Each of the three criteria above cannot be considered in isolation. You will need to make a judgement about the net benefit from the new drug in the context of existing treatments. For example, a small increase in efficacy compared with the standard treatment may not be acceptable if the new drug also has a worse adverse effect profile; a new drug might seem safer, but only because there is less experience with it or perhaps because it is being used at a lower (and so less effective dose) than alternatives (see Box 8.5 for examples); the new drug might have a different safety profile (e.g. because it is excreted by the kidney rather the liver, or is involved in different kinds of interactions), and so it might be considered useful in certain well-defined circumstances.
Box 8.5. Judging the net benefit from a new drug
Example 1. Apparently safer, but also less effective
In a short-term clinical trial, etodolac (a nonsteroidal anti-inflammatory drug first marketed in 1986 in the UK), was reported to cause less damage to the stomach than naproxen. But later the UK Drug Safety Research Unit (DSRU), using a Prescription-Event Monitoring system, rated etodolac effective in only 56% of 9109 patients. The DSRU concluded that the average daily dose of etodolac (400mg) was too low to be effective.17
Example 2. Trading in one type of adverse effect for another
Coxibs seem possibly less likely than conventional nonsteroidal anti-inflammatory drugs (NSAIDs) to cause dyspeptic-type symptoms and are associated with fewer endoscopically visible gastroduodenal ulcers or erosions. Epidemiological data also suggest a lower likelihood of upper gastrointestinal bleeding with celecoxib than with conventional NSAIDs. However, long-term outcome studies have not demonstrated a significant reduction in major ulcer complications (such as bleeding or perforation) with celecoxib compared with commonly used NSAIDs. An important unexpected finding of the VIGOR study was a significantly higher incidence of myocardial infarction with rofecoxib than with naproxen. Also, as experience with the recently withdrawn rofecoxib has demonstrated, there may be a 'trade-off' between better gastrointestinal tolerability with coxibs and a possible increase in the risk of serious cardiovascular events.
Source: Taking stock of coxibs. Drug and Therapeutics Bulletin 2005;43:1-6.