GMC's proposals for revalidation. Purpose of revalidation process must be agreed on.

Editor—Southgate and Pringle outlined the General Medical Council's proposals for revalidation, the reasoning behind them, and how they might be implemented in general practice and, by implication, other specialties.1 The proposals adhere to international guidelines,2 but Wakeford has challenged them as being potentially unfair, unacceptable, and expensive.3 The portfolio based assessment described by Southgate and Pringle and advocated by the government and royal colleges has face validity. If blueprinted to provide evidence about all aspects of good medical practice, it should also be valid in content. However, Wakeford questions the fairness and reliability of the judgments to be made by the team of assessors. This is a credible objection because, while portfolio based assessment is a valuable technique for formative assessment, evidence of its reliability as a summative assessment tool is not yet convincing. Advocating this approach when the stakes are potentially so high seems to create a problem with the advent of evidence based medicine and best evidence medical education. Wakeford's solution also has flaws. Although an exercise administered on paper or computer as the basis for revalidation might answer some concerns about reliability, it would raise serious questions about validity. Such testing could not provide a measure of performance in practice and at best could only provide evidence about some aspects of competence. It would not provide information on many parts of clinical practice that worry the public and discredit the profession. The method for revalidation should be both practical and technically defensible. To achieve this, the purpose of the revalidation process must therefore be agreed on. Currently, the two purposes are being confused. The first is to identify seriously underperforming doctors. The second is to support all doctors in striving constantly to improve their performance. The methods of achieving these purposes need to be different. The prevalence of seriously underperforming doctors is believed to be low (<5%), so identifying them requires an inexpensive screening procedure with high sensitivity but low specificity. Doctors so identified would undergo a second stage set of procedures with high specificity. An option for this screening might be the use of peer and patient ratings as developed by the American Board of Internal Medicine.4 The second stage could be the current performance procedures already used by the GMC, which include an assessment of performance in practice and the combination of a written and practical test of clinical competence.5 The second purpose needs more discussion to establish the most fitting combination of methods, taking into account the best evidence available. There also needs to be a response to the concerns expressed by Wakeford about the feasibility and cost of planned approaches.