<International Circulation>: Warfarin as a classical anticoagulant drug is widely used in clinical practice. If the dose of warfarin is excessive， hemorrhage may occur and if the dose of warfarin is insufficient， embolism may occur. So clinicians need to be careful and prudent in using warfarin. Predicting warfarin dosage using pharmacogenetic algorithms is more scientific and reasonable however this technique is still strange to many clinicians. Can you outline the measures involved in applying this technique?

    Prof. Lee: This is very true. You don’t even have to mention algorithms; just the fact that you need to genotype the patients was a novel concept five years ago in Taiwan. When I started my first clinical study five years ago， I met a physician， who said， “What’s the point?” I think the first step required to move this into general clinical practice is to have some form of regional study in each country because the key will be to educate physicians first and they will need to see a real benefit from those studies. Once physicians can see the benefit of incorporating genetic information， and eventually dosing algorithms in predicting warfarin dosage， then those physicians will head out to meetings like this and spread the message. Physicians are more likely to accept the word from other physicians than from scientists like me， because they have hands-on experience. That is the first step. Another critical issue is the availability of genotyping facilities. We can have the best algorithms predicting 70% accuracy but if you cannot do genotyping quickly， then they are relatively limited in their effectiveness. So currently， patients and physicians still do not have access to the benefits of pharmacogenomic algorithms. My first strategic goal is， that by conducting clinical trials in major medical centers， physicians will start to understand pharmacogenomics; and secondly， by conducting studies in participating Asian medical centers， I am setting up infrastructures. This is important， because in order for this to work， we not only need the involvement of physicians， we also need diagnostic labs to perform genotyping and to do it fast enough. There are various technologies， so we need to choose in terms of speed and cost and find a balance， and we have to pass on the genotyping results to physicians so that they can understand and easily utilize the information. Therefore， the last link is that we need some form of bioinformatics. If you look at algorithms， no physician is going to take out a calculator and perform the maths. The best way is that once the data is generated from a molecular diagnostic lab， all that is required， is to enter the genotype information， and virtually all pharmacogenetic algorithms include clinical factors such as height， age， weight and so on. Simply by entering this information into a web interface， the physician can be advised directly in terms of dosage rate. This whole introductory process though takes time， and we need education at every single step. The first stage is convincing physicians and I think we have convinced many more than before and there is interest now， which is why I have the opportunity here at this meeting. The next obstacle is spreading the word amongst the general public. Even in the United States， many patients do not have access to genotyping not even the major hospitals perform such procedures routinely. Once we overcome this， physicians will feel more comfortable and eventually the price of genotyping will drop through mass demand and hopefully the insurance companies will pick it up. It doesn’t matter if it is public insurance as in China or Taiwan， or private cover as in the US， the cost of a procedure or product has to come down to a certain point before it will be covered by an insurer. I believe we are getting there. Earlier this year， around April/May， an American company called Medco which handles drug prescriptions for many insurers conducted a study themselves which indicated that using genetic information would reduce hospitalization by at least 30%. To an insurance company， that is a significant saving. And if a private entity is interested at that level， then it should tell us something.