Digitizing Drug Development: How much can it save?

The potential of data/ IT to transform drug development is a debate unfolding across healthcare and technology. Emerging cracks in pharma’s pricing power may force healthcare companies to embrace change sooner than many expect. Technology could disrupt the traditional R&D process, finds a Morgan Stanley Investment Management report. Excerpts

Healthcare, specifically biopharma, remains one of the industries least affected by digitization. While scientific breakthroughs such as genomics and immuno-oncology have created better medicines, the industry has yet to fully exploit technology to create more efficient processes in drug development and the supply chain. High regulation and low transparency into pricing and profitability pools are responsible. As soaring drug prices stress the healthcare system, political and public pressure may force biopharma to seek efficiencies and lower costs. Technology can potentially reduce prices for individuals while improving the efficacy of R&D and enabling more medical breakthroughs.

Do cash-flush tech companies spy opportunity in healthcare inefficiencies? IBM has invested ~$4 billion in healthcare acquisitions and Google’s parent company, Alphabet, has set up two separate healthcare ventures, Verily and Calico. Interestingly, technology executives hold an estimated ~7 per cent of healthcare company board seats (over 40 per cent of whom have joined since 2014). From consumer-focused wearables to machine-learning-driven, evidence-based medicine, digitisation is emerging across all segments of healthcare.

Our global therapeutic, healthcare services, and technology teams have developed a matrix to frame potential outcomes along two vectors. One registers the impact of digitisation on price transparency from low to high. With low transparency, companies enjoy historical levels of pricing power, while high transparency lowers price inflation to zero and then negative single digits, as seen in European markets. The other vector indicates digitisation’s impact on innovation, with improvements that include enhancing R&D yields, finding new patients to treat, and increasing prescription fill rates, adherence, and persistence. The resulting quadrants yield four scenarios that we explore in depth: Balancing Act, Rainbow’s End, Status Quo, and “Doomsday.”

We see consistent trends across industries that have experienced increased digitisation, such as Media, Hotels, and Airlines:

• Despite concerns over stock performance and potential early underperformance, the industries outperform the S&P over the long term.
• New disruptive companies always win.
• Incumbents also outperform if they are able to adapt. Intellectual property is a driver of incumbent success, which could favor much of healthcare, considering the scientific and regulatory knowledge required.
• In most cases, pricing pressures occur but are more than offset by resulting higher utilisation rates and cost efficiencies from digitisation.

What’s in the Stocks?

Analysing the potential impact of increased technology use across drug development companies and the supply chain suggests as much as ~10 per cent more downside to pharma, biotech, and distributor market capitalisation in our Balancing Act scenario. This implies the industry needs to further sharpen its pencils to lower cost structure. Contract research organisations (CROs) should thrive from increased innovation but risk being disrupted if they are unable to keep up. Healthcare initiatives aren’t yet embedded in technology company valuations. Apple, Fitbit, Alphabet, and IBM have the biggest potential upside.

We explore potential outcomes through a quadrant analysis, with technology impacts on one vector and drug price changes on the other. Our four scenarios are Balancing Act, Rainbow’s End, Status Quo, and “Doomsday” (see Exhibit 1) which we outline below and explore in detail later in the report. For each scenario, we identify stocks to own and avoid. In the process of developing this analytical framework, our teams’ thinking has evolved.

• On pricing, our views have converged (Balancing Act and “Doomsday”). We expect drug companies to raise list prices less aggressively due to political scrutiny as they continue to face rebate and competitive pressures. At this point, divisions in Washington remain an obstacle to broad legislation to lower drug prices, though this situation could change.
• On the level of technology integration, we diverge into two camps. One group believes that over the longer term, increased use of technology will improve the clinical trial process (Rainbow’s End or Balancing Act). The other group thinks technology will contribute to medical cures as in the past and that technological progress can enhance innovation, but it will not drive a step change in cures. R&D productivity will continue to be dominated by “traditional” factors such as genetics and biological innovation, with technology aiding discoveries but not becoming the major driver (Status Quo or “Doomsday”). Our internal debate mirrors what we have learned from communicating with industry participants. Some drug companies believe that data can transform clinical trials and are pursuing partnerships, while others are pushing back.

The drug industry has historically been relatively insulated from pricing pressure, but change is coming. Over the past three to five years pressure from payors (in the form of higher rebates) has materially increased and is unlikely to abate for the foreseeable future. Escalating drug price increases are creating systemic stress. Drug companies, pharmacy benefit managers (PBMs), and plan sponsors are passing along gross price increases to consumers, who are unable to pay. The resulting erosion in the quality of healthcare is attracting public scrutiny, and as pressure mounts, high drug prices are becoming a bipartisan issue. External forces may sharpen drug companies’ focus on costs, including the shift toward a value-based reimbursement system and more attention from policy makers (the government is the largest payor at 45 per cent of US healthcare expenditures).

Technological “transformation” has come later to biopharma than other sectors, but is gaining a foothold. First, strict guidelines in the FDA approval process have constrained innovation. Second, the business model of taking risks to develop drugs (or purchasing companies that have developed them) has been immune to price pressure from competition and regulation. Finally, unresolved questions in fundamental biology have limited the scopefor technological efficiencies – if you don’t fully understand a problem, fixing it is a challenge. However, the FDA is taking its first steps toward adopting changes. The agency has expanded the use of wearable technology for patient reporting as part of clinical trial design and provided clearer guidelines for the use of risk-based monitoring.

Recent activity at the periphery of the healthcare drug development space may be a harbinger of change. Inefficiencies and large investment budgets have motivated tech giants to expand the field of opportunity of IT to healthcare. IBM has invested ~$4 billion in healthcare acquisitions, and Google’s parent company, Alphabet, set up two separate healthcare ventures – Verily and Calico. The changing composition of boards across Tech and Healthcare companies offers further evidence of change. We estimate ~ seven per cent of healthcare company board seats are held by technology executives and a slightly higher percentage of technology company board seats, by healthcare executives (over 40 per cent of whom have joined since 2014). Most recently, Apple’s marketing head, Phil Schiller, joined Illumina’s board of directors in July. IBM and Pfizer lead in board positions assigned to healthcare and technology executives, respectively. In evaluating a company’s governance, we view board experience as a key pillar. As healthcare and tech companies team up in symbiotic partnerships, it will be increasingly important for relevant “cross-pollinating” of skill sets at the board level. At the executive level, companies like Alphabet’s healthcare subsidiaries have been aggressively hiring talent in the biotech and life sciences industries, including Arthur D. Levinson, former CEO of Genentech and current Apple Inc. chairman, who is Calico’s CEO; Tom Insel, the former head of the National Institute of Mental Health, who is at Verily; and Jessica Mega, a former cardiologist at Harvard Medical School and Brigham and Women’s Hospital, now at Verily. Additionally, Alphabet’s Sidewalk Labs recently posted job listings for engineers, a chief health officer, and a head of community health in order to digitally transform how care is delivered.

Potential win-win for individuals and the society at large. Increasing both cost efficiency and price transparency can make healthcare more affordable for consumers – a social responsibility pressed upon healthcare companies, especially lately. In addition, pharma and biotech company R&D spending (together with government funding) benefits society through important medical knowledge and discoveries. Technology can potentially improve affordability for individuals while improving the efficacy of R&D and making more medical breakthroughs possible.

Better health outcomes are relevant for “impact investors,” who seek exposure to business models that solve social problems – the healthtech intersection could expand this investable universe. Asset managers increasingly seek to identify impact investing opportunities in order to address growing interest from asset owners. Framing the debate through quadrant analysis and scenarios we offer a quadrant analysis resulting in four scenarios: Balancing Act, Rainbow’s End, Status Quo, and “Doomsday.”

• Balancing Act: In this scenario we assume a shift towards a tighter price environment by 2024, triggered by either a move towards value-based pricing or price controls. This will be offset by R&D efficiencies in the clinical trial process from a combination of: 1) electronic health record mining, 2) electronic source data verification, 3) mobile technologies (for data capture), and 4) at-home testing.  Savings on Phase I-III development costs could reach ~30 per cent based on analysis by the Eastern Research Group. Based on our analysis of hypothetical profit streams, we see more downside to Pharma, Biotech, and Distributors’ market value assuming no other cost-cutting measures or further upstream (preclinical) enhancements. Contract Research Organisations (CROs) benefitting from improved productivity and increased outsourcing stand out as a group with upside opportunity, although we caution that more involvement from traditional tech companies has the potential to shift some of the value creation away from the CROs.
• Rainbow’s End: Innovation benefits from a combination of advances in biology and technology. Basic science has historically led to advances in therapeutic approaches to disease treatment, whether the advent of antibiotics, sequencing the human genome, or harnessing the body’s own immune system in new oncology therapies. The combination of technology and biological innovation leads to differentiated drugs which command a premium price without the pricing pressures experienced in the Balancing Act scenario. In this scenario, all subsectors stand to benefit, with upside ranging from 49 per cent for PBMs to +23 per cent for Biotech. Under such a scenario, pharma and biotech multiples could expand to a ~10x premium to the S&P 500, in line with levels seen in 1999-2001 when the mapping of the genome drove multiple expansion. This would imply a 27.5x PE based on current S&P multiples, leaving room for even more upside for stocks.
• Status Quo: The Status Quo scenario is based on Morgan Stanley global team estimates assuming price inflation continues at its current pace, new innovative products enter the market at premium prices, and the cost structure remains unchanged. Under thisscenario, PBM shares are most undervalued (20 per cent upside) while CROs would see least upside (+2 per cent).
• Doomsday: In this scenario we examine the potential impact on industry profits if price regulation takes its toll on industry profits with no corresponding improvement in the current cost structure, productivity, or risk profile. The Biopharma profit pool could be cut almost in half, likely resulting in players exiting the market and stifled innovation. In this scenario, our net present value of Pharma companies implies ~39 per cent downside to current market cap and ~34 per cent downside for the Biotech group.

In this scenario, pharma and biotech multiples could contract to 4x vs. the S&P500, in line with trough levels experienced in 2009-2010 when the entire sector derated on patent cliff fears. This would imply a 13.5x PE based on current multiples.

Our analysis holds constant cost of goods and selling and marketing as a percentage of revenues across two scenarios (Balancing Act and Rainbow’s End) but assumes no adjustments to cost structure in the “Doomsday” scenario.

How does technology lower R&D costs?

Our analysis in this report is focused on the clinical phases where we can identify tangible opportunities for technology to lower cost. We believe there is room to further optimise lead generation in pre-clinical stages ultimately resulting in lower downstream failure rates, but given a lack of transparency as to the specific technologies deployed, we do not include such potential savings in our analysis.

Sizing the opportunities

According to the Tufts Center for the Study of Drug Development (see here), the estimated average cost per approved drug is $2.6 billion. A material cost driver has been the high failure rates for drugs tested in human subjects: ~57 per cent of cost is spent on phases I to III vs. ~30 per cent in discovery and ~13 per cent in pre-clinical testing.

We note Biopharma companies include the time value of the opportunity cost foregone when a drug is in development as a capitalised cost on their balance sheets (on average an incremental $1,259 million to out-of-pocket costs per approved drug). While our individual company analysis is based on out-of-pocket costs, we note that if technology were to indeed lower failure rate/ risk, the opportunity cost will be reduced as well.

Analysis of the potential for late-stage technology efficiencies based on academic publications and data from Medidata points to the potential to save ~13 per cent to 32 per cent of Phase I-III costs, or $59 million to $151 million in savings per drug candidate or ~ $191 million to $470 million per approved drug  (see Exhibit 3). According to the FDA, the US agency has averaged 28 novel drug approvals per year from 2006 through 2014. While trials could vary meaningfully in scale and savings opportunities, assuming the average cost savings outlined above (~$330 million) per approved drug would yield an average $9.3 billion in savings over the course of the clinical trial process (~8-15 years from discovery to marketplace) for approved drugs.

The Eastern Research Group working under contract to the US Department of Health and Human Services (HHS) identifies sources of technology-driven savings (more detail and further cost savings can be found in the Eastern Research Group report, here). Our analysis focused on those which we think are the most relevant within the context of this report.

• Home testing and flexible facilities: Flexibility of patient follow-up during clinical trials could reduce labor costs in centralised hospitals (both in hours and skills needed to collect data), increase the frequency of data collection, and improve patient retention in trials. Home testing could improve the quality of data due to continuous tracking of certain parameters. One example is Apple’s acquisition of Gliimpse. The platform enables users to record theiremotional state of mind, track laboratory results, and record levels of pain to inform a physician. Medtronic as well works through its Integrated Health Solutions division to expand into patient monitoring (e.g. acquisition of Cardiocom in 2013). Similarly, point-of-care testing has increasingly become a focus for healthcare stakeholders, with diagnostics companies developing services for these purposes (e.g. Qiagen acquires point-of-care test maker Amnisure in 2012). Despite these moves, we have yet to see penetration in the home testing market. As a result, in our view, this remains an opportunity and we estimate savings from home testing and flexible facilities could save between seven per cent – 22 per cent per phase or ~$114 million and ~$267 million per approved drug.
• Mobile technologies for electronic data capture (EDC): Wearable devices is one of the fastest and most flexible new technologies to penetrate healthcare, creating the opportunity to gather larger quantities of biometric information on a continuous basis. They are becoming more mainstream, with 23 per cent ownership penetration of US consumers vs. 58 per cent owning traditional watches, according to our survey in 2015. Additionally, Fitbit has been used in over 200 clinical trials, ranging from obesity to diabetes to cystic fibrosis to cancer, according to the company. These devices can serve as a data source for drug trials, which is more precise than previously self-reported information by phone surveys, and provide management of data in the trend toward real-world evidence for reimbursement both pre- and post-trial. Potential cost savings from integrating mobile technologies is estimated to be 3 per cent – 12 per cent per clinical phase or ~$66 million and ~$160 million per approved drug.
• Electronic health records (EHR): Over the last seven years since the HITECH Act (Health Information Technology for Economic and Clinical Health Act), EHR penetration has reached ~80 per cent, establishing an infrastructure of data and technology for drug companies to tap into to analyse patient information to enhance the clinical trial process. Despite this progress, only 58 per cent of physicians report the ability to generate patient lists by disease via their electronic system. This shows the disconnect between the infrastructure and the ability to analyse complex data sets in real time for clinical analyses (where we see the potential for cost savings). In order for a system to efficiently find patients for clinical trials, it will have to comb through much more complicated data than an ICD-10 code, including lab results and medical history, something few systems have the ability to do today. IBM Watson has been a leader in this area, partnering with contract research organisation ICON and the Cleveland Clinic and Mayo Clinic to interpret clinicians’ notes and lab results in 80 million patient medical records to match patients who qualify for a study’s inclusion and exclusion criteria. Initial efforts are focused on breast, lung, colon, and rectal cancer trials. Potential costsavings from mining the data in EHRs is estimated to be ~2 per cent per clinical phase.
• Source data verification (SDV): The process of data collection in clinical trials requires regular and frequent site visits and source data verification (every 4-8 weeks on average). This process usually requires physical presence, which can be reduced by using technology such as smartphones and tablets, as well as platforms such as Medidata’s Targeted Source Document Verification and targeted monitoring through analytics. According to estimates from contract research organisations (CROs) we have spoken with, ~50 per cent of Biopharma companies have adopted some form of risk-based monitoring, but we estimate that only between 15 per cent and 25 per cent of clinical trials use risk-based monitoring currently. Accordingly, we see a significant runway of savings as penetration accelerates. Potential cost savings from technology- enabled risk-based monitoring are up to ~1 per cent per clinical phase, according to the Eastern Research Group.Across all therapeutic areas, we estimate that IT-driven savings could represent on average 18 per cent of Phase 1 costs, 27 per cent of Phase 2 costs, and 22 per cent of Phase 3 costs. Phases 2 and 3 unsurprisingly offers the highest potential for savings at ~$130 million on average per approved drug.

We acknowledge that 100 per cent adoption of the addressable savings we discussed is likely unrealistic or would take a long time to fully materialise due to regulatory constraints and potential imitations associated with specific therapeutic categories and patient populations. Regulatory requirements mean not all costs associated with drug development are variable, and fear of nonapproval makes drug companies less likely to adopt new methods. That said, regulatory agencies do appear to be catching up. For example, one CRO we spoke with highlighted a shift in client interest in risk-based monitoring for source data verification based on recent FDA guidance to the industry for how to implement.

Which therapeutic categories are more levered to potential cost savings?

Hematology, Respiratory, and Pain/Anesthesia are most levered to benefits from higher technology penetration. While this report highlights opportunities, we note that some of the results are counterintuitive (e.g., Cardiovascular is a low cost savings area despite very large patient numbers in trials and Hematology is highlighted as having cost-savings potential but Oncology is not). This is not to say that Cardiovascular and Oncology will not also benefit from emerging technology, but that based on the 4 IT/Data opportunities outlined above, they are not at the top of the list.