By Ed Biller, Editor
You have “big data,” but are you really taking advantage of its potential?
By Ramon Chen, VP Marketing, Reltio
It’s easy to see how the flow of data has transformed the retail business, revolutionized the travel industry, and upended the financial services sector. Operations in these and other vital areas of the economy are nothing like they were a decade ago.
Life sciences and, more specifically, the medical device industry, has seen its share of changes, too. But have those changes gone far enough?
A comprehensive survey conducted this summer by Reltio — the first research effort dedicated to the priorities and the use of data across medical device, pharma, and biotech in life sciences — paints a murky picture at best. At the very least, survey results suggest there’s still a long way to go in truly embracing big data to improve the design, manufacture, and sales of life science products.
So what can be done about it?
The challenges are well-documented: Reduced government, health insurer, and public hospital budgets, lower reimbursements for private facilities, industry consolidation through M&A, and global competition from companies delivering products that are “good enough” and competitively priced. Finally, despite a massive influx of data — and the promise of much more to come, thanks to wearable technologies and the Internet of Things — many processes remain mired in the past. Why is there such a disconnect between the data and the applications that could benefit from its use?
Here’s a high-level view: More than half of all life sciences organizations (55 percent) now describe themselves as “very data-driven,” which seems to indicate progress. However, while almost 70 percent of biotech and pharma companies say they're now using data very effectively, only 30 percent of medical device companies say the same.
“Data-driven” operations use current and historical information about customers, organizations, products, places, and even internal teams to generate relevant insights, and to recommend actions that positively impact business.
Most healthcare data — greater than 70 percent overall — now comes from external sources, including third-party providers, public and social media feeds, etc. When this data is combined with internal data, the sheer volume of information necessitates an infrastructure that can guarantee better data reliability before correlating relevant insights across datasets. But Reltio’s survey indicates this isn’t happening often or well.
A majority of the respondents, 74 percent, worry that their data is incomplete or missing, while 50 percent say the insights are not actionable. More than a quarter, 28 percent, say they still have siloed data (incompatible or not integrated with other data systems), even with millions of dollars invested in solutions that were designed to solve the problem. Finally, only percent of respondents have put any mission-critical data into the technical terms du jour — big data lakes or Hadoop — a clear sign that, in spite of the perceived momentum, there’s a major gap between what’s possible and what’s real.
Go Beyond Intended Purpose
In life sciences, there are severe fines and penalties for marketing products beyond their intended purpose. There is no such restriction placed on use of data within an enterprise, yet all traditional applications focus on collecting and managing data in silos, restricting their use for one primary focus.
But utilizing data beyond its initial intended purpose requires complex machinations from a company. It must create new data warehouses, buy new data integration and analytics tools, and procure expensive hardware and infrastructure. This exercise applies to data for clinical trials, customer relationship management (CRM) for healthcare provider (HCP) and healthcare organization (HCO) management, enterprise resource planning (ERP) product data for manufacturing, and HR or expense data for open payments compliance.
Even though technology has advanced and costs have come down dramatically, adoption of expanded data usage has been slow. Why? A little context helps explain the scope of the problem.
The healthcare industry has a long history with the discipline of Master Data Management (MDM). For the last ten years, this is how poor data quality was rectified, and information correlated. This method requires bringing together data across siloed applications such as CRM, ERP, HR, and financials to get a consolidated, cleansed view of name, address, and other attributes.
Companies with adequate resources set up the infrastructure, resources, and skills, and then created identifiers back to source applications so that they could mine the associated transactions — including scripts written, speaker payments, clinical trial payments, and so forth — with a 360-degree view of all physician activities being the goal. Armed with this information, companies had an index that could generate more accurate reports for compliance, increase sales and marketing efficiency through more accurate contact data, and foster better understanding of the customer.
But today’s 360-degree view has to be more comprehensive. HCPs are increasingly affiliated with integrated delivery networks (IDNs), have multiple places of business, and do more than just write and recommend prescriptions. Some are key opinion/thought leaders (KOLs), sit on committees with significant influence, and keep an eagle eye on medical device vigilance.
Every company understands the need for reliable HCP/HCO data, but affiliations, hierarchies, and relationships between people, products, and organizations must be captured, modeled on the real world, and constantly updated to be truly relevant.
Moving to the cloud is a good start, as it helps increase organizational agility, reduces cost, and democratizes access to clean, reliable data for companies of all sizes. But that effort falls short without the ability to manage big data — transactions and interactions, social and machine-generated — which threatens to overwhelm the senses and capacity of life sciences companies.
There is no shortage of tools intended to address each distinct data management issue and, indeed, business teams spend billions on standalone business intelligence (BI) and analytics tools, while IT spends billions on big data lake and Hadoop infrastructures. To be truly data-driven, though, enterprises need a new breed of data-driven applications.
Take A Cue From LinkedIn
On LinkedIn, you get up-to-date statistics on first-level connections and the vast potential network that lies just beyond. If you want to engage with someone you don’t know, LinkedIn guides you to use a shared connection. The site presents jobs relevant to your current profession, weeding out the best options from the millions listed; It realizes the goals you have, and it’s there to help.
Recruiters on LinkedIn can easily filter skills and profiles before reaching out to candidates. Marketers and sales professionals can manage and track leads, getting help from peers along the way. LinkedIn has built-in data quality: It’s both self-governed and crowd-enhanced through recommendations and endorsements. All of this is built on top of a single pool of data that continues to grow with perfect scale and performance. That’s a perfect data-driven application right there.
By contrast, process-driven applications such as CRM act as recording devices, with no recommendations or suggestions on what to do next. If you’re in marketing, you can’t really leverage the same data that sales uses in CRM without integration or transfer to a separate data warehouse.
An enterprise data-driven system won’t be an exact replica of LinkedIn’s system, of course. Enterprise data-driven applications are more difficult to develop because they need to incorporate third-party sources to augment and enrich the data. This has traditionally been a cumbersome, back-office process. However, current data-as-a-service capabilities offer direct connectivity between applications and third-party industry data sources. They’re pre-integrated and pre-aligned for one-click onboarding, enabling R&D and commercial business users to easily shop for and blend data together.
Thus, enterprise data-driven applications can now start with a modern data management platform that embodies the best of MDM, big data, data-as-a-service, analytics, and machine learning. They are being deployed strategically to treat business challenges in each part of an organization. Data used to solve one problem is now free to flow across the enterprise and be used by other apps. Teams access data to meet their goals, with recommended actions that are enhanced through a closed-loop that learns from its outcomes.
Examples: Real World Data-Driven Evidence
A global diagnostic device and service provider was experiencing gross margin pressure and multiple product recalls. Its customer information was siloed across direct and distributor channels, a problem further magnified by inconsistent views of the company’s own products and hierarchies. Through a modern data management platform, the company eliminated information silos and uncovered relationships across the supply chain, becoming more aware of and eliminating its bottlenecks.
A large distributor engaging in M&A used the same technology to bring together a single integrated view, accessible by select approved teams pre-merger. Analytics could now be executed to respond to government inquiries for divestitures, as well as to identify growth strategies through better understanding of customer sales and purchasing influence, together with associated relationships between distributors and products. This upgraded technological capability also allowed the combined entity to get a jumpstart on post-merger integration planning.
Take Corrective Action
With billions being spent on new technologies, companies can be more data-driven, but leadership must think holistically about what the entire enterprise will need. Corrective action must be taken, and that means saying “no” to siloed tools and services when trying to solve a siloed problem. Treating data — all data — as an asset to be seamlessly shared across the company will ultimately benefit every group, delivering better bottom-line results and supporting a more agile business.
The most important question you can ask your technology vendor or partner is: “I get that you can solve my problem. How can your technology share data and address other problems across my company, some as yet unforeseen?” This is the medical device industry’s big (and better) data opportunity.
About the Author
Ramon Chen is a data management expert currently responsible for worldwide marketing at Reltio. Prior to Reltio, he held positions at Veeva Systems, RainStor, Siperian, GoldenGate Software, MetaTV, Evolve Software, Sterling Software, and Synon, Inc.
He holds a B.S. in computer science from Essex University.