Abstract
Introduction
Reimbursement appraisals of medicines should emphasize outcomes clinically relevant to patients and clinicians. This study aims to characterize the outcomes selected in oncology reimbursement appraisal reports in Portugal.
Methods
A cross-sectional analysis of public INFARMED pharmacotherapeutic appraisal reports for oncology medicines published in 2023–2024 was conducted. For each included report, we extracted efficacy and safety outcomes, their clinical relevance rating (critical/important), whether data were presented for analysis, and which outcomes were prioritized in the final decision rationale. Descriptive statistics were employed.
Results
Fifty-four oncology reimbursement appraisals met the inclusion criteria. Overall survival was rated critical in 52/54 reports (96.3%) and prioritized in 29 (53.7%). Progression-free survival was rarely rated critical (n = 2, 3.7%) and was more often deemed important (n = 49, 90.7%), being prioritized in 15 (27.8%) reports. Quality of life was rated as critical in 52 (96.3%) reports but prioritized in only one (1.8%) decision. For safety, grade 3 to 4 AEs and mortality were graded as critical in 51 (94.4%) reports each, followed by discontinuation due to AEs (n = 50, 92.6%), although none of the appraisals prioritized safety outcomes in the final reimbursement decision.
Conclusions
Overall survival predominates as the most influential outcome in Portuguese oncology drug reimbursement decisions. Progression-free survival is frequently used as a practical surrogate. Although quality of life and safety outcomes are commonly prioritized during scope-setting, they are not consistently reflected in decision rationales, underscoring the need to strengthen outcome reporting and evidence generation to better inform patient-centered reimbursement decisions.
Keywords
Introduction
Oncology medicines have improved outcomes for many cancers, yet they remain a major driver of pharmaceutical spending and a frequent source of evidentiary uncertainty at the time of reimbursement.1–4 New oncology therapies often enter appraisal with immature survival data, limited follow-up, reliance on surrogate endpoints, and heterogeneous evidence on symptoms and functioning.5,6 These features create a tension between timely access to innovation and the need for robust evidence to support sustainable, equitable financing decisions.3,7–9
Health Technology Assessment (HTA) is a multidisciplinary, systematic process that evaluates the value of health technologies across domains such as clinical effectiveness, safety, economics, organization, and ethics, with the aim of informing fair and efficient decision making. 10 In the context of medicines, HTA commonly synthesizes comparative evidence, considers unmet need and therapeutic alternatives, and supports evidence-to-decision processes that balance health gains with opportunity costs.11–14
In Portugal, the National Authority of Medicines and Health Products (INFARMED) coordinates the national HTA system for medicines and leads pharmacotherapeutic appraisal to inform reimbursement and financing within the National Health Service (NHS).15,16 Scope-setting for pharmacotherapeutic evaluation is structured through a PICO (Population, Intervention, Comparison, Outcomes) framework. 16 It is followed by an evidence appraisal based on the GRADE (Grading of Recommendations Assessment, Development and Evaluation) methodology.17,18–20
In oncology, overall survival (OS) is generally regarded as the most direct and patient-relevant endpoint.5,6 However, OS is often challenging to demonstrate during appraisal because of trial design features such as crossover, subsequent lines of therapy, and insufficient follow-up.5,6 Consequently, progression-free survival (PFS), response rate, and other intermediate endpoints are frequently used to support early decisions, although their relationship with OS and patient quality of life (QoL) can vary substantially by indication and setting.4,6,21
Patient-reported outcome measures (PROMs), including health-related QoL instruments, can capture symptom burden and functional impact that are not fully reflected in traditional clinical endpoints.22,23 Regulators and HTA bodies increasingly encourage the use of PROMs and the development of electronic PROM systems.22,24 Yet reporting completeness and interpretability remain inconsistent, and PROM evidence is not always translated into decision rationales.6,23,24
Within this context, understanding which outcomes are prioritized in Portuguese HTA scope-setting and how those priorities are reflected in final reimbursement decision rationales is essential.17,25,26 In particular, it is important to assess whether outcomes classified as critical in PICO matrices are the same outcomes that ultimately drive funding decisions, and whether safety outcomes function as determinants of positive decisions or mainly as threshold criteria.18–20,25,27
The aim of this study was to characterize the clinical efficacy and safety outcomes reported in Portuguese HTA public appraisals for oncology medicines and to examine the extent to which these outcomes are prioritized in final reimbursement decision rationales.
Methods
Study design and data source
This is a cross-sectional analysis of publicly available INFARMED reimbursement reports for medicines indicated to treat oncological diseases. Reports were retrieved from INFARMED's public repository and covered assessments published in 2023–2024. 28
When multiple documents were available for a single appraisal (e.g., an updated version), the most complete document was used as the unit of analysis.
The Portuguese HTA follows the GRADE methodology and includes scope-setting (assessment matrix), evidence appraisal, and formulation of a recommendation that supports a financing decision.18,29 Public evaluation reports provide a written synthesis of this process and therefore offer a consistent source for examining which outcomes are formally prioritized in the final decision rationale.
Eligibility criteria
We included reports referring to medicines with an oncology indication (solid tumors or hematological malignancies) and describing a reimbursement/financing decision within the Portuguese National Health Service. Reports without a final decision (archived), not accessible, or without explicit data supporting the reimbursement decision were excluded.
Data extraction
For each eligible report, the following data was extracted: publication year; medicine and therapeutic indication; comparator(s); outcomes listed in the scope-setting, grading of the certainty of evidence (per outcome and overall), therapeutic value rating, and the final reimbursement decision (approved/rejected). Added therapeutic value (ATV) can be rated as major, moderate, minor, or non-quantifiable. The assessment can also conclude for therapeutic equivalence or for no therapeutic benefit.
Extraction was conducted using a standardized data collection template, and data was recorded at the report level.
Outcomes and definitions
Efficacy outcomes were recorded as stated in each report as well as patient-reported outcomes, including QoL. Safety outcomes were captured as reported, including mortality, grade 3–4 adverse events, and discontinuation due to adverse events.
Outcomes were classified according to the importance rating assigned during scope-setting (critical vs important). An outcome was considered prioritized for reimbursement when it was explicitly referenced in the final decision rationale as a key driver supporting approval or rejection.
To support interpretation, we also recorded whether each efficacy outcome was a direct endpoint (e.g., OS, QoL) or a surrogate/intermediate endpoint (e.g., PFS, response rate).
Analysis
Descriptive statistics were applied to describe the results. Microsoft Excel® was used to conduct analysis.
Results
Oncology accounted for 40.3% (25/62) and 30.7% (34/114) of INFARMED reimbursement reports published in 2023–2024, respectively, yielding a total of 59 oncology public reimbursement reports. Of these, five were excluded, three because the reimbursement assessments were not concluded and two because fundamental data was unavailable. A final sample of 54 reports with accessible full text and a stated reimbursement decision were included in the main analysis (Appendix A-Table A.1). The ATV from the drugs assessed was rated as non-quantifiable in most of the reports (n = 32, 59.3%), followed by moderate (n = 6, 11.1%), major (n = 5, 9.3%), and minor (n = 4, 7.4%) (Table 1). Four (7.4%) reports concluded for therapeutic equivalence, and 3 (5.6%) reimbursement requests were rejected due to no therapeutic benefit being identified. Overall, the quality of the evidence across the reports was predominantly low (n = 21, 38.9%) or moderate (n = 15, 27.8%), with fewer rated as very low (n = 6, 11.1%) or high (n = 3, 5.6%) (Table 1). Nine (16.7%) did not report/assess the quality of evidence. The most evaluated cancer type across the set of reports, based on the medicines’ therapeutic indications, were digestive (n = 12; 22.2%), hematological (n = 10; 18.5%), and breast cancer (n = 9; 16.7%) (Table 1). Pembrolizumab (n = 10, 18.5%), nivolumab (n = 6, 11.1%), and avelumab (n = 3; 5.6%) were the most frequently assessed drugs during the study period (Table 1).
Summary of the findings from the reimbursement reports.
Efficacy outcomes
Table 2 describes the clinical relevance rating, and prioritization in reimbursement decisions for efficacy outcomes. OS was classified as a critical outcome in 52 (96.3%) reports, and it was prioritized in 29 (53.7%) reimbursement rationales. QoL was classified as critical in 52 (96.3%), yet it was prioritized in only one (1.9%) decision. Both outcomes were not classified as important in any report. PFS was classified as critical in two (3.7%) reports and as important in 49 (90.7%) reports. This outcome was prioritized in 15 (1 important + 14 critical; 27.8%) decision rationales. Response rate was classified as critical in two (3.7%) reports and as important in 38 (70.4%) reports, but it was not prioritized in any decision rational. Invasive-disease free survival was classified as important in four (7.4%) reports and prioritized in three decisions (5.6%). Major molecular response (important), complete regression of actinic keratosis (critical), event-free survival (important), and recurrence-free survival (important) were prioritized in one (1.9%) decision each (Table 2).
Efficacy outcomes selected in the reports, clinical relevance, data availability, and priorization in decision-making.
Usually, the frequency at which data was presented for analysis was lower than the frequency at which the outcomes were selected in the reports: overall survival (n = 37, 68.5%), quality of life and progression free survival (n = 38, 70.4%, each), and response rate (n = 28, 22.2%) (Table 2).
Certainty of evidence for overall survival was most frequently assessed as moderate (=17, 31.5%), followed by low (n = 10, 18.5%) and high (n = 7, 13.0%) (Appendix A-Table A.2). For quality of life (QoL), moderate predominated (n = 16, 29.6%), followed by low (n = 12, 22.2%), very low (n = 9, 16.7%), and high (n = 1, 1.8%). For progression-free survival (PFS), moderate was the most common (n = 21, 38.9%), followed by high (n = 8, 10.8%), low (n = 6, 11.1%), and very low (n = 3, 5.6%). For response rate, moderate led (n = 17, 31.5%), followed by low and very low (n = 4, 7.4% each) and high (n = 3, 5.6%). Similar shares of not reported/ assessed the certainty of evidence was noted for these outcomes (approximately 22–27%) (Appendix A-Table A.2).
Safety outcomes
Table 3 describes the clinical relevance rating, and prioritization in reimbursement decisions for safety outcomes. Mortality was reported as a critical safety outcome in 51 reports (94.4%). Grade 3 to 4 adverse events (n = 51; 94.4%) and discontinuation due to adverse events (n = 50, 92.6%) were also commonly highlighted as critical. Nonetheless, the former two outcomes were classified as important in two (3.7%) reports each. Adverse events of special interest were classified as critical in four (7.4%) and important in two (3.7%) reports. Overall adverse events were rated as an important outcome in all reports (n = 54; 100%). None of the appraisals prioritized safety outcomes in the final reimbursement decision (Table 3).
Safety outcomes selected in the reports, clinical relevance, data availability, and priorisation in decision-making.
The frequency at which data was presented for analysis was lower than the frequency at which the outcomes were selected in the reports: mortality (n = 38, 70.4%), overall AEs (n = 39, 72.2%), and grade 3 to 4 AE (n = 40, 74.1%) and discontinuation due to AE (n = 40, 74.1%, each) (Table 3).
Certainty of evidence for mortality was most frequently assessed as moderate (n = 25, 46.3%), followed by low (n = 8, 14.8%), very low (n = 4, 7.4%), and high (n = 1, 1.8%) (Appendix A -Table A.3). For overall adverse events (AE), moderate ratings predominated (n = 27, 50.0%), followed by low (n = 5, 9.3%), very low (n = 4, 7.4%), and high (n = 3, 5.6%). For grade 3–4 AEs, moderate was the most frequent (n = 25, 46.3%), followed by low (n = 7, 13.0%), very low (n = 4, 7.4%), and high (n = 3, 5.6%). For discontinuation due to AEs, moderate evidence was the most common (n = 27, 50.0%), followed by low (n = 7, 13.0%), very low (n = 4, 7.4%), and high (n = 2, 3.7%). Similar shares of not reported/assessed certainty of evidence were noted for these outcomes (approximately 22–28%) (Appendix A -Table A.3).
Discussion
This cross-sectional analysis of INFARMED public evaluation reports for oncology medicines indicates that scope-setting routinely defines a broad set of outcomes, but reimbursement rationales emphasize a narrower subset. OS was the most influential outcome in written decision rationales, while PFS frequently served as a surrogate decision driver when OS evidence was immature. In contrast, quality of life and safety outcomes were commonly prioritized during scope-setting but rarely foregrounded as decisive in final funding rationales.
These findings have implications for transparency, evidence generation, and patient-centered decision making. They also suggest that evidence maturity and data availability may shape which outcomes become salient in the final reimbursement rationale, even when the formal scope assigns high importance to other endpoints.
Outcome prioritization and evidence maturity
OS emerged as the predominant decision driver, consistent with international expectations that OS is the most direct measure of clinical benefit in oncology. 6 However, OS evidence is frequently incomplete at the time of appraisal due to long follow-up requirements, crossover, and subsequent therapies. 6 In our sample, OS was not prioritized in almost half of decision rationales despite being classified as critical in most PICO matrices, suggesting that decision-makers sometimes faced substantial uncertainty about the magnitude or maturity of OS benefit.
The limited prioritization of QoL is also noteworthy. Although QoL was frequently classified as critical, it was prioritized in very few decisions. This may reflect limited availability of interpretable QoL data, heterogeneity of instruments, missingness, or uncertainty about the clinical meaning of observed differences.6,23 It may also indicate that QoL evidence is often treated as supportive rather than decisive when survival or disease-control endpoints dominate deliberations. 23
Improving the completeness and interpretability of QoL evidence could strengthen patient-centeredness in oncology HTA. 30 Practical steps include careful selection of PROM instruments, transparent handling of missing data, reporting results in clinically interpretable terms, and broader use of electronic PROMs to support real-world data generation relevant to HTA.30–32
Surrogate endpoints in oncology reimbursement decisions
PFS was seldom classified as a critical outcome but was frequently prioritized in reimbursement rationales. This suggests that, in practice, PFS often functions as a key indicator of benefit when OS evidence is unavailable, confounded, or immature. 33 While surrogate endpoints can enable earlier decisions and improve access, their validity is context-specific and they do not always translate into longer survival or improved QoL. 6
The observed gap between scope-setting priorities and decision rationales may therefore be partly explained by differences in the maturity of endpoints at the time of appraisal. Outcomes with strong patient relevance may be designated critical, but if data are incomplete, decision-makers may rely more heavily on available surrogate evidence. This reinforces the importance of post-launch evidence generation, including managed entry agreements and follow-up studies designed to confirm clinical benefit and capture patient-relevant outcomes.
Implications for reporting and evidence generation
The findings suggest that written reimbursement rationales may under-represent outcomes that are formally prioritized during scope-setting, particularly QoL and safety. This can have practical implications, since manufacturers and researchers may prioritize generating evidence on endpoints that are most likely to influence funding, potentially reinforcing a cycle in which patient-centered outcomes remain under-developed and under-used. These findings also highlight a gap between outcome selection and the availability of analyzable data within reports: outcomes frequently designated as critical were not consistently accompanied by presented results suitable for appraisal. Moreover, GRADE certainty ratings across key efficacy and safety outcomes were most often moderate or low, with high-certainty evidence uncommon, underscoring the need for clearer reporting of certainty assessments and for targeted evidence generation to reduce decision uncertainty. HTA bodies can mitigate this by strengthening methodological expectations for patient-reported outcomes and by specifying how QoL and safety evidence will be weighed alongside survival and disease-control endpoints. Clearer guidance on minimum reporting standards, clinically meaningful change thresholds, and integration of PROM data could support more consistent interpretation across appraisals. 32
In addition, increased use of real-world evidence after initial funding may help address uncertainties identified at appraisal. 34 Structured post-launch evidence plans that include survival follow-up, toxicity monitoring, and patient-reported outcomes can help confirm benefit, refine positioning, and support reassessment when initial decisions rely heavily on surrogate endpoints.35,36
Safety outcomes as decision drivers
Safety-related outcomes exemplified the divergence between evaluation and decision prioritization. Although frequently included and often classified as critical, none were prioritized as the primary basis for reimbursement decisions. This pattern suggests that safety functions primarily as a threshold requirement: therapies must demonstrate an acceptable benefit–risk profile, but positive funding is justified mainly by evidence of efficacy. 37
In oncology, where toxicities can be substantial and treatment goals vary across disease stages, the limited explicit use of safety outcomes in written rationales may reduce transparency about how tolerability is weighed against efficacy, particularly when survival benefit is uncertain. Clearer articulation of benefit–risk trade-offs, including the clinical relevance of adverse event profiles and their impact on QoL, could improve the interpretability of reimbursement decisions.
International context
HTA bodies tend to patient-relevant outcomes such as OS when assessing the clinical effectiveness of oncology medicines instead of surrogated outcomes.4 Regarding safety, the requirements for assessment vary considerably across countries. 4 However, countries diverge in their final reimbursement decisions due to local budget constraints, thresholds for clinical benefit, and heterogeneity in evidentiary standards.4,38,39
Our findings are consistent with several countries where real-world decision-making practices diverge from formal guidelines.40–42 Although HTA agencies generally acknowledge OS as the gold-standard clinical outcome, the practical challenges associated with generating robust OS data, such as long follow-up periods and confounding subsequent therapies, often limit its use in reimbursement decisions.40–42 Consequently, PFS is frequently accepted as a surrogate endpoint, despite the uncertainty surrounding its predictive value for OS.40–42
Also, QoL is frequently selected as principal clinical outcome, however it is often treated as a secondary endpoint with limited influence on reimbursement decisions. 40 Studies across 6 European countries and reached conclusions similar to ours: although QoL is theoretically considered an important outcome in the assessment of new oncology medicines, it contributes minimally to decision-making in practice.42,43 This is largely due to the frequent unavailability of QoL data or concerns regarding its methodological robustness, resulting in limited impact on final recommendations.42,43
Comparative research indicates that reimbursement rates for oncology medicines vary widely across jurisdictions despite similar clinical trial evidence, underscoring the role of local decision thresholds and budget impact considerations.4,38,39 Portugal's emphasis on OS and frequent practical reliance on PFS are therefore best interpreted within a broader context where HTA systems negotiate common evidentiary challenges but may articulate them differently in published rationales.
Atezolizumab for advanced hepatocellular carcinoma is an example of a drug included in our analysis that received similar recommendations across European HTA agencies (INFARMED, National Institute for Health and Care Excellence (NICE), the Haute Autorité de Santé (HAS), and Institute for Quality and Efficiency in Health Care (IQWIG)), despite uncertainties in overall survival data, illustrating alignment when evidence is interpreted similarly.44–47 In contrast, ramucirumab for advanced gastric cancer, although supported by robust evidence including overall survival, demonstrated only a limited survival benefit, leading to divergent decisions, while some jurisdictions accepted it as INFARMED and HAS, others such as NICE did not recommend it due to cost-effectiveness concerns.48–50 Together, these examples highlight that reimbursement decisions are not determined solely by the type of outcomes assessed, but also by the magnitude of benefit, associated uncertainty, and economic context.
Limitations
This study is limited to what is explicitly reported in publicly available appraisal reports and cannot capture deliberations that occur outside the written rationale. Identification of “prioritized” outcomes relied on explicit wording in the final motive(s), which may underestimate implicit considerations. One eligible report could not be accessed, and a small number of appraisals were excluded because the rationale was described as predominantly economic.
Despite these limitations, the analysis provides a structured overview of how outcomes are framed during scope-setting and how they are reflected in written reimbursement rationales, using a consistent extraction approach across a complete two-year set of publicly available oncology appraisals.
Finally, the study focuses on appraisals published in 2023–2024 and therefore does not address potential changes over longer periods or the influence of evolving methodological guidance. Future analyses could extend the time horizon and examine whether the integration of PROMs and real-world evidence increases over time.
Conclusions
In Portuguese oncology HTA appraisal reports, OS is the most influential clinical outcome supporting the reimbursement decision rationales, while PFS frequently serves as a surrogate driver when OS evidence is immature. QoL and safety outcomes are routinely prioritized during scope-setting but are rarely foregrounded in the decisions.
Strengthening the availability, quality, and interpretability of QoL and safety evidence, and making explicit how these outcomes contribute to the final therapeutic value judgement, could improve transparency and patient-centeredness of reimbursement decision making. Future research should assess how outcomes are operationalized in managed entry agreements and how real-world evidence (including PROMs) is incorporated after initial funding by NHS.
Supplemental Material
sj-pdf-1-opp-10.1177_10781552261449185 - Supplemental material for Clinical outcomes in HTA appraisal reports of oncology medicines in Portugal: A cross-sectional analysis
Supplemental material, sj-pdf-1-opp-10.1177_10781552261449185 for Clinical outcomes in HTA appraisal reports of oncology medicines in Portugal: A cross-sectional analysis by Pedro Teodoro, Catarina Ribeiro, Diogo Mendes and Carlos Alves in Journal of Oncology Pharmacy Practice
Footnotes
Acknowledgments
The authors declare that there are no acknowledgments.
Author contributions
Conceptualization: PT, DM, CA; Data curation: PT, DM, CA; Formal analysis: PT, CA; Investigation: PT, CR, DM, CA; Methodology: PT, CA; Validation: CR, DM, CA; Supervision: DM, CA; Writing – original draft: PT, CA; Writing – review and editing: CR, DM, CA.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Catarina Ribeiro is supported by a PhD studentship from Fundação para a Ciência e a Tecnologia, I.P. (FCT) (2024.05098.BDANA).
Fundação para a Ciência e a Tecnologia, (grant number 2024.05098.BDANA).
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Supplemental material
Supplemental material for this article is available online.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
