Drug-development efforts are always fraught with  promise and peril. Today, more than half of all investigational drugs in Phase III clinical trials fail 

, due to factors such as flawed design, inappropriate endpoints, safety issues, lack of funding, failure to demonstrate clinical efficacy and under-enrollment. Delays create economic consequences for drugmakers in terms of sunk costs and foregone earnings, and clinical consequences for physicians and patients who would have potentially benefitted from the novel therapy.

Meanwhile, clinical study costs continue to skyrocket. A recent study

that examined 138 clinical trials that supported the FDA approval of 59 new therapies found that the median estimated cost per trial was $19 million. Digging deeper, those studies that had no control arm incurred a mean estimated cost of $13.5 million per trial, while for those that included a placebo-controlled or comparator-drug study arm, the mean cost per trial of $35.1 million.

Against the backdrop of high failure rates and untenable clinical trial costs, drug developers are sharpening their focus on the promise of advanced data-analytics methodologies. At the same time, there is growing willingness among regulators and payers to evaluate a strong data-driven value case when considering regulatory submissions—for new medications and label expansions—and negotiating drug pricing and reimbursement contracts and formulary status (for recent examples from pharma and biotech, see sidebar further down page).

Improving the odds of clinical and commercial success

“From an accessibility point of view, there is no single place where drug developers can get all the real-world data (RWD) they would potentially need,” says Fernando Schwartz, PhD, AVP, global data science head, commercial, Merck & Co. Instead, drug developers must buy RWD from many disparate sources.

In this era of pervasively computerized healthcare, RWD that reflects both patient-level and population-level insights can be found in:

Billing and claims records related to healthcare and pharmacy activities

Disease- and drug-specific patient registries

Medical-imaging and lab-testing databases

Wearable medical-monitoring and consumer health-tracker devices

Information collected during telemedicine encounters

Historically, such repositories of RWD were thought to be too vast or too complex to be used effectively. But today, advanced data-analytics and modeling capabilities are able to explore and exploit the insights that are buried there, and enable relevant analyses, simulations and predictions to be carried out, yielding actionable, evidence-based findings, or real-world evidence (RWE).

“Collectively, efforts to build and exploit predictive algorithms (that can benefit from a longitudinal view through all aspects of a patient’s treatment and disease progression) offer a lot of promise, and early successes suggest a strong and growing future role for them. But there have also been  a lot of strong claims out there with regard to what all of these 

efforts can produce, and much of the full potential has yet to be seen,” says Schwartz. “Nonetheless, forward-looking drug developers are setting themselves up to explore these data-analytics technologies to fully understand the molecule and its potential clinical and financial impact.”

“Having data is not enough—the trick is to derive insights from the data, and that is where the bulk of the innovation is happening today,” adds Roman Casciano, senior vice president and general manager, Certara Evidence and Access. This helps drug developers to better understand the impact that factors such as the distribution of disease severity have on the measurable differences in treatment effect between the study arms, and develop predictions related to outcomes and probability of success as a function of trial design, he notes.

Modeling algorithms are first established and trained by humans, and then validated by advanced data-science methodologies, including artificial intelligence (AI), machine learning (ML), natural language processing (NLP) and others. Such algorithms are able to analyze massive data sets, find patterns, evaluate competing trial-enrollment criteria and endpoints, and identify safety signals and other anomalies earlier in the drug-development process.

“It is truly the upside and downside of a free-market society—competition leads to much more innovation when it comes to commercially available sources of valuable RWD,” says Ed Ikeguchi, CEO of AiCure, “But it creates persistent interoperability issues that must be reconciled if we are to ever make full use of the rich RWD sources available.”

“With so much fragmentation in the realm of data generators and data aggregators today, you literally cannot lay the data side by side,” adds Schwartz. “It’s like trying to assemble a jigsaw puzzle where some pieces are built from cardboard while others are built from wood or metal. This complicates the way you can build models.”

Today, considerable time and effort is being expended to clean and harmonize RWD aggregated from disparate sources, and each data set requires labor-intensive custom onboarding. “Such data must be linked through a common token while maintaining HIPAA compliance, and then standardized, normalized and integrated into a data set that can drive business analytics that are actionable,” explains Theresa Greco, chief commercial officer for Prognos Health, whose platform allows clients to securely analyze billions of HIPAA-compliant laboratory-testing data and health records related to more than 325 million de-identified patients. “Such efforts can index and process billions of records in seconds, making it possible to provide the answer to key healthcare questions in minutes instead of months,” she says.

According to Aaron Galaznik, MD, Head, RWE Solutions, Acorn AI by Medidata, the company is able to analyze a million data points and identify 4,000 data patterns in less than an hour, “with the combination of smart data engineering, extensible cloud-based solutions and advanced analytic capabilities.”

Meanwhile, patient centricity has emerged as the rallying cry in drug development these days. Examining PROs and endpoints helps drug sponsors understand not just the patients’ unmet medical needs but their preferences and values (related to quality of life), too. By creating clinical outcome assessments (COA) at scale, based on RWD, drug developers can ensure that the patient perspective is pervasive in all drug-development and commercialization efforts. Today the industry is just scratching the surface on which components of the COA should be added to trials.

While no one disputes the inherent value of the randomized controlled trial (RCT) process—which is routinely referred to as the gold standard for drug development—many would argue that the process is deeply flawed by design, and should no longer be the only method by which investigational therapies move from the bench to the bedside.

RCTs routinely use strict inclusion and exclusion criteria —based on age, gender, racial/ethnic status, pregnancy, the presence of comorbidities and genetic status (such the presence or absence of specific biomarkers that correlate with clinical outcomes)—to create an ideal, homogeneous group of trial participants. However, such strict gatekeeping leaves many patient subgroups out in the cold and narrows the generalizability of the clinical findings in real-world healthcare settings.

In addition, a variety of structural barriers related to race/ethnicity, age, gender, educational and socioeconomic status and literacy challenges often limit patient access to healthcare in general, and trial opportunities in particular. Persistent diversity issues during trial enrollment result in serious gaps in the knowledge base, leaving physicians to extrapolate the RCT findings at the point of care. “Over time, these disparities often lead to lower-than-expected clinical efficacy, uptake and adherence, and ultimately worse outcomes in the real world once new drugs and medical products launch,” says Galaznik.

Data-driven studies can help fill in the missing pieces and greatly expand the findings in the trial to make them more relevant—at larger scale—to the heterogeneous patient subgroups in the real world, says Dan Riskin, MD, CEO of Verantos. Two types of RWE studies are particularly useful to address persistent diversity challenges in RCTs:

These compare the clinical effectiveness of several therapies, in order to identify which performs better within specific patient subgroups. The findings can help to differentiate a therapy’s performance in a crowded therapeutic space.

These involve studies of how a given therapy performs within a targeted patient subgroup, to develop new clinical and safety insights that can inform clinical care and address important knowledge gaps.

“If investigators are able to compellingly show that certain subgroups fare better on their therapy (using high-validity subgroup analytics and comparative effectiveness studies), they can help to improve the standard of care for prescribers, improve clinical outcomes for patients, and help drug companies to differentiate their products in ways that can help to increase market share,” says Riskin.

To that end, it is critical for pharmaceutical companies to think through all of the opportunities—both clinical and financial—as they are designing their trials, and not focus only on the immediate outcome of the study. “Sometimes we see a trial protocol call for blood to be drawn to analyze for the presence or absence of one or more biomarkers—not because such information is required for existing treatment, but rather with the hope of being able to look back and mine that data later to identify unanticipated clinical correlations,” says Ikeguchi.

“Drug developers are increasingly having to generate more evidence, in addition to the RCT. This is particularly notable in oncology, for example, where standards of care are rapidly evolving, or where today’s relapsed patients are different from yesterday,” adds Akiko Shimamura, senior director of product, medical analytics and outcomes research, Acorn AI by Medidata.

Today, large commercial companies pool and curate data from parallel sources, so that relevant data elements can be mined using appropriate eligibility criteria. In principal, that sounds simple; in practice, it is not.

Adding to the challenge is the fact that RWD typically exists in two formats—structured and unstructured. Structured

 data, such as that claims and billing data found in a payer or pharmacy database, and clinical data captured in an EHR system, is typically gathered and managed in a highly formatted structure. Such organizational logic makes it relatively easy to access and curate the information using data-mining techniques.

However, due to rigid coding requirements used to impose uniformity during data entry, the information contained in structured claims and billing databases may not adequately express the types of clinical details investigators need. “From a study standpoint for drug developers, the richer clinical details are those found in the form of unstructured physician notes that are housed within the EHR system,” says Schwartz.

While the type of unstructured information found in EHR systems and other sources does not fit neatly into highly formatted rows and columns, it provides a gold mine of clinical information to be prospected. However, the lack of structure, coupled with other confounding factors such as the use of medical jargon, shorthand, abbreviations and misspellings, has traditionally consigned unstructured data to the margins of the big data universe.

Fortunately, in recent years, ongoing advances in NLP, a subset of data analytics, has helped to address this challenge. “NLP techniques try to evaluate the vast amount of clinical information that resides in physician notes related to millions of patients and extract the most relevant clinical details to enable deeper clinical insights,” says Schwartz. NLP is also being used to automate literature reviews and glean clinical insights from notes collected by nursing call centers and other telemedicine encounters, databases of patient-reported outcomes and more.

“Our NLP processes, initially trained on large publicly available datasets and then finetuned on our specific dataset, are able to outperform humans,” says Adam Petranovich, chief data scientist, Prognos Health.

For instance, given an arbitrary anatomical pathology or molecular diagnosis report, “this process is able to automatically infer its interpretation, site, subsite and other meta information to be used or discovered by other systems,” he points out.

The use of a placebo-controlled clinical study arm is a necessary yet controversial aspect of RCT design, driving up trial costs while undermining recruitment efforts.

In certain therapeutic catagories, particularly oncology, the notion of randomly assigning some of the patient volunteers into a placebo-controlled study arm and thus squandering precious time is considered unethical. In other cases, such as those targeting rare diseases, it can be hard to find enough patients to populate the study and control arms. And while single-arm, open-label trials, where all patients are treated with the therapy being studied, can yield relevant findings, they provide no option for comparison to a placebo or to the standard of care.

To address these challenges, the use of synthetic control arms (a term trademarked by Medidata, which is also called an external control arm) is on the rise. Such studies give today’s advanced data-analytics methodologies based on big data a chance to shine.

Synthetic control arms are small, standalone studies that are conducted using de-identified, population-level data mined from various sources of RWD. They enable the use of pooled clinical trial data “that has rich covariates typically not available in the real world, and allow for side-by-side comparisons of the RCT population with the real-world population, to contextualize the trial and demonstrate relevance of the trial findings with the payer population,” explains Shimamura.

These external control arms, based on data analytics, serve as proxies for traditional control arms and can help to elucidate and expand the findings of the current trial and allow for head-to-head comparisons of the investigational drug versus the standard of care—while lowering the costs and reducing both patient burden and ethical considerations.

“The application of synthetic controls is not a new phenomenon, though it is certainly on the rise, especially in the formal regulatory approval context,” says Casciano.

 The traditional contracting model between manufacturers and payers, typically defined by rebate payment levels based on volume, “is quickly becoming outdated and unsustainable,” says Galaznik. By contrast, so-called value-based contracts (VBCs), which are sometimes called outcomes-based agreements (OBAs), strive to tie reimbursement rates to demonstrated clinical outcomes.

Such data-driven contracts are designed to compensate drugmakers—with more favorable pricing or reimbursement terms or formulary positioning—when specific agreed-upon patient outcomes or other clinical measures are achieved during real-world use of the therapy over time and are expected to grow over time.

Using the latest data-analytics efforts, “drug sponsors can and should use their findings to demonstrate both the clinical and economic value of their new medications using data-driven studies, and communicate those opportunities to payers,” says Ikeguchi.

But the stakes are higher than ever. “We used to be asked to provide ‘regulatory-grade data’ for regulatory submissions. That is now being categorized as ‘research-grade data,’ which supports high-validity evidence to make clinical assertions,” notes Riskin. “The term research-grade implies use by not only regulators, but also by payers and clinicians, to inform their decision-making with regard to drug pricing, reimbursement, formulary placement and care pathways.”

Meanwhile, for VBCs to work from an operational standpoint, drug companies and payers will need to create robust systems that can do two things—provide longitudinal tracking of outcomes over time among the insured patients and adjudicate the financial terms to allow for reimbursement per the contract times. Over time, industry experts expect to see the use of blockchain technology —which can create “one single version of the truth” in terms of tracking clinical factors and other performance metrics with complete traceability—to become the technical backbone that will enable such agreements to flourish at scale.

“As an industry, we have to do a better job of improving our models, and, importantly, validating them and communicating our successes, and we must continue to advance common standards for data collection, aggregation and sampling methods to assure representativeness and quality of the data underneath our most sophisticated analyses,” adds Casciano. “There has certainly been movement in that direction, and this trend needs to continue.”  

1. Hwang T.J., et al., Failure of investigational drugs in late-stage clinical development and publication of trial results. 

. Med. 2016;176:1826–1833; https://pubmed.ncbi.nlm.nih.gov/27723879/

2. Moore T., et al.. Estimated costs of pivotal trials for novel therapeutic agents approved the US Food and Drug Administration, 2015-2016 [published online September 24, 2018]. J

; https://pubmed.ncbi.nlm.nih.gov/30264133/

Articles

How analytics-driven insights can help drugmakers navigate the leap from trial to market—and add clinical and financial value at every step

Advanced analytics: Powering pharma’s most critical data mission

“Encourage your kids to be data scientists,” Jane Quigley, senior VP, digital health, at PRA Health Sciences, offered. Considered one of the “godmothers of real-world data (RWD),” Quigley was underscoring the need today—and great opportunity amid the convergence of technology and accessibility—to surface valuable business- and patient-impacting insights in areas such as clinical trial design, operational execution, regulatory and payer discussions, and asset differentiation.

“It’s about keeping up with both the information and the ability to do deep analytics on it; not just analyzing it with a spreadsheet,” notes Quigley, who leads a team of 300 people globally at PRA; her group, tapping into advanced analytics and programming languages, such as Python and R, focuses on using patient- and physician-level data to support pharma sponsors in disease modeling, RWD/real-world evidence (RWE), and drug forecasting.

“We have petabytes of data [today],” she points out. “A large size of real-world data is good, but if you can’t put context to it, then it’s not as valuable.”

aught up with Quigley recently to discuss the progress and challenges in achieving this dynamic in the life sciences and public health settings, its importance amid shifts in engagement and care delivery during the Covid-19 pandemic, and her career transition from academic medicine to the world of medical informatics.

What do you believe are the biggest opportunities—or where the most untapped potential exists—for biopharam companies and their CRO partners to take advantage of RWD/RWE? 

I think RWE is having its moment right now, because of the virtuality and the nature of how healthcare is being delivered today. It’s not necessarily new what we’re doing, as far as it relates to things like clinical trial protocol evaluations, or informing the overall product lifecycle. But it’s about using the data, which represents in the US about 300 million people, and representing it in a regulatory-grade fashion. When you think about the information that we’ve collected traditionally in clinical trials, that data is coming directly from the same sources that a lot of the RWE data is coming from—[medical] claims data, electronic medical records (EMRs), etc.

The interesting piece in all this, is a drug is only effective if a patient takes it. And if we’re basing that safety and efficacy off of an N of, say, 150 or 180 people in an oncology trial, versus a big mega cardiovascular trial, are you having the right representation of people by age, race, gender, ethnicity?

It’s about exposing our pharma partners and biotech clients to the value that real-world data can bring to help them inform where they should actually be adjusting or pivoting.

Drug and vaccine trials for Covid-19 have particularly spotlighted these patient representation issues. Are there sources of RWD emerging that can glean this kind of information to aid in protocol design?

Right now, when you look at the incidence of Covid outside of nursing homes—we run a Covid tracker that updates weekly—the number of people that are non-white, who are actually dying, is almost three times that of somebody who is white. So how do we have such a huge delta and disparity there? When you start digging into the data, and layer in all of these other kind of attributes, it really gives you an identity to the rest of the world.

There’s actually a lot of new streams and sources that are coming in. We pull in quite a bit from the social and digital space. People who self-identify as white or non-white, for example, we can incorporate that data in and add it as a layer on top of the true clinical real-world data. We also layer on the census data.

One of the things that Covid has exposed is the hard truth about the health inequities that occur in the US particularly. The uninsured and underinsured populations have an outsized number of non-white people. And when you look at the mortality rates and differences [for Covid-19], it’s staggering. With these disparities in healthcare, if you aren’t incorporating in the social determinants as a major factor in how you’re guiding your R&D program, it will be difficult to enroll. That mindset needs to be there, because the setting of care has changed.

Overall, is the synergy between healthcare data and clinical data becoming more seamless today in helping companies make practical business decisions, such as go/no-go determinations for clinical trials?

Yes. Understanding the delivery of healthcare has probably been one of the biggest under representations in the R&D space. You had a lot of bench and traditional scientists, who were looking at one molecule to see its utility across a single target—inflammation, for example. And then it narrows down to, say, rheumatoid arthritis, and it may look like it has some of the same signaling that would be in Crohn’s [disease], for example.

But when you think about that molecule now, when you understand how healthcare is transacted—and that would be claims data and how it’s billed and paid for—the molecule has to incorporate all those value-based care elements. We’ve always installed metrics within the trial itself for pure clinical outcome, but you have to incorporate them into the overall lifecycle now, because if what you’re representing in your trial has no realistic representation in the real world, it’s going to be difficult to get off the ground.

With health technology assessment bodies in the US and Europe expanding their reliance on RWE, do you see this area inching more to “real-world applications”—that is in the decisions made routinely by payers to back treatments and physicians to prescribe them?

Yes and no. Europe has always placed a great deal of value on RWE, due to the nature of a nationalized healthcare. If you have two very expensive medicines—monoclonal antibodies, for example—they’re going to go with the one that has the best price that they can achieve, and the one that’s going to get consistent outcomes.

Here in the States, because we have a different healthcare delivery system, if you have two similar agents that are both exerting basically the same effect—a good example is the tyrosine kinase inhibitors in CML (chronic myelogenous leukemia); when the first landmark drug was launched, it became one of the first billion-dollar products in oncology, but then similar treatments came out—then the payers could suggest the first-line on the second-generation drug because there is enough evidence to support it, and likely receive a better price for it.

RWE is here today as it relates to the commercial use, but can be stronger as it relates to the R&D. That’s due, I believe, to the underlying nature still in our drug development process, our molecule-to-market—and that a lot of agents may be in the lab for five to 10 years before they ever get to a trial patient.

How has RWE influenced the decisions of payers and drug pricing watchdogs when assessing the gene and cell therapies that have emerged in recent years? Is there too little of that type of real-world and observational data available yet to apply substantial insights?

That’s actually one of the best scenarios to use the real-world data—these heavy-hitting CAR-T and other immunotherapies. In fact, in a clinical trial for a pediatric B-cell acute lymphoblastic leukemia treatment, the control arm—the observational arm—was real-world data. A rare disease and a very rare population, they knew they couldn’t do a placebo-controlled trial. So they used an external control arm just using the real-world data.

For gene therapies especially, people are getting treated with them as a child. It’s about being able to longitudinally find these patients over time in the real world; where is somebody if they got treated at age 7 and in 15 years, they’ll be 22? That RWE mechanism to be able to look backwards in time to find these patients and reach out to their current physician to get a patient-reported outcome, is critical. And it’s also a regulatory requirement.

Does the need remain to educate payers more about these types of data sources to increase their comfort in using them in coverage determinations? 

Payers have always had a great deal of value of what the real-world data is bringing to them. But, typically, they couldn’t compare themselves to other insurers. So by democratizing real-world data, payers can get a better sense of where they fall; what did their patient populations look like, for example? They can get a better understanding of the behaviors of the folks that are in their enrollment.

They’ve always had tons of actuaries and data scientists. What they’ve had to do is groom them to ask some different questions. It’s shifting away from the pricing and utilization review sorts of metrics to more toward, “what are our patients looking like? How often are people going to the physician or the nurse practitioner? Do you have a subpopulation that’s driving x-percent of your costs? If so, why?” That’s when you start getting into the use of disease management programs and similar things to help people better manage their healthcare journey.

Are there still system constraints, at times, for payers in collecting and acting on RWE from sources such as claims and EMRs? 

The thing about an electronic medical record is that no data field is mandatory. With claims data, every data field is mandatory. Because what is it answering? One, claims data, is answering the finances and what the patient was billed for and what the physician billed for the nurse practitioner, etc. The other one, EMR, is more clinical, whether it’s laboratory, tumor staging, or more about the phenotype of that patient.

The marriage of them both is really your richest source, and that ability to link multiple different data sources. Whether it’s opt-ins plus disease registries, for example. Tying those all together gives you that full picture of the patient, and it can typically give it to you longitudinally over time.

8. With significant disruptions to clinical trials due to Covid, how has RWE helped keep some development projects on track?

We’ve been tracking Covid from day one. We’re providing that data to both pharma and government clients to be a good corporate steward. But the other piece to it is we’re trying to use that data to better understand the disease. We know some people with Covid can get very sick, very quickly. So if we can create a group of people that we know have tested positive, but aren’t requiring hospitalization, and thus ventilation, we can use them as a control arm.

Then what we’ve been doing is pulling them forward in time to uncover things like their time to recovery—what did they need after that? Did they go on steroid drugs or not? So by mining the data we have from the real world, especially from the first wave of the virus, we’ve been able to model a lot of clinical trials in relation.

9. For Covid-specific trials or studies in general? 

Actually for everything. It’s really about turning all the levers on. With Covid, it was impacting enrollment across all our trials, because the clinics were closed. People were still going into, for example, the oncologist office to get their chemotherapy. But the whole nature of the visit changed. You went in, you were spaced apart, your time changed, everyone with a mask, gloves. There was never an opportunity for a physician to say, would you 

be interested in a clinical trial? We had to look at all of that data as it related to Covid impacting the entire healthcare space, not just people diagnosed with the virus. People weren’t going to the doctor, so all of those were missed opportunities for any sort of discussion about a trial enrollment. And we also had to pivot on things like, what does that conversation look like in a remote visit?

We’ve had to be really creative, because you have to access people in a completely different fashion now; that’s why we’ve exploring looking at in-home assistance, for example.

And, most importantly, you need to have all of the appropriate mechanisms in place around privacy, security, and compliance. If you don’t, as it relates to people receiving care traditionally in the clinic, you’re years behind. You need to have the tools to be able to both digest and ingest information as it’s coming in, in real time.

10. You began your career actually in academic medicine, and were formerly the clinical nurse specialist for hematology-oncology at Georgetown Lombardi Comprehensive Cancer Center, where you also directed the bone marrow transplantation program. What drove your switch to the data insights and analytics side?  

At Georgetown, we had a very small number of people coming into the clinic that were on a trial; everybody else wasn’t on a trial, but we weren’t collecting data. At the time, I was working with a postdoc, and I said, “let’s create a database.” Sure enough, mining through that data, it actually helped to support a client’s safety initiative. So I kind of evolved into the data and technology and consulting space.

I went on to create a company that did disease management, and established an oncology program. We sold that to Optum. I was part of the team with [IntrinsiQ LLC, later acquired by AmerisourceBergen] that created the first oncology EMR [in collaboration with GE Healthcare]. Then I got into the heavy-hitting analytics and patient-level data, first with a company (SDI) that would be sold to IMS Health (where Quigley would spend three years before joining PRA in 2014).

If I stayed in academic medicine, I would see probably 700 or a thousand patients a year. Right now, I see 300 million people in the US every day.

A conversation with Jane Quigley, senior VP, Digital Health, PRA Health Sciences

Clues and connections from the real world

Driverless cars and doctorless diagnosticians must work flawlessly every time. Likewise, their users must be confident that such artificial intelligence (AI)-driven innovations will work as expected. Adoption of these technologies—and, hence, their tremendous benefits—hinges on trust. It is a catch-22.

As AI continues to show great promise for commercial life sciences, trusting its suggestions is paramount to widespread adoption and growth. In fact, as AI evolves into the more advanced, contextual version that is emerging today, its predictions are becoming increasingly pivotal to treatment decisions. Sales reps, brand teams and medical science liaisons (MSLs) are better equipped than ever to deliver the right information to healthcare providers (HCPs) at the right time-—for instance, highlighting a new therapy with fewer side effects to a recently diagnosed patient. The problem is most users have little visibility into how AI-powered systems make the decisions they do, and this lack of transparency inevitably casts doubt. Many of the algorithms used for machine learning (ML)—particularly the popular deep learning approaches in use—do not transparently convey how and why a decision has been made.

The lack of rationale hampers our ability to let go of misgivings and accept the accuracy of machines whose mechanics we cannot easily see or comprehend. As such, human beings want computer systems to produce 

transparent explanations for their decisions—a form of science known as explainable AI (xAI). Fortunately, the most advanced AI solutions in commercial life sciences now incorporate xAI with a blend of other analytics technologies and human insight to improve decision-making by building real-life context and plain-language explanations into recommendations.

As trust between man and machine builds, AI 2.0 will become foundational to deepening HCP engagement in today’s digitally driven commercial model. Here is a look at how next-generation AI builds trust between sophisticated systems and the users they support.

Complexity demands more humanity in modern AI

The life sciences industry is incredibly complex. There are more channels, data, regulations, stakeholders and pricing pressures than nearly any other industry--—making trust both hard-earned and simultaneously worth its weight in gold.

Companies across many industries have become increasingly comfortable leveraging the insights provided by ML algorithms to aid decision-making. For example, GE Power is using advanced analytics and ML to enable predictive maintenance, operational efficiencies and business optimization to move toward its vision of a digital power plant. American Express, which processes $1 trillion in transactions and has 110 million credit cards in operation, relies on ML to help detect fraud in near-real-time, saving millions of dollars in losses. AI is even helping the farming industry, as John Deere uses ML Farmsight system to guide decisions on things like pest-control or what crop to plant when.

While exciting, these AI examples do not have as serious consequences if AI gets it wrong. For the life sciences industry, an errant recommendation could inadvertently prevent potentially life-changing information from getting into the hands of doctors.

In addition to commercial AI’s potential weighty impact, AI in the life sciences industry is particularly complex. Consider a comparison to online retail. Amazon offers personalized recommendations based on browsing and shopping history. Most data reside within Amazon and make very direct connections—for instance, you shopped for this, so you may also like this. In the life sciences, however, there are many more variables to consider, such as the variety and location of relevant data sets, varying HCP interactions across therapeutic areas, data privacy requirements and multi-region brand strategies.

An environment with so many twists and turns—made even more challenging by the COVID-19 pandemic—requires an evolved form of AI if users are going to embrace its full potential and overcome go-to-market complexities. AI 2.0 must feel more human to garner confidence. By incorporating specific feedback from field and brand teams as well as MSLs, AI 2.0 incorporates context from real-world situations to produce recommendations that feel like suggestions from an experienced peer.

“We now see the power of AI, but my early experiences with AI at previous companies were not as successful. Users were hesitant to trust the recommendations as they seemed synthetic,” explained Joel VanderMeulen, senior director of commercial strategy and operations at EMD Serono, a Merck KGaA company. “AI 2.0 combines different analytics technologies like geospatial targeting that knows where a rep will be and provides pertinent suggestions at the right time, so users don’t feel like they are interacting with a robot. It brings the machine to life for a more natural experience they can trust.”

Context ensures hard-earned trust is not easily lost

AI 2.0 leverages the best of both humans, who are good at making decisions in complicated situations, and machines, which excel at data processing but can struggle to make basic judgments, to create more meaningful relationships with HCPs. Early commercial AI offerings were hampered by the one-note nature of their approach. Most solutions fell neatly into one of two buckets: ML or expert systems (an if-then, rules-based approach). Neither technology, by itself, does a great job replicating the way humans really think. Without necessary context, ML technologies can generate conclusions that are not practical in the real world—causing users to doubt their validity. At the same time, it would be impossible to catalog every possible scenario a user might encounter with rules alone—resulting in less accurate conclusions.

Modern AI combines advanced analytics technologies—business logic, ML, explainable AI and optimization—plus human intelligence to transform nascent forms of AI into a trustworthy assistant. Its unique blend of business rules and ML creates real-world guardrails to prevent recommendations that contradict commonsense.

Further, third-party data sets and the insights captured in a company’s customer relationship management application, marketing automation platform and other tools provide a foundation for understanding how HCPs typically interact with a brand.

Ultimately, it is the overlay of human intervention that ensures recommendations are rational and, therefore, more likely to be trusted by end users.

AI yields optimal results when every interaction and data point are analyzed together, which is possible with the right blend of analytics technologies. However, explainable AI (xAI) may be the most important analytics technology of all when it comes to building trust from its users.

Explainable AI is a key part of AI 2.0 that is helping life sciences companies drive greater adoption among MSLs, brand teams and sales reps. It is an emerging field in ML that tries to address how decisions are made by looking at the steps involved in decision-making. It hinges on clear articulation of how ML generates their outputs. For example, xAI answers questions like: Why did the AI system make a specific suggestion? Why didn’t it suggest something else instead? It explains how conclusions were reached to improve understanding.

In the life sciences industry, xAI might recommend a sales rep send an email instead of making a virtual visit, supported by the reason that this HCP is most responsive via email. In another example, a suggestion to an MSL to schedule a meeting with an HCP today might come with an explanation that says the reason for the urgency is because the doctor has not been engaged with in 120 days.

Thanks to xAI, ML is no longer a mysterious, black box. It provides transparent explanations for every recommendation to foster enough confidence, so end users trust that AI is making good decisions. It is crucial for continued AI deployment in life sciences and to build upon today’s successes for a more intelligent future.

“Field reps and MSLs who have been interacting with their customers for years tenaciously believe in their own intuition,” said VanderMeulen. “It can be hard to convince them to do something that goes against the grain or change their minds if, say, an AI engine suggests that they call on a customer they don’t think is a good target. It’s a matter of trust, but once reps see the value of AI, they believe.”

AI that evolves evokes ongoing user confidence

Development cycles may be long and predictable in the life sciences, but commercial market factors are not. They are dynamic and unpredictable with competitive threats, regulatory changes and ever-changing HCP preferences. The pandemic has piled on more challenges, as digital outreach from all customer-facing groups, including sales reps, nurse educators, key account managers and MSLs, has dramatically increased, demanding companies find unique ways to cut through the noise. In fact, 58% of HCPs surveyed in August 2020 said they have been spammed by a pharmaceutical company—receiving high volumes of digital content that misses the mark.

AI 2.0 can help—if commercial teams learn to trust that today’s AI systems are sophisticated enough to navigate dynamic markets. Today, AI 2.0 is no longer making educated guesses because it is constantly learning. As campaign elements succeed or fail, next-generation AI incorporates those learnings to continuously optimize execution. AI that can apply what it learns from one environment to the next is the difference between AI that “thinks” like a human and AI that performs ROTE calculations like a machine.

Additionally, next-generation AI makes multi-faceted assessments at the individual level, analyzing every HCP characteristic—from specialty and patient coverage data to when they graduated medical school—against a large dataset of HCPs to determine communication preferences and likelihood to engage on a one-to-one level. AI 2.0 not only knows when an HCP has historically accepted visits or opened emails, but also predicts the optimal time to engage when a new channel is added to the mix.

“AI innovation that incorporates human context is incredibly valuable,” added VanderMeulen. “Combining primary data, such as unstructured call notes from field teams, with secondary data, such as script writing and other data from market research companies, is the next level. It’s extremely exciting—and, equally as important, it helps to foster trust in the accuracy of its recommendations.”

AI will transform the life sciences commercial model

Physicians’ increasing expectations for personalization coupled with decreasing access have driven a slow-drip of changes to the commercial model in recent years, but the ultimate catalyst has been Covid-19. As one colleague put it, the future just showed up. Even if it is not clear what HCP engagement in a post-pandemic world will look like, the need for commercial transformation could not be more apparent. If AI 2.0 paves the path to transformation, trust in AI ensures we walk it.

Fortunately, the groundwork is well underway. After years of steady, incremental progress, most life sciences commercial teams now have a strong data and analytics infrastructure in place and are already turning on the right digital and non-personal communication channels at scale. Users, such as sales reps at EMD Serono, are seeing the value of AI and quickly building confidence in its recommendations.

As the industry navigates today’s increasingly complex and high-pressure environment, a contextual AI approach ensures that the information HCPs receive is tuned to their specific needs and preferences. AI 2.0’s blend of analytics technologies and human insight digitizes, analyzes and incorporates every piece of data and feedback from the field—from prescribing habits and patient demographics to research interests and event attendance-—to personalize the customer experience. Now, with xAI, users can finally trust its output to expand AI’s adoption across the organization and industry.

1. Accenture, Research Report, “Is Covid-19 Altering How Pharma Engages with Reps?” [Published online August 4, 2020]

Embracing AI 2.0 and modern machine learning as foundational tools to deepening HCP engagement

Next-gen AI: Widespread adoption hinges on trust

A recent report by Deloitte projects that the artificial intelligence (AI) drug discovery market will grow 53% by 2025, citing the increased investment of  labs in AI as a way to boost research but also to collect, analyze, and produce real-time, accurate data. Of course, such information, including when also applied to the health outcomes side to demonstrate drug value—whether we call it real-world evidence, real-world data, predictive modeling/forecasting, etc.—has been, at different levels, part of broader pharma practice for several years now, as the industry, some might say, begrudgingly compelled itself to get on the digital health train.

But pharma and its partners have taken advantage, as we examine in our 

, focusing on strategies to build and sustain. Yes, the industry has its long-rooted regulations and risk assessments and compliance and data-protection measures, that in the delicate world of medicine, you can’t as easily adapt to next-gen trends and technology as other enterprises; but who can really ignore the strides made today in advanced analytics and digital health? Tools such as natural language processing, data and text mining, data visualization, wearables, system checkers, and so on—and the opportunities they offer in clinical operations, patient and physician engagement and support, and also in health economics and the market access/reimbursement realm.

Which brings us to Covid-19 and how this whole notion of data integration has been reframed. Edward Hensley of AssistRx points out in his 

 this month that Covid sparked the acceleration of technological change and adoption, as daily routines were altered. A change he says was already afoot but now is requiring, out of necessity, the mass use of existing digital tools. A survey by Optum supports this, finding that 56% of healthcare executives said their companies accelerated AI plans in response to Covid. Generally, however, it seems progress remains in embracing AI. In a survey by MMG, doctors gave a score of 4.65 out of 10 to the level of development of AI; pharma/healthcare executives rate it at 4.79, and medical students at 5.11.

Perhaps acceptance simply comes down to trust, as Aktana’s Derek Choy 

 in our December coverage, outlining ways he says AI 2.0, or “explainable AI,” can actually build trust between user and machine. He argues that the complexities of the life sciences “demands more humanity in modern AI.” Doesn’t sound too unreasonable to consider a little bit of “new” in these not-normal times.

Mike Christel is Editorial Director of Pharmaceutical Commerce. He can be reached at 

Worthy of further analysis

Since March, when the majority of the US was placed on lockdown and began limiting in-person interactions, the healthcare and pharmaceutical industry has learned much about prioritizing patients and healthcare providers (HCPs). Meeting these populations’ needs, especially in times of crisis, can significantly impact their lives, building brand loyalty long after the crisis has ended. Covid-19 served as a catalyst to accelerate the technological change and adoption that was already happening, as everyday routines were altered indefinitely.

First amplified by the need to overcome the unique challenges of remote working environments, tech-enabled solutions facilitate meaningful human contact and reduce time-intensive administrative tasks so HCPs can focus on the actions that will most improve patient lives. For pharmaceutical manufacturers and life sciences companies, striking the balance between technologically enhanced service for patients/HCPs and minimizing the impact on the bottom line, both now and in the future, is top of mind.

Telehealth: Evolving patient, payer, and HCP preferences

At the start of the pandemic, telehealth services exploded, with claim lines increasing more than 3,000% nationally from July 2019 to July 2020, according to data from 

FAIR Health’s Monthly Telehealth Regional Tracke

r. As stay-at-home restrictions loosened, utilization decreased, but telehealth options remain attractive to patients and HCPs. Telehealth significantly decreased no-shows, 

eliminating the barrier to access of transportation

 that HCPs hadn’t had access to since the days of physician house calls. Patients with pre-existing conditions or compromised immune systems that make them particularly susceptible to Covid-19 continue to use telehealth to interact with their doctors.

However, major insurance payers are rolling back the terms of telehealth coverage they launched this spring. According to a report in 

, HCPs face challenges in managing the wide variety of rules by payers, which have varying deadlines for reimbursement changes. This may change again though, with some predicting an uptick in usage as Covid-19 cases resurge in winter months. Even after the pandemic, 

 and demand for digital healthcare is likely to drive changes across the industry.

Similarly, certain HCPs will lean more heavily on telehealth than others post-pandemic. Pharmaceutical companies will need to make long-term investments in tech solutions to continue to serve patients and HCPs virtually. Offices that elect to have at least some of their staff work remotely will need access to solutions that don’t require in-office fax machines or non-cloud-based electronic medical records (EMRs).

In addition to adopting telehealth as a result of reduced in-person contact, HCPs are limiting and may continue to limit in-person visitors who are not patients. This means, in addition to a patient’s family or support system no longer being able to accompany them to appointments, pharmaceutical manufacturer field teams are also no longer able to enter healthcare offices.

Your patient solutions provider should be able to support you in the new, virtual normal by providing virtual education materials and access to resources and tools available online. By looking at their field team roles through a virtual lens, manufacturers can support them with more tools and ways to address workflow challenges. Leveraging the right suite of tools, field teams can continue to provide value for your brand, including coordinating product distribution, streamlining the missing-information process, providing real-time alerts and guidance for prior authorization requirements, decisions and follow ups, as well as guiding the patient journey.

In addition to the pandemic’s societal and industry impact, it has had a significant impact on the economy, with many patients experiencing lapses in insurance due to furlough or job loss. Therefore, more patients qualify for patient assistance programs (PAPs), which can strain a manufacturer’s ability to keep up with the need. A knee-jerk reaction some manufacturers may have is to narrow the number of patients who are eligible for free drug programs. However, turning away patients in-need could hurt your brand’s reputation.

To accommodate an increase in PAP-eligible patients, support programs should leverage more tech-enabled solutions to balance the necessity of hiring additional of full-time employees. These solutions automate processes to eliminate back-and-forth phone calls, allowing technology to drive processes and people to focus on the tasks that drive the most value for patients.

Moreover, life sciences organizations need to design PAPs that deliver exceptional experiences at every touchpoint, whether interacting with the program for the first time or completing reverification. Product, patient, provider and payer profiles should all be taken into consideration when deciding how to integrate tech-enabled solutions into PAP design. For example, increasing patient and provider self-service and automated processes upfront can improve the patient solutions team’s efficiency.

However, life sciences and healthcare organizations need to take into account variables like the tech-savviness of their patient and provider populations, as well as the electronic capability of the product’s payer mix to determine whether a mostly talent-enabled PAP would deliver the best experiences for providers and patients. 

Tech in the new normal: Driving cost savings and better outcomes

Though the healthcare industry has historically been slow to adopt new technologies, HCPs and manufacturers alike need to utilize technology to enable processes, drive cost savings, and deliver better outcomes for patients with more obstacles in place than ever before.

The healthcare technologies making the biggest impact on the patient and provider experience and elevating manufacturer brands are:

– Solutions with limited connectivity, such as fax machines, and EMRs not connected to the cloud, have been especially limiting in the months since social distancing, remote work, and telehealth appointments became the norm. By  leveraging secure, cloud-connected tools, prescribers and office staff can limit their time in the office while still providing timely services for their patients.

– These self-serve tools can provide critical education and training for HCPs and patients that meet them where they are, at their convenience. e-Support solutions, such as e-consent or e-enrollment, can be embedded in these websites for ease of use for both patients and HCPs.

– Automated electronic solutions and technologies for submitting prior authorizations, enrolling patients into programs and checking patient benefits save the most time for HCPs and patients. When integrated into patient support program customer relationship management sysyems (CRMs), patient solutions and hub teams streamline their processes as well, resulting in improved speed to therapy and reduced program costs, as well as alleviated administrative burden.

– Field sales and reimbursement office visits are no longer feasible. Providers and patients seek virtual support for their reimbursement and clinical programs via live chat, phone, or video. For programs offering nursing services, such as in-office or in-home injection training, pharmaceutical manufacturers should be able to offer these virtually through video conferencing.

– Tech-enabled solutions can identify patient drop-off points and opportunities to streamline PA and appeals processes to get more patients on their prescribed therapies.

Many of these healthcare technologies have been around for some time, with mass adoption the only barrier to a transformed industry landscape. The changes Covid-19 has set in motion are unlikely to completely disappear, making a significant impact on how we conduct business in the industry, as well as in our personal lives. These long-standing technologies, now adopted en masse out of necessity, will make the difference between a program that meets patient and HCP expectations and programs unable to adapt to changing preferences.

Edward Hensley is chief commercial officer for AssistRx  

Using technology to enable processes, drive cost savings, and deliver better patient outcomes

Elevating brand outreach in the age of Covid

The quest to simulate human intelligence in machines has introduced numerous industries to the power of artificial intelligence (AI). In the healthcare and pharmaceutical settings, the past few years have witnessed the advancement and adoption of AI platforms, most notably, chat and voice solutions, which improve patient/physician engagement, streamline access to resources and promote patient self-efficacy throughout the care journey. Other integral AI-powered resources include electronic health records (EHRs), wearable devices such as Fitbit, symptom checkers and more.

This remodeling has triggered a high demand (and even an expectation) for chat and voice solutions beyond patient and provider engagements. Today, 69% of consumers prefer to use conversational AI due to the quick communication it offers. Over the next three years, a projected 70% of consumers will replace visits to the dealer, store and bank with voice assistants. In healthcare, 51.9% of consumers are interested in using voice assistants. 2020’s unprecedented convergence of political, economic and regulatory forces has triggered wide-spread and progressive conversations around tech adoption. The Covid-19 pandemic and subsequent recession have increased the necessity for innovative and cost-effective solutions, while still maintaining the security and privacy of health data.

Prior to the pandemic, the stringent HIPAA regulations and additional privacy requirements introduced in 2003 severely limited the exchange of health data to protect patients’ personal information. The easing of these regulations in response to Covid, compounded by growing patient and physician demand for virtual care and engagement, has cleared the path for chat and voice solutions—which have now demonstrated their ability to support patients in the provider and payer settings.

Unfortunately, pharma has yet to see any similar leeway from FDA, which has heightened the years-long debate: do the strict regulations hinder the creation of tomorrow’s solutions?

From clinical research and through to FDA post-market safety monitoring, it is mandated that pharma companies must report all adverse events (AEs) with strict fines and penalties if they fail to do so. Serious AEs must be reported within 24 hours and alert reports must be submitted within 15 calendar days. This has limited pharma’s adoption of AI and reduced chatbots to button-driven experiences.

Pharmacovigilance is expansive and requires broad communication with patients and healthcare providers (HCPs). Many pharma companies have teams of people monitoring the various points of communication to detect, confirm and escalate AEs. This is an expensive undertaking, and also faces heightened challenges to keep up with the rest of healthcare’s transition to adopting virtual solutions.

As the Covid crisis has evolved and reliance on chat and voice tools remains firmly entrenched, key players have entered the market to deliver solutions that use the power of AI and elevate pharma to meet patients and HCPs in this digital space, while adhering to FDA regulations. With the advancement of machine learning and stronger natural language processing algorithms, pharma can now address at least one part of these AE reporting requirements: detection. Enabled as a monitoring capability within chat and voice experiences, AE detection modules are an initial step for the pharma to enable immersive conversational experiences for patients and HCPs alike.

While there is a gradual acceptance of this movement and adoption of chat and voice solutions, the climate from every angle is becoming more progressive and favorable toward change. Innovative entities outside of the pharma realm can support these companies by introducing and augmenting chat and voice services to maintain the integrity of the pharmacovigilance process while adjusting to the major technology shifts that were accelerated by the changes brought on during 2020.

Elise Whitaker is VP of Customer Success, Practice Lead, Pharmaceuticals & Life Sciences, at Orbita. She would like to thank Hayden Meinero for his contributions in putting together this article.

Are regulations hindering pharma innovations?

AI ‘voice’ grows louder in 2020

 prescription drug products as approved for marketing, and there are approximately 

 active physicians in the US who can prescribe those drugs. Accounting for all of the pharmaceutical products that ultimately reach patients requires a massive effort, and this complexity directly impacts pharmaceutical manufacturers who contract with, and sell to, middlemen and prescriber organizations.

The key to untangling the web of complexity between drug companies, wholesalers and prescribers is effective master data management (MDM). For pharmaceutical companies, “master data” goes well beyond clean and up-to-date records of products, prices and customers. In order to effectively and profitably contract with prescribers, pharmaceutical manufacturers must maintain complicated sets of master data, including ever-changing group membership rosters. They must track group purchasing 

organizations (GPOs), integrated delivery networks (IDNs) and health systems and ascertain their eligibility across highly regulated government health insurance programs such as the 340B drug pricing program, Department of Defense (Tricare) and the Veteran’s Administration (VA).

Mismanaging master data can mean lost revenue for companies, incorrectly calculated rebates and government prices and even legal non-compliance. In other words, the consequences are significant and can directly affect the success of an organization.

Complexity drives the need for master data management

For the purposes of this article, master data management (MDM) means a true understanding of the parties a company is doing business with, what it sells and what price is charged. This “master data” information is also sometimes referred to as the “material master,” and there are also associated “masters” for identifying and tracking related data, including customer and pricing masters.

However, for pharmaceutical manufacturers, things are much more complicated, because basic customer or product information reflects only “ground-level,” or baseline, information required to manage complex prescriber and payment data.

This complexity is driven by the numerous different entities that purchase drugs, including independent and community pharmacies, large chains, hospitals, independent doctor’s offices and clinics, VA facilities, long-term care facilities and even military bases. And given the regulations that govern pharma companies, the way each of these discrete customers purchases drugs can be very different. As for government customers, regulations are even more strict and, therefore, complex.

When it comes to drug purchasing, larger entities obviously have more power, and, for that reason, many medical practices will join with others to obtain group and volume discounts. A common example is GPOs, which leverage the buying power of multiple entities to obtain better pricing, as well as relevant and legally-appropriate discounts. However, keeping track of which purchasers are eligible for what discounts, as well as ensuring they’re correctly enrolled in various discount programs, requires an incredible amount of tracking and management. Many pharma companies try— and fail—to manage their master data using manual methods like spreadsheets, and this critical business function often gets deprioritized.

In addition, manufacturers must keep track of wholesalers and their relationships with various provider organizations. For example, a long-term care facility might be working with a wholesaler that is unaware of the discounts that the facility is entitled to, or that wholesaler might be giving the facility an excessive and/or incorrect discount based on incorrect data.

Unfortunately, the onus is on the manufacturer to manage customer data and appropriate discounts, because there is little financial incentive for wholesalers to carefully track class of trade and other master data; their prime motivation is to make sales. Their lack of concern for the financial consequences that manufacturers face is because, in cases where manufacturers do not have strong MDM, these middlemen are made financially whole, whether they sell a given drug at full price or discount it and recover costs via chargeback.

Ultimately, pharma companies must deal more effectively with MDM. This is because mismanaging master data can lead to unhappy customers, as well as lost revenue, a result of incorrectly calculated rebates and government prices. In addition, failure to manage master data creates compliance risk and potential legal liability. What’s more, drug pricing has become a politically-charged issue, and the need for highly-scrutinized pharma companies to “get it right” has never been more important from a public relations standpoint.

Best practices and technology: Taming master data

Effective management of master data is critical to mitigating business risk, and, most important to manufacturers, to capturing maximum revenue. Given the complexity of the overall system of prescribers and payers, the most effective way to handle MDM and capture all of a company’s rightfully-earned revenue is by leveraging best practices and technology.

Comprehensive master data is far more than basic customer names, addresses and billing information. Proper master data is built on accurate and up-to-date visibility into products and what pricing and discount terms a provider organization should receive, taking into account factors such as relevant memberships and government regulations. Beyond a basic grasp of “what pills we sell,” firms must have visibility into the solutions they offer, as well as the critical details associated with those products, such as their size, shape, dimensions and weight. Comprehensive master data must also capture which regulatory processes those products have gone through, including their availability or lack thereof.

Go-to safeguards for manufacturers in avoiding mismanagement of their business-central records

Pharmaceutical Commerce - December 2020