Breaking Big Data Barriers in Pharma Research

Leveraging big data for medical research and advancement has proven to be one of the more challenging use cases owing to the fact that medical data is surrounded by privacy and security concerns that make it difficult to pile it together. Sam Marwaha, Director of McKinsey & Company, addressed the big data issue as it pertains to pharmaceutical research and development in an essay and accompanying video.

According to the McKinsey Global Institute, big data implementation to optimize and inform decision making could have a $100 billion impact on the healthcare industry in the United States alone. Such value would result “by optimizing innovation, improving the efficiency of research and clinical trials, and building new tools for physicians, consumers, insurers, and regulators to meet the promise of more individualized approaches.”

Individualizing medicine has been a key theme when big data analytics and healthcare are mentioned in the same paragraphs. The problem generally lies in data collection and, according to Marwaha, too many institutions are rushing into deals that end up being sub-optimal.

 “We’ve seen a lot of people that run off and start to build partnerships for getting data or trying to do a bunch of analytics,” Marwaha said, “which can be a wasted effort if you’re not clear how and where it’s going to be used.”

Rushing into a big data analytics system for the sake of running said analytics is not a problem unique to healthcare. However, research, especially that of the pharmaceutical sort, can be significantly set back if not enough time and care is taken to identify the proper procedure. As a result, Marwaha recommends in the video below that institutions understand the underlying quality of data they’re getting. “If you’re a provider that’s suddenly at risk for the quality of care, then let’s understand what are the drivers of that quality and let’s understand the data that we need and the analytics that goes into supporting that.”

Not unlike scientific research, pharmaceutical research and development is best when conducted over a clinical trial supported by a peer-reviewed research paper. Marwaha addressed the concern that data collection and analysis may be replacing that process, noting that the quality of data can be circumspect. “Now we’re suddenly reviewing data, we don’t know what quality it is coming off of clearinghouses and lab results, there’s lots of errors, there’s a lot of pushback on whether that is voodoos science. So this is not a substitution of clinical research that’s randomized, it’s a complement to it.”

Part of the problem is that there lacks a standardized methodology for addressing big data. Further, the data that is available is subject to barriers that discourage proper collection. “There’s a risk that data gets balkanized,” Marwaha said, “that’s in different places and people to start to erect barriers around it or you get a little bit of a monopoly around it and you’re not able to use it in a way that makes sense because of privacy concerns.”

The solution for Marwaha would be to implement a common ground for anonymously sharing data such that medical practitioners and researchers can act as scientists do when working on data-intensive questions. “The ability to get to a place where there’s a form of data commons where you can address the private concerns on the one hand and yet create the right value proposition for the users including the patients so that this is seen as a positive outcome.”

Big data analytics has significant potential in the healthcare industry, provided enough care is taken to address security concerns in the effort to share and standardize medical data.

Related Articles

What Can Enterprises Learn From Genome Sequencing?

Preventing Brain Injuries with Predictive Analytics

Big Science and the Gift that Keeps on Giving