More claims for genomics and rising health care costs

Mar 14, 2012


medicine, research


There is a nice summary of the relation between technology and health care costs in JAMA by Katrina Armstrong

There are 2 perspectives on the effect of genomics on health care costs. One view is that genomics is a form of new technology and it is well established that new technologies increase health care costs.3? For example, identification of the ERBB2HER2/neu (formerly) gene and the development of trastuzumab has increased the cost of breast cancer care, in part because of the additional cost of ERBB2expression testing (approximately $150), but mostly because of the cost of trastuzumab (approximately $45 000) for the 20% to 30% of women who testERBB2-positive. The total cost of this genomic advance, with substantial clinical benefit, has been estimated at more than $750 million per year.4 Other examples include the implantation of defibrillators in patients who are found to carry a mutation linked to a hereditary cardiomyopathy or the use of breast magnetic resonance imaging screening for women with a BRCA mutation. From this perspective, advances in genomics will extend life but exacerbate the current growth in health care expenditures.

The author then goes on to say why genomics is different. The gist of the argument being that genomics can be used to determine who to treat, and reduce unnecessary treatments. The examples quotes are for gene expression profiling in breast cancer, and K-ras testing in colorectal cancer. I am underwhelmed. Of course, it is nice for people to openly write about how many technologies increase health care costs. From much of the press relating to medical research, one might be forgiven for thinking that ‘research’ will always reduce expenditure. However, after observing the hype around the various fashionable approaches to medical science over the last quarter century, I am now sceptical of virtually all claims from the translational biomedical research bandwagons. Many of the new treatments are far too marginal in effect and whenever clinical  choices open up practictioners and guideline committees will tend to favour small effects even though the secureness of these effects has been based on data in which the mass of clinical variance is unexplained. The numbers therefore are usually messy, and the system will always promote marginal effects that may or may not be real. This comes with enormous cost. The author of course is arguing that genomic information will have the opposite effect of what I think will happen.

A large chunk of health care funding is inevitable and technology will have limited effects on controlling it. Care by humans is expensive and doesn’t scale well in terms of factory models of cost-efficiency. Many people think the funding of technologies has removed appropriate funding from ‘care’. I think they are right. One thing of course that might save some modest expenditure is getting doctors to make the right diagnosis in a bigger proportion of cases. You can’t improve health care without putting doctors under the spotlight.

Post by Jonathan Rees

Clinical academic and skin watcher at the University of Edinburgh

Comments are closed.