Healthcare's Biggest Data and Analytics Challenges in 2018

^{Images have been resized. Courtesy of KamiPhuc, Flickr.}

Just as 2018 sows promise for the more widespread use of big data and analytics in healthcare, the year also brings challenges. What stands in the way of efficiency and effectiveness? And how might various segments of healthcare tackle these issues?

Healthcare Analytics News™ asked its editorial advisory board members those questions, hoping to yield insights that could guide coming coverage. Their answers, however, hold value not just for health-technology journalism but also key players in this space. Here are their thoughts, which have been lightly edited.

John Giannouris, vice president of specialty pharma services at ValueCentric

The biggest challenge I see in the pharma space as it relates to specialty products is connecting their limited distribution network specialty pharmacy data with other data, such as information from retail and medical claims, electronic medical records, labs, biometrics, and consumers. The challenge is linking the patient across these data in a longitudinal manner, ensuring HIPAA [Healthcare Insurance Portability and Accountability Act] compliance and data interoperability.

The way to overcome this challenge is having a singular HIPAA-compliant patient de-identification engine being used across all the data.

Jean Drouin, MD, CEO and founder of Clarify Health Solutions

The problems: too much poorly processed data dumped onto users; not enough actionable insight; difficulty getting the insights into the workflow at the point of decision/point of care. And paradoxically, there are not enough data to drive artificial intelligence and machine learning in the same ways as in other industries, like retail and banking. Regulation is also getting in the way of linking data in the most useful ways.

Brenton Fargnoli, MD, medical director of value-based care and director of product marketing and strategy at Flatiron Health

In oncology, the biggest challenge is unstructured data. Next-generation sequencing results, clinical visit notes, pathology reports, and more all make up the patient story and all are primarily unstructured and disconnected. This creates a problem with the quality and usability of this information to drive new clinical insights.

Perhaps this is not surprising, as the electronic health record is a technology that was never intended for evidence generation.

This problem can be overcome by curating unstructured data. Before any unstructured data can be used in a clinical research setting, it should go through a rigorous process to assign logic and parameters to all relevant data points. Once these data are organized, only then can they be leveraged efficiently by the oncology community.

Frank Baitman, senior adviser for healthcare at DeSilva & Phillips

There are multiple challenges, but at the root of many is HIPAA. This well-intentioned law preceded the rise of the internet, big data, and precision medicine. The protection of personal health information must remain paramount, but HIPAA uses a sledge hammer to unceremoniously solve this problem. Its impact is broad-based: Individual health data are siloed (and often lost); caregivers are challenged to provide holistic health since specialties are occasionally walled off; pharma is challenged to track post-clinical trial usage; and public health departments have limited visibility into the communities they are charged with protecting.

John Quackenbush, PhD, professor of biostatistics and computational biology, professor of cancer biology, director of the Center for Cancer Computational Biology at Dana-Farber Cancer Institute, and professor of computational biology and bioinformatics at the Harvard T.H. Chan School of Public Health

The biggest problems are incomplete data, incomplete meta-data, and the lack of appropriate analytical tools that integrate our knowledge of the biology of the systems we are studying.

Elena Alikhachkina, PhD, worldwide head of analytics technology, strategic insights, and data solutions for Johnson & Johnson

On the commercial side of the house (in marketing and digital), a lot of data are with external vendors or agencies. There are barriers to access the data, poor data quality, high costs for data integrations, and more. This problem can be overcome by enabling a single source of data truth with the company, driving the data governance and data quality processes.

Also, commercial, research and development, finance, and others are still working in silos. Enabling a single source of data will help here as well.

Christopher Sellin, head of operation excellence and analytics, US commercial, at Shire

I agree with the thoughts on challenges with warehousing, processing, and blending data. That said, I wanted to bring up the challenge of developing consultative skills in your business analyst/data scientist resources. How do we maximize our ability to have people get to the right business question? Also, post analysis, how do we built our team’s ability to synthesize and communicate the narrative in a meaningful way that optimizes pull through?