Systematic review and meta-analysis: What’s next – embracing complexity and improving patient care!

Abstract

At the outset, the whole Editorial Board Team of the Indian Journal of Anaesthesia (IJA) wishes everyone a very happy new year. This special themed issue of IJA presents various systematic reviews and meta-analyses (SRMAs) on important clinically relevant topics and narrative reviews conceptualising different aspects of SRMAs. While finalising this January 2025-themed issue of IJA, I was thinking about what’s next in SRMAs. Yes, we agree that SRMAs, with their robust methodology, provide the highest-level evidence and are thus useful for bringing change in our clinical practice, but are the three core pillars of ‘evidence-based medicine’ integrated into the conclusions of these SRMAs? SRMAs use high-quality research to synthesise results and culminate in a comprehensive conclusion and thus guide clinical practice.[1-3] With a better understanding of various aspects of SRMAs and the availability and understanding of various advanced statistical tools, SRMAs are increasingly being published. The SRMAs are not ‘just SRMA’; SRMA with trial sequential analysis (TSA), pooled analysis, network meta-analysis (NMA), individual participant data (IPD) meta-analysis, etc., are also being published. In the future, the role of artificial intelligence (AI) in clinical research and SRMAs may not be underscored. In recent times, the traditional concept of SRMAs involving the pooling of data from published research to reach a conclusion has been challenged in real-world clinical scenarios. This is primarily because of many limitations in the applicability of these SRMAs’ conclusions. Emerging clinical concepts, technologies, strategies, drugs, and tools also require updating the SRMAs. Various analytical techniques are used to synthesise data from multiple studies with their own limitations and advantages [Table 1].[1-10]Table 1: Meta-Analytical MethodsThe availability of multiple gadgets, tools, and drugs limits the conventional SRMA for a definite conclusion, comparing only a few of them rather than all. This is where NMA has a definite role, comparing multiple gadgets, tools, and drugs to create a hierarchy of therapeutic/management/selection options. Even using this strategy may not work to the full extent as, at times, individual published studies used for meta-analysis may have different criteria for inclusion and exclusion. This issue needs the use of IPD meta-analysis, where raw participant-level data from multiple studies are sought from individual study authors, and a meta-analysis is conducted to enable personalised subgroup analyses. Is the conclusion stated from these conclusive? Is it sufficiently powered to conclude? What about the risk of random errors and false-positive findings? Herein, the need for TSA needs to be emphasised. TSA ensures the futility or need for further studies. EVIDENCE-BASED MEDICINE (EBM) AND SRMAs While SRMAs provide high-level evidence, it is imperative to integrate the three core pillars of evidence-based medicine—best research evidence, clinical expertise, and patient values—to ensure comprehensive and patient-centred care [Table 2].[10]Table 2: Integrating the Pillars of Evidence-Based MedicineIt needs to be emphasised that the core pillars of evidence-based medicine and existing challenges in evidence-based medicine implementations need to be integrated and addressed through SRMAs. It is the time for methodologists and researchers to incorporate tools that actively engage these pillars of evidence-based medicine, ensuring that SRMAs remain clinically relevant and patient-centred. ARTIFICIAL INTELLIGENCE AND SRMAs With the evolution of AI and its role in medical sciences, AI may also have a role in SRMA reporting.[11,12] However, the use of AI in research has its concerns and limitations, but its involvement in SRMAs needs to be further considered with due diligence. The various features, such as algorithm-based automated literature search and natural language processing for data extraction and synthesis, need further exploration. Predictive modelling can also identify gaps in existing research and guide future clinical trials. In addition, AI may be explored further for automated updating systematic reviews as new evidence emerges, creating ‘living reviews’. FUTURE DIRECTIONS FOR SRMAs The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Reporting guidelines ensures clear, transparent, and complete reporting of SRMAs. Not only research but also SRMAs need to focus more on patient-centred outcomes, including functional recovery, quality of life, patient/caregiver satisfaction, and return to intended oncological management [Table 3].Table 3: Future needs in SRMAsIt is the need of the hour to ensure and match the research priorities with real-world clinical priorities. Continuing with this parlance and a better understanding of the multidisciplinary approach, research and SRMA are needed with multi- and cross-disciplinary involvements.

Ähnliche Arbeiten