Narrative approaches to reviews

Garg et al. have used a narrative review to summarize the effect of non-operative therapies in the management of low back pain.¹ The results of the review have to be approached with caution for the following reasons. Narrative reviews or literature reviews are criticized for being neither systematic nor transparent in their approach to the synthesis of the literature.² Although the authors claim to have conducted a literature review, the search terms, databases and duration of the literature research have not been reported. The review must make sure that all or the majority of the pertinent research on the subject has been identified and included after clearly outlining its objectives. Along with parameters linked to variability, such as critical analysis of the articles included in the initial reviews, the quality and consistency of the evidence must be addressed.³

The next major drawback is that the authors have not defined the clinical question for which the review was done, which is a prerequisite for Grading of Recommendations Assessment, Development, and Evaluation (GRADE).⁴ The outcomes and comparator agents have not been specifically reported for many of the reviewed interventions. It is also not clear how the authors graded the quality of evidence using GRADE as the effect size for each outcome in absolute terms was not provided.⁵ Though this may be partially explained in terms of diverse characteristics of the studies considered, the authors could have considered alternative synthesis of results such as summary statistics of intervention effect estimates, vote counting based on direction of effect, and combining p values, and could have presented the results as a synthesis without meta-analysis.⁶

The results of some of the studied interventions have not been stated clearly. For example, the section on education does not specifically mention the outcomes and the comparator agent that were considered to determine if patient education was successful. Further, certain recommendations are based on old references (e.g. section on exercises). A recent meta-analysis has concluded that pilates, McKenzie therapy and functional restoration were more effective than other types of exercise treatment for reducing pain intensity and functional limitations.⁷ The results of a network meta-analysis had concluded that the pilates, resistance and stabilization/motor control, and resistance and aerobic exercises are effective against true control only for pain, physical function and mental health, respectively.⁸ It may therefore be ensured that reviews adhere to preferred reporting items that help in the decision-making process.