Quantitative Framework

Net Clinical Benefit


1. Description

NCB is a quantitative framework that compares the overall change in the benefits and risks of a drug over a comparator. NCB is the sum of the change in expected benefits minus the change in expected risks as a result of treatment. The benefits and risks must be placed on a common scale, such as using the health-state related utilities. The expected benefit is calculated by multiplying the benefit, assuming it is realised by the patient, by the probability of its being realised, with a similar calculation for expected risks.

2. Evaluation


2.1 Principle
The NCB framework has three steps:
  • define decision problem and data sources
  • establish the functional form of NCB equation
  • estimate the NCB
  • The final step is divided into smaller pieces for added transparency when estimating the net clinical benefit of a treatment compared to an alternative treatment

2.2 Features
  • Sensitivity analysis can be performed on individual parameters in the NCB equation but may be impractical.
  • NCB framework was proposed as a Bayesian model, and therefore inherits its features.
  • An extension of NCB to estimate MERT (minimum event rate for which treatment is favourable) is also available.

2.3 Visualisation
The visualisations that are associated with NCB are:
A variation of this method was tested in the Natalizumab case study.

2.4 Assessability and accessibility
  • NCB as a number of resultant metrics, namely NCB, NNT, and MERT, which can increase the interpretability of the results when communicating with different stakeholders.
  • The exact functional form of NCB and the underlying assumptions determine whether the results are acceptable.
  • The Bayesian version of NCB and the model flexibility may help address many questions of uncertainties surrounding the evidence synthesis and benefit-risk assessment.

3. References


[1] Sutton AJ, Cooper NJ, Abrams KR, Lambert PC, Jones DR. A Bayesian approach to evaluating net clinical benefit allowed for parameter uncertainty. J Clin Epidemiol 2005 Jan;58(1):26-40.