The healthcare industry has a finite amount of money that must be divided among drug discovery, treatment, and basic research. The industry needs clear financial incentives to invest in the IT and informatics infrastructure necessary to achieve information-based medicine. Three such incentives are:

New blockbuster drugs. One concern with personalized medicine is the cost of developing drugs that treat small populations of individuals. But what if the outcome were a reclassification of diseases because certain drugs were discovered to be effective for treating multiple diseases previously thought unrelated? An early example is cyclo-oxygenase inhibitors, designed to treat chronic inflammation, which have demonstrated effectiveness against some lung cancers because they act as inhibitors of angiogenesis. As diseases are reclassified according to their molecular-level characteristics and gene-expression patterns, a new class of phenotype-independent blockbuster drug is likely to emerge.

Lower risk. A study by the Institute of Medicine estimates that medical errors cost the United States $37.6 billion annually by injuring 1 out of every 25 patients. Fatal mistakes by 700,000 U.S. doctors killed 119,000 patients in 2000. Virtually all these mishaps were unavoidable because the doctor lacked data that would have prevented the mistake. Likewise, information-based medicine would remove much of the trial-and-error approach to treating disease. Most cancer treatment protocols, for example, still rely on a hit-or-miss regimen that results in escalating costs downstream associated with treatment.

Fewer trial failures and "drug rescue." Availability of a large statistical basis for drug discovery will reduce the risk of a new lead compound failing trial — the principal reason drugs exit the pipeline. Indeed, many failed drugs sitting on shelves could be revived once they can be tested in patient populations precisely substratified at the gene-expression level. Neurological and psychiatric diseases, for example, are notoriously difficult to target because they are classified solely on the basis of phenotype. And dozens of promising epilepsy treatments have been sidelined because of low efficacy. Precise substratification of patients would allow better targeting of these compounds, with the possible discovery of high efficacy in certain patient populations.

Despite these financial incentives, the deployment of information-based medicine is hostage to myriad technical challenges and the cooperation of numerous organizations. Early projects have been led by large research medical centers, and it's likely this trend will continue. As more payers, providers, and IT companies offer their support, these projects will bear fruit in the next few years.

Back to Making Information-Based Medicine Work