Authors’ response

We appreciate the author of the letter to editor¹ for diligently reading our article.² We wish to clarify that the figure presented in our article should be interpreted as a composite conceptual framework, rather than two independent pyramids. Our reference to the traditional pyramid was intended only as background context. The central argument of our article concerns the distinction between two fundamentally different scholarly activities: evidence generation and evidence synthesis.²

Our generation-integration framework clarifies the distinction between evidence generation and evidence synthesis. The left side represents evidence generation RCTs, cohort or case-control studies, and observational and qualitative research, each functioning independently to produce primary data. The right side represents evidence integration, including SRMA that synthesizes findings from these primary studies, which are mostly interlinked and dependent on other evidence generation methods. The framework does not rank methods but highlights their complementary roles: generation produces new evidence, while synthesis integrates existing knowledge.²

We agree that research hierarchies should be research question-based rather than rigidly hierarchical, as different questions require different methodological approaches. Our argument, therefore, promotes a question (research)- based perspective asking who should conduct SRMAs, when, why, where, and how?³ In practice, research rarely follows a linear path from observational studies to randomized trials; instead, multiple study designs often evolve simultaneously, guided by feasibility, available data, and the urgency of clinical questions.³

As per the claim, ‘approximately 140 randomised controlled trials and nearly 80 systematic reviews are published daily.’⁴ However, a robust systematic review typically requires 20–30 RCTs, suggesting that only about 20–30 high-quality SRMAs per day could be supported even when considering all types of SRMAs across disciplines. This disparity indicates a potential imbalance, raising concerns about redundancy and dilution of meaningful evidence synthesis.

We also note the rapid proliferation of workshops and short training programs on SRMAs.⁵ While such training is valuable, their expansion may create the perception that SRMAs offer a quicker route to publication than primary research. We caution that this trend may unintentionally reinforce the perception that desk-based synthesis is inherently superior to primary evidence generation.²

We reiterate that our article does not question the value of systematic reviews or meta-analyses.² Rather, it highlights the complementary relationship between evidence generation and evidence synthesis, emphasizing that robust synthesis ultimately depends on the availability of diverse, high-quality primary evidence.^6,7