Data Sharing and Simulated Clinical Trials

As the availability of electronic clinical data becomes more accessible and the use of artificial intelligence algorithms become more widely-used, the drug development process becomes the next big industry to harness the power of big data analytics. Pharmaceutical companies on average spend $2 billion over a period of 15+ years to develop, manufacture and bring to market a single drug or therapy.  Additionally, once the drug has been approved by the FDA and brought to market, drug companies receive a patent that expires after 20 years. Factors such as these, as well as evolving political landscapes [that I clearly don’t have time to elaborate on in this post] contribute to the overwhelming prices of drugs.

With critically important therapies like Daraprim and the Epipen rising to unbelievable prices, the topic pharmaceutical drug prices is becoming an increasingly controversial issue that constantly makes the headlines. Apart from political action, how can the issue be addressed? The answer lies in the production of the drug. As stated early, these massive price hikes are mostly due to the fact that drugs cost billions to produce and take almost two decades to bring to market...In other words, each day of production costs $300,000+ in production costs. By reducing the time period to bring these drugs to market a mere couple of weeks, massive reductions in drug costs can be attained.

A solution can be found in the use of data sharing and the application of simulated clinical trials. Within that 15 year timeline for drug development, 10-13 years consist of preclinical and clinical trials...or in other words, conducting experimental treatments on animal and human subjects. In fact, many drugs don’t even progress beyond that stage, costing drug companies an even greater number of costs. Fortunately due to the recent availability of large quantities of electronic clinical data, the potential of creating simulated clinical trials is becoming a reality. This means that by aggregating large datasets that culminate variables from numerous trials over decades of research, artificial intelligence based algorithms can be used to predict the outcome of specific trials without them needing to be conducted in a physical laboratory. These trial simulations have the potential to reduce the drug development timeline by numerous years.

However, the hesitation to share data among different pharma companies hinders this type of progress. Since these past trials, whether successful or not, cost drug companies millions of dollars to conduct, they’re usually very apprehensive to share their data due to their proprietary nature. Additionally, drug companies and researchers are hesitant to share data because if a third-party analysis finds flaws within the original data/analysis, their studies and, professional reputations in general, can be invalidated. Fortunately for the consumer, nonprofit organizations, who have regulatory and federal support, are bridging the gap by uniting companies in non-competitive arenas to share data for the greater good. As this sort of drug manufacture approach become more relevant, drug timelines and costs associated with them will begin to experience drastic declines.

Questions to ponder...

1. Although there will be significant savings in the manufacture process, will this incentivize big pharma to lower prices, or will they just pocket additional revenue?
2. How do you think companies will overcome the fear of error exposure?
3. Do you think this type of approach could become more integrated in a political agenda, or will governmental lobbying just hinder progress as it does in more industries?