Number needed to treat

The number needed to treat ( en. number needed to treat NNT ) is a statistical measure that indicates how many patients per unit of time (eg 1 year) must be treated with the test substance or test method to achieve the desired goal of therapy in a to achieve patient or an event to prevent (eg, heart attack). It is in addition to the absolute and relative risk reduction a means to illustrate the benefit of treatment in and corresponds to the reciprocal of the absolute risk reduction ( ARR ). A larger NNT consequently means a smaller risk reduction.

NNT is also used in preventive medicine and sometimes as the number of necessary operations screen ( en. numbers needed to screen NNS ) refers. In practical application, the NNT should be set with the NNH (s, numbers needed to harm ) in relation, that is, the number of treatments needed to achieve the desired goal of therapy in a patient with the necessary number of treatment processes, to cause damage in a patient are compared.

Calculation

You can see the relationships very well with the help of a fourfold table illustrate. This provides a way of evaluating a scientific investigation dar. Ideally, it is a prospective, controlled, randomized clinical trial.

Where:

  • A is the number of patients who achieved with step 1, the objective
  • B is the number of patients in which the objective was not achieved with step 1
  • C for the number of patients who achieved with measure 2, the target,
  • D is the number of patients in which the objective was not achieved with Measure 2.

The ARR ( absolute risk reduction) is then calculated as follows:

A positive absolute risk reduction indicates a superiority of one measure against measure 2 Has the Absolute risk reduction a negative value, it means a inferiority of measure 1 for the validity of the ARR is still their confidence intervals important.

The NNT is calculated as follows:

Or

A negative ARR would be represented not as a negative number needed to treat, but as a positive number needed to harm ( NNH ).

Rounding: If you want to specify the NNT without decimal places, so it is recommended to round up. example:

  • 4.2304 patients must be treated to a patient dies less
  • (at least) 5 patients must be treated to a patient dies less

The usual rounding normally would result in that the therapeutic effect would be overestimated.

Valence

It is important that the values ​​of the NNT or ARR strictly initially apply only to the study, in which they were determined. The external validity of the study describes the transferability to the outside - that is, to a particular case or a general statement. Therefore, it is important that in addition to the NNT, the groups, the period and the destination point are described in detail. example:

See also: Relative risk reduction, absolute risk reduction

71455
de