What if everyone's health history was recorded and analyzed? What could be done with this information to improve treatment regimens or develop new medicines?
A couple of recent pairings are working to answer these questions. IBM (NYSE:IBM) Watson Health announced on December 10, 2015 that it was partnering with Novo Nordisk (NYSE:NVO). Not much earlier, Google (NASDAQ:GOOGL) Life Sciences (Alphabet) announced on August 31, 2015 that it was partnering with Sanofi (NYSE:SNY). And both Pfizer (NYSE:PFE) and Novartis (NYSE:NVS) have teamed up with Amazon (NASDAQ:AMZN) to take advantage of Amazon Web Services. I take a look at what these partnerships might provide from both the healthcare and revenue standpoints, with a focus on the IBM/Novo Nordisk collaboration.
What can Big Data analytics contribute to healthcare, especially as it relates to Big Pharma? IBM described its Big Data analytics approach to healthcare in a presentation. This summary is representative of how all Big Data analytics services will contribute.
In general, Big Data analytics takes large amounts of raw information, distills it into what's needed, and converts it into useful, actionable data. In healthcare, this means taking health records of various sorts along with related ancillary information and finding relationships and trends that can impact health-related decisions. From a patient's perspective, the wealth of information from other people with similar or related conditions can be brought to bear on that individual's specific situational treatment.
This personalization of care can, to the extent possible with the known and extrapolated data, inform doctors of treatment options and potential outcomes. The expectation is that such optimized treatment personalization will improve provider decision-making and patient outcomes, which will, in turn, reduce payer expenses.
How does the data analytics part work? IBM's offering is, of course, Watson, and the division that provides these services is IBM Watson Health. Watson is "a technology platform that uses natural language processing and machine learning to reveal insights from large amounts of unstructured data." Natural language processing provides interpretation to determine meaning to input questions and facilitates answer searches that are smart enough to address ambiguities within the data.
The Big Data analytics on a platform like Watson can be described using a three-phase process. First, it takes the raw, complex and convoluted information from multiple sources, extracts the relevant information and organizes it. In healthcare, this can yield an anonymous representation of a patient based on that person's health history, relatives' health history, genetic information, etc. Second, these representations are parsed for pertinent, specific features that can feed the analysis engine. Third, analysis is performed using complex statistical algorithms and a variety of predictive models to reveal new insights, patterns and relationships between patients.
Along with the potential for improved patient care through individualized treatments, there are several other Big Pharma benefits from Big Data analytics. Here are some more specific benefits:
- Permits early identification of at-risk subjects using connections between health, genetic, ancestral and other databases, which supports preventative lifestyle changes and treatments, and contains population-wide healthcare costs.
- Permits improved subject targeting. Data yields early identification of new patients for Big Pharma and potentially improved outcomes, which should lead to more efficient pharma sales forces and higher sales.
- Provides appropriate actions to be taken in complex cases, including correct tests and properly selected, dosed and timed treatments. Better treatment selection and regimen will lead to better patient compliance, thereby improving sales.
- Enables on-person, automated and controlled electronic drug-delivery devices that are connected to the data analytics engine through the internet to provide personalized treatment regimens.
- Informs R&D investments, including drug development, by providing a means for computational studies and advice on treatable populations, trial design, subject selection (e.g., best-responding patient groups), delivery method (e.g., oral versus injectable), etc.
- Provides the opportunity to use results from information provided and acted upon as inputs into the analytics engine. Such feedback improves multiple aspects of the Big Data analytics engine, including accuracy and run-time durations.
Diabetes is The First Focus
The ideal application of Big Data analytics to Big Pharma healthcare is with complex chronic diseases that have large populations. Here are the top five chronic disease categories in the U.S. Note that nearly 10% of the population, or 30 million people, have diabetes.
The International Diabetes Federation (IDF) estimates that 415 million people have diabetes today, which would go up to 642 million by 2040. They identify the complications from poorly managed diabetes as including cardiovascular disease, kidney disease, nerve damage, and blindness. The healthcare costs of diabetes and its complications account for 12% of the global health expenditure, which corresponds to $673B. The sizable market is a strong reason for the initial focus on diabetes.
Both the IBM-Novo Nordisk and the Alphabet-Sanofi partnerships are first targeting diabetes care, citing the expectation that managing this disease would be well served by artificial intelligence. On an individual level, the treatment complexity arises owing to the personal interplay between insulin levels, food intake, and exercise.
Diabetes is also an excellent first target for Big Data analytics because of the complexity of the disease's progression within its large population, including the occurrence of comorbidities. Only around 5% of patients with diabetes have it as their only chronic condition. In fact, the vast majority of patients with chronic diseases have multiple comorbidities.
Correspondingly, there is a wide range of diabetes-related issues, treatments and concomitant health records available, including many of the standard sources just discussed. IBM Watson already has 50 million anonymized American health records, making it a valuable data resource already. There is a need for organizing this information and extracting insights from it in order to improve these complex care situations.
New electronic sources of diabetes data are also becoming available that could provide continuous blood sugar monitoring. Mobile applications that provide near real-time digital data regarding diabetes condition, treatment and compliance are already being introduced in the marketplace. One example is WellDoc and its BlueStar type-2 diabetes platform, which analyzes current data, such as blood glucose levels and medicine usage, contextualizes with past data, provides summarized and curated data to a healthcare team and provides a dynamic self-management plan update.
IBM and Novo Nordisk
IBM's press release states that "the companies will explore possibilities for improved diabetes care via insights from real-time, real-world evidence of Novo Nordisk diabetes treatments and devices. By harnessing the potential of the Watson Health Cloud, Novo Nordisk aims to further advance its offerings to people living with diabetes and their healthcare professionals."
Novo Nordisk expects to leverage the "combined capabilities to improve the lives of people with diabetes by making the management of the condition more simple, effective and measurable." The companies will achieve this with, among other things, personalized care initiatives.
Such individualized care regimens will be developed from knowledge gained from population data in combination with day-to-day personal situational data. Using Watson, for example, they may be able to learn from data how to improve regimens such as dosages and timing, and eventually provide treatment notifications or even automate insulin delivery.
There is also another interesting, potentially related partnership at hand. IBM is connected to Apple (NASDAQ:AAPL) in a healthcare capacity. Using iPad and iPhone apps to provide inputs, the companies are working with Japan's largest health insurer, Japan Post, to improve healthcare and wellness in Japan's seniors. This an offshoot of the IBM and Apple global business applications collaboration formed in 2014. So, Watson is essentially becoming an integral part of the Apple healthcare ecosystem.
At this point, there is no acknowledged linkage of Apple to the Novo Nordisk collaboration, but it's certainly possible this could come about. It really is just a small step to bring Novo Nordisk into the fold, e.g., using Apple iOS devices with an IBM-developed app linked to Watson to collect and analyze diabetes data for Novo Nordisk and transmit actionable information back to the patient via the iOS device. Note that Apple has already equipped Eli Lilly (NYSE:LLY) in an enterprise initiative, providing iPads through its mobility partner program.
Alphabet and Sanofi
The head of Alphabet's life science team said the company wants to build "smart insulin delivery devices, smart measurement devices, and an interface and an integrating platform that helps physicians and patients see how they're doing." Working through cloud-based data will allow it "to move away from the reactive and episodic, towards the proactive and preventative." Alphabet's partnership with Sanofi is also intended to develop ways to improve diabetes treatment regimens and eventually automate them.
Amazon and Pfizer, Novartis
Pfizer has teamed up with Amazon for general cloud services. It is using Amazon Web Services "to provide a secure environment in which to carry out computations for worldwide research and development, which supports large-scale data analysis, research projects, clinical analytics, and modeling."
Novartis teamed up with Amazon Web Services to run "a project that involved virtually screening 10 million compounds against a common cancer target in less than a week." They conducted 39 years of computational chemistry in 9 hours and saved tens of millions of dollars.
How big is the market, and how will revenue be generated in these collaborations? (I exclude straight drug development-related services here, because I expect this revenue to be relatively small.) There are at least four possibilities of large revenue-building contributions: preventive services, improved treatment services, active devices and real-world proof of value for payers. I'll address these in terms of the IBM-Novo Nordisk collaboration, though these could be generalized to any such collaborations, especially Alphabet/Sanofi.
We can get a rough estimate of the value of the IBM-Novo Nordisk collaboration by considering the revenue impact of the contributions just discussed. I'll make a lot of simplifying assumptions and do some rounding to get there.
It's been estimated that the healthcare savings from the implementation of effective Big Data analytics generated preventative actions is at least $450B, or 12-17% of U.S. healthcare costs. So, one would expect there are tens of billions of dollars that could easily be subsumed by the Big Data/Pharma participants to identify cost-saving preventative care methods and still leave a significant savings cut for the insurers/payers.
However, currently in the U.S., wellness programs are absorbed by large employers under trivial costs. For example, these efforts are typically limited to health consultations, promotional activities and minor preventive care like flu shots. And many of these programs are non-specific, as they cannot be related to specific disease states. Many chronic diseases have common risk factors.
Consequently, new ways would need to be developed, from which payment for these preventive services would be made. An example would be an additional collaboration with a payer with whose cost savings on their coverage universe would be estimated and shared with the data/pharma co-collaborators.
What dollar amount might a payer be able to afford for Watson-based services? In evaluating this benefit, I don't separate out diabetes from other chronic conditions, and so, also don't separate out a percentage cut for Novo Nordisk.
If the average health insurance premium is around $4000 and savings are on the order of 10%, or $400/person, it would be reasonable to estimate around 25%, or $100, of the savings could be used to fund Watson-based services and still be quite beneficial to the payers. Alternatively, a premium increase of just 2.5% could pass this amount directly to the insured. If IBM made such a deal with health insurer Anthem (NYSE:ANTM) and its 50 million policyholders, this would produce $5B revenue annually for IBM. For the limiting case, if all 300 million Americans were covered under Watson's purview, that would rise to $30B.
Improved Treatment Services
There are two outcomes to improved treatment service: increased pharma sales and decreased healthcare costs. To address these opportunities, the companies could share in medicine-specific sales increases to pay for the former and, again, make cost savings sharing deals with insurers and pharmacy benefit management companies to pay for the latter.
First, let's link Novo Nordisk sales increases to payment for IBM Watson services. Novo Nordisk has $15B annual revenue. Assume Watson can increase sales some reasonable amount like 20%, or $3B. With profit margins around 30%, that leaves around $1B for Novo Nordisk to use to support Watson services. If they're willing to split this evenly, IBM would see $500M in revenue.
Second, let's link insurance or pharmacy benefit management companies and providers, with a per-person fee, to obtain compensation for healthcare savings. This would essentially be a fee paid for inclusion in IBM Watson's data analytics engine and Novo Nordisk's treatment regimens. In the end, structural changes in the healthcare system will be required to take advantage of, and pay for, these kinds of collaborative population-wide advancements.
Such service costs would be in addition to preventative care costs once a chronic disease diagnosis has been made; however, the costs would be distributed among all policyholders. One way to account for them would be by relation to the relative cost savings to treatment of the disease, and there would need to be some split between the partners.
Annual U.S. diabetes healthcare costs are around $200B. At 10% cost savings, that would amount to $20B, or $70/person, nationwide. A 25% cut to such partnerships, with the rest going to the payers, would be around $20/person. Using the Anthem example, this would produce $1B for the partnership, which, assuming an even split with Novo Nordisk, would leave $500M for IBM.
Active Treatment Devices
The development of active electronic treatment devices is probably the most straightforward for determining revenue, since a simple sales revenue sharing agreement could be put in place. Such devices would require connection to real-time analysis and decision making, and so, may require a subscription-type service with Watson.
Cost competition would likely not allow too much value-added pricing to such products, but they could reach blockbuster status. Let's assume a 20% premium to get this active service and $2.5B in pre-premium sales for these products. That would mean around $500M for IBM.
Real-world Proof of Value
A major issue facing pharmas is getting pharmacy benefit management companies and insurers to include their products in formularies. Some formularies are being decided on outright cost alone, cutting out competing products.
What if you could prove that your drug conferred a health advantage and long-term cost savings from a population standpoint over a cheaper competing product? This value proposition could be investigated using Watson and used to develop proof that could influence payer decisions. This has been stated as one of Novo Nordisk's goals in its partnership with Watson.
AstraZeneca (NYSE:AZN) has already implemented such a relationship with Anthem's data and analytics group. They plan to conduct real-world studies to determine best treatments, including some for chronic diseases, as well as to guide R&D investment decisions. AstraZeneca is working with payers to ensure it has the evidence to obtain coverage for its drugs.
Additionally, it has been reported that Novartis has a deal with insurers whereby the company will get paid extra if patients receiving its new heart-failure drug stay out of the hospital. We are entering a new era of healthcare where monetary payments based on in-situ, real-world evidence is coming to the fore. It's a small step from direct impact, like that with Novartis, to indirect impacts, such as population-based reductions of comorbidities. Watson and other data analytics services will help monetize these.
While essential, I wouldn't expect the pharma valuation of this service to be more than a couple percent of earnings. For Novo Nordisk, this would put its payment to IBM under $100M.
A Look at the Numbers
Putting this all together, an optimistic look at the Novo Nordisk collaboration could produce up to $2B annual revenue for IBM, which is apart from preventative services. Because much of the infrastructure is already in place for Watson applications, I would expect these revenues to be supported by relatively high profit margins, well above the current sub-20% IBM corporate numbers. Consequently, we could see overall profit margins move up over the next several years as Watson gets leveraged more broadly and IBM's business mix changes. Longer term, preventative services, if proven out and tied to a payer, could produce $5B+ additional revenue.
With IBM's revenue at $80B, these numbers only make a small addition, but they could, in part, be replicated over many industries with a multitude of companies.
The relative revenue benefit for Novo Nordisk could be far greater in both the additional revenue and the protection of its market share. My simple estimate supports a potential revenue increase of $5B from these efforts, along with potential protection of its current $15B and growing revenue stream.
Decent revenue and profit potential exists in the Big Data analytics/Big Pharma partnerships. These pairings will revolutionize healthcare by simultaneously improving lives and reducing population-wide healthcare costs. Such savings will be achieved by implementing personalized care that is only possible through such Big Data/Pharma collaborations. There is, however, a significant challenge in the means by which Big Data/Pharma will obtain a portion of those savings to pay for services. And the near-term payoff for IBM will likely depend in large part on Novo's estimation of the partnership's value.
Perhaps most importantly for Big Pharma will be the ability to produce lifetime cost-benefit data for treatments that can be used to justify drug prices. In a world where pharmacy benefit management organizations are making selective decisions, excluding some drugs on price alone, these collaborations will provide opportunities for pharmas to directly include lifetime cost savings as well. This may well be the only way for Big Pharmas to remain competitive in indications where there are many options. Hence, I expect Big Data/Pharma collaborations to become commonplace in future.
Novo Nordisk is a worldwide leader in diabetes treatments, and I really like this Watson partnership. This particular application is just a small piece of what IBM hopes to be a major turnaround story. I believe it is on the right track.
If this holds, IBM could become a very good investment. However, I expect IBM's valuation to fall further in the near term owing to overall market weakness, so there may be better opportunities coming up. Investing in this particular diabetes story would best be done through the purchase of Novo Nordisk stock, since the expected revenue benefit will be a much greater percentage of current revenue.
The success of Watson bears watching carefully over the next year or so. The large potential, and seemingly essential need for Big Data analytics by Big Pharma, makes it a worthwhile collaborative endeavor even though it may not contribute significantly to the bottom line in these companies for at least a few years. However, if IBM develops many similar arrangements with hundreds of companies across a multitude of industries, Watson collaborations could foreseeably lead to very meaningful future revenue and big profits.
Disclosure: I'm indirectly invested in Novo Nordisk through its partnership with Emisphere Technologies (OTCQB:EMIS), which would likely benefit from the IBM Watson collaboration.
Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.
I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.
Editor's Note: This article covers one or more stocks trading at less than $1 per share and/or with less than a $100 million market cap. Please be aware of the risks associated with these stocks.