Understanding the risks and benefits of available treatments represents an essential element of clinical practice. Previous work has demonstrated that knowledge of net benefits and net risks can relate to our decisions on whether or not to administer a particular treatment or order a diagnostic test. A wider application of this model has been difficult because data on net benefits and net risks are not directly reported. We used more frequently reported data on treatment efficacy (E) and risks (Rrx) to obtain an equation for the treatment threshold probability above which treatment should be given and below which it should be withheld. The diagnostic test should only be performed if the probability of a disease is between the testing threshold and the treatment threshold. We first described a theoretical background for these calculations. We then used a JavaScript programming language to write a computer program which physicians can use to calculate these threshold probabilities effortlessly through the Internet. In most clinical situations we do not have to achieve maximum diagnostic certainty in order to act. However, we should never treat or order a diagnostic test if the risk of the treatment is greater than its efficacy. The minimally required E/R ratio of a particular treatment is equal to the reciprocal value of the mortality/morbidity of untreated disease. Similarly, the lowest number of patients needed to be treated (NNT) for therapy to be worth administering is equal to the reciprocal of the treatment risk. We show how evidence-based summary measures of therapeutic effects, such as the treatment efficacy, harms, and NNT, can successfully be integrated within a decision analytic model. This in turn will facilitate wider use of the quantitative benefit-risk analysis. Accessing the Internet for direct and immediate approach to the formulas described here should make this task even easier in everyday clinical decision making.