Defining Biological and Clinical Plausibility: The DICSA Framework for Protocolized Assessment in Survival Extrapolations Across Therapeutic Areas

Abstract

Background

Numerous health technology assessment guidance documents emphasize the importance of biological/clinical plausibility of modeled lifetime incremental survival without clearly defining it.

Objectives

This paper defines biologically and clinically plausible lifetime survival extrapolations and proposes a framework to systematically assess this by comparing survival expectations estimated premodeling, with the final modeled survival extrapolations. This framework is embedded in a survival extrapolation protocol template, which ensures that both the expectations and extrapolations are based on unified, comprehensive evidence.

Methods

A targeted review was conducted of 29 guidance documents from National Institute for Health and Care Excellence, Pharmaceutical Benefits Advisory Committee, Haute Autorité de Santé, Canada’s Drug Agency, and European joint clinical assessment, focusing on survival analysis, evidence synthesis, cost-effectiveness modeling methods, and use of observational data.

Results

Survival extrapolations are biologically/clinically plausible when “predicted survival estimates that fall within the range considered plausible a-priori, obtained using a-priori justified methodology.” These a priori expectations should utilize the totality of evidence available and take into account local target setting (i.e., survival-influencing aspects such as patient population, treatment pathway, and country). Pre-protocolized biologically/clinically plausible survival extrapolation was operationalized in a five-step DICSA approach: (1) Describe the target setting as defined by all relevant treatment and disease aspects that influence survival; (2) collect Information from relevant sources; (3) Compare survival-influencing aspects across information sources; (4) Set pre-protocolized survival expectations and plausible ranges; and (5) Assess how trial-based extrapolations align with the set expectations by comparing modeled survival extrapolations to the range of values a priori considered to be plausible.

Conclusion

The definition of plausibility of survival extrapolations, the operationalization of its assessment, and the corresponding extrapolation protocol template can contribute to the transparent development of biologically/clinically plausible survival extrapolations.

Supplementary Information

The online version contains supplementary material available at 10.1007/s40273-025-01485-0.

Key Points for Decision Makers

Health technology assessment guidance documents emphasize that survival extrapolations should be clinically and biologically plausible, but they do not provide definitions of plausibility.

It is crucial to prospectively elicit the opinions of experts to validate a model’s plausibility, since eliciting them retrospectively may result in subjective judgement of the model outcome and potential bias.

In this paper, biologically/clinically plausible lifetime survival extrapolations are defined as “predicted survival estimates that fall within the range considered plausible a-priori, obtained using a-priori justified methodology.”

A five-step approach to assess biological/clinical plausibility is proposed and embedded in a new survival extrapolation protocol template.

The definition of plausible survival extrapolation, the operationalization of its assessment, and the corresponding survival extrapolation protocol template can contribute to the transparent development of biologically/clinically plausible survival extrapolations.

Introduction

The modeled impact of treatments on lifetime survival is a key driver for health technology assessment (HTA) of new medicines and therefore for reimbursement decisions. Accordingly, many HTA agencies have developed guidance documents that directly or indirectly reference modeling lifetime survival and incremental survival. Many of these guidance documents highlight the importance of generating biologically/clinically plausible survival extrapolations and of making underlying biologically/clinically plausible modeling choices. However, they do not clearly define and operationalize this. National Institute for Health and Care Excellence (NICE) Technical support document (TSD) 21 provides the most concrete guidance, stating that ensuring the extrapolations are realistic often requires external data, clinical expert input, or consideration of biological plausibility—such as expected survival function shapes, patient characteristics, or disease-specific characteristics [1]. It further states that retrospectively assessing the plausibility of extrapolations is inherently subjective and may result in personal bias [1, 2]. Understanding the plausibility of extrapolations using immature survival data or earlier endpoints is important for reimbursement decisions, as most of the modeled survival benefit is generated in the extrapolation period and therefore uncertain. This was illustrated in Tannock et al. [3] who found that many new drugs that showed an improvement in progression-free survival (PFS) failed to demonstrate a corresponding improvement in overall survival (OS) in later data cuts, highlighting that expected survival benefits may never materialize. Willigers et al. [4] illustrated that extrapolating immature trial survival data using standard parametric survival models—such as in chronic kidney disease—often led to implausible projections, emphasizing the uncertainty inherent in model-based approaches and the value of incorporating external information to improve plausibility. Building on this, NICE TSD 19 cites multiple instances where committees identified potentially implausible survival extrapolations due to biased modeling assumptions [5]. Incorrect survival extrapolations likely cause incorrect incremental survival predictions, incremental cost-effectiveness ratio's (ICERs), and finally reimbursement/pricing decisions, underscoring the importance of plausible survival extrapolations. Therefore, Daly et al. [6] suggested living HTAs for diabetes products, as they questioned the validity of surrogate outcomes used in diabetes models for predicting survival and incremental survival.

A first step toward creating biologically/clinically plausible survival extrapolations is made by Palmer et al., 2023, who proposed an algorithm that provides a systematic, evidence-based approach to justify the selection of survival extrapolation models for cancer immunotherapies, covering the standard and flexible parametric extrapolation approaches discussed in NICE TSDs 14 and 21. The authors, however, stopped short of offering a definition of biological/clinical plausibility [1, 7, 8].

There is a need for a quantifiable definition of biological/clinical plausibility of modeled lifetime survival extrapolations and for a prospective protocolized process to assess plausibility. Therefore, the objective of this article is threefold: (1) to define biologically/clinically plausible lifetime survival extrapolation; (2) to operationalize the prospective plausibility assessment of modeled lifetime survival extrapolations including presetting lifetime survival expectations before generating modeled lifetime survival extrapolations; and (3) to present a extrapolation protocol template embedding the definition and the operationalization of biological/clinical plausibility. While our paper addresses survival (as the time-to-event endpoint often given the greatest focus in HTA), many of our recommendations are equally applicable to addressing the plausibility of other time-to-event endpoints.

Methods

A targeted search was conducted in June 2024 of the websites of the following HTA agencies: NICE,1^,2 Pharmaceutical Benefits Advisory Committee (PBAC),3 Haute Autorité de Santé (HAS),4 Canada’s Drug Agency (CDA),5 and EU joint clinical assessment.6 These were selected because they are HTA bodies that frequently evaluate cost-effectiveness and offer detailed guidance on survival extrapolation. The search aimed to identify guidance documents directly or indirectly referencing survival extrapolations, and thus, plausibility with a focus on the following topics: survival analysis, evidence synthesis, cost-effectiveness modeling methods, and observational data.

All relevant HTA guidelines were reviewed in their entirety. Our recommendations were informed by a broad analysis of the content of these documents, including implicit guidance on model plausibility and recommendations directly or indirectly regarding survival extrapolations found within the documents. The terms “plausible” and “plausibility” were used to identify plausibility considerations and definitions of plausibility.

This information was combined with the overview of more general biological plausibility definitions applied in epidemiology presented by Whaley et al. [9] and the authors’ knowledge and experiences to inform the definition of biologically/clinically plausible survival extrapolations, the operationalization of its assessment, and the creation of a survival extrapolation protocol template.

Results

A total of 29 guidance documents were identified, including 5 on general modeling [10–14], 3 on cost-effectiveness modelling approaches [5, 15, 16], and 1 on observational data in general [17]. Regarding evidence synthesis, 13 considered indirect treatment comparisons (ITC) [18–30], 3 survival analysis [1, 8, 31], 3 treatment-switching methodologies [30, 32, 33], and 2 surrogacy [34, 35].

The review showed that besides the plausibility of survival extrapolations, plausibility of underlying model choices is also important for HTA agencies when assessing these models. The review further showed that the terms biologically plausible and clinically plausible are used interchangeably. In our view, biological plausibility is broadly defined by disease processes and treatment mechanisms of action, whereas clinical aspects are mostly defined by human interaction with the biological process. In general, biological and clinical aspects will jointly influence survival, therefore, in the remainder of the manuscript, the term plausibility will refer to both aspects.

In the sections below, we will define plausibility of survival extrapolations and propose a process for the operationalization of its assessment. Next, we will propose a survival extrapolation protocol template embedding the definition and the process for assessing plausibility of modeled survival extrapolations. Please note that the steps described in the next two paragraphs should be conducted partly in parallel (e.g., the last step of the assessment process should be conducted after the protocol is finalized).

Defining and Operationalizing Biological/Clinical Plausibility of Survival Extrapolations

We define clinically and biologically plausible survival extrapolations [1, 5, 8, 15], as “predicted survival estimates that fall within the range considered plausible a-priori, obtained using a-priori justified methodology.”

Expectations should align with the totality of evidence on biological and clinical aspects that may influence survival and with local HTA guidance, such as guidance on plausibility on trend in hazards, model approaches, and ITCs. They should be generated for the target setting, defined by, for instance, patient population, treatment pathway, and country. The target setting may reflect current medical practice or the trial setting, depending on the nature of the decision, available evidence, and the HTA agency [14, 17]. Guidance documents usually suggest that the target setting should reflect clinical practice [10, 11, 14].

For operationalization and quantitative assessment of biological/clinical plausibility, we propose a five-step process called DICSA building on the model development steps by Chilcott et al. [36].

DICSA Step 1

The first step of DICSA is to Define, describe, and understand the target setting of interest in terms of survival and the survival treatment effect and in terms of aspects influencing survival and the survival treatment effect. Here, the aspects may be related to disease processes, treatment pathway, or patient characteristics.

DICSA Step 2

The second step of DICSA relates to Information collection to support setting survival expectations. TSD 13 highlights the importance of reporting the data sources and the searches, so that users of the model can understand how sources of evidence came to be incorporated into the process and can judge whether the model is based on a plausible or acceptable set of evidence [15]. The following sources may provide valuable information:

• Clinical guidelines and clinical input (qualitative and quantitative)

• Existing health economic models

• Existing (systematic) reviews

• Historical (phase 1–4) trials

• Real-world evidence (RWE) studies

• Network meta-analysis and surrogacy evaluations

• Routine monitoring sources

Data derived from the sources above do not necessarily need to relate to comparator- or active-treatment-specific information; they can also concern the general disease process or information from the same drug class in the same indication, later lines, and or other indications or may just describe the population in the target setting.

Data may be in the format of previously observed or expected survival estimates and the uncertainty around these may be more qualitative in nature (for example, whether hazards are expected to increase or decrease over a particular time period given the nature of the disease).

DICSA Step 3

The third step of DICSA involves comparing patient-, treatment-, and disease-related aspects between relevant information sources and the target setting. The relevance of information sources is related to maturity of survival data, comprehensiveness of reported aspects, and applicability of these aspects to the target setting (sources that closely align with the target setting are preferred). Preference may be given to data sources that establish a clear link between the reported aspects and mature survival, such as those featuring published risk equations or individual patient-level data with mature survival data.

If there are no published survival data over the disease course for standard of care (SoC), one may need to fall back to epidemiological publications reporting published standard mortality ratios (SMR). Please note, when relying on SMR-like metrics, one likely needs to assume the disease course has the same shape as the general population mortality. If these types of metrics are not available (e.g., for orphan diseases), the authors recommend the sponsor to start retrospective survival data collection, via databases or chart reviews, as it is unlikely health authorities will accept modeled survival benefits if the disease course is fully unknown.

For active therapies, the focus should be on the survival treatment effect, defined by, for example, hazard ratios (HR) over (different segments of) the disease course, instantaneous hazards, or incremental survival. Typically, prior to pivotal trial readout, limited information on relative treatment efficacy of the investigational therapy is available. Therefore, one may need to rely on phase 2 studies, historical trials in the target setting that ideally compare the same drug classes as those in the pivotal trial, later lines of therapy, or other similar indications.

Experts need to assess how the differences in factors, such as treatment pathway or patient characteristics, between information sources and defined target setting impact survival and associated treatment effect. For this, we suggest developing disease process and service pathway problem-oriented conceptual models, as suggested in NICE TSD 13. These conceptual models should inform the qualitative discussions with clinical experts on life time survival and incremental survival [15]. As an extension to the problem-oriented conceptual models suggested in TSD 13, we suggest developing a problem-oriented conceptual model for the treatment effect. The treatment-relevant aspects and corresponding considerations influencing the lifetime relative or incremental survival are presented in Table 1. PBAC also provides relevant guidance on factors influencing the treatment effect [37]. Figure 1 was developed to help capture the longitudinal impact of different aspects on the treatment effect. Patients enter the treatment effect conceptual model while starting treatment, either the investigational therapy of the pivotal trial or SoC. Clinical experts then need to qualitatively describe how the anticipated treatment effect might evolve while on treatment, during potential off-treatment periods (e.g., due to finite therapy), and while on the subsequent treatment lines. It is important to start with a reference treatment effect (e.g., OS HR) from the study protocol, historical trials, and or risk equations. If the ongoing trial might not be generalizable to the target setting, effect modification also needs to be considered.

Table 1.

Aspects important for developing a problem-oriented conceptual treatment effect model

Main aspects	Considerations	Experts, please indicate and justify how is this anticipated to impact the survival treatment effect over the disease course?
Initial treatment (mechanism)	Mechanism of action (including disease suppression, disease modification, and cure)
	Treatment duration (e.g., finite, infinite versus until progression or loss of effectiveness)
	Impact of being off treatment
	Concomitant medication
	Impact of rescue medication
	Anticipated discontinuation rate and reasons for discontinuation (e.g., lack of efficacy, safety, patient/physician choice) and impact thereof
	Impact of compliance and impact of potential difference in compliance patterns between trials and adherence in clinical practice
	Mode of administration (oral and long-acting formulations) if influencing survival
	Impact of patient characteristics (e.g., treatment effect modifiers, natural aging process)
Surrogacy associations (especially important for chronic, low-mortality diseases)	Impact of primary endpoint on survival
	Impact of treatment effect on primary endpoint and treatment effect on survival
Prior (to pivotal trial) treatment	Impact on treatment pathway
	Impact on efficacy of initial and subsequent treatments
Subsequent treatments	Impact of initial treatment on choice of subsequent treatments
	Efficacy subsequent treatments affected by initial treatment (e.g., due to frailty, drug–drug interactions, and/or disease modification by initial treatment if applicable)
	Differences between trial and clinical practice in terms of subsequent treatments (choice, time of initiation, adherence/compliance, % of patients receiving subsequent treatments)

Fig. 1.

Illustrative problem-oriented model for factors influencing relative effectiveness in lifetime survival estimation. HR, hazard ratio; Tx, treatment

Where the treatment effect on an early endpoint is carried forward to a treatment effect on OS by assuming surrogacy, it is important that the experts consider the potential that an overall survival benefit might not materialize, e.g., surrogacy paradox [38]. For instance, patients achieving a pathological complete response (pCR) might be less aggressively treated in subsequent treatment lines than those who did not achieve a pCR, which might explain that pCR is not a valid surrogate and that a treatment effect on pCR does not automatically mean a treatment effect on OS [39].

DICSA Step 4

The fourth DICSA step concerns Setting expectations about survival in the target setting and corresponding plausible ranges by means of a structured expert elicitation (SEE) process. The information generated in DICSA steps 1–3, potentially formalized in an early model in combination with SEE techniques, can be used to generate predicted lifetime survival and incremental survival and corresponding plausible ranges [4, 10, 31, 40–50]. For all relevant treatment arms, SEE can include questions such as:

• What percentage of patients are expected to be alive at 2 years, 5 years, 10 years, and 20 years in the target setting for SoC?

• Given the percentages for SoC, what are the corresponding percentages for the active treatment?

For designing high-quality SEE, including expert selection, we refer to Bojke et al. and to the work of Ren et al. specific to the use of SEE for assessment of the plausibility of survival extrapolations [40–42, 51]. SEE can be used either directly within model extrapolations (such as is proposed in the survextrap package) or to set expectations for comparing extrapolations that do not directly use the data [52]. DISCA step 4 addresses the latter of these options. For more details on the SEE approach, please see the Supplementary Materials.

DICSA Step 5

The final step of DICSA is to Assess how trial-based extrapolations align with the set expectations by comparing modeled survival extrapolations to the range of values a priori considered to be plausible in DICSA 4. If they are aligned, the survival extrapolations are plausible. If not, the model extrapolations seem implausible. However, there are instances where model extrapolations might be plausible and DICSA 4 predictions may prove to be implausible, as plausibility depends on the knowledge of the day. For instance, for diseases with rapidly changing clinical practices, and trials investigating a treatment with a new mechanism of action, expectations formed at DICSA step 4 may no longer be current at DICSA step 5. Therefore, it is important to update the literature review (DICSA 2) at this point. When modeled extrapolations go outside the plausible ranges, it is crucial to assess why this might be the case. Here it is important to start with the standard of care arm survival expectations and extrapolations, as the survival extrapolations and expectations of the active arm depend on those for standard of care. Such an assessment may justify, despite violations, a conclusion that the modeled survival extrapolations are indeed plausible. For example, trial recruitment may not have reflected the predicted population, subsequent treatment use may be different from expected, or a therapy with a new mechanism of action might over- or under-perform the expectations.

If model extrapolations are implausible, one should consider using a different survival extrapolation model, adjusting for subsequent treatments, applying population adjustment techniques, and/or using SEE techniques to directly inform survival extrapolations.

The assessment and any potential justifications should be transparently reported in the HTA technical report submitted to health authorities as well as the original set expectations.

Figure 2 provides an overview of how the authors came to their findings.

Fig. 2.

Flow diagram of how HTA review and more general definitions of biological plausibility [9] led to the findings presented in this study

An infographic of DICSA is presented in the Supplementary Materials.

Survival Extrapolation Protocol Template

For HTA agencies, it is essential that both the underlying model assumptions related to survival extrapolations and the resulting survival extrapolations are plausible. Therefore, DICSA should be ideally embedded in an extrapolation protocol. Table 2 provides a survival extrapolation protocol template. The survival extrapolation protocol template was developed to reflect and integrate the key considerations outlined in the HTA guidance documents identified from NICE, HAS, PBAC, and CDA and supplemented by the authors’ experience.

Table 2.

Proposed survival extrapolation protocol template

Step	Section	Description
1	Target setting	Define the target setting in terms of clinical practice/pivotal trial and decision-making parameters, specifying prior, current, and subsequent treatments as well as patient characteristics. Address any significant gaps with expert input
2	Pivotal trial setting	Outline key attributes of the pivotal trial, including population, design, intervention, and comparator details, as well as the anticipated maturity of the time-to-event endpoints (the importance of maturity is explained below). Describe generalizability of pivotal trial to target setting
3	Information collection	Document data sources and the (systematic) search strategy supporting DICSA steps. Including PICOs for HTA-mandatory literature reviews in the document. If external data are anticipated to be needed to populate the modeled survival extrapolations, document (systematic) approach to identify and select the needed data sources (see also TSD 13) [15]
4	External data for survival extrapolation	Identify external data sources that may inform survival extrapolations, including data from the evidence network and RWE sources. This includes data that may inform subsequent modelled health states. Assess the generalizability and relevance of these data to the target setting in the health states they are applied to. List the assumptions made when applying these data in the model. Any missingness in terms of baseline characteristics, subsequent treatments, outcomes, and consequences also need to be described
5	DICSA 3 and conceptual modeling	Select information sources that present survival data and aspects influencing survival from onset of treatment/model. Describe differences in aspects between selected information sources and target setting. Describe disease process, service pathway, and relative-effectiveness conceptual models, and qualitatively describe how these differences in aspects may influence absolute and relative treatment survival. The conceptual modelling may help to provide a qualitative understanding of whether there is a causal association between exposure to treatment and survival benefits
6	DICSA 4 and 5	See paragraph 3.1. DICSA 4 can be completed as part of the protocol. DICSA 5 cannot, but DICSA 5 can be described in the protocol template in terms of how predictions and modeled extrapolations are compared and how to handle deviations
7	Model approach and structure [5, 15, 16]	Specify the modeling approaches (partitioned survival analyses [PartSA], state transition models [StateTM], and patient-level simulations) on the basis of the model approach considerations reported in the relevant guidance documents. Please note CDA has a preference for StateTM [31]
		Define model structure on the basis of patient-relevant clinical events and treatment pathway, ensuring compatibility with conceptual modeling described in step 5
8	Adjusting model data for target setting	Describe generalizability and missingness mitigation strategies for all data sources informing the model (see protocol steps 2 and 4). Please consider, e.g., but not limited to: population adjustment [25], censoring adjustments [5, 53], treatment-switching adjustments [32, 33], and imputation techniques [17]. Any issues that cannot be solved, e.g., due to lack of data, requiring assumptions, may require specification of structural uncertainty [10] parameters as suggested by Strong et al. [54]
9	Survival Extrapolations	Describe whether there is a need to consider general population mortality within a relative survival approach. If not, describe another approach to account for general population mortality [1]
		If relevant historical trial data identified in the previous step include survival information, the following approach may be informative for selecting plausible (flexible) parametric survival extrapolation methods (as outlined in TSD 14 and 21) once the IPD from the pivotal trial becomes available:
		1. Create pseudo-IPD for trials in the evidence network.
		2. Adjust the follow-up duration of these trials to reflect the anticipated follow-up period for the pivotal trial at the time of HTA submission, by censoring events beyond that timepoint.
		3. Fit parametric and flexible parametric survival models to the reduced follow-up data from these trials.
		4. Compare the accuracy of the extrapolations in the extended follow period of these historical trials by analyzing the difference in area under curve (AUC) of observed survival beyond anticipated follow-up for pivotal trial with the extrapolations in the extended follow-up based on steps 2 and 3. [55, 56]. The smaller the delta in AUC, the more plausible it is that the extrapolation method will accurately predict long-term survival
		• For selecting standard and flexible survival models for the pivotal trial, we also refer to Palmer et al., 2023 [7, 57]. However, we prefer flexible parametric models with clinically interpretable parameters, such as mixture models and piecewise models, over more complex mathematical models such as cubic splines and fractional polynomials
		• For state transition models, once pivotal trial data are available, to understand/inform modeled relative efficacy beyond the first health state, assess and compare post-progression survival in both treatment arms [5]
		• Pre-protocolize selection of ITC approach in terms of, e.g., standard versus population-adjusted, anchored versus unanchored, or individual versus aggregate data, and align with parametric survival extrapolation approach [18–30, 58]. Consider the ongoing debate on population-adjusted ITCs related to marginal versus conditional treatment effects and shared effect modification assumption [59–62]. Please note that for population-adjusted ITCs, the selection of effect modifiers also needs to be pre-protocolized and clinical plausible [10]
		• Related to generalizability of subsequent treatments in health economic models, we refer to the individual recommendations by treatment-switching methods presented in NICE TSD 24. [32] This is an update of TSD 16, which described the treatment-switching problem and introduced a selection of adjustment methods that may be used, such as rank preserving structural failure time models (RPSFTM), iterative parameter estimation (IPE), marginal structural models (MSM) with inverse probability of censoring weights (IPCW), and two-stage estimation. [33] PBAC also has treatment-switch guidance [30].
		• For meta-analytic surrogacy analyses, it is important to assess biological plausibility of the association between treatment effects on the surrogate and on survival by assessing generalizability of underlying surrogacy data to the pivotal trial and target setting [35]. Any discrepancies should be reported and a description of how this affects the structural uncertainty and corresponding mitigation strategies should be presented. In addition, describe the underlying statistical model for assessing the surrogacy association and the corresponding uncertainty. Here NICE prefers bivariate models [34]
10	Sensitivity analyses	• Describe sources for all parameters and corresponding distributions, variance covariance matrices considered in the PSA [10, 11, 14, 31]
		• Describe structural uncertainty parameters [10] and corresponding distributions and mitigation strategies for consideration in the PSA, e.g., discrepancy approach or weighing approach [54, 63, 64]. We refer the reader to Strong et al. [54] and Jackson et al. [64] for a description on how to include structural uncertainty in the PSA
		• Describe all univariate sensitivity analysis [10, 11, 14, 31]
11	Scenario and subgroup analyses	Describe the subgroup analyses and the scenario analyses, such as mandatory treatment-effect waning scenarios [10, 11, 14, 31]
12	Model validation	Plausibility of modeled survival depends on quality of the underlying the model program. Many models submitted to HTAs include programming errors [65]. Therefore, it is important to describe model program validation strategies, incorporating established validation tools (e.g., AdViSHE for Australia and Coyle et al., 2024 for Canada) to detect and address technical errors in economic model submissions [66, 67]

NICE, National Institute for Health and Care Excellence; TSD, technical support document; RPSFTM, rank-preserving structural failure time models; IPE, iterative parameter estimation; MSM, marginal structural model; IPCW, inverse probability of censoring weights; PSA, probabilistic sensitivity analysis; AdViSHE, Assessment of the Validation Status of Health-Economic decision models; ITC, indirect treatment comparison; AUC, area under curve; IPD, individual patient-level data; PartSA, partition survival analyses; StateTM, state transition model; PICOs, population, intervention, comparators, outcomes, study design; HTA, health technology assessment; RWE, real-world evidence

Discussion

Many HTA agencies have highlighted the importance of modeling the impact of treatments on lifetime survival on the basis of biologically/clinically plausible extrapolations without clearly defining or operationalizing this. This paper defines biologically and clinically plausible lifetime survival extrapolations and proposes a framework called DICSA to systematically assess their plausibility. The assessment involves comparing prospective preset survival expectations with the final modeled survival extrapolations. DICSA is embedded in a survival extrapolation protocol template, which ensures that both the expectations and the extrapolations are based on a unified protocol and comprehensive evidence.

The strength of this protocolized approach is that it is structured, transparent, and is grounded in totality of evidence. Prospectively setting survival expectations may help reduce subjectivity and personal bias, which are associated with retrospective plausibility assessment of survival extrapolations [1, 2]. However, prospective plausibility assessment may not necessarily fully avoid personal bias. Therefore, to further reduce the risk of subjectivity, we recommend discussing the survival extrapolation protocol template including the preset survival expectations and plausible ranges in early HTA consultation meetings.

In the DICSA approach, SEE is used to validate the plausibility of survival extrapolations. This is different from using SEE in the estimation [4, 46]. Considering expert opinion in the estimation process implies that SEE cannot be used to validate the plausibility of survival extrapolations, as that would result in circular reasoning. However, Bayesian methodology could be applied to combine set DICSA survival expectations and modeled survival extrapolations after validation.

Plausible survival extrapolations are defined for a specific target setting during core model development. However, this target setting may vary by country due to differences in factors such as subsequent treatments and patient characteristics. For local model adaptations, it is essential to evaluate the DICSA report to determine whether the target setting remains representative. If not, DICSA 1–4 can be focused on identifying differences between the original and new target setting, and in collaboration with local experts, establishing locally plausible survival expectations.The DICSA framework is aimed to be generalizable to any therapy for any disease for which a lifetime survival benefit is anticipated to be predicted by the model for HTA purposes. For diseases with little survival data available (e.g., orphan diseases and therapies with a new mechanism of action), setting plausible lifetime survival expectations is likely more complicated compared with more prevalent indications and subsequent in-class therapies. In these situations where lifetime disease course, underlying process, and corresponding impact of therapies is not well understood, it is particularly important to generate retrospective SoC survival data (e.g., based on chart reviews or database studies) to help inform the DICSA process and the health economic model.

A potential limitation of the proposed approach is that the SEE may result in large plausible ranges, especially for diseases with limited survival data or for therapies with a new mechanism of action. With wide plausibility ranges, DICSA will not be very discriminating as to whether the eventual modelled extrapolations are plausible.

A complicating factor for diseases with very limited survival information available is that there is also limited HTA guidance on how to model these diseases, whereas the number of assumptions required to model lifetime survival are much bigger compared with diseases where mature survival data are observed in the trials. The following gaps in guidance were identified: description of minimal evidence base needed to justify modeling a survival benefit; development of generalizable risk equations based on external data including the considerations for inclusion of covariates; development of generalizable associations without explanatory parameters; and quantification of the structural uncertainty related to linking trial outcomes with external data for modelling survival [54, 64]. In addition, since CDA-AMC has recommended StateTM over PartSA, more guidance is needed on developing HTA-ready StateTMs, addressing the issues related to StateTMs documented in TSD 19 [5].

We acknowledge that the recommendations reflect our own perspectives on the basis of the review conducted. Future work could involve collaboration with broader stakeholder groups or expert panels to validate and refine these recommendations, increasing their generalizability and acceptance. In addition, the framework can be further refined using case studies involving models for actual HTA submissions. A useful starting point for these case studies involves treatment in indications where a survival benefit is expected during the extrapolation period but has not yet been observed in the initial trial. These situations have less guidance, more choices, and consequently greater uncertainty in model extrapolations related to biologically/clinically plausible survival extrapolations.

The definition of plausibility of survival extrapolations, the operationalization of its assessment, and the corresponding model survival extrapolation protocol template can help increase transparent development of biologically/clinically plausible survival extrapolations.

Supplementary Information

Below is the link to the electronic supplementary material.

Declarations

Funding

This research received no external funding.

Medical Writing

Medical writing support was provided by Colleen Dumont who is a writer employed by Cytel, and by Jane Adam, who is one of the co-authors.

Conflicts of Interest

The authors declare that they have no conflicts of interest relevant to the content of this article.

Data Availability

Statement on availability of data and materials is not applicable.

Ethics Approval

This study did not require ethics approval as it did not involve human participants or animals.

Author Contributions

All authors contributed to the conception and design of the study. Bart Heeg conceptualized the paper’s content and drafted the manuscript. Dawn Lee provided critical review from an ERG perspective. Jane Adam offered critical revisions, particularly on HTA, clinical aspects, and biological plausibility. Maarten Postma provided critical revisions from an academic perspective. Mario Ouwens helped conceptualize the papers and contributed critical revisions. All authors read and approved the final manuscript.