Architects and Gardeners

Most leadership offsites I’ve attended have included some flavor of personality assessment – not so much to formally classify us, but rather to make the point that different people have different styles, and to emphasize that you can’t assume everyone you work with approaches the world the same way you do. 

In this spirit, I wanted to share a simple binary personality framework that resonated with me, even though I learned about it from a surprising source — several articles about Game of Thrones author George R. R. Martin (see here, here).

According to Martin, writers tend to be either architects or gardeners. 

Architects, he explains,

“plan everything ahead of time…They know how many rooms are going to be in the house, what kind of roof they’re going to have, where the wires are going to run, what kind of plumbing there’s going to be. They have the whole thing designed and blueprinted out before they even nail the first board up.”

Harry Potter author J.K. Rowling, for instance, has been described as “a great example of an architect.”

Martin says that gardeners, in contrast,

“dig a hole, drop in a seed and water it. They kind of know what seed it is. They know if [they] planted a fantasy seed or mystery seed or whatever. But as the plant comes up and they water it, they don’t know how many branches it’s going to have. They find out as it grows.”

Martin describes himself as “much more a gardener than an architect.”

While of course most writers — and most of us — have attributes of both architects and gardeners, these two archetypes can help explain some of the stylistic differences one encounters in biopharma and healthtech. 

For instance, large pharma companies seem to select quite strongly for architects, particularly in senior leaders. Corporations value architect-like approaches to problems, generally wanting to see a complete, comprehensive plan that has been agreed to by all stakeholders before getting to work. (Safi Bahcall has discussed this nicely in Loonshots – see here, here).

In contrast, I suspect many startups are heavily populated by gardeners, innovators who begin with a promising idea or thesis, and then are comfortable (provided that their funding holds….) building it out, seeing where it goes, making agile adjustments on the fly as needed. The image of Mark Zuckerberg founding The Facebook comes to mind (see here).

In practice, this distinction isn’t a binary matter of either/or. Succeeding at scale in complex fields like biomedical R&D ultimately requires both sets of skills.

From the startup world, consider the example of Incredible Health co-founder Iman Abuzeid, a physician, management-consultant, and MBA by training. In a fascinating March 2020 podcast interview on 20VC, the highly recommended podcast hosted by Harry Stebbins, Abuzeid describes an approach that seems to combine the two. 

After spending a year cultivating, perhaps gardener-like, a different sort of company (“a SaaS platform for small and medium healthcare business”), Dr. Abuzeid says her team discovered they “just could not get it to grow.”  

At this point, the startup seemed to enter full-bore architect mode, switching to an approach that would have seemed at home in any large pharma (or management consulting slide deck).

The company conducted an incredibly comprehensive analysis including “coming up with a hundred different ideas” then “applying multiple filters.” After narrowing these down to 10 ideas, they did extensive customer research on each of these to wind up with the company that is now Incredible Health, a hiring platform for nurses, with a private company valuation (as of August 2022) of $1.65 billion.

If Dr. Abuzeid exemplifies how architect qualities can be helpful in a startup, others highlight the value of gardeners in big corporations. Ed Catmull, former CEO of Pixar, makes the case for why large enterprises should support gardeners, particularly businesses working in domains where creativity is critical, like movies – and, presumably, coming up with imaginative new medicines.

Discussing Catmull’s book Creativity, Inc in Forbes in 2014, I wrote,

“Catmull emphasizes that most creative works, such as movies, are emergent phenomena, the result of a multiplicity of interacting forces into which we only have limited understanding.   Great projects aren’t born, but develop over time, and are inevitably associated with a vast number of false steps and dead-ends.”

I’ve also discuss the application of Catmull’s perspective to biopharma here, highlighting Catmull’s key insight that (as I distilled it) “Each project requires its own journey; each film is the result of an often-difficult, frequently chaotic, and largely unpredictable evolutionary process that (with considerable coaxing) reveals itself over time.”

Particularly in today’s difficult funding environment, I suspect at least some venture investors are placing more emphasis than usual on the sort of highly architected analysis and approach that Abuzeid and her team adopted. 

I continue to hope that large pharmas will eventually learn to recognize the value of emergent discoveries, and the gardeners who cultivate them, as Catmull has vividly described and passionately championed. However, I’d also caution young gardeners to think twice before staking their careers on this still distant possibility. For those who are determined, learning the language of architects can help. Inside large corporations, substantive architect/gardener partnerships can prove particularly productive and offer the organization access to the best of both mindsets.

The Elegance of Developmental Biology

Developmental biology is an intrinsically captivating field, focused on how a single cell gives rise to a complex, fully-formed organism. 

As Doug Melton and I wrote in Cell, back when I was a post-doc:

“As we study the life and lineage of a particular adult cell, we ask the same questions that a biographer asks of her subject: what were the critical decisions that defined the trajectory of this life, and when were they made? What was the contribution of neighbors, and what role was played by more distant influences? What was the role of chance? At what point was the final fate initially specified, and when was it ultimately sealed? In essence, we would like to understand the molecular biography of the cell.”

It turns out that a colleague of mine, Dr. Ben Stanger, a physician-scientist and gastroenterologist now at the University of Pennsylvania, and who was in Melton lab at the same time I was, has now written a book, From One Cell, about the history of developmental biology, and it’s absolutely captivating.

In a style that’s engaging but not self-absorbed, Dr. Stanger steps us through the fascinating history of the field. Perhaps because of his own laboratory experience, he has a particular appreciation for the technical difficulty of key experiments, a perspective that brings a sense of immediacy to his descriptions of the work and allows us to appreciate the magnitude of the achievements.

Even when he moves to more contemporary advances, like the discovery of induced pluripotent stem cells (iPSCs), his explanations are clear, simple without being simplistic. You’re left with a visceral appreciation for the beauty of development, the virtue of basic research, and the promise this field continues to offer for the understanding and treatment of disease. It’s a magnificent and inspiring read.

Admittedly, it is perhaps a less than ideal listen; I had purchased the audiobook, and while the professional audiobook narrator has a fine voice, he mispronounces many essential terms (like “Drosophila”). 

Hopefully, if Dr. Stanger writes a second edition, he will narrate the audiobook himself.

While we’re on the subject of audiobooks, two that are, in fact, narrated by their respective authors, and which I really enjoyed are:

• Bruce Springsteen’s autobiography, Born To Run. From start to finish, a gritty and gripping account of Springsteen’s rise to stardom. His success wasn’t inevitable. He worked incredibly hard, and when he did manage to catch a break, he fought fiercely (including acting like a bit of a control freak) to ensure these hard-won opportunities were leveraged fully. Like so many of The Boss’s songs, it’s an essential listen.
• Easy Money: Cryptocurrency, Casino Capitalism, and the Golden Age of Fraud by Ben McKenzie and Jacob Silverman. The OC’s Ben McKenzie (an econ major in college) and journalist Jacob Silverman were early skeptics of cryptocurrency. They describe – with conspicuous delight, and perhaps just a tad too much smugness, crypto’s meteoric rise, devasting fall, and many shady champions. An important cautionary tale – and a wild ride.

The Inspiring Immigrant Story of a Stanford AI Leader

While most a16z podcasts relating to AI tend to emphasize the technology, the latest episode of “Bio Eats World” features an inspiring discussion with Stanford AI pioneer Fei-Fei Li, and focuses mostly on her remarkable life story. (Her journey also features prominently, we learn, in her just-published book entitled “The Worlds I See.”)

Li is perhaps best known for her foundational work in computer vision, and specifically her critical early recognition of the need for a massive robust dataset to train AI models. Her development of ImageNet (first described in this 2009 paper) enabled the training of deep learning models that dramatically accelerated and energized the field.

Li was born in Beijing and raised in Chengdu. As a child she was a “STEM kid,” and particularly loved physics, she tells Vijay Pande, Andreessen-Horowitz general partner, Bio+Health lead, and host of “Bio Eats World.” 

When Li was 15, her family immigrated to Parsippany, New Jersey, where she attended public high school. She says she barely spoke a word of English at the time and took ESL (English as a Second Language) classes in every subject but math. 

A “profound” influence during high school, she says, was a math teacher named Mr. Sabella, who really took her under his wing, even setting up an individual advanced calculus class for her, which he then taught.

She credits her parents for supporting her pursuit of her passions. In particular, she says, they never pushed her to be a doctor or lawyer. She tells Pande that when her parents’ friends and neighbors learned she was majoring in physics, they told her parents to discourage her from pursuing such a “useless” area of study. Her mom, she relates with pride, simply responded, “She likes it.”

Li describes a remarkable double life as an undergraduate at Princeton. During the week, she says, she was a typical student, who enjoyed college life and “loved physics.”

Yet because her parents’ jobs had become precarious, she scraped together money to buy a dry cleaner in Parsippany, which she would manage in person from Friday night through Sunday, before returning to student life at Princeton.

Pande astutely observes this may have served as useful entrepreneurial training, adding “you weren’t playing around – it was survival.”

Ultimately, Li’s career in computer vision took off, leading her ultimately to Stanford University, where, among her other responsibilities, she is the founding co-director of the Stanford Institute for Human-Centered AI. 

AI, Li says, is at a real inflection point. Like DeepMind founder Mustafa Suleyman (whose new book, The Coming Wave, I just reviewed in the Wall Street Journal – see here), Li emphasizes that AI technologies are powerful tools that can be used for both good and ill. She explains that she’s less concerned about the rhetoric of human extinction (a preoccupation of some AI “doomers”), and more interested in identifying pragmatic ways to ensure these powerful tools contribute positively. Li says this requires us not only to embrace the tools, but to recognize our responsibilities. 

She emphasizes that policy should involve not just guardrails but also good incentive structures, and she would like to see more public sector investment in the technology so that all the resources and talent don’t wind up at private companies. 

In applying AI to healthcare, she notes that in both healthcare and medicine, there is a robust regulatory framework already, and says we don’t need to reinvent the wheel. Rather, she suggests, we should update the framework where needed, and “inform, educate, and communicate with regulators” regarding AI. 

She acknowledges the existence of a “catastrophic layer” of concerns, and worries about disinformation, in particular the potent combination of AI and social media (Suleyman’s similar concern around deepfakes is highlighted in the review). 

More hopefully, and sounding a bit like John Maynard Keynes, Li says she “dreams about” a “dignity economy,” where productivity is increased so much by tech that we work because “we feel it gives us agency and dignity and reward,” rather than representing a daily grind we endure so we can put food on the table. 

She acknowledges that attaining this nirvana will require hard work and require us to navigate many potential pitfalls; In particular, she worries about AI’s impact on “communities that are underserved and underrepresented.”

Li’s remarkable journey, and her reflections on it as elicited by Pande, engage and inspire. While Li’s science clearly centers around AI, her story serves to highlight the promise and possibility of human potential.

I can’t stop thinking about a recent series of poignant blog posts, written by an emergency room physician affiliated with the Mayo Clinic. Her husband has been battling a terrible cancer – recurrent/metastatic head and neck squamous cell carcinoma. 

Given what she does for a living, the author, Dr. Bess Stillman, is about as well-positioned to be a savvy patient advocate as anyone could possibly be.

And yet, she describes in earnest, eloquent, horrific, harrowing, infuriating detail her Herculean efforts to find and enroll her husband, Jake Seliger, in a suitable clinical trial. Jake has also written a thoughtful companion piece offering specific, constructive suggestions – an incredibly generous contribution from a person who has suffered so much.

While I plan to highlight a few key points, these riveting first-person perspectives should be required reading for everyone in medicine, everyone in biopharma, everyone involved in the regulation of clinical trials (both FDA and Institutional Review Boards [IRBs]), and every entrepreneur eagerly trying to develop “solutions” for clinical trials. 

So often, those of us in medicine, as well as those of us in biopharma, talk amongst ourselves about clinical research, and the challenges of conducting clinical trials. Typically, we review this information from a physical and emotional distance – perhaps in the context of a scientific result, or a discussion of an organizational process we seek to improve.

Physicians tend to absorb trial results from medical journals and at medical conferences and think about how to apply the findings in their day-to-day practice. Biopharma companies extensively discuss the many challenges associated with study operations and are motivated to improve the performance of discrete metrics like site activation time and speed of patient enrollment.

But most of these discussions tend to be – pardon the phrase – exceedingly clinical, and ultimately quite detached from the human experience. Neither the discussions of journal data nor the tactical review of study conduct begin to convey the visceral, excruciating frustration and anguish that comes across in the writing of Dr. Stillman and Jake.

So please, read Dr. Stillman’s pieces first, starting with part I, here. Then please read Jake’s companion discussion, here.

Context: Why Clinical Trials Matter

Clinical trials represent the foundational tool of medical progress; they are the standard by which potential interventions are evaluated by the biomedical community.  When I was a third-year medical student, and learning about patient care in the hospital, it was routine for published clinical trial results to form the basis of our team’s discussion. It’s the way we weed out the treatments that don’t work and elevate the ones that meet the standard of safety and efficacy.

Many academic leaders have made their name through their leadership of well-controlled, carefully conducted clinical studies that advanced patient care. Sometimes, these involve huge, sprawling enterprises, as is often the case in cardiology; the TIMI study group comes to mind, founded by the legendary Dr. Eugene Braunwald, and now led by Dr. Marc Sabatine (a former MGH colleague). But trials can also involve smaller, more focused efforts – here I think of the pioneering work in reproductive endocrinology led by Dr. Bill Crowley at MGH. 

Clinical trials are of course critical for biopharmaceutical companies as well; a candidate therapy needs to pass the exacting standards of FDA review in order for the company to be able to market it in the United States, for example. Most R&D dollars in big pharma are invested in the execution of clinical trials, particularly “late phase” (phase 2 and 3) studies. 

Not surprisingly, the industry is constantly seeking to improve the way trials are done.  A wide variety of vendors, from large contract research organizations (CROs) like IQVIA and prominent software-as-a-service (SaaS) platforms like Veeva, to a galaxy of niche startups, contribute, or aspire to contribute to the process. It’s a massive ecosystem, all ostensibly built around the idea of improving the trial process and accelerating the development of promising emerging medicines.

As clinical researchers recognize and constantly try to address, very few patients in the US actually participate in trials. The number I’ve often heard quoted is 4%; for cancer trials the fraction may be somewhat higher (see here). Dr. Stillman writes that according to “Judy Seward, head of clinical trial experience at Pfizer, only 8% of adult cancer patients take part in trials.”

Adds Dr. Stillman, “Given our recent experience, I’m amazed it’s that many.”

The Patient Perspective

Jake’s challenges, as described by Dr. Stillman, began from the moment he was failed by available (FDA-approved) medical treatment. That meant his only options were to enroll in a clinical trial. 

The first problem was Jake’s “lackadaisical” and “sluggish” oncologist at the Mayo Clinic, whose approach seemed characterized more by resignation than by urgency. 

This struck a chord with me. As I wrote in an op-ed for the New York Times a few years ago, we had a similar experience with an oncologist at Memorial Sloan Kettering Cancer Center who was treating my uncle for pancreatic cancer. This doctor wasn’t keen to prescribe antibiotics for my uncle’s lung infection. My aunt had to fight exhaustively to ensure her husband received appropriate treatment. The medicine worked, and he enjoyed (yes, enjoyed) a number of additional months of life. 

I can’t imagine what it must be like as an oncologist focused on the treatment of patients with terrible cancers. Perhaps some degree of emotional detachment is required to help these physicians do their unusually difficult job day after day.  But what many patients so fervently want and need is a doctor who will engage with them through a litany of treatment failures, and stay with them, continuing to urgently, aggressively, energetically search (if this is what the patient prefers) for the best possible path forward.

ClinicalTrials.gov – a Mess

Dr. Stillman’s second challenge was the website that lists all clinical trials – clinicaltrials.gov. The good news is that all studies are listed in this database. The bad news is that it’s not especially user-friendly – which makes finding relevant trials surprisingly difficult for patients.

For starters, the search function is miserable: “If you change the wording of the search subtly,” she writes, “you’ll get different outcomes.” Searches for “Squamous cell carcinoma of the head and neck” and “Head and Neck Squamous Cell Carcinoma,” she writes, “all yield different results. There’s no apparent standardization.”

It’s also difficult to keep the different treatment options straight. For example, in some trial listings, a medicine might be referred to by the company’s internal designation (she points to “MCLA-158” as one example), or by the standardized name (“petosemtamab,” e.g.) in others. Inconsistency in nomenclature makes it difficult to keep even identical therapies straight.

Then there are the various inclusion and exclusion criteria associated with different trials. “It’s easier to find a dress to my exact specifications out of thousands on H&M.com,” she explains, “than it is to find a clinical trial” that a patient might qualify for.

How does a patient sort through such a list? Ideally, an oncologist might be able to help. But Dr. Stillman found that most oncologists had no idea how to help, other than a general suggestion that the patient go to a large cancer center that runs a lot of trials.

Far more helpful, Dr. Stillman and Jake discovered, was a skilled individual named Eileen Faucher. A biochemist who had spent years consulting for pharmaceutical companies, Faucher has now set up her own business, serving essentially as a private wilderness guide for clinical trials.  

“Eileen charged a lot per hour but her fees were worth every penny: she was efficient, and without her I’d have been lost,” Dr. Stillman writes.

Similar services are cropping up to help people (who can afford the fees) make their way through the staggering complexity of our healthcare system.

Other examples include concierge physicians (who provide you their cell phone so you don’t have to endure endless phone trees to leave a message at your doctor’s office and pray someone eventually calls you back) and patient navigators, who help sort through hospital bills and insurance coverage. See also this WSJ review I wrote a few years back on a book, The Patient’s Playbook, describing the need for a “medical quarterback.”

And what about the many AI services increasingly pitching their wares to patients, hospitals, and pharmas? Unfortunately, Dr. Stillman and Jake discovered that several AI services purporting to help patients sort through clinical trial results – ostensibly, computerized, scalable version of Eileen — weren’t useful or helpful in practice. “Based on what we’ve seen so far,” Dr. Stillman writes, “the ‘AI’ tech isn’t there yet.”

The Struggle To Enroll

Once Dr. Stillman, working with Eileen, developed and refined a list of the most promising trials for Jake, she discovered how difficult it was to determine if he could enroll.

For starters, the contact info listed for many trials was inaccurate or not useful. Other times, when she reached a person at the clinical trial site, she was told the only way to receive information about the trial was by voice, read over the phone – all other modalities were prohibited. And that was when information was provided at all.

More commonly, she discovered that even when she reached someone at a trial site, almost no one would even discuss whether Jake was eligible without him first showing up for a doctor’s appointment to “establish care.” 

Generally, this wasn’t allowed via telemedicine. That would constitute practicing medicine across state lines, a big no-no. However, in some places, it seems like a “don’t ask/don’t tell” system exists in which patients can assert they are calling from in-state.

The alternative to telemedicine is a lot of travel, which is expensive, time-consuming, and exhausting for cancer patients. As Dr. Stillman observes, it’s “very hard for sick patients to travel and ‘establish care’ at multiple sites.”

Even the process of “establishing care” involves multiple parts. First, a chart must be created in the hospital system (without which you don’t exist, Dr. Stillman writes). Then, you need to be seen by a physician. This physician “can’t do anything for you — even answer questions — until after your first visit (even though they can review information you send),” she explains.

It’s also at the in-person visit that you are finally told what trials are open and what you might qualify for.

Even after a patient seems like he or she might qualify, they need to demonstrate they are at just the right point in their disease; the implicit message, Dr. Stillman writes, is “please be dying, not too quickly.”

Please be dying, not too quickly

In Jake’s case, he needed to demonstrate his disease was progressing rapidly enough to qualify. But his disease progression clock would start over each time he had a dose of chemo (which wasn’t expected to be curative — he was taking it simply to try to stay alive long enough to make it into a clinical trial). On the other hand, if tests showed he had metastases to the brain, he would be considered too sick to qualify.

Strikingly, Dr. Stillman and Jake discovered considerable site-to-site variability for what was considered sufficient progression. Even at different sites associated with the same clinical trial (and thus with identical entry criteria), some physicians would decide Jake was progressing fast enough to qualify. Others thought he wasn’t – yet all these decisions were based on identical clinical and imaging information.

Lessons Learned – Digital and Data

Both Dr. Stillman and Jake discussed a number of challenges in getting access to data. Much of what they had to say reminded me this National Academy of Medicine-sponsored paper that I co-authored during the COVID pandemic. 

My co-authors and I wrote this passage about connectivity and last-mile issues in the context of COVID, but similar points could be made today regarding clinical trials:

“[D]uring the initial stage of the pandemic in the U.S., decision-makers were essentially flying blind. Electronic health record (EHR) systems were mired in a sea of codes, few of which pertained to COVID-19, due to its novelty. These systems were not connected to enterprise resource planning systems, and thus lacked the ability to correlate relevant patient encounters with human resources and physical capacity. The utilization of testing, PPE, beds, and ventilators varied within and across each and every health system (and often varied even across departments within a single hospital or clinic). Public health departments in charge of implementing rules, policies, public guidance, and contact tracing operations each operated within their own data silos—often taking the form of piles of spreadsheets—and were almost always unconnected and non-interoperable with any other health care information technology (IT) system. In too many cases, the only effective communication of data between health care delivery systems and public health agencies was through a fax machine.”

Dr. Stillman shines additional light on several of the critical information gaps that collectively result in a shocking lack of awareness and coordination:

• Lack of trial awareness by oncology researchers: “Researchers didn’t have their own spreadsheet or databases of active trials,” she writes. “They didn’t know what trials were going on down the street. Sometimes they didn’t know what trials were going on in their own (very large) departments.”  She also cites a damning survey from the Tufts Center for Drug Development, reporting that although “about 90% of physicians surveyed said they feel comfortable discussing clinical trials with their patients,” nevertheless, “less than 0.2% actively refer them to studies. These doctors reported that they lack access to trial information (54%), do not know where to refer patients (48%), or do not have the time to learn about active trails (33%).” 
• Lack of trial awareness by researchers in the same network. Even within a given research network, such as Sarah Cannon or MD Anderson, there seems to be a lack of awareness of other trials occurring under the same organizational umbrella.  “A researcher at Sarah Cannon in Denver told me that even the Denver site doesn’t know what the Nashville or Florida sites are doing…. there’s no easy internal system between hospitals, or regions,”  Stillman writes, stunned by the degree “information siloing.”
• Lack of site-to-site referrals within the same study: “If one hospital thinks a patient would be a good fit for a study, but their study arm is closed, wouldn’t it be beneficial to the drug company to recommend other sites?” Dr. Stillman asks. “Shouldn’t sites coordinate with each other? That way, the overall trial can more quickly fill a given study with eligible patients.”
• Lack of information for patients from sponsors about site availability: “I tried contacting various drug company trial contacts directly, to see if they’d provide me with a list of all sites and openings,” Dr. Stillman writes, adding “I was only able to get contact information. I couldn’t get real-time information about availability, not because it didn’t exist, but because no one would give it to me.”
• Inability of a PI (“Principal Investigator” – researcher leading a study at a site) to know, in their own study, if spots are available: “Even the PI hosting a trial often don’t know in real-time if a participant spot is available,” Dr. Stillman writes. She shares an example of requesting a spot for Jake in a trial and then finding out two days later the study was closed by sponsor to new participants.

Regulatory Challenges and Opportunities

Dr. Stillman’s frustration extends to regulators, both at the federal level and with local Institutional Review Boards. 

For instance, she learned about data suggesting that a fecal transplant procedure can make some non-responders to the PD-1 inhibitor pembrolizumab (Keytruda – see here for more the remarkable history of this blockbuster product) start responding to the drug.

She pursued this idea as a course of treatment for Jake. Yet, she struggled to get permission to do this, she reports, explaining that the FDA cracked down on the fecal transplant procedure after two deaths (out of what she estimates are 10,000 procedures). 

Her local IRB, she says, also hasn’t been eager to approve this procedure. With understandable acidity, she writes, “Luckily for Jake, the FDA and IRB will protect him from being one of the 2 in 10,000 patients who died from a fecal transplant by denying him this potentially life-saving therapy so that he could have a 99%+ chance of dying from his cancer. I feel safer.”

I’ve discussed the need for an approach to regulation that places a greater premium on the harms that can be done by delaying access to promising experimental medicines, and the value (see here) of a more personalized approach to regulation.

Glimpsing the Underbelly of Clinical Trial Operations

The challenge of contemporary evidence generation has been discussed very thoughtfully by Matt Herper at STAT, here. I’ve emphasized the need to learn from real world experience here.  But if you really want to understand some of the gnarly operational issues associated with getting clinical trials up and running, you owe it to yourself to read this (now deleted but archived at the link provided) blog post, from an anonymous (but credible) clinical trials operations expert.  

The author offers an unvarnished, behind-the-scenes view of how the trial process works from the in-the-trenches, operations side, starting with clinical site selection, which the author notes is “superficially quite reasonable,” yet “actually going through this process is an extended exercise in sheer absurdity.”  

And on set-up process for clinical sites: “There are so many examples of horrible, atrocious inefficiencies that I encountered while dealing with the logistics of clinical trial setup that I could easily triple the length of this section if I wanted to.”  

Much like Dr. Stillman’s articles, the anonymous author’s depiction of the messy reality of clinical trials operations and set-up offers urgently needed visibility into what’s actually going on.  Uncomfortable as it is to read, the description provides an essential alternative to the abstracted, idealized version of trials so often shared on PowerPoint, and discussed in planning meetings.  Everyone — including (perhaps especially) senior leaders and decision-makers — deserve, and would likely benefit from, a candid look into how the sausage actually gets made.

Aiming To Do Better

Both Dr. Stillman and Jake strive to be constructive. They highlight what patient-centricity (the core value expressed by every stakeholder in the clinical trial process) actually looks like from the patient’s perspective. 

What’s needed, they suggest, is a far more convenient way to identify, prioritize, learn about, qualify for, enroll in, and participate in relevant clinical trials.

What’s needed, they suggest, is a far more convenient way to identify, prioritize, learn about, qualify for, enroll in, and participate in relevant clinical trials

Most of the people I know involved in the clinical research enterprise aspire to deliver this, yet it’s clear from Dr. Stillman and Jake just how far from this ideal we collectively seem to be. 

As Dr. Stillman concludes, hopefully,

“There’s a future that can support the interests of patients who are willing to take part in trials, the patients who will benefit from the results, the doctors who are trying to do what’s right for their patients, as well as the interests of the pharmaceutical companies supporting the trials.

There’s a better path. We’re just not on it. Yet.”

My Ground Truth

Every other week, I stroll across the Longfellow Bridge from Cambridge to Boston. It can be a magnificent walk in the fall and spring, when the weather is temperate and the skies clear. You can see the deep blue of the Charles River, the Esplanade on your left, with the Boston skyline behind it, and the Citgo sign in the distance. 

On the other side, you see the biotechnology cluster of Kendall Square and a large swath of MIT, expanding out from the river’s northern shore. Boston University’s campus lies a bit further upstream, on the south, while Harvard’s campus is a few bridges further down, to the north.

For me, the regular walk to Boston provides me with critical grounding – partly because I’ve spent most of my adult life within the vista before me, but mostly because of the hour-long teaching conference at Massachusetts General Hospital (MGH) I attend, learning about the complicated challenges faced by a patient receiving care on internal medicine service. The discussion is often inspirational and is invariably humbling.

The conference, and the activities that support and enable it, were originally conceived by two MGH medicine residents, Dr. Lauren Zeitels and Dr. Victor Fedorov.  In 2016, they founded the Pathways program, “dedicated time during residency for house staff to connect with scientists and delve into the fascinating biological questions that arise at the bedside.”  Particularly in a busy hospital like MGH, with many acutely ill patients requiring focused attention, such time to reflect can be fleeting, yet remains vital, as Dr. Denny Ausiello and I discussed here.

Tragically, Drs. Zeitels and Fedorov died in 2017, in an avalanche while snowshoeing in Canada.  The pathway program endures as a legacy to the power and urgency of their vision.

Pathways conferences are led by talented medicine residents who spend two weeks focused intently on the biological questions underlying a patient’s illness, with the hope of identifying key scientific questions that, if answered, could point the way to improved diagnosis and treatment.  Like most clinical cases discussions, the conference presentation is oriented around the patient, starting with a review of how the patient came to be in the hospital, and generally ending with a discussion of the patient’s hospital course and anticipated trajectory.  In between these essential bookends, there’s invariably an intriguing discussion around the often-mysterious biology underlying a patient’s illness.

By design, the team focuses on complex or confusing cases – at times, it feels like it might be called the “idiopathic conference” since many of the diagnoses featured tend to fall into this category (“idiopathic” is the fancy medical term for “we don’t understand what’s going on”). 

Unfailingly, these deep dives into biology expose a central truth that has long impressed me about medicine: how little we still understand about how the body works, about what causes many diseases, and how limited are our arsenal of tools to diagnose and effectively, precisely treat many ailments.  I felt this profoundly when I was a medical student seeing patients on the wards at MGH, and again when I was an internal medicine resident and later endocrinology fellow at MGH.  I continue to experience these exact emotions today.

When you look at the examples of progress in medicine and biology, the list is impressive.  The modalities available to diagnose and treat disease continues to expand, while the terrible diseases that we’ve learned to prevent (like polio) or more effectively manage (like cystic fibrosis) speak uncontestably to the promise of science. 

Yet when you take your measure by your ability (or lack of ability) to diagnosis and treat the patient in front of you – much less prevent the disease from occurring in the first place – you are quickly overwhelmed by profound humility. 

But you also feel something else.  Not only do you leave these conferences with a deep appreciation for the limitations of our existing knowledge, you invariably emerge with a palpable sense of urgency, a determination that we must do better, and drive harder to ensure that nascent technologies, both biological and digital, are brought to bear, relentlessly as well as thoughtfully, in service of patients.

Level-setting expectations for emerging tech

If the MGH Pathway conferences regularly remind me of how urgently improved diagnostics and therapeutics are required, recent events have reminded us how difficult the journey will be.

Several recent examples stand out:

Precision Medicine

Consider this question from Andreessen-Horowitz general partner and Bio+Health lead Vijay Pande, during his podcast interview of Olga Troyanskaya, a computer scientist and geneticist at Princeton. 

With remarkable candor, Pande asks,

“There was so much excitement about precision medicine with genomics, right? I think the idea was that, okay, we could sequence a patient and maybe sequence a tumor and from that we would be able to do so much. We’d be able to figure out which cancer drug to give or which drug in general to give. And that hasn’t quite come to fruition yet, it feels like — or maybe it has happened so gradually that we’ve lost track. What’s your take on precision medicine?”

The idea that precision medicine hasn’t quite lived up to the admittedly extravagant expectations is a familiar perspective on the wards and in lab, but to hear it from within the reliably exultant investment community feels like real progress.

Troyanskaya’s response offers similar candor.  While pointing out that “we have seen a lot of successes,” she also recognizes that “a lot of it was hype,” adding “None of us, I think, quite appreciated quite how complex it is. We sort of thought we’ll find these few genes, they’re the drivers, we’ll develop targeted therapies, we’re done.”

To be sure, she points out, there are “true revolutionary successes of precision medicine.” She cites our increasingly sophisticated approach to breast cancer and some subtypes of lung cancer as two examples.  

But, Troyanskaya adds, “What I think we underestimated is that it’s not enough to actually just know the genome, that it’s not, you know, it’s not just a few causal genes.”  We also need to understand the non-coding 98% of the genome, she says, and to capture other types of data, including proteomics and metabolomics.

AI for Drug Discovery

Coming up with impactful new medicines is remarkably difficult, and the idea that AI can improve the odds offers obvious appeal. 

But at least some of the bloom is off the rose, as it’s now clear that many of the so-called “AI discovered” drugs have not made it through development. 

“The first AI-designed drugs have ended with disappointment,” writes Andrew Dunn in Endpoints. 

He continues,

“Over the last year-plus, the first handful of molecules created by artificial intelligence have failed trials or been deprioritized. The AI companies behind these drugs brought them into the clinic full of fanfare about a new age of drug discovery — and have quietly shelved them after learning old lessons about how hard pharmaceutical R&D can be.”

He also points out that while the last few years have been notoriously challenging for small biotechnology companies (the XBI biotechnology stock index has dropped by 50% over the last two years), it’s been even worse for companies focused on AI drug discovery, dropping by 75-90% over the same time.  

[Chart courtesy of A. Dunn, Endpoints News]

(It’s also been a challenging time for digital health, see this report from Rock Health, and for healthtech – see this report from Bessemer Venture Partners. Interestingly, investor focus in these areas seems to be shifting from the aspiration of radically disrupting healthcare and biopharma to the appreciably more modest goal of making existing processes, especially non-clinical or back-office, somewhat more efficient.)

As many commentators have pointed out, the obsession with “AI drugs” is a little silly, in that AI is a tool that may inform a complicated, multi-step process. Moreover, a range of computational techniques are often involved in the development of medicines. Thus, the idea that one product is an “AI drug” and another isn’t feels more than a little arbitrary.

There’s also the issue of expectations – even if AI is able to contribute to the efficiency of early activities in drug discovery, at best this might only improve the overall process slightly given the multiple downstream hurdles that remain (especially since it’s not clear, as we’ve discussed, that AI can offer particularly useful predictions around late phase clinical studies, where most of the time and dollars are spent). 

VC Patrick Malone (cited by Dunn) may be exactly right when he says,

“If you take the hype and PR at face value over the last 10 years, you would think [the probability of creating a successful drug] goes from 5% to 90%. But if you know how these models work, it goes from 5% to maybe 6% or 7%.”

In a separate post on LinkedIn, Malone (citing Leonard Wossnig, Chief Technology Officer at LabGenius) riffs on a key underappreciated challenge in drug discovery, particularly drug discovery that involves algorithms: figuring out what to optimize for – the so-called “objective function.” 

For instance, you might try to identify a compound that binds incredibly tightly to a specific receptor, say. Yet it turns out that in many cases, such an approach isn’t quite right – empirically, you may discover you need a molecule that binds with moderate affinity to multiple receptors. (I’ve discussed this nuance in the context of phenotypic screening and the discovery of olanzapine, a medicine for schizophrenia, for example; see also this captivating recent commentary about phenotypic drug discovery, penned by researchers at Pfizer).    

Generative AI in the Enterprise

Generative AI captured our collective attention with the November 2022 release of ChatGPT, as this column has frequently discussed.

Despite the insistence of consultants that if your company is not already leveraging generative AI, it’s behind, I’ve not seen many examples of generative AI in actual use at large companies, particularly biopharmas.

It turns out, this perception is well-founded.  According to Andreessen-Horowitz co-founder Ben Horowitz – a great champion of AI – “We haven’t seen anybody [involved in generative AI] with any traction in the enterprise.”  (“Enterprise” in this context refers to the information technology systems and processes used by large companies, like pharmas.) 

According to the technologist Horowitz was interviewing — Ali Ghodsi, co-founder and CEO of the big data analytics company, Databricks – there are at least four reasons generative AI hasn’t yet entered the enterprise:

• Big companies invariably move slowly and cautiously.
• Corporations are terrified that they will lose control of high-value proprietary data.
• Companies typically need output that is exactly right, not sort of right.
• There is often a “food fight” (as Ghodsi puts it) at companies about which division controls generative AI.

From what I’ve seen across our industry to date, I can only say that Ghodsi seems to deeply understand his enterprise customer base.

I was also struck by how Ghodsi – whose business embraces and enables generative AI – readily acknowledged the limitations of this technology.  “It’s stupid and it makes mistakes,” Ghodsi says, adding “it quickly becomes clear that you need a human in the loop. You need to augment it with human. Look, there’s no way you can just let this thing loose right now.” (The need for a human in the loop was also a key assertion Harvard’s Zak Kohane made in his recently-published book about ChatGPT and medicine – see here.)

Ghodsi was also skeptical about performance benchmarks cited for generative AI – including around their supposedly high scores on medical examinations.  He suggested that the models might have picked up the test questions during the course of their training (training that’s shrouded in mystery, as Harvard’s Kohane has pointed out), and thus, their high scores may be deceptive — like an exam where you got the answers the night before.

To be sure, Ghodsi remains excited about the potential of generative AI; he’s just realistic about its present limitations.

III – BioTechno-Optimism?

The talk of tech these days is the latest manifesto penned by Andreessen-Horowitz co-founder, Marc Andreessen, this time on “Techno-Optimism.” In case you missed it, his simplified thesis is that technology=good, and anyone who would constrain it=bad.

The piece has triggered an energetic and largely critical response from writers like Ezra Klein in the New York Times (“buzzy, bizarre”), Steven Levy in Wired (“an over-the-top declaration of humanity’s destiny as a tech-empowered super species—Ayn Rand resurrected as a Substack author”), science fiction author Ted Chiang (“it’s mostly nonsense”), and Jemima Kelly of the Financial Times (“is the billionaire bitcoin-backing venture capitalist OK?”).

Levy’s analysis was perhaps the most resonant:

“[Andreessen] posits that technology is the key driver of human wealth and happiness. I have no problem with that. In fact, I too am a techno-optimist—or at least I was before I read this essay, which attaches toxic baggage to the term. It’s pretty darn obvious that things like air-conditioning, the internet, rocket ships, and electric light are safely in the ‘win’ column. As we enter the age of AI, I’m on the side that thinks that the benefits are well worth pursuing, even if it requires vigilance to ensure that the consequences won’t be disastrous.”

But if I’m being honest, I thought there was something visceral in Andreessen’s screed – or perhaps more accurately, in the impulse behind it — that really touched a chord for me, emphasizing both the value emerging technology can bring and the incredible challenge of nurturing new technologies through their growing pains (what Carlota Perez might describe as the Installation phase), to the point where the demonstrated value is unassailable (the Deployment phase). 

Not only is it intrinsically difficult to figure out how to implement new technology effectively, necessitating all sorts of incremental innovations (see here), but there’s exceptional resistance all along the way (at least outside of public relations, where new technology is championed relentlessly). 

Incumbents – particularly in healthcare — resist technology for years before eventually (perhaps) absorbing it.  In part this reflects inertia – we’re most comfortable with what we already know. 

Another factor, as I’ve written, and as Andreessen observes (with his usual bluster), involves the  precautionary principle.  Stakeholders (in my view) are appropriately concerned about doing harm with novel technologies, but routinely overlook the harm that’s done by inhibiting the adoption of promising new approaches.

Moreover, new technologies often bring significant uncertainties, as Dr. Paul Offit has nicely described in You Bet Your Life (my WSJ review here). This is true not only for new medical treatments, as Offit nicely discusses, but also in how high-stakes research is conducted. 

Consider a typical example from the world of clinical trials, the lifeblood of biopharma and academic clinical researchers.  Imagine that you hear about a new tech-enabled approach for a key aspect of this process — patient recruitment, say, or decentralized data collection, an approach which is claimed to work better than traditional methods. 

If you are leading a clinical development team in a pharma company, where everyone in the business is already breathing down your neck with advice and suggestions and telling you that you’re already behind and reminding you how important it is to be successful, the last thing on earth you want is to take on more risk. 

If you use established procedures, at least you know what you’re getting into.  But if you try a new approach (even an approach perhaps piloted under far more forgiving conditions), you could get yourself in deep trouble.  This actually happens.  Not surprisingly, faced with these sorts of decisions, many rational actors within pharma prefer to be “fast followers” rather than adventurous pioneers, a mindset Safi Bahcall has described in Loonshots (see my WSJ review here, and this additional, biotech-focused discussion in Forbes, here).   

Concern about assuming new risk (beyond intrinsic risk of new molecules) is also why Gregg Meyers, the Chief Digital and Technology Officer at Bristol-Myers Squibb, says “this is an industry that will not adopt something unless it is really 10x better than the way things are historically done.”  See also this generally accurate perspective on the challenges of selling innovative tech into pharma, written from perspective of a startup that’s developing the technology.

(Bio)Technology: Critical Enabler of Medical Progress

Ultimately, for me, it comes back to the patients we discuss every two weeks at MGH.  We need far better diagnostics, far better therapeutics, and we need to understand health and disease at a far more sophisticated level.

Historically, improved tools – the microscope, ultrasound, MRI, DNA sequencing – have radically redefined the way we understand and approach disease.  Today, emerging modalities and tools — like antibody-drug conjugates (which are enjoying a banner year), cell and gene therapy (enduring real growing pains but with demonstrated successes and tantalizing possibilities), spatial biology techniques (quite exciting), and of course artificial intelligence (which is getting more capable by the second) offer exceptional promise, even if in many cases we’re just beginning to contemplate how to leverage these new technologies effectively.  Safety and ethics – appropriately — remain paramount consideration as well; see my recent discussion of Ziad Obermeyer’s important research revealing hidden bias in medical algorithms, and my WSJ review of Brian Christian’s essential The Alignment Problem, here.

While I disagree with Andreessen’s denunciation of skeptics and naysayers, I feel his frustration, and share his belief that technology (guided, I would add, by inquisitive researchers, practitioners, and other “lead users” who are passionate, thoughtful, and deeply empathetic) offers the most promising path forward in our drive to improve the human condition in the area of health and medicine.

Sometimes, it seems, we get so caught up in our clever criticism and overwrought angst that we get in our own way, stall our own momentum, and impede the progress upon which the health of our patients, and of our future patients, will so critically depend.