Guidelines to understand and compute the number needed to treat

Abstract

ObjectiveWe aim to explain the unadjusted, adjusted and marginal number needed to treat (NNT) and provide software for clinicians to compute them.MethodsThe NNT is an efficacy index that is commonly used in randomised clinical trials. The NNT is the average number of patients needed to treat to obtain one successful outcome (ie, response) due to treatment. We developed the nntcalc R package for desktop use and extended it to a user-friendly web application. We provided users with a user-friendly step-by-step guide. The application calculates the NNT for various models with and without explanatory variables. The implemented models for the adjusted NNT are linear regression and analysis of variance (ANOVA), logistic regression, Kaplan-Meier and Cox regression. If no explanatory variables are available, one can compute the unadjusted Laupacis et al’s NNT, Kraemer and Kupfer’s NNT and the Furukawa and Leucht’s NNT. All NNT estimators are computed with their associated appropriate 95% confidence intervals. All calculations are in R and are replicable.ResultsThe application provides the user with an easy-to-use web application to compute the NNT in different settings and models. We illustrate the use of the application from examples in schizophrenia research based on the Positive and Negative Syndrome Scale. The application is available from https://nntcalc.iem.technion.ac.il. The output is given in a journal compatible text format, which users can copy and paste or download in a comma-separated values format.ConclusionThis application will help researchers and clinicians assess the efficacy of treatment and consequently improve the quality and accuracy of decisions.

PAX-D: study protocol for a randomised placebo-controlled trial evaluating the efficacy and mechanism of pramipexole as add-on treatment for people with treatment resistant depression

“…Furthermore, the trial will examine the acceptance, tolerability and cost-effectiveness of pramipexole treatment, and results from this study will inform clinical practice. 30 …”

“…The design of the trial also allows investigation of the mechanistic effects of pramipexole on reward sensitivity and whether changes in reward sensitivity can predict response to treatment. Furthermore, the trial will examine the acceptance, tolerability and cost-effectiveness of pramipexole treatment, and results from this study will inform clinical practice 30…”

“…on November 28, 2021 6 examining pramipexole monotherapy in nontreatment resistant unipolar depression found that 1 mg was superior to placebo, while 0.375 mg was ineffective and 5 mg caused very high drop out. A very small study by Franco-Chaves et al 7 (n=13 per group) of unipolar patients resistant to a single previous antidepressant randomised to receive pramipexole monotherapy (target dose=2.25 mg), pramipexole combined with escitalopram or escitalopram monotherapy, found no statistical difference between the study groups, although only four patients in the combined group completed the study.…”

Capturing adolescents in need of psychiatric care with psychopathological symptoms: A population-based cohort study

“…Fifth, psychopathological symptoms were categorized as severe if they were in the top 10th percentile of symptoms (otherwise not severe), leaving our study vulnerable to the limitations of dichotomization (e.g., loss of information [ 43 ]). However, in clinical and observational studies, cut-offs are widely used to ascertain severity [ 44 ], and to compute widely understood clinically values like the Number Needed to Treat [ 45 ]. Also, the current study analysis showed consistent results across cut-off thresholds, indicating that these are quite robust thresholds worthy of future research.…”

The number needed to treat adjusted for explanatory variables in regression and survival analysis: Theory and application

“…Notably, we compare point estimators, the lengths, and the coverage rates of the four CIs methods. Finally, we refer readers to the nntcalc R package 23 and the corresponding web application 24 to calculate the conditional and the harmonic NNTs with their corresponding CIs. The R package and the web application are made available for users online * .…”