Kelly H. Zou, Jim Z. Li, and Nikuj (Nik) Sethi

Kelly H. Zou, vice president and head of medical analytics and insights at the Upjohn Division of Pfizer Inc.; Jim Z. Li, senior director of medical analytics and insights at the Upjohn Division of Pfizer Inc.; and Nik Sethi, business technology lead at the Upjohn Division of Pfizer Inc., taught a three-part web-based lecture on generating and harnessing real-world evidence (RWE), health information technology (HIT), and artificial intelligence (AI). The lecture was jointly sponsored by the Biopharmaceutical, Health Policy Statistics, and Statistical Learning and Data Science sections. Here, Richard Zink from the Biopharmaceutical Section asks the lecturers a few follow-up questions.

Kelly, describe your recent work since the interview on ASA Biopharm’s Podcast in 2017.

Kelly: My colleagues here and I have been busy looking for ways to improve treatment for noncommunicable diseases, known as the NCDs, around the world. According to the World Health Organization (WHO), “noncommunicable or chronic diseases are diseases of long duration and generally slow progression.” According to the WHO, there are four major NCD categories: cardiovascular diseases (like heart attacks and stroke), cancer, chronic respiratory diseases (such as chronic obstructed pulmonary disease and asthma), and diabetes.

Give us a brief overview of your current roles and how you became interested in statistics, data science, and business technology.

Jim: Kelly and I both work in medical analytics and insights, research, development, and medical in Pfizer’s Upjohn venture.

Nik: I am the BT lead for Upjohn’s research, development, and medical functions and a peer of both Jim and Kelly. Upjohn’s mission is relieving the burden of NCDs with trusted, quality medicines for every patient, everywhere. We seek to leverage our portfolio, experience, and expertise to become the trusted partner of choice to all stakeholders committed to improving patient health. We aim to significantly expand access to our high-quality medicines. Our goal is to treat 225 million new patients by 2025, and it appears to be an achievable goal to make a global impact in health care.

Jim: While Kelly is a statistician by training, our career paths took us from medicine, health care policy, [and] business technology to health care innovation. In each step of the way, statistics, data science, and analytics are not only important tools and methodological bases, but also infrastructure and enablement building. Therefore, our diverse backgrounds and capabilities are complementary, especially for interdisciplinary and collaborative work in the spirit of being fast, focused, flexible, and fun.

Describe some of the opportunities and challenges of RWE for noncommunicable diseases, known as the NCDs.

Jim: According to the World Health Organization, there are four major non-communicable disease categories. Here, the word noncommunicable means these diseases are not contagious. The first is cardiovascular diseases (like heart attacks and stroke). The second is cancer. The third is chronic respiratory diseases (such as chronic obstructed pulmonary disease and asthma). The fourth is diabetes.

Kelly: Some startling statistics: NCDs are the leading cause of death around the world, killing 41 million people per year, equivalent to 71% of all deaths globally, with more than three quarters (32 million) occurring in low- and middle-income countries. The devastation of such a “silent killer” increasingly becomes a global health burden. In the US, cardiovascular diseases accounted for 840,678 deaths in 2016 (i.e., approximately 1 of every 3 deaths). In 2010, the global cost of CVD is estimated at $863 billion, and it is estimated to rise to $1,044 billion in 2030.

Nik: Given the prevalence of NCDs globally—including emerging markets, Greater China, and developed markets—developing new capabilities that are practical and built on accessible technologies is key to identifying patients that may be unknowingly impacted by an NCD, educating health care providers on treatment options for NCDs—above and beyond a single disease—and working with payers to demonstrate through RWE and technology capabilities how addressing NCD burden broadly can provide significant value in the health care system. These opportunities make our work ever more important, particularly in developing countries in the emerging markets.

How can RWD be useful for diagnosing and treating NCDs?

Kelly: For medical researchers serving in industry, government, and academia around the world, RWD, big data, and digital capabilities and technologies must be enabled and elevated globally, especially when randomized controlled trials (RCT) become increasingly costly. There are tremendous opportunities around the world to gain insights through RWE and digital information outside RCTs.

Jim: For example, in developing countries in Asia, Latin America, Africa, and the Middle East, there is a large population that can benefit from breakthroughs that change patients’ lives. The expansion of the middle class, the desire for quality medicines and therapies, and the rise of NCDs are the realities these regions face every day. Furthermore, the advances in information and telecommunication infrastructures have also drastically improved, which has enabled a massive amount of RWD to be generated from many sources outside the RCT framework. Such data sources include electronic health records, patient registries, patient-reported outcomes, and digital apps.

Nik: Speaking of opportunities but at the same time challenges, cutting-edge innovation can be enabled when data are available or in aggregate. In this new era of data as precious as oil, data integration, access, interoperability, standardization, quality control, security, and privacy protection are important issues where data is available. A major barrier, however, is simply accessing data due to a diverse set of sources and repositories of data. As we move into a new era of RWD in the health care landscape, a key challenge is going to be how to make data accessible to the right parties at the right time to only be used as prescribed. Finally, once data is accessible, deriving algorithms that are robust and tailored toward the specific patient populations will enable patient identification and help address challenges around the patient journey. The challenge we are facing as an industry is the shortage of talent in both data translators, who are like shepherds with subject-matter expert knowledge, and data scientists.

AI is a commonly used term these days. Most statisticians likely think of it in terms of predictive modeling. What other uses of AI are there? Are there examples of AI that are being used in practice? What does the future hold for AI?

Nik: From a consumer’s perspective, AI has in many ways become pervasive in enabling mobile devices and applications running on the mobile devices to be more intelligent. For instance, I recently got a text message from a number that was not saved in my contact list, and the phone (searching through my other messages and emails) suggested who the person could have been and was correct. That has a direct impact on me, saving me time and giving me a more personal relationship.

From a health care perspective, AI is having a profound impact across the spectrum, from research and drug development to patient care—including health care management. A great example is in medical imaging and diagnostics, in which technologies are being used to detect skin cancer from photos taken by patients. The key question for leaders in the industry is how to put such a complex set of information together within the unique country-specific health care systems and data repositories, which will improve diagnoses and treatment.

Jim: The different languages in those countries may make variables and formats in databases harder to standardize. These countries also tend to have increased data and privacy protections, as well as diverse data formats sitting in patient registries without being standardized and connected (see the British Medical Journal website).

Kelly: Let me give some final closing remarks here. It is certainly an exciting era for statisticians and data scientists to make meaningful and impactful contributions to combat these chronic and debilitating NCDs. Skillful statistics and data science capabilities are critical in gaining insights from RWE. For digital and technological innovations that can target therapies and optimal treatment strategies, it is imperative to foster collaborations and partnerships.

EDITOR’S NOTE: The views expressed here are the authors’ and do not necessarily represent those of Pfizer Inc.

Listen to Kelly Zou’s interview for ASA Biopharm’s Podcast in 2017.