The Numbers Needed to Treat for Neurological Disorders

• Published in: Canadian journal of neurological sciences 2005-11, Vol.32 (4), p.440-449
• Main Authors: Bussière, Miguel, Wiebe, Samuel
• Format: Article
• Language: English
• Subjects: Biological and medical sciences; Canada; Confidence Intervals; Degenerative and inherited degenerative diseases of the nervous system. Leukodystrophies. Prion diseases; Evidence-Based Medicine; Health Services Needs and Demand; Humans; Medical sciences; Meta-Analysis as Topic; Nervous system (semeiology, syndromes); Nervous system as a whole; Nervous System Diseases - prevention & control; Nervous System Diseases - therapy; Neurology; Original Article; Outcome Assessment (Health Care); Randomized Controlled Trials as Topic; Treatment Outcome
• ISSN: 0317-1671; 2057-0155
• DOI: 10.1017/S031716710000442X

Numerous therapeutic interventions have been developed in the neurosciences. Clinicians need summary measures about efficacy of therapies that derive from the best available evidence, and that can be readily extrapolated to clinical practice. The number needed to treat (NNT) is intuitive and clinically applicable. We provide clinicians with a single source that summarizes important therapies in the main neurological and neurosurgical areas. Critically appraised evidence about therapies in the neurosciences was obtained from meta-analyses in all neurosciences groups in the Cochrane library, and from critically appraised topics at the University of Western Ontario. Therapies were included if they were deemed relevant and if outcomes were dichotomous. For each therapy, we obtained absolute risk differences and their 95% confidence intervals (CI), the corresponding NNTs, control and experimental event rates, and the time-frame of the outcome assessment. We assembled a table of NNTs for 87 interventions in ten disease categories, deriving from meta-analyses (70%) or randomized controlled trials (30%), and assessing surgical interventions (7%), procedures (9%) or pharmacological treatments (84%). The NNTs varied widely, ranging from 1 in the use of epidural blood patch for post-dural puncture headache to 4608 for meningococcal vaccination. Preventative interventions had substantially larger NNTs. Time-frames were inappropriately short for many chronic conditions. Large collections of NNTs provide useful, updateable summaries of therapeutic effects in the neurosciences, an increasingly interventional clinical field.