Health technology assessment in low- and middle-income countries: a case study of trastuzumab for early and locally advanced HER-2 positive breast cancer in Tunisia


 The role of health technology assessment (HTA) in low- and middle-income countries (LMICs) has become increasingly important in order to inform resource allocation and wider health policy decision-making. Using the case study of Tunisia’s first published HTA on trastuzumab in early and locally advanced HER2 positive breast cancer, we aim to evaluate how the key principles and elements of HTA might be applied in the assessment of health interventions undertaken by emerging HTA agencies in LMICs, in light of likely capacity and resource constraints. We critically evaluate the Tunisian HTA in the context of widely cited ‘best practice’ guidelines to determine how the assessment conducted by the Tunisian National Authority for Assessment and Accreditation in Healthcare aligns with recommended HTA methods. We find the Tunisian trastuzumab HTA to be an encouraging example of the ability of a newly established agency to conduct a transparent and scientifically rigorous evidence driven assessment. We draw on lessons learned in other countries and present a number of ongoing research and policy initiatives to promote the implementation of HTA, given resource constraints; in order to formulate recommendations for future HTA practice in Tunisia and LMICs, respectively. We discuss the sharing of evidence and resources and the potential of stakeholder engagement in building and strengthening HTA capacity worldwide.


Background
In 2014, the World Health Organization (WHO) reiterated the critical role of health intervention and technology assessment to inform priority setting and resource allocation, reduce inefficiencies, and ultimately achieve universal health coverage [ 1,2 ]. In the past decade, global stakeholders from all sectors of healthcare have embraced evidence-based health policy development and decision-making, resulting in a growing number of initiatives to promote health technology assessment (HTA) particularly in low-and middle-income countries (LMICs). Initiatives such as the Guide to Health Economic Analysis and Research (GEAR), Decide-Health Decision Hub, the International Decision Support Initiative (iDSI), Thailand's Health Intervention and Technology Assessment Program (HITAP) international unit or the National Institute for Health and Care Excellence (NICE) international services provide a collection of online resources and tools to share, establish and strengthen HTA capacity worldwide [2][3][4][5].
A key principle of HTA, widely acknowledged by international agencies-both established and emerging, is the use of scientifically rigorous and evidence driven methodology to ensure quality and consistency of assessments, over time and across health interventions, and to present these in a transparent manner [ 6 ]. In May 2020, the International Network of Agencies for Health Technology Assessment (INAHTA) released a new definition of HTA which notes the process of HTA should be: "formal, systematic and transparent, and uses state-ofthe-art methods to consider the best available evidence" [ 7 ]. However, to date no universally comprehensive framework or 'how-to' methods guide for HTA currently exist, or indeed is ever likely to exist, and the practice of HTA remains intrinsically shaped by a range of country-specific factors, policy agendas and institutional contexts, as well as, social and cultural values [ 8 ]. In addition, inherent challenges of implementing HTA methods in LMICs, namely limited resources, capacity and capabilities in the key disciplines of HTA, and potential lack of local data to inform decision-making, can curtail efforts by new agencies to adhere to HTA best practice [ 9,10 ].

Case study: trastuzumab in Tunisia
In Tunisia, breast cancer is the primary cancer in women and represents the most common cause of cancer mortality [ 11,12 ]. The standardised incidence of breast cancer in Tunisian woman has nearly tripled since 1994, reaching 50.17/100,000 persons years in 2017 [ 13,14 ]; it is projected to grow even further in coming years to eventually match the incidence of HICs. HER-2 positive breast cancer patients represent between 15 to 25% of all cases [ 15 ]. Trastuzumab-a targeted cancer therapy-was granted market authorisation by Tunisian regulatory authorities for early or locally advanced breast cancer in 2008. Although it is considered the 'gold' standard in the adjuvant setting and already used in clinical practice in Tunisia, its funding in its licensed indication represents the largest drug cost burden for the national health service, raising concerns regarding its cost-effectiveness and 'economic acceptability' [ 13,15 ]. The health care system in Tunisia includes both public and private sectors; however, the publicly run Caisse Nationale d'Assurance Maladie (CNAM)-the compulsory national social health insurance schemeprovides care for the majority of the population [ 16 ]. In 2013, the WHO reported more than 90% of the Tunisian population were covered through the CNAM or the free medical assistance programme for the most vulnerable [ 16,17 ]. Public sector coverage under the CNAM includes health centres providing primary care, district and regional hospitals, and university hospitals.
In May 2017, INEAS was commissioned to conduct an HTA of trastuzumab for early and locally advanced HER2 positive breast cancer; in July 2018, it published its first guidance online [ 15 ]. The scope of the INEAS assessment was to conduct a pharmacoeconomic analysis and evaluate the net clinical benefit of trastuzumab for early and locally advanced HER2 positive breast cancer. Their evaluation included: (1) a review of the literature to summarise the net clinical benefit of trastuzumab compared with standard chemotherapy; (2) a comparison of international clinical practice guidelines recommendations; (3) an international drug price comparison; (4) a survey of international HTA agencies on the assessment of trastuzumab; and (5) a CEA.

Objectives and methods
We use the case study of the first published HTA by INEAS-trastuzumab (Herceptin®) in early and locally advanced HER2 positive breast cancer [ 15 ]-to critically compare the approach and methods used by an emerging HTA body to widely cited good practice guidelines in HTA [ [18][19][20][21][22], given likely resource and capacity constraints. The degree to which this is possible is limited by a country's willingness to share detailed descriptions of the methods used and results of their technology assessments. The Tunisian National Authority for Assessment and Accreditation in Healthcare (INEAS) was created in 2012 with an overarching goal to regulate the national health system by promoting quality and efficiency [www.ineas.tn/fr]. Within its newly acquired mandate, HTA is at the core of its mission. In this case, we refer to the WHO's definition of HTA as: "the systematic evaluation of properties, effects and/or impacts of health technologies and interventions" [ 1 ]. This multidisciplinary process of clinical and economic assessment is taken as distinct from the appraisal or decision-making, informed by this assessment, but which inevitably involves a range of other considerations and which is typically less transparent.
The HTA guidelines used were identified via the INAHTA website, the GEAR health economic evaluation guidelines database, and the ISPOR Outcomes Research Guidelines Index. These include the landmark report from the European Collaboration for HTA/Assessment of Health Interventions (ECHTA/ECAHI) project promoting best practice in undertaking and reporting HTA [ 18 ], the Centre for Review and Dissemination (CRD)'s guidance for undertaking systematic reviews in health care [ 22 ], the WHO guide to cost-effectiveness analysis (CEA) [

Critical assessment HTA protocol and research questions
The decision problem put forward by INEAS focused on a therapy already in common use in Tunisia, as opposed to a new health intervention. The stated rationale for this first assessment was the growing financial strain trastuzumab presented on the national health system and the need to explore the 'efficiency' of the drug in the given indication [ 15 ]. The report by INEAS clearly stated the policy question motivating the assessment and provided succinct, yet sufficient, background information on the condition and target population, technology, and current practice to contextualise the research questions posited. However, the report could have benefited from an a priori defined HTA protocol, describing explicitly how each stated objective was to be addressed, how and by whom [ 18,35 ].
In particular, which methodology was to be followed, based on which data were to be available, and how the retrieved evidence was to be synthesised. It should also be noted that despite the relatively high burden of disease and considerable percentage of health expenditure devoted to breast cancer in Tunisia [ 14,15 ]; at the time the assessment was published, the WHO had launched a pilot procedure for prequalification of biotherapeutic products including trastuzumab [ 36 ] and the market entry of biosimilars was likely to change the distribution and/or pricing of trastuzumab in the country.

Findings and methodology: safety, efficacy/effectiveness
In accordance with good practice guidelines, a systematic approach to identify relevant clinical evidence was presented by INEAS [ 18,20,22,37 ]. This included a pre-defined search strategy combining free-text and keywords, a stepwise study screening process by two independent reviewers, and a quality assessment of the included INEAS is indeed more pragmatic, perhaps given time and/or resource constraints; however, this was not explicitly stated or justified by assessors and potential sources of bias in their approach were not reported.

Findings and methodology: economic issues
One of the key elements of the INEAS report was also acknowledged by assessors as its biggest limitation, that is the lack of a Tunisian-specific economic model and the absence of 'contextualised' data to inform a CEA.
INEAS stated that despite an appeal to manufacturers to provide a CEA from the perspective of the Lastly, it should be noted that wider psychological, social, and ethical considerations, the patient perspective, as well as organisational and professional implications referenced in a number of good practice guidelines, were not explicitly reported in the INEAS assessment but may have been discussed by appraisers [ 4,18 ].

Conclusions and recommendations
We find the first HTA by INEAS to be an encouraging example of evidence-based practice to inform decision-making in an LMIC. In fact, we commend INEAS for making their assessment publicly available, allowing this critique, and providing a unique opportunity for shared learning. Overall, the conduct and reporting of the assessment of trastuzumab's net clinical benefit for early and locally advanced HER2 positive breast cancer in Tunisia was well-structured and transparent. However, our review also highlighted important limitations in the assessment of the CEA of trastuzumab compared with standard chemotherapy for the Tunisian health system. We acknowledge the learnings from this case study are limited as only reflect a single HTA report by one emerging HTA agency in a LMIC; however, our recommendations are more far-reaching and could be generalisable to other LMICs. Following their survey of researchers who have experience in conducting economic evaluations in LMICs, Luz et al. find similar technical issues were reported across a number of assessments: "the lack of quality local clinical data, poor reporting, insufficient data to conduct the analysis from the chosen perspective (i.e. the lack of cost data), lack of context-relevant standards for economic evaluation, and lack of local health-related preference data, respectively." 10 Adapting or borrowing evidence across countries and HTA agencies can be a practical and efficient 'shortcut' to conducting HTA in LMICs, but it may not provide sufficient context to answer the policy and research questions set out by local agencies [ 50 ]. It may be misleading to assume that relative treatment effects are transferable, as the majority of randomised clinical trials for a particular condition may not be undertaken in LMICs and case-mix might differ between HICs and LMICs [ 51 ]. However, it may be challenging to develop or identify evidence regarding relative treatment effects in the local context. More important, resource utilisation, unit costs and quality of life data (and utilities using country-specific valuation sets) may differ substantially between LMICs for societal and cultural reasons; therefore care must be taken when comparing model inputs across jurisdictions, and ideally local data should be used [ 9,44,52 ].
The paucity of local data, or difficulties to obtain it, are well documented barriers to the implementation of HTA in LMICs [9,40]. Efforts to promote local data collection, especially through collaboration, should be prioritised.
Considering the resource and time implications, obstacles to the gathering or sharing of data already being collected (e.g. cancer registries, medical cost data) could be removed; issues with missing or poor quality data The challenges of local data collection also raise an important question regarding the 'burden of proof' and who bears the cost of making the case for the funding of a new technology, in the broader sense of, generating relevant and generalisable evidence, capturing contextualised data and modelling. In certain countries, the manufacturer is responsible for submitting this case for newly developed technologies. The degree to which manufacturers will be willing to submit assessments will depend on their own local resources and their perceived return on investment in such submissions. One consideration will be how the local agencies proceed if manufacturers decide not to make a submission. In this case study, INEAS decided to pursue the assessment using published evidence. It should also be borne in mind that for many technologies, such as those that no longer enjoy patent protection, there may not be a single identifiable manufacturer with a strong incentive to develop a submission.
The WHO and iDSI promote multiple stakeholder engagement as a tool to build capacity to support HTA Lastly, better reporting is an efficient way of improving the methodological quality of an HTA without necessarily adding to the resource burden of such assessments. Similarly, pre-defining a HTA protocol, data requirements, and preferred methods for HTA in a tailored methods guide produced by local HTA agency would provide further clarity and consistency to future assessments. In our case study, we highlight a number of instances were HTA elements were not substantiated enough or not reported by INEAS. However, it is likely that these were captured or discussed and could have been included, even in Appendix, for completeness and in line with best HTA practice, such as search strategies and data extraction, psychological, social, and ethical considerations, or organizational and professional implications.
A stated caveat of our review is that specific issues surrounding the appraisal of new health technologies in LMICs, intrinsically linked to HTA, were not considered. Learnings from other countries demonstrate that understanding how evidence generated during the HTA process is translated into resource allocation decisions, and how HTA is implemented as part of a wider health policy change, particularly in LMICs, is critical [ 57 ]. Decisionmaking regarding resource allocation to specific healthcare technologies leads to inevitable 'trade-offs' and these may be contentious, if restricting funding and access to potentially useful, cost-effective treatments [ 68 ]. It is important that the process used to arrive at such decisions is seen as being socially just, otherwise both individual decisions and the evaluation process itself are likely to be contested. In this respect, although stakeholder engagement may be perceived as 'slowing down' and increasing the 'cost' of HTA, it is an important component of a socially just decision-making process and may contribute to the acceptance and increase the impact of HTA based decision-making [ 69 ]. The 2020 roadmap for systematic priority setting and HTA, developed by Management Sciences for Health (MSH) and the USAID Medicines, Technologies, and Pharmaceutical Services (MTaPS) Program, echoes a number of our recommendations for the implementation of HTA in LMICs. In addition to identifying key challenges for LMICs, this roadmap also provides tools and approaches to support HTA efforts and help navigate the process of institutionalising HTA-"from agenda setting through formulation, adoption/implementation to impact evaluation" [ 68 ].
Our critical review compared the approach and methods used by INEAS in their assessment of trastuzumab to a number of published HTA methods guidelines. However, using traditional HTA elements and methods, without consideration for capacity and resource constraints, as a benchmark for 'best practice' in LMICs is inconsistent. It may be more appropriate to focus on how useful an assessment is, given an agency's resource constraints. Luz et al. find that the lack of standard practice that is relevant to LMICs was frequently reported as an issue by countries surveyed regarding the conduct, reporting and use of economic evaluations [ 10 ]. Despite challenges, it is promising that the core principles underpinning good HTA practice are endorsed by newly established HTA agencies such as INEAS [ 4,9,24 ]; and that ongoing research and global initiatives are continuously extending and adapting HTA guidelines to advance HTA practice in both HICs and LMICs [ 2,68,70 ].