modelling provides mathematical tools for evidence synthesis and the assessment of consistency among data sources),

• to make predictions (“interpolations”, forecasts, “extrapolations”, prioritisation and planning),

• to support causal explanations and infer the impact of interventions, or

• to inform decision-making (about patient-level care, drug or device licensing, health care policy or the need to conduct additional research (Dahabreh et al., 2017).

Although this specific scenario of modelling and simulation based on the combination of already existing evidence/data could be considered an in silico methodology, it will not be included in this chapter as there are good and updated reviews on that (Dahabreh et al., 2008, 2017; Jalali et al., 2021).

6.4 In Silico Methodologies as a Source of Evidence

Science generates evidence through observation, deduction, and induction. Simulation, like deduction, starts with specified assumptions regarding a proposed system and generates data suitable for analysis by induction. However, this data does not come from direct observation in the real-world (Stahl, 2008).

These assumptions can be designed according to observed data and predicted as a function of the experimentally observed variability (phenomenological) or by leveraging some pre-existing knowledge about the physics, chemistry, physiology and biology of the phenomenon being modelled (Viceconti et al., 2020b).

In silico methodologies can be a source of evidence when developing or validating a health technology, a pharmaceutical product or a medical device (model-based medical results or computational modelling and simulation results). These are predictive computer models that are used to provide evidence in support of the safety and/or efficacy/performance of a medical product during its marketing authorisation process. It can also be used during any assessment phase through the technology lifecycle and, thus, become part of the evidence to be used for HTA as well.

Methodologies and tools used to produce regulatory evidence are usually qualified by the regulator or certified according to a specific technical standard such as, for example, the ASME VV-40 for the use of computational modelling to evaluate medical devices.³

6.4.1 Medical Devices and Interventions

Computational modelling and simulation can help to increase the scientific evidence for evaluating high-risk medical products and interventions, especially when they enable replacing, reducing and refining nonclinical in vitro/ex vivo experiments, nonclinical animal studies or clinical human studies in case of ethical issues and, time or costs constraints.

It is also particularly significant with the new medical device regulation⁴ of the European Commission where scientific evidence used to assess high-risk medical devices must be based on methodologically sound trials, which may be supplemented with alternative evidence sources such as computational modelling and simulation (Olberg et al., 2017).

6.4.2 Pharmaceutical Products

The clearest indication for using simulation methods is when direct experimentation via randomised controlled trials (RCT) is impossible due to cost, time, or ethical constraints. In this regard, RCTs can be considered a form of simulation as it represents and simplifies the system under study. However, computer simulations of these trials typically decrease time and cost, besides overcoming some ethical restrictions of experimentation on humans. These ethical limitations can mainly be found when a question needs exploring (effects of exposure), but conducting the trial would require exposing a vulnerable group to unacceptable risks (Stahl, 2008).

Computational methods aim to complement in vitro and in vivo tests to minimise the need for animal/human testing, reduce the cost and time of toxicity tests, and improve toxicity prediction and safety assessment. In silico toxicology encompasses simulation tools for biochemical dynamics and modelling tools for toxicity prediction. They are useful in drug design to determine how drugs should be altered to reduce their toxicity. In turn, this knowledge can be used for the evaluation of pharmaceuticals and to enrich clinical evidence.

For example, there are methods for predicting outcomes based on chemical analogues with known toxicity. On the other hand, researchers also use dose–response or time–response models, which establish relationships between doses or time and the incidence of a defined biological effect (e.g., toxicity or mortality) (Viceconti et al., 2017).

6.5 In Silico Methodologies: Product Life Cycle and HTA

At the cost of oversimplifying, any health technology’s development and assessment cycle can be reduced into different macro-phases: design/discovery, pre-clinical and clinical assessment, regulatory assessment, market access and post-marketing assessment. Decision-maker uncertainty is high in the discovery and design phase when new and emerging health technologies have not yet generated any evidence regarding the future value they could bring to the health systems. The more we move through the diffusion curve of technologies, the more evidence is generated and uncertainty reduced. In silico methodologies have the potential to support all steps of the product life cycle (see Table 6.1).

Table 6.1 Potential applications of in silico methodologies along the product life cycle and suitable HTA modality

6.6 Methodologies for in Silico Clinical Studies

6.6.1 HTA Health Technology Assessment

Throughout the product life cycle, the industry increasingly relies on computational modelling and simulation to speed development and provide additional performance and safety assurance. Still, such use is currently limited (Viceconti et al., 2016). According to the results of a 2014 survey of medical device companies, computational modelling and simulation were more commonly utilised in the early stages of product development or after product commercialisation, but rarely to simulate the interaction of the device with a laboratory animal or a patient.⁵

When in silico methodologies are used as a source of evidence for health technology development, extending the traditional HTA that informs coverage/reimbursement decisions to early HTA that informs early research, development, and investment decisions (Tummers et al., 2020) could be of great importance, especially for medical devices, where the development process is a costly and uncertain undertaking (Ijzerman & Steuten, 2011). Failed development not only results in a lack of economic return for the company, but also in higher costs without healthcare improvements for society. There are multiple reasons for failed device development, but one important factor is the lack of an early evaluation of the device potential in healthcare practice, usually only after the prototype design is finalised. The aim of an early assessment is to reduce the failure rate at each stage of the development process, while enhancing R&D efficiency and of limiting the use of resources, through prioritisation of the innovations most likely to succeed among others. It may also be used to support reimbursement claims by providing quantitative input for developing risk–benefit sharing agreements (Markiewicz et al., 2014). With improved confidence in modelling results and a better-established regulatory framework, the use of in silico evidence as part of the regulatory submission process is becoming more common, but it has not yet entered the HTA arena and evidence from in silico methodologies is seldom used in HTA.

6.6.2 Discovery, Design and Pre-clinical Stages

Using in silico methodologies in the discovery and design stage can potentially streamline target identification, secure proof of concept, and identify those drugs/devices worthy of progressing into pre- and clinical development. In silico methodologies can also streamline the finding of which new and emerging health technologies have the potential to satisfy identified health system unmet needs.

Compared with in vitro, ex vivo, and in vivo experiments, in silico simulations are fast, cheap, safe, easy to implement and free of experimental errors. Consequently, they are becoming increasingly helpful in designing new technologies and strategies.

Simulations and computational models allow the effect of the interactions to be examined not only at the local level but in the context of the entire pathway in which the target interacts. To include all the features of these complex systems in these pathways, simulation at the biochemical level may be a suitable foundation for simulation. In this sense, different computational models have been proposed to simulate intercellular interaction at the biochemical and physical levels. By means of this type of model, information on the impact of the target on metabolism can be obtained.

Preclinical in silico assays can potentially minimise problems in the translation between experimental and clinical research. Moreover, preclinical data can be a valid source to include in a computational model to gain additional insight into the factors that modulate the response in later clinical phases. For this reason, in silico experiments have the capacity to reveal and formalise the underlying mechanisms.

The potential use of in silico methodologies can be particularly important in the chemo-prevention and toxicology (Benigni et al., 2020; Valerio, 2009). In silico methodologies are used effectively in preclinical studies to optimise dosage administration and predict the overall performance of the optimised schedule (Pappalardo et al., 2019). Because the number of chemicals marketed for human use is rapidly increasing, computational toxicology models have been developed that estimate the event probability of a molecule based on its chemical structure (Quantitative Structure–Activity Relationship or QSAR).

Using in silico experiments to predict toxicological outcomes of drugs and hazard and risk assessment is widespread. Such experiments can determine the priority of molecules for in vivo or in vitro testing. This prioritisation optimises the testing strategy, potentially minimising the need for animal testing (Benigni et al., 2020).

In this regard, Passini et al. have recently developed software which runs in silico drug trials in populations of human cardiac models, simulating populations of human action potentials. Designed to predict drug safety and efficacy, the software simulates the effects of drugs on the action potentials of cardiac cells. After conducting variable drug-dose response studies, this software provides statistics of biomarkers of drug action and adverse drug effects, such as arrhythmias, with good clinical accuracy. For example, an in silico trial of 62 drugs showed that in silico simulations predicted clinical risk with 89% accuracy (Passini et al., 2017, 2021). In 2011, the US Food and Drug Administration (FDA) approved the first in silico diabetes type 1 model as a possible substitute for pre-clinical animal testing for new control strategies for type 1 diabetes. The European Medicines Agency is also considering in silico approaches as an alternative to animal testing to protect animal health and the environment.

6.6.3 Clinical Development

6.6.3.1 Medical Devices

Computational models of the heart based on data obtained from medical imaging of patients have made it possible to use simulations to view different strategies for cardiac rhythm configuration. They have also enabled the identification of the optimal region for localising cardiac pacing.⁶ These models are not yet widely accepted as medical devices for clinical decision-making. However, this example illustrates how such models and simulations can be applied in personalised medicine.

An example of these developments in patient-level simulation is blood flow simulation using MRI images combined with blood pressure and blood flow information. With the CRIMSON software (Arthurs et al., 2021), 3D models of the arterial system are created and used to determine a prognosis and then to perform an intervention that best preserves blood flow (Ahmed et al., 2021).

The oncNGS pre-commercial procurement⁷ procedure aims to develop novel, affordable solutions to provide the best Next Generation Sequencing (NGS) tests for all solid tumours/lymphoma patients. The call for tender,⁸ launched in December 2021, is challenging the market to address their identified unmet needs through the provision of an efficient molecular DNA/RNA profiling of tumour-derived material in liquid biopsies using a pan-cancer tumour marker analysis kit. This analysis includes NGS analysis integrated with an in silico decision support system that also provides analytical test interpretation and reporting. The oncNGC PCP contract is structured in three phases:

• Phase 1: Design of the oncNGS solution

• Phase 2: Technical, analytical and clinical performance validation of the oncNGS complete solution prototype at the Supplier’s site

• Phase 3: Technical, analytical and clinical performance validation of the oncNGS solution in the clinical samples in Supplier’s sites and real clinical settings.

To ensure suppliers keep working on the sustainable dimension of the novel solutions across the three phases, they are required to keep up-to-date in silico simulations of their novel panels during both Phase 2 and Phase 3 to demonstrate their solutions are affordable ensuring sufficient and homogeneous coverage of all the targets in agreement with the business case to be applied in routine basis, at each (chemo)therapy cycle to follow clinical response and inspire adaptive therapies.

6.6.3.2 Pharmaceuticals

Phase III clinical trials evaluate a new drug in terms of its clinical value (efficacy and safety), its most appropriate dose and dosage (posology), as well as other aspects such as adherence and tolerability. These in vivo studies are expensive and challenging to conduct, as a large sample size is required.

By providing a reliable prediction of the Phase III outcomes based on the data collected during the Phase II clinical trial, in silico methodologies may increase confidence in investing at this late stage of the pre-commercial process. High-quality in silico methodologies using subject-specific models could be proposed as valid evidence to complement the information from these trials, possibly with the requirement to carry out studies to confirm the simulated post-marketing outcomes with real-world data.

By using in silico methodologies to predict outcomes for potential phase III trials, it is possible to optimise both the experimental design and the required sample size. As a result, the development cost could be reduced, as well as the time to market (Pappalardo et al., 2019).

6.6.4 Market Access and Post-marketing Assessment

FFR_CT software, developed by the US medical firm Heartflow to provide a non-invasive quantification of the fractional flow reserve in coronary stenosis, was the first clinical technology based on subject-specific modelling to get marketing authorisation from FDA. The software has also received CE marking and regulatory approval in Japan.

What might be relevant from the perspective of decision-makers is the possibility of testing and identifying in advance which patients’ subgroups are likely to benefit the most from a novel technology (i.e., enhanced patient population stratification) or to investigate and provide empirical evidence of safety issues that could emerge as a result of the implementation of the technology with consequent streamlined recommendations for a safer and effective indication of use (Ciani et al., 2017).

Another relevant example is the stratification of patients with infectious diseases due to multi-drug resistant (MDR) organisms. Thanks to the provided research and development services contracted through Anti-SUPERBUGS pre-commercial procurement,⁹ ANTI-SUPERBUGS PCP Buyers’ Group aims to:

• Reduce both the costs and the operational impact resulting from infections caused by multi-drug resistant organisms;

• Improve the appropriateness of antimicrobial medicine usage;

• Improve the quality-of-care processes in hospitals;

• Reduce the community and social care impact of MDR infections acquired in hospitals by procuring pre-commercial technologies that will transform current Surveillance and Infections control systems into new comprehensive systems.

The 2019 call for tender is challenging the market to address their identified unmet needs through the provision of an anti-superbug in silico solution comprising a bundle of technologies offering different approaches and outputs at a different levels of infection management (as surveillance, environmental safety, first patient screening and patient early diagnosis).

Subsequent public procurement of innovative solutions (PPI), already under preparation, will need to consider that the current COVID-19 pandemic is exacerbating antimicrobial resistance. Data from some EU countries suggest that 6.9% of COVID-19 diagnoses are associated with bacterial infections (3.5% diagnosed concurrently and 14.3% post-COVID-19), with higher prevalence in patients who require intensive critical care (Strathdee et al., 2020). In silico methodologies offer the advantage of increasing the cohorts, refining clinical validation, and taking into consideration this new potential use case, including intensive care unit (ICU) patients infected by COVID-19 and MDR organisms to be able to demonstrate value to the buyers by predicting the real-life benefits and the optimal target use case.

6.6.5 Post-marketing Assessment

In silico studies should also be part of adaptive licensing and reimbursement pathways, where access and coverage are gradually extended as the evidence-based evolves and benefits are demonstrated in clinical research for broader patient populations. This overlapping interest from regulatory bodies, industry, clinics, academia, and even animal-welfare groups has led to the establishment of networks and initiatives worldwide to promote developing, validating, and using in silico medicine technologies.

6.7 Critical Assessment of the in Silico Approach and Limitations

Attention should also be drawn to current in silico simulation tools’ limitations to provide a balanced perspective regarding their potential role in future HTA.

One of the primary limitations is that these techniques should be considered when considering their use in HTA. Primarily, it should be noted that these techniques do not currently allow adequate predictions for all chemicals and outcome variables. Of particular relevance is that there are currently no models for certain systems or components.

Model adequacy is particularly interesting when evaluating complex systems, such as drugs with multiple mechanisms of action or the interaction of different drugs in poly-medicated patients.

These limitations are partially a result of the reliability or transparency of the data used to design the model on which the simulations will be performed. For example, incorrect training data describing the relationship between dosage and adverse events would amplify these errors in the prediction model.

Using in silico techniques may add greater uncertainty if it replaces in vivo experimentation. This is because assessing the simulation results’ external validity is desirable. Recognising the limitations of the technology, there is an increasing interest in combining real-time generated biological data with in silico predictions using a rational approach to integrating computational tools with the experimental setting (Jolivette & Ekins, 2007). Using in silico evidence to reduce or refine in vivo or in vitro experimentation can reduce such uncertainty if reliable and valid models are available.

6.8 How to Assess Evidence from in Silico Methodologies?

In product development and evaluation, in silico models of increased complexity are often used for similar applications as the ‘simpler’ pharmacometrics models, e.g., trial design optimisation, dose-finding/selection, extrapolation of drug efficacy and safety, etc.

For this reason, considering that the model validation processes described in the previous chapters have been carried out correctly, it is logical that requirements for their acceptability follow the same standards as those already established for models currently included in the regulatory dossier or parallel HTA requests.

It is important to document these experiments thoroughly to ensure that the in silico technology is properly evaluated. This documentation should allow independent evaluation by HTA bodies in the specific CoU of the technology.

When evaluating these experiments, it is important to assess the reliability and relevance of the models used, particularly when the models could pose a risk to patients, involve complex systems, or when there is a considerable distance between the nature of the input (for example, chemical-physical parameters) and the nature or dimension of the output (health symptoms).

Last but not least, HTA bodies might require to assess the credibility of the predictive model, as discussed in Chap. 4, as well as the quality of the software artefact implementing it, as described in Chap. 3. In particular, it is necessary to perform sensitivity analyses on those parameters with the highest uncertainty and which have a moderate or high impact on the results. In turn, granting access to the HTA bodies to the models and the data processing schemes would be useful to facilitate the assessment of the model in the specific contexts under evaluation.

6.9 Challenges for the Future

As stated at the beginning, in silico technologies could also be a health technology or part of health technology, that is, Digital Patient or Digital Twin technologies. These are predictive computer models used as decision-support support systems by a clinician in treating an individual patient. From a regulatory point of view, these are considered “Software as a Medical Device” or “Medical Device Software”. Still, in addition to the specific requirements for Software Quality Assurance (see Chap. 3), such medical devices should be certified for their model credibility (see Chap. 4). Computational modelling and simulation results might eventually be included in regulatory submissions. In that case, incorporating this predicted evidence needs to follow data/evidence generation, analysis, and reporting standards to enable the regulatory bodies (and HTA agencies) to assess the submitted material efficiently.

In 2021, the Horizon Europe Framework Horizon program envisaged a line of action to provide regulatory agencies and HTA bodies with the necessary tools to exploit the potential of synthetic data¹⁰ for decision-making in the field of regulation and health technology assessment. One of the challenges of research in this area is determining the evidence value of this information source. Overall, there is a need for rigour and transparency in the methods used for in silico model development and validation, as well as their wider acceptance as a valuable source of evidence by the scientific community, including academic researchers, the pharmaceutical industry, regulatory bodies and HTA/payers (Musuamba et al., 2021). A need exists for documenting the available tools, the manners they are being used, the conditions for their adequate use and the challenges encountered. The current hurdles for the wider acceptability of in silico models as a reliable source of evidence for high (HTA) impact applications in drug/medical devices development include:

• lack of common standards and best practice documents commonly accepted by all relevant stakeholders,

• the lack of important digital infrastructure to carry out the in silico methodologies (e.g., fast communication networks and high computing power and storage capacity) that could compromise the cost-effectiveness of the resulting health technologies and the coverage, reimbursement or pricing decisions by the public healthcare systems (Leo et al., 2022),

• the protection of individual citizens from harmful use, also due to security breaches, of their personal data. An approach to solving the challenge surrounding big health data sharing is the generation of synthetic data created from real data by adding statistically similar information,

• biases in algorithm definition and poor training of analysts may pose risks to equity,

• poor communication between stakeholders in that regard,

• the deficit in the skills and knowledge essential to perform HTA based on in silico methodologies along the technology life cycle, and

• relatively slow development of regulatory science and HTA as compared to commercial solution developments.

Also, there is currently an unmet need for HTA guidance/best practice documents clearly describing standards for mechanistic in silico model development, evaluation and reporting considering the specificities not only in their structure and the data sources for their construction and evaluation but also in the software and algorithms used for their implementation.

Finally, further research is needed to understand the promises of the use ofin silico methodologies for the development and evaluation of health technologies, to improve their reliability, acceptance, and diffusion and to understand their expected impact on licensing and reimbursement decisions, as well as the role that HTA can have in the various phases of the application of in silico methodologies.

6.10 Definitions of Various HTA Modalities

Horizon scanning (Simpson and EuroScan International Network, 2014): this is the systematic identification of new and emerging health technologies that have the potential to impact health, health services, and society; and which might be considered for an HTA. Identification can be:

Proactive: where a range of sources are searched for information on new and emerging health technologies.

Reactive: where systems are in place that allows stakeholders, health professionals, developers and/or consumers to inform the Early Awareness and Alert (EAA) system on new and emerging health technologies

Pre-commercial procurement (PCP),^{11 12}: public procurers can drive innovation from the demand side by acting as technologically demanding customers that buy the development and testing of new solutions from several competing suppliers in parallel to compare alternative solution approaches and identify the best value for money solutions that the market can deliver to address their needs. PCP consists of a procurement of Research & Development (R&D) services that involves risk–benefit sharing at market conditions and in which a number of companies develop in competition new solutions for mid-to-long-term public sector needs. The needs are so technologically demanding and in advance of what the market can offer that either no commercially stable solution exists yet, or existing solutions exhibit shortcomings which require new R&D. R&D is split into phases: solution design, prototyping, original development, and validation/testing of a limited set of first products.

Early scientific advice (early dialogues)¹³ (Ijzerman & Steuten, 2011; Tummers et al., 2020): is a non-binding scientific advice, before the start of pivotal clinical trials (after feasibility/proof of concept study), in order to improve the quality and appropriateness of the data produced by the developers in view of future HTA assessment/re-assessment. Early HTA is increasingly being used to support health economic evidence development during early stages of clinical research. Such early models can be used to inform research and development about the design and management of new medical technologies to mitigate the risks, perceived by industry and the public sector, associated with market access and reimbursement.

Initial HTA: the early phase HTA helps technology owners or investors make evidence‑informed decisions about further investment in the development of medical device and other health technologies, especially with expected public reimbursement or procurement. It attempts to provide appropriate value judgement and assessment of health financing scenarios of innovative technologies before moving ahead with the development process or investing in technology.

Public procurement of innovative solutions (PPI)¹⁴: PPI happens when the public health systems bodies and providers use their purchasing power to address their identified challenges acting as early adopter of innovative solutions which are not yet available on large scale commercial basis, that are nearly or already in small quantity in the market and don’t need new R&D.

Mainstream HTA (Reuzel and Van Der Wilt, 2000): mainstream HTA entails scientific research into the effects and associated costs of health technologies and should support the decision-makers to decide on questions as ‘Is this technology better than the technology currently used?’, ‘How does it compare with alternatives in terms of effectiveness, appropriateness and cost of technologies?’ (see Sect. 6.3).

Coverage and Reimbursement policy¹⁵: in decision-making processes regarding the reimbursement of medicines, it needs to be established whether a medicine should be considered eligible for reimbursement. Subsequently, if the medicine is classified as ‘reimbursable’, it needs to be assessed how much of the price the public payer should (or can) cover. Therefore, setting a price (pricing) and deciding on the level of coverage by public payers (reimbursement) are strongly interlinked. The assessment process usually includes criteria such as efficacy, effectiveness, safety, ease of use, and added therapeutic value, besides cost-effectiveness. In some European countries, the same decision–making process is now used for digital therapeutics.¹⁶

Value-Based Public Procurement¹⁷: public contract based on the value it generates across the whole healthcare provision chain (from the patients to the healthcare professionals, the healthcare providers and the payors).

Re-assessment HTA¹⁸: Re-assessment assesses changes that may occur in medical technologies as they mature, as well as any new evidence available or other factors that can diminish past HTA findings and their utility for health care policies. As such, HTA can be more of an iterative process than a one-time analysis. Coverage and reimbursement policies and subsequent value-based public procurement contracts shall consider the results of HTA reassessments.

6.11 Essential Good Simulation Practice Recommendations

In silico methodologies can provide evidence to be used in HTA for:

• demonstrating value to payers by predicting the real-life benefit and the optimal target population for drugs or medical devices

• transposing Phase 3 trial results into virtual populations representative of specific geographies and context

• Benchmarking competing health technologies by also considering the market access of new technologies and the achieved effectiveness in the real world.