Before substituting an existing standard treatment with a new drug, prescribers need to be sure that the new treatment has advantages over what is already available. To be considered an advance when compared with existing treatment(s),2 a new treatment should have:

• better efficacy, and/or
• fewer or less severe adverse effects, and/or
• safer or more convenient administration.

Therapeutic advance should not be seen in isolation. When reviewing a new drug, cost is obviously an important consideration, although it has no bearing on whether a drug is a therapeutic advance. Section 8.6 of this chapter discusses how cost is considered in the evaluation.

Questions to ask about the new drug

At the start of a review of a new treatment, there are three main questions to consider:

1. Is treatment really necessary for the patient? That is, does the drug improve the appropriate outcome(s) in the targeted population, and is it relevant to the health of your country’s population? There is usually little or no available evidence from randomised controlled trials that have used appropriate outcome measures with which to answer the question (see Boxes 8.1 and 8.2 and Section 8.4 for a discussion of hierarchy of outcome measures and evidence). However, non-randomised studies, such as observational follow-up studies, with mortality as the outcome measure may be available and can be useful. For example, serum cholesterol concentrations of 6.21-6.71 mmol/L, which are considered high in western countries, occur in the healthiest people with the lowest mortality and living a normal life in Japan, and so it follows that in Japan such people should never be regarded as ‘patients’ with hyperlipidaemia needing treatment.³

2. If a treatment is necessary, is there an effective non-drug treatment?

3. If drug treatment is needed, what are the standard, or already available, drug treatments?

The next step will be to examine whether the new treatment is more effective, safer or more convenient to administer than what is already available.