Using the Likelihood of Being Helped Versus Harmed Metric to Inform Neuropathic Pain Treatment Decisions

Choosing the most appropriate treatment for neuropathic pain is often challenging. Clinicians must weigh modest average treatment effects, heterogeneous patient responses, and the risk of adverse events, all within a therapeutic landscape that remains far from optimal. Systematic reviews and meta-analyses are essential in guiding these decisions, as are summary measures such as the number needed to treat (NNT) and the number needed to harm (NNH). Yet interpreting these metrics side by side is not always straightforward, and a single measure that integrates both benefit and harm could simplify decision-making for clinicians and patients alike. One such measure is the Likelihood of Being Helped versus Harmed (LHH). Defined as the ratio of NNH to NNT, the LHH expresses how many times more likely a patient is to experience a clinically meaningful benefit than a treatment-limiting adverse event [1]. This metric becomes particularly informative when applied to large, high-quality evidence syntheses. In their recent systematic review and meta-analysis, Soliman et al. [2] synthesized efficacy and safety outcomes across 313 randomized controlled trials. NNT was defined as the number of patients who must receive treatment for one additional patient to achieve a clinically meaningful pain response, based on the proportion of responders achieving ≥ 50% or ≥ 30% pain reduction or moderate pain relief. NNH, by contrast, was defined as the number of patients who must be treated for one additional patient to withdraw due to adverse events. Because this definition captures treatment discontinuation rather than the full spectrum of adverse effects, the resulting LHH values reflect the balance between benefit and withdrawal-related harm rather than overall tolerability. Using these definitions and the results provided in the meta-analysis, I calculated the LHH and corresponding 95% confidence intervals to illustrate how effectively this single metric can summarize benefit–harm trade-offs (Table 1). For example, an intervention with an NNT of 8.9 and an NNH of 26.2 (as reported for α2δ ligands) yields an LHH of approximately 2.9, indicating that benefit is nearly three times more likely than harm. Conversely, treatments with LHH values close to or below 1 suggest a finely balanced or potentially unfavorable benefit–harm profile. Some of the highest LHH values arise for second- or third-line interventions such as botulinum toxin A, capsaicin 8% patches, and rTMS; these large ratios primarily reflect low rates of treatment-related withdrawal in relatively small datasets and should not be interpreted as evidence of superior clinical effectiveness. The LHH also makes visible the uncertainty inherent in harm estimates. Wide confidence intervals for NNH naturally translate into wide LHH intervals, signaling that the true benefit–harm balance may vary substantially. This is particularly evident for interventions with very large NNH values and upper confidence limits extending to infinity, which reflect imprecision rather than overwhelmingly favorable profiles. In this way, the LHH can support more person-centered and evidence-guided treatment choices. However, its interpretation depends on the outcomes being compared, the nature of adverse events, and the certainty of the underlying evidence, which is often low to moderate. The LHH should therefore be viewed as a complementary interpretative tool rather than a metric capable of independently guiding treatment selection. Incorporating the LHH into future analyses, guidelines, and clinical discussions could strengthen individualized care by providing a simple, clinically meaningful summary of benefit–harm trade-offs. Pilot studies evaluating its use in real-world shared decision-making, as well as validation using individual-participant data, would further clarify its practical utility. Routine reporting of NNT, NNH, and LHH in future trials and systematic reviews would enhance transparency and support informed decision-making in a field where therapeutic choices often hinge on subtle and highly individualized considerations. Francesco Brigo is the only author of this article. He designed and conceptualized the study, collected and analyzed data, and drafted the manuscript. The author has agreed to the final submitted version. The author has nothing to report. The author has nothing to report. The author has nothing to report. The author has nothing to report. The author declares no conflicts of interest. Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.