The Evolution of Pharma Depends on Data Integration and Application

NICE INSIGHT OVERVIEW: The Cloud

The pharmaceutical industry must respond to increasing pressure to rapidly develop cost-effective, fairly priced medicines. New business models must be adopted to leverage the massive amounts of data generated across the pharma/healthcare continuum. Pharma companies must also pursue new collaborations to enable optimal digital workflows that provide maximum efficiency and productivity.

Pricing and Performance Pressures

Tradition can add real meaning to our personal lives, but, for many businesses, holding on to traditional practices can inhibit much-needed progress. This is certainly true in the pharmaceutical industry, which faces tremendous pressure on many fronts. Only those companies willing to replace historical traditions with new ways of operating — on the manufacturing floor and in the digital world — will evolve to meet the needs and expectations of all future stakeholders.

Those stakeholders — governments, insurers, physicians and patients — no longer accept whatever products drug makers choose to offer at whatever price they set. Armed with readily accessible data and support from advocacy groups and others through social media networks, patients are deciding which drugs are most valuable. They also expect to be included in the drug development process. Their desires are in line with those of regulators, who are beginning to include patient concerns and opinions in the drug approval process. New drugs are expected to exceed the efficacy of existing products and provide measurable, real-world value: addressing unmet medical needs, enhancing quality of life, extending lifetimes and improving patient adherence.

With this move to value- and outcome-based pricing, stakeholders are balking at rising prices. Insurers are demanding discounts and price protection in their formulary contracts. New drugs that carry hefty price tags are being protested by physicians, patients and insurers. Overall, public perception of the pharmaceutical industry has reached a low point. Drug development, however, continues to be a lengthy, complex and expensive process. In addition, the durations of patents and market exclusivity are shrinking, at a point when branded drugs face growing competition from generics and biosimilars. Clearly, the traditional business model is no longer sufficient.

So Much Data from So Many Sources

There is hope. Drug manufacturers today have access to vast quantities of data generated throughout the healthcare ecosystem, from the results obtained from fundamental research projects to data generated in preclinical to late-stage clinical trials and post-marketing studies, not to mention the patient information documented in electronic health records (EHRs).

Basic research generates genetic, proteomic, metabolomic, transcriptomic, spectroscopic, phenotyping and other data. Drug development efforts provide data on pharmacokinetics, stability, toxicity, efficacy and more. Post-marketing (post-approval) studies, which the FDA has routinely required since it was given the authority by the 2007 FDA Administrative Amendment Act (FDAAA), provide evidence of product safety and effectiveness in real-world settings (hospitalization rates, drug utilization patterns, health outcomes and overall treatment costs) for a range of patients, including patient populations that did not participate in clinical trials.¹

Additional data on patients and patient populations comes from a number of sources, such as vendors, collaborative research partners, insurers and patients that freely share their information. EHRs generated during routine doctor visits contain demographic and other personal information, insurance-related medical and pharmacy claim data and data generated during treatment or from monitoring devices. The scope and format of this data vary internationally because it is generated by many different sources, including primary physicians, emergency rooms, inpatient hospitals, insurance companies and even social media platforms.² In addition, it can change over time as technologies and policies evolve. 

Many Applications

The potential applications for this vast quantity and variety of data are nearly endless. When properly accessed and utilized, these data can be used to identify disease mechanisms and drug targets and build predictive models, accelerating drug discovery and identifying preventive approaches.³ It can enable the targeting of potential study participants⁴ and facilitate improved clinical trial designs,³ reducing their length and cost and providing more reliable results. It can be used to find new applications for existing compounds, both those that are approved and those that failed to make it out of the clinic (drug repurposing).⁴

The data in EHRs, for instance, can be used for much more than enhancing patient care.⁵ They can provide data for post-marketing, registry-based randomized studies and comparative effectiveness studies. Drug utilization, epidemiology and safety information can also be obtained from EHRs. Data can be applied to establish more effective hypotheses, parameters and endpoints for clinical trials and facilitate patient recruitment and potentially to populate electronic case report forms (eCRFs). EHR data can also be used to identify possible new indications for existing drugs.

Available digital data of all kinds can even be used for the development of more effective marketing strategies. Pharmaceutical companies can gain insight into patient needs and concerns and use that information to create targeted marketing messages and build long-term relationships.⁶ Digital technologies, such as apps for mobile devices and social media, can further enhance marketing activities. Post-marketing studies, in particular, provide information on payer and provider perceptions and thus allow drug makers to identify areas for product improvement with respect to safety, tolerability, adherence, outcomes, ideal patient populations and overall cost/savings.¹

Why Data Integration Is Important

Data will not become information on its own. The vast quantities of data in disparate formats available to pharmaceutical companies must be combined, analyzed and interpreted rapidly in a way that generates actionable information.⁷ Data must also be integrated in a way that maintains the highest standards of data governance, data quality and data security while addressing the challenges of compliance and globalization.³

If properly integrated, different data sets can be analyzed multiple times to support different goals. They can also be accessed from a central location by machine learning, artificial intelligence and natural language processing (NLP) programs, allowing for autonomous decision making and response.⁷

Integration of data generated during discovery and development with post-approval study results and patient records is equally essential for pharmaceutical companies from a business perspective.⁸ Accelerating the development of cost-effective drugs that provide measurable value is fundamental and can only be achieved by leveraging the myriad data sources to more rapidly identify drug candidates with the greatest likelihood to provide real-world performance.

With effective data enrichment, integration and management capabilities, pharmaceutical companies can generate value from historical and real-time information, including insights into disease mechanisms, ways to optimize clinical trials and achieve manufacturing efficiencies, new formulation approaches to prevent fraud and improve patient adherence and treatment outcomes and actions that will provide competitive advantages.³

Data Silos Still Exist

However, data integration is a challenge for most pharmaceutical companies. Protection of patient privacy (and the associated legal issues) can be a significant hurdle for some firms.⁶ A deficiency of staff with the training and expertise required to implement and manage comprehensive data integration and management platforms is another. In many cases, data are still collected manually.

Perhaps the biggest challenge is the fact that the disparate types of data are generated and maintained in different locations by different groups that have not historically communicated with one another.⁹ Structured and unstructured information on different diseases, drug candidates, formulations, dosages, preclinical results, clinical trial results and patients are stored in different databases.⁹ When drug manufacturers collaborate with academic researchers, insurers, physicians, advocacy groups and other organizations, the complexity of data integration — and security risks — increases dramatically.

Need for New Business Models

To realize the opportunities presented by the avalanche of data available today, pharmaceutical companies must move away from this traditional siloed approach to a new business model that enables data sharing within the enterprise and with other stakeholders and encourages consideration of real-world performance from discovery through commercialization.¹⁰

The application of digital technologies and analytics within existing modes of operation that were not designed to leverage them adds costs while limiting the potential benefits that can be gained. New business models must be built to enable data sharing and digital innovation and allow for the use of predictive analytics and the incorporation of stakeholder (particularly payer and patient) expectations for performance at the earliest stages of discovery. As importantly, it requires “breaking down functional silos and adopting agile ways of working.”¹¹

Initial Moves

Most pharmaceutical companies are taking steps in the right direction. According to the results of Deloitte’s 2017 real-world evidence (RWE) benchmark survey, the use of RWE in many different applications is increasing, with over 50% of respondents investing in RWE programs to improve R&D efforts.¹⁰ The survey also revealed, however, that there is a need to access the right real-world data, which may require the establishment of new types of partnerships. One encouraging result was the interest shown by respondents in investing in new technologies that impact workflows and increase efficiencies — perhaps representing initial movement toward new business models.

IBM’s Watson cognitive computing technology is one such technology being embraced by several pharmaceutical companies for a number of different applications. IBM has configured a version of Watson to specifically support life science research through incorporation of medical literature, patents, genomics and chemical and pharmacological data. Cognitive solutions are attractive because they are specifically designed to integrate and analyze big data sets and can be “trained to understand technical, industry-specific content and use advanced reasoning, predictive modeling, and machine learning techniques.”¹² Watson has been used to accelerate drug target identification and drug repurposing projects.

Collaboration is Essential

Collaborations like those with IBM and other software/technology firms are essential if pharmaceutical companies are to successfully move to new business models based on leveraging data and driven by digital innovation. Strategic relationships with insurers, patient advocacy groups and other “customers” are also needed to ensure access to the right real-world data. All partners can win from such collaborations.¹¹ For instance, insurers can identify patients at the highest risk for disease and urge them to seek care; those patients can then receive the targeted care they need, and the drug manufacturers can gain access to data that can accelerate the development of drugs that are safer and more effective. In addition, these collaborations can help all stakeholders better understand the value of drugs and their ability to provide desired outcomes.⁸

References

1. Numerof, Rita. “The Role of Post-Market Studies in Market Access.” eyeforpharma. 19 Jul. 2017. Web.
2. Banerjee, Amitava, David Mathew, Katherine Rouane. “Using patient data for patients’ benefit,” BMJ; 358:j4413 (2017).
3. “Big Data for the Pharmaceutical Industry.” Informatica. 2013. Web.
4. Keshava, Nirmal. “Opportunities for Data Science in the Pharmaceutical Industry.” IEEE Pulse. 17 May 2017. Web.
5. Cowie, Martin R., et al. “Electronic health records to facilitate clinical research.” Clin Res Cardiol. 106(1): 1–9. (2017).
6. Patel, Ritesh. “Data, Marketing, and the Pharmaceutical Industry.” Pharma VOICE, Sept. 2016. Web.
7. Cassese, Vita. “Delivering Value from Data Analytics.” Pharmaceutical Executive, 2 Mar. 2018.
8. “21st Century Pharmaceutical Collaboration: The Value Convergence.” Price Waterhouse Cooper. Jul. 2015. Web.
9. Palgon, Gary. “The Pharmaceutical R&D Process and the Inherent Data Challenges.” Liason. 7 Apr. 2017. Web.
10. “Getting real with real-world evidence (RWE): 2017 RWE benchmark survey.” Deloitte. n.d. Web.
11. Marwaha, Sam, Michael Ruhl, Paul Shorkey. “Doubling Pharma Value with Data Science.” Boston Consulting Group. 9 Feb. 2018. Web.
12. Chen, Ying, Elenee Argentinis, Griff Weber. “IBM Watson: How Cognitive Computing Can Be Applied to Big Data Challenges in Life Sciences Research.” Clinical Therapeutics. 38: 688-701 (2016).
13. Sennaar, Kumba. “AI in Pharma and Biomedicine – Analysis of the Top 5 Global Drug Companies.” Techemergence. 1 Feb. 2018. Web.