2023 Healthcare and Life Sciences Predictions
Macroeconomic uncertainty, clinical workforce shortages, and public health concerns continue to challenge healthcare and life sciences companies. Here are ten reasons why we feel optimistic about this sector in 2023.
Health tech is fast approaching its teenage years. The cottage industry that commanded about $1 billion of venture capital funding in 2011 raised nearly 30 times more in 2021. With this milestone, we recommit to what matters most: building enduring and equitable healthcare solutions that improve clinical outcomes, reduce costs for the system, and enhance the experiences of our healthcare workers.
Tomorrow’s health tech startups are saving lives—and making lives easier. In bio, on the back of stellar pandemic responses from pharma, rapid advancements in genomic medicines and other novel modalities are enabling scientists to discover treatments for our most intractable diseases. On the frontier, we welcome the growth of burgeoning categories like fintech for healthcare and computational biology. And despite many years of debate on the role of artificial intelligence in healthcare and life sciences, we’re encouraged by recent step-function improvements that we predict will transform its utility in the field.
Guided by the policy forecast, technological advancement, and close counsel with builders in the ecosystem, these are the ten themes and trends that underpin how we’re thinking about healthcare in 2023.
In 2023, we’ll witness new company formation at scale, spawned by the healthcare and life sciences sector’s equivalent of the PayPal Mafia. Whereas companies like Castlight, Flatiron Health and Oscar created the first cohort of health tech alumni networks, the talent that has matured alongside the industry over the last decade will usher in the next wave.
Health tech now boasts a class of veteran talent—leaders with over a decade of domain expertise in both health tech as a sector and in a particular functional area (e.g., product managers who have built technology in both pre- and post-FHIR eras). In response to the industry’s record growth in venture investment and company formation over the last few years, we’ve also welcomed significant talent from other verticals bringing fresh ideas from big tech, consumer internet, fintech, and more. As the shiny veneer of health tech wears for our newly acquainted, bright-eyed colleagues, we’re hopeful that a non-trivial percentage will make themselves comfortable, roll up their sleeves, and stay for a little while longer.
As tighter capital markets force companies to grapple with the decisions to “roll or be rolled,” we predict that the veterans of health tech will take the entrepreneurial plunge in droves and in partnership with the fresh talent that has entered the sector over the last few years. With more than $160 billion of venture funding allocated for new investments sitting on the sidelines, high-quality teams and concepts soon entering the market, and an exciting policy forecast, the seed stage will remain active in 2023, and our seasoned colleagues will bring new perspectives to the table.
Are you joining healthcare or bio’s next major alumni network? We’d love to hear from you.
In 2021, we predicted that the pandemic would remove the scarlet letter from diagnostics, as we saw how billions of people around the world became intimately familiar with rapid diagnostics nearly overnight. As expected, what followed was the meteoric rise of companies like Everlywell, Rupa, LetsGetChecked, and others driving the distribution of existing diagnostic technology, creating high-quality consumer experiences for at-home testing. These organizations will likely continue to thrive and invest in novel diagnostics to expand their offerings.
In 2023, diagnostics will remain center stage and we will see a renewed focus on innovation in the core diagnostic technology and regulation. Although 70% of health care decisions are based on clinical lab tests, diagnostics remain a challenging area for investment plagued by the capital intensity of development, long, and arduous paths to commercialization, and oftentimes, binary risk profiles. Over the last few years, however, the surface area for diagnostic capabilities has expanded exponentially thanks to advancements in artificial intelligence (AI) and computer vision, which is now being applied across medical imaging exams including colonoscopy (Iterative Health), coronary angiograms (Cleerly), retinal imaging (AI Optics), and ultrasound (Caption Health). Furthermore, molecular diagnostics and clinical microbiology are advancing expeditiously as genomics, proteomics, and other diagnostic markers offer combinatorial approaches for phenotyping disease.
With science advancing more rapidly than regulation, Congress is considering policy provisions that could modernize how the FDA oversees diagnostic tests via the VALID Act. If passed, the policy could create a consistent standard for all diagnostic lab tests, and would shift regulation of diagnostic tests from the FDA’s oversight of medical devices to a novel framework designed specifically for diagnostics. If the VALID Act materializes, we would expect a flurry of activity in molecular diagnostics with potential implications to streamline approval for software-based diagnostic counterparts. Should the bill fail to pass, it’s possible the FDA will begin regulating diagnostic lab tests under the same process as medical devices—an outcome that might inhibit many early-stage startups from navigating the payer commercialization pathway within venture timeframes and ultimately requiring alternative go-to-market motions.
Whether or not the VALID Act becomes law, we are bullish on next-generation diagnostics. The technology for combinatorial approaches (e.g. imaging and multi-omics) is here. And we’re encouraged by the commercialization pathways paved by companies like Iterative Health and PathAI that have partnered with biopharma to provide better, faster, cheaper clinical trial identification in exchange for R&D resources. Though reimbursement will remain critical to unlocking the potential of these technologies in the clinic, starting with biopharma de-risks the development paths of emerging companies for founders and investors alike, and underscores the massively underestimated market sizes for these platforms across both research and clinical applications.
The generative AI models that are transforming consumer social, gaming, and art will have an outsized impact in healthcare and life sciences, starting with non-clinical tasks. Generative models represent a recent breakthrough in AI whereby algorithms can utilize existing data to create new, plausible data. Builders now have access to high-quality, cheap, and fast AI models that generate text, images, videos, software code, music, voice, 3D models, and more. These models are data-hungry, however, as ChatGPT’s training protocol leveraged the near-entirety of the pre-2022 Internet.
To meet the data appetites of generative AI models in healthcare and life sciences, the explosion of machine-readable biological and clinical data and increasing data liquidity has kept pace, driven by two forces:
- Greater than 10x decreases in the costs of sequencing since 2015 as well as new sequencing methods (e.g. single-cell RNA sequencing), which have increased access to digitized, machine-readable chemical and biological data.
- Critical regulations such as the HITECH Act, which digitized more than 90% of clinical data, and more recently, the 21st Century Cures Act, which mandates healthcare data interoperability across stakeholders.
We predict that generative AI will first impact non-clinical tasks in healthcare and life sciences that require lower accuracy thresholds to deliver value such as call center automation, drug discovery, privacy preservation via synthetic data, and code generation for bioinformatics, to name a few. As foundation models like GPT-3 (and soon GPT-4, which will launch in 2023) continue to improve, and as we understand how to infuse biomedical domain expertise to improve the accuracy of foundation models, clinical and scientific use cases, such as clinical decision and scientific research support, will gain traction. Though we don’t expect the accuracy of these models to achieve these benchmarks in 2023, the right question is “when, not if.” And, as we saw with ChatGPT, the first machine learning model to go viral after capturing one million users during its first week, we believe generative AI technology is well-suited for the product-led growth strategy for clinicians and scientists.
To illustrate the power and application of large language models, Ansible Health published research on ChatGPT’s potential for AI-assisted medical education; the paper, authored by our portfolio company in collaboration with our team, explores how ChatGPT performs at or near the passing threshold for all three US Medical Licensing Exams, without any specialized training or reinforcement.
We’ve believed in the potential for AI to transform healthcare and life sciences since we launched our Deep Health Seed Program in 2018, and expect this technology will become near-ubiquitous in the years to come. We made our first investments in companies deploying generative AI in the industry in 2018 and 2019 with Subtle Medical (medical imaging) and Abridge (clinical transcription), respectively, and expect to make more over the coming year.
It’s no surprise that the payment infrastructure and financing for healthcare have struggled to keep up with other sectors like retail and banking. Although the majority of healthcare payments travel via Automated Clearing House (ACH), approximately one-quarter of all healthcare payments to payers or providers are still sent via physical mail. Consequently, payments are delayed often with little visibility, exacerbating the already tenuous relationship between payers and providers under fee-for-service (FFS) arrangements. Did we mention how poor the patient experience has become due to such an underperforming payment infrastructure?
The revenue cycle management industry has long been a major driver of wasteful spending in healthcare, an estimated 75% of which is spent on human labor pushing paper and fax. We predict that digitizing and automating the FFS payments stack through the adoption of best-in-class fintech infrastructure, software, and artificial intelligence will address these inefficiencies.
A confluence of trends make us optimistic that fintech will make progress in healthcare in 2023:
- Regulatory tailwinds catalyzed by the CMS Proposed Rule to Expand Access to Health Information and Improve the Prior Authorization Process could require health plans to digitize and automate prior auth, creating a clear “why now” moment for the procurement of new technology, such as our portfolio company, Verata Health, which was acquired by Olive in 2020.
- Emerging from the height of the pandemic and finding new normal, providers are prioritizing investment in revenue cycle management technology as a top area of software spending over the coming year.
- Greater price transparency, catalyzed by the 2021 Price Transparency Ruling and accelerated by data providers like our portfolio company, Turquoise Health, is shining a light on reimbursement variance. In addition to litigation, we think price transparency might inspire new and different payment logic.
- Patients are demanding digital payments. When surveyed, nearly two-thirds of patients claim they would switch doctors based on a negative payment experience.
We also see immense opportunity to build healthcare-specific fintech applications designed for value-based payments, which require tight integration of clinical and financial data. Some companies are tackling this new category by starting upstream, helping providers and payers gather the documentation to execute value-based care arrangements by providing pay-for-performance measures (Stellar Health), next-generation risk adjustment analytics (Juxly), and quality outcomes monitoring (Clarify Health).
We expect to see more innovation across the healthcare payments lifecycle with solutions that offer custodianship and execution of value-based contracts, incorporate relevant clinical data and performance measures, provide real-time analytics, and facilitate automated financial transactions among payers, providers, and the burgeoning class of “payviders.”
Regulation is a reliable catalyst for innovation in healthcare and life sciences. But it can cut both ways, and so with any stroke of the pen, there is both opportunity and risk. The Inflation Reduction Act (IRA), passed in August 2022, is one such example. Despite its lofty goals, the act includes a contentious provision that allows Medicare to negotiate rates for a set of small molecule drugs nine years after they launch, and some biologics 13 years after they launch. A potential near-term win for patient-consumers and payers, the pharma industry has voiced concerns that the legislation may stifle biotech innovation by reducing incentives to invest in the development of small molecules. By shortening the patent life of small molecules to nine years effectively, the act could cap the potential terminal value for negotiated drugs and as a result decrease investor appetite for small molecules at the earliest stages.
Though we’re still at the tip of the iceberg in understanding the impact the IRA will have on the biopharmaceutical industry, we predict the legislation will increase focus on and investment in novel modalities that might be shielded most from the policy such as gene editing, gene therapy, and biologics. And while the direct impact of the IRA in drug price negotiations won’t come into effect until 2026, we expect biopharma companies to make serious changes to strategy and operations in 2023 in anticipation including re-prioritizing pipeline programs, innovating in commercialization, and investing in new drug channel distribution strategies.
In 2022, we saw increased antitrust pressure on Pharmacy Benefit Managers (PBMs) given the role they play in the ongoing drug pricing debate and how they’re vertically integrated with payers. We predict that 2023 will be a pivotal year during which PBMs and the pharmacy supply chain are tasked with providing more transparency in pricing and operations. There are a few factors at play driving massive changes against PBMs in 2023:
- First, the Inflation Reduction Act (IRA), which we described will allow Medicare to negotiate the prices of medications under Medicare Part D beginning in 2026, could be viewed as directly encroaching on the role of the PBM in Medicare Part D and could force PBMs to recoup lost Part D revenue elsewhere. Though we expect the IRA to impact PBMs, the magnitude of the impact will depend on which drugs Health and Human Services selects for direct negotiation. Certain drugs, such as high-cost medications with little competition, typically aren’t massive rebate opportunities for PBMs. We’ll learn more over the coming months.
- We also expect increased scrutiny of PBM business practices following the FTC’s 2022 probe into the industry. We expect everything to be on the table—from gross-to-net pricing gaps, to prior authorizations, to formulary design fees. We’re also keeping a close eye on the Pharmacy Benefit Manager Transparency Act of 2022, which, if passed, could ban PBMs from profiting on pricing spreads and clawbacks.
- With all eyes on PBMs, we are also watching closely as biosimilars (the equivalent of ‘generics’ for biologics) take off in the new year. Humira, the highest-selling drug to treat multiple autoimmune disorders, is expected to have 6-7 biosimilars in the market by mid-2023. Historically, biosimilars have struggled to gain market share due to challenging formulary rebate negotiations with PBMs and the interchangeability of drugs that has historically benefitted branded drugs (e.g. Semglee, an insulin biosimilar). Given the cost burden of Humira for payers, we expect to see more support for the adoption of biosimilars in 2023 and diminishing tolerance for the great rebate game.
We see immense opportunity for new platforms emerging from the interstitial space between pharma companies, payers, providers, and patients as it relates to pharmacy. Already, there is a cohort of companies providing increased transparency into the flow of funds from the drug industry including Kalderos (drug discount management), Lyfegen (value-based contracting platform), and CuraFi (specialty medication access). We expect new categories of software and tech-enabled services to take this ecosystem head-on with the intention of driving savings and clinical outcomes, such as our portfolio company HouseRx (tech-enabled specialty pharmacy), as well as other platforms specializing in rebate tracking, new pharmaceutical payment models, financing, and analytics.
Biosecurity will “enter the chat” in 2023. We can now program biology to explore new chemical and biological space that is not found in nature. While we are excited to use this technology for drug discovery applications, it could be seemingly repurposed for other use cases that could threaten society. In March 2022, researchers demonstrated that it took less than six hours for drug-developing AI technology to invent 40,000 potentially lethal molecules. The saying goes, “every action has an equal and opposite reaction,” and so in light of our immense progress using AI in drug discovery, with 18 assets discovered via AI currently in clinical trials (up from zero in 2020), we need to interrogate potential risks posed by these powerful new approaches.
More companies focused on biosecurity will emerge this year and will require interdisciplinary founding teams that marry deep biology and chemistry domain expertise with defense, cybersecurity, government contracting, and machine learning acumen. These companies could take shape as core lab infrastructure, software, hardware, or developer tools, for example. Already, we’ve seen interesting platforms emerge tackling the de-identification of genomic data and biosafety for synthetic biology developers.
Similar to other markets in healthcare and life sciences, opening the market for new technology adoption and venture investment in biosecurity may fall on the back of a regulatory catalyst, which could also materialize in 2023. In October 2022, President Biden signed a memorandum detailing a new National Biosecurity plan, which includes launching an inter-governmental effort across 20 federal agencies to detect, prevent, prepare for, respond to, and recover from biological incidents and foster deep public-private partnerships to implement robust solutions. Simply put, the government announced a biosecurity #RequestForProduct.
While we expect biosecurity technology to become investable for venture capital in 2023, we anticipate it will take many years to develop into a mature technology “category.”
In 2023, we’ll separate hype from reality in techbio. Advancements in computational biology grabbed headlines in 2022, from DeepMind’s AlphaFold to Meta’s ESM Metagenomic Atlas, to the flurry of pre-prints capturing attention on open-access services each week. Yet while some were busy in the lab (of both the wet and dry variety), semantic debates about the differences between biotech and techbio companies captured others. We joined the conversation in good faith, sporting our best “TechBio Mullets,” a term we use at Bessemer to describe companies that look like traditional biotech companies from the outside looking in, but leverage computational approaches on the backend.
Regardless of your position on the great biotech vs. techbio debate, we predict that in 2023, all bio companies will increasingly be held to the standards that matter: whether they can make new effective medicines efficiently and safely. Both techbio and biotech companies will demonstrate these abilities, and as the younger sibling of biotech, we predict techbio platforms will build upon already promising clinical traction this year. During the last two years, the number of AI-discovered assets currently in clinical trials has grown from zero in 2020 to 18 in 2022, and we expect this number to triple by 2025. We also expect greater clarity around how computational biology companies are built and scaled, how their business models evolve, and ultimately, how they should be valued. Meanwhile, traditional biotech companies will continue to integrate artificial intelligence into discovery and development practices.
The new year will bring novel methods, shiny new tools, and ever-greater compute. The companies that make transformative medicines—whether they be “techbio” or “biotech”—will use whatever best-in-class technologies exist at the time. The bio companies that build the infrastructure to trial, adopt, and integrate new technologies as they become available will find the most success.
By 2040, we expect that artificial intelligence will have played a role in the discovery of all drugs approved each year, with a non-trivial percentage of drugs only made discoverable by computational approaches. We don’t think we’ll still be having the same semantic debates by then, either.
This year, we predict that organoid models will flourish. Alternative methods to culture human organs, such as organoid and cell-based models, have been around for a long time, dating back to the early 20th century when researchers began experimenting with culturing cells from various tissues and organs with the first “brain in a dish” developed by the 1980’s. The FDA Modernization Act 2.0 passed in 2022 eliminates the federal mandate for animal testing for new drugs and in exchange, will support the use of alternative models grounded in human biology to demonstrate the safety and effectiveness of new drugs.
This regulation coincides with breakthroughs in systems biology, stem cells, engineered tissues, and computational biology approaches, each of which can improve our ability to predict risk and efficacy of promising new drugs. During the height of the pandemic, human organ chips played an essential role in rapid drug repurposing for COVID-19 and demonstrated an ability to accelerate pre-clinical testing.
Though we don’t expect animal models to disappear overnight, we do predict accelerated development and adoption of organoid and cell-based assays that increasingly resemble derivative tissue and/or organ systems as an alternative to ~60,000 dogs, >100,000 non-human primates, and >100 million mice and rats used in experimental research each year. We are cautiously optimistic these alternative models will increase efficiency in assessing patient safety, reduce the duration of drug interaction studies and associated costs, and ultimately, improve predictions of efficacy in patients.
Every system is perfectly designed to achieve the outcomes it gets. It is perhaps unsurprising that in healthcare, an industry with such historical and stark incentive misalignment, we’ve struggled to produce quality outcomes with consistency or cost efficiency.
For a long time, we’ve evangelized value-based care (VBC) as the solution to these problems. First trialed in the 1990s and formally introduced in the early 2000s, we need to be intellectually honest about where we’ve landed with the model: nearly thirty years since its conceptual inception, only 15% of physicians participate in value-based payment models and another 26% participate in episodic, bundled, or capitation payments. Moreover, of providers participating in value-based payment programs, about one-third report that quality has improved, whereas roughly half cite that financials have improved slightly or stayed the same. There is much work remaining to demonstrate the viability and durability of value-based models in the American healthcare system.
All hope for value-based care is not lost, however. This year, we simply predict more acceptance that VBC is not a monolithic solution to healthcare’s shortcomings, inspiring a renewed interest in alternative structures free from the tyranny of a two-party system: fee-for-service vs. fee-for-value. We expect the industry to accept that not all disease states or specialties are amenable to value-based models, especially those for which we don’t have a deep understanding of the disease etiology and that polychronic patients cannot be managed by five different vendors successfully. We suspect payers and their partners will let the data tell the story, doubling down where VBC models have shown promise such as episode payments in cardiology and orthopedics.
Though value-based care in its current form may not be the silver bullet we hoped for in healthcare, we support continued experimentation with risk-based models that deliver valuable results. There are many green shoots that will sprout fresh leaves in 2023, from new CMS Specialty Care Models to the continued growth of ACO Reach. Greater price transparency and more robust technology for managing value-based care contracts will underpin a growing corpus of learnings. Irrespective of the payment model, we remain compelled by care delivery organizations that can demonstrate economies of scale at a regional level and empower providers to focus on patient care. We’ll remain value-based care optimists, but not absolutists.
If one or more of our 2023 predictions resonated with you, reach out! You can connect with each of us by email: Steve Kraus (email@example.com), Morgan Cheatham (firstname.lastname@example.org), Andrew Hedin (email@example.com), and Sofia Guerra (firstname.lastname@example.org).
You can find our 2022 predictions here.