Better Together: The Impact of Joint Clinical Assessments on Systematic Literature Reviews

The Health Technology Assessment (HTA) process in Europe currently mostly involves individual assessments conducted across different member states of the European Union (EU). The EU Joint Clinical Assessment (JCA) aims to provide a coordinated and efficient assessment process at an EU level and marks a significant milestone in harmonising the clinical assessment element of HTA across EU member states.

Systematic literature reviews (SLRs) play a vital role within the HTA process, underpinning a robust evidence-based approach to the numerous analyses that inform assessments. The new unified JCA approach will result in the need for a singular SLR that aggregates evidence that is relevant for all participating countries and adheres to consistent methodological and reporting requirements. To ensure the SLR scope addresses the needs of each member nation, the Patient, Intervention, Comparison and Outcome (PICO) criteria used in the search strategy must be methodically considered with reference to the predicted and actual PICO scope of the JCA following the scoping process (Figure 1). The influence of the JCA scoping survey on the JCA PICO criteria was evidenced in its application to 11 hypothetical pharmaceutical JCAs, with the process resulting in significant adjustments to various PICO elements. The level of agreement was greatest for the intervention and patient population, whilst the list of relevant comparators and outcomes were subject to the most discussion.¹ It is inevitable that many JCAs will require multiple PICOs within one SLR to accommodate for disparities among member states.

However, herein lies the major challenge. At the same time as requiring larger, more resource-intensive SLRs that can address diverse country market needs in one single review, the timelines of the JCA process impose only a narrow time window between knowing the final PICOs that need to be addressed and needing to deliver the SLR (and any related evidence synthesis) for inclusion in the JCA dossier submission. These two factors together pose a significant challenge to conducting SLRs for JCA on a practical level. In our view, this will necessitate careful planning and execution of each SLR used for a JCA:

• To ensure efficiency throughout the SLR, strategic integration of available technology to automate tasks will play a pivotal role in reducing the resource burden. This may include leveraging artificial intelligence and machine learning to streamline activities like article screening and data extraction.
• The final PICOs to be addressed in the JCA dossier will be available around three months before the submission is due. The cut-off data for running the SLR searches also needs to be a maximum of 3 months before dossier submission. Other than very specific circumstances (e.g. a known small literature base and confidence that no indirect treatment comparison is required), starting the SLR at this 3-month timepoint will not be possible – completing a robust SLR, digesting results and running any necessary indirect treatment comparisons based on SLR results in 3 months is not realistic. Therefore, to ensure the SLR identifies the most up-to-date evidence and meets the diverse needs of participating countries within the strict timelines of the JCA process, an iterative approach should be adopted for the SLR. This will involve starting with an initial SLR based on a broad predicted scope, followed by an SLR update using the adapted, and potentially narrower, scope created from the finalised PICO. For this approach to be feasible, meticulous attention must be paid to the initial SLR scope, ensuring cautious inclusivity with respect to the eligibility criteria, alongside categorisation of studies according to the predicted JCA PICO they apply to, to facilitate reporting within the dossier.

While the transition to JCAs may present immediate challenges for conducting sufficiently comprehensive SLRs that address all relevant PICOs in a timely manner, there could be long term benefits for the medical and HEOR community. The requirement for extensive, high-quality SLRs that capture a comprehensive evidence base across a disease area should result in efficiencies by minimising the production of numerous overlapping SLRs, often with limited scopes, that may be used for sequential local HTA submissions currently. These comprehensive JCA SLRs also offer a valuable opportunity for the significant investment of resource that goes into the review to contribute to the knowledge base of the disease area among the scientific and HEOR community at large. However, this relies on the methods and results of these SLRs being made public. We would advocate:

• That manufacturers should ensure prospective registration of their protocols, for example, on PROSPERO.
• That the JCA process makes SLR protocols and results publicly available alongside the submission dossier, whilst manufacturers are encouraged to promptly publish these reviews in peer-reviewed journals.

As SLRs are based on public data, there should be no concerns around making this element of the JCA dossier public, and this would provide benefit in that the investment made in this systematic research could then be leveraged more widely to support quality and efficiency of future similar reviews, for example by allowing sharing of database search terms.

Figure 1. Process steps for the development of the PICO question(s)

Figure adapted from the EUnetHTA PICO Concept Paper.¹ Abbreviations: PICO, Population Intervention Comparator Outcome.

References

1. EUnetHTA Joint Action 3 WP4 (2020). PICO Sub Group: PICO Concept Paper. Available here.