How BMS Uses Data To Provide Better Pharmaceutical Care

We spoke to Kamayini Kaul, Head of Data Lakes and Integrations at Bristol-Myers Squibb about data, healthcare and the future of them both


The American healthcare system is a complex industry. While it has some of the most advanced innovations in the world, it also has many flaws. And few are as devastating and inexplicable as the ever-rising cost of healthcare. In this respect, the US stands head and shoulders above the rest of the developed world. The most recent Peterson-Kaiser Health System Tracker Health statistic regarding health care spending per person in the U.S. was $10,348 – 31% higher than Switzerland, the next highest per capita spender.

However, the hope here, as with many of the other flaws in our society, is that technology will pull us out of this downward spiral. AI, machine learning, wearables, IoT etc. have all been integrated into the healthcare industry in some fashion in recent years. However, unlike other industries where the only barrier to entry is the ability to make additional profit, the healthcare industry has always had miles of red tape to circumvent. This is obviously because it deals with people at their most vulnerable. While we can expect several and unending ethical debates around the use of patient data and how involved humans/machines should be in the diagnosing/treatment/care of patients, the technologies enabling these will continue to become pervasive over the next decade.

Technology has brought both healthcare and, more specifically, the pharmaceutical industry into somewhat of a transformational period. More and more companies are now attempting to not only make a profit, but use all this tech and data to find ways to actually better the quality of life of their patients in meaningful ways. To further examine this, I spoke to Kamayini Kaul, Head of Data Lakes and Integrations at Bristol-Myers Squibb (BMS). BMS is a global biopharmaceutical company focused on discovering, developing and delivering innovative medicines for patients with serious diseases.

Kamayini Kaul joined Bristol-Myers Squibb in 2017 as the Global Head of Enterprise Data Lakes and Integrations. She leads global IT capabilities for BMS’s data lakes, master/reference data and middleware integrations platforms and applications. She’s a key lead in the execution of BMS’s digital health vision enabling faster drugs from discovery to patients with proven and targeted clinical outcomes.

As one of the speakers in the upcoming Big Data & Analytics in Healthcare Summit happening this May 22-23 in Philadelphia, we discussed the direction the industry is going and the impact she expects technology to make on it in the near future.

Do you think wearables will play a larger role in healthcare or are they too intrusive? Is healthcare equipped to exploit the data produced by wearables and other IoT devices?

  • Wearable devices and IoTs present a very critical source of capturing patient-reported outcomes and attitudinal data, both in RCT and Real-world settings.
  • Most organizations are at various stages of developing a business strategy for harvesting IoT and wearables data. While their IT organizations are likely well positioned, with most organizations having a ready architecture point-of-view on IoT data, this is likely to remain an area in life sciences where the business use cases will drive the pace of adoption.
  • Adoption in select use cases will be a basis of competitive differentiation from an end-patient value creation perspective.

Do you think data, in general, is being utilized effectively in healthcare? What can be done to better the situation?

I believe we have only started to chip away at leveraging ‘data as the new currency’ in healthcare. While there has been a lot of historic emphasis on high-quality data being leveraged for conducting pre-clinical and clinical study work, leveraging data across the traditional silos of Discovery, Clinical and RWE continue to be a challenge and underutilized.

Hence, there are a number of areas that need foundational work to be conducted over the next few years:

  • Finding a way to link data sitting in functional silos of discovery, clinical & RWE.
  • Building in minimum data completeness and governance to ensure accountable use of data for problem-solving and insight generation.
  • There is massive room for integrating patient and caregivers alike into ensuring optimal clinical outcomes leveraging digital capabilities offered by wearables, social media, AI for AE reporting, etc.

How do you balance the concerns patients have for their confidentiality against the need for data to innovate?

There are a couple of criteria, which we have to account for, while designing patient-interaction oriented digital assets and processes:

  • Plan for informed consent from the onset.
  • De-identify data whenever possible.
  • Secure PII from non-PII data separately.
  • Use Honest Brokers when in doubt when exchanging patient data between organizations.
  • Design for the right-to-forget as recently identified by GPDR.

Most patients are willing to engage in the journey to medical discovery and invention, as long as the above-stated criteria are addressed sufficiently and transparently.

Do you feel there needs to be more regulation or less when it comes to utilizing healthcare data?

In general, the healthcare industry, especially when talking about medical devices and in-vivo products and services, has always been highly regulated. I believe what is needed is to regulate the use of this data by adjacent industries to ensure consumer protection. For instance, disability insurance premiums being increased and possibly denied based on a patient’s medical history.

How much use of visualizations in your enterprise? Do you find it an effective way of communicating ideas to the less technically minded?

  • There is a significant use of visualization in the enterprise. It was always prevalent in R&D and Manufacturing but is now also mainstream in Commercial and Shared Services Functions like Procurement, Audit & Compliance, Finance, etc.
  • Visualization as means of communication has less to do with being technically minded and more to do with an individual’s ability to learn and absorb information, especially insight heavy content. Often, a person is innately orientated to being visually or quantitatively attuned. The power of visualization is more in knowing when to leverage a visual communication technique versus a traditional tabular or quantitative/qualitative representation. Also note, when vast amounts of data are in question, most human brains resort to visual approximations to assimilate a higher order roll-up of information.

What technologies do you foresee changing the way you use data over the next five years?

Key technologies I see dramatically changing the way we do business in healthcare are:

  1. AI and ML being embedded in scientific discovery and development processes
  2. Blockchain for streamlining the supply chain of pharmaceutical manufacturing
  3. Robotic Process Automation being applied to operations heavy sub-niches of healthcare such as claims processing.
  4. NLP for deriving information from unstructured data.

And what strategies do you have in place to prepare yourself to use them?

Among the key initiatives most large enterprise are pursuing today to advance the enterprise understanding of these digital capabilities and prepare a battery of use cases of immediate value are:

1. Digital Acorn seed funding a basic starting pool for innovation using emerging technologies for specific use cases.

2. Innovation LightHouses - strategic listening and evaluation functions to shape the enterprise’s response and strategy for emerging technologies.

3. Digital agenda embedded into IT enterprise portfolio – the standard IT/digital capability evolution via a classic IT portfolio of investments.

For similar insights into how companies are improving patient outcomes with data analytics, attend our Big Data & Analytics in Healthcare Summit happening in Philadelphia, May 22-23.


Read next:

Why We Need Data Visualization To Understand Unstructured Data