The Pharmaceutical Journal Vol 267 No 7170 p566-568
20 October 2001

This article
Reprint
Photocopy

UK Medicines Information summary

Why NICE is not nice for cancer therapy?

By evaluating medicines thoroughly in a “real world” way, pharmacists empower others to make important decisions about funding choices. Dr MAX Summerhayes, principal oncology pharmacist for Guy’s and St Thomas’ Hospitals presented a local ranking system developed in the south-east London area for prioritising funding decisions on oncology drugs. “In our system, benefit and evidence are ranked together” he said, both aspects being given a rating according to separate scales (see ranking Panel). Participants heard that although the results from the local ranking system are sometimes in agreement with the verdicts delivered in National Institute for Clinical Excellence appraisals (eg taxanes), they are often at odds.

Local ranking systems for efficacy and evidence of oncology drugs Local ranking of efficacy A Prolongs median survival by more than 9 months and improves quality of life B Prolongs median survival by 3–6 months and improves quality of life C Improves quality of life, but no survival benefit D Minimal quality of life impact and no survival impact Local ranking of evidence a+ Data from meta-analysis or two randomised controlled trials a One high quality randomised controlled trial and supporting non-randomised data b One poor quality randomised controlled trial and/or several phase II studies g Single phase II study only

Dr Summerhayes highlighted the cases of gemcitabine (Gemzar) and temozolamide (Temodal) which were quickly processed by NICE and given a positive appraisal. The local system had given these low ratings when compared with 30 other potential oncology developments. Another example was rituximab (Mab Thera). This was rated highly by the local system, he said, “but NICE has taken over nine months to deliberate on the drug and it is currently subject to an appeal which suggests that an unfavourable NICE verdict may be imminent”. “NICE appears to recommend, without problems, things that locally we rank low, and to have problems with things we rank high,” he said (see NICE Panel).

Dr Summerhayes then began to explore some of the possible reasons for this and drew attention to the fact that NICE appraises products in isolation. He explained that this means their relative value cannot be assessed, and there can be no mechanism for consistency. He gave an example of two separate oncology appraisals: one in which phase II clinical trial data had been used as part of the appraisal process, and another in which phase II data had been disallowed. “No recently licensed product is likely to be all bad,” he said, “and clinicians seldom say that they will never use a product.” This means that looking at individual oncology drugs in isolation is somewhat unrealistic.

Another problem Dr Summerhayes highlighted is that the appraisal process is predominantly conducted by those not working in oncology. He explained that the report teams may be familiar with large trials that have single end-points, but that these are unusual in oncology. He explained that, in oncology, many end-points are difficult to assess using quality of life measures, and various other features make them unusual compared with trials in other clinical areas. This could result in NICE report teams not understanding end-points, not comprehending the trial strategy, or failing to spot flaws in the trial design.

Dr Summerhayes questioned how NICE could assess the worth of a drug without set criteria against which to make a judgement. “What is the value of a saved or extended life” he asked. He was also concerned that there are no explicit guidelines on how to reject evidence. Finally he said that NICE is in real danger of becoming discredited by its decisions, since it does not seem to know how to turn a product down.

He concluded that the revised NICE process may be more satisfactory in terms of appraisal. The new “scoping” process will involve more far-reaching consultation and at an earlier stage so that real expertise will be enrolled, he said. However, this will be at the expense of a greatly slowed turnover time for each appraisal (at least a year), and this will probably be too long for all concerned.

Back to Top