A New Moment in Healthcare; Reducing Complexity with AI

We’ve arrived at a rare frontier moment for healthcare today. LLMs have created key leverage points to chip away at some of the industry’s biggest logjams. As if suddenly gaining the ability to speak a new language, we’re confronted with a level of legibility, utility, and fluency in healthcare data (unstructured and structured) that has never been previously available. Let’s explore why Costanoa believes there’s no better time to be building in healthcare – beyond complexity – and where we’re investing.

Healthcare Is Complex Like No Other Category

Healthcare, an industry steeped in risk aversion, has historically been highly resistant to change. For good reason. The consequences of getting it wrong are severe in human terms, unlike in other sectors where rapid experimentation might be tolerated (or preferred).

One salient example is the promising Value Based Care (VBC) movement. Operationalizing VBC is difficult. Stakeholders struggle to establish agreed-upon “high quality” metrics for fear of unintended consequences of gaming performance or under-delivering on care. Further market distortion abounds in healthcare payments, such as Pharma Benefit Manager pricing mechanisms creating disparities in cash payments and insurance co-pays. 

What makes this moment in history different? An ability to ingest, understand, and link fragmented structured and unstructured data with AI. Software can now tackle operationally intensive and detail oriented tasks across critical healthcare processes: documentation, coding, billing, data management and more, and start automating larger chunks of bottlenecked workflows. 

Butterfly Effects With AI: Why They Matter in Healthcare

These seemingly small operational changes have far reaching consequences. A single accelerated clinical trial could bring life-saving treatments to market years sooner. Increased administrative automation can re-direct countless hours of repetitive paperwork back to elevating patient care, where medicine’s true impact is realized. 

At Costanoa, we believe increased efficiency in healthcare operations can considerably improve downstream care delivery and scientific discovery. Here are three themes that can have outsized impact:  

Theme 1: Next Generation Real World Data Management 

The global healthcare analytics market size speeds toward ~$70Bn USD by 2030. The richness and increasing availability of real world data (RWD), and consequently real world evidence (RWE), has already unlocked profound impact across pharma R&D, regulatory process, long term follow-up, and contracting/pricing. These are big ticket problems. Over the next three to five years, an average top-20 pharmaceutical company could unlock more than $300M/year by adopting RWE (Mckinsey).

What’s next? AI improves ingestion, speed, and data normalization. Integration of various sources of unstructured RWD and proprietary data can accelerate drug discovery, enhance population health management, and improve clinical decision-making. 

New ways to capture, appropriately orchestrate, clean and ensure data quality will be more important than ever. Here are some areas we’re excited about and examples of companies:

• Transformation of Data into Research Grade Compliant Assets: novel methods of evidence generation to produce fit for purpose regulatory grade assets (Crescendo Health; Atropos Health)
• RWD Marketplaces: comprehensive data aggregation including images and unstructured sources like clinical notes to train healthcare models (Dandelion Health)
• Data Cleaning, Curation, and Normalization: cleaning RWD effectively in less time and annotating data to make it queryable (Cornerstone AI; Gradient Health)
• Data Accessibility: expanded access to patient clinical histories to inform treatment decisions (Metriport)

Theme 2: Purpose-Built Fintech for Healthcare

Vertical fintech is required for the healthcare industry. A lack of integration between clinical systems of record and financial revenue cycle management systems leads to black-box payments and delayed reimbursement. Further, the increased consumerization of healthcare and greater employee “skin in the game” drives a need for modern consumer-grade healthcare finance products. 

Payments are tied with risk. When risk is shifted at decision points, this challenges the system to embed new capabilities into financial flows. We’re betting in new types of health insurance, ways to modulate and increase flexibility in plan design, and patient payment and financing products, such as:

• Revenue Cycle Management: handling coding, billing, and claims submissions through AI RCM (Akasa for health systems; Zentist for DSOs).
• Benefits Operations: building modern benefits administration for enrolment (Noyo)
• Consumer Directed Healthcare Payments: full-stack digital solutions for flexible consumer spending like HSA management (Lively) and patient billing (Ecton.io; PayZen)
• Revenue Reporting and Accounting: producing audit ready financials and reporting suites for providers (Level Health; Flychain)
• Modern Health Insurance: AI-native approaches to reimagine Third Party Administrator processes (Avant), insurance underwriting (Arlo), and operating customizable health plans (Yuzu).
• Cost and Pricing Intelligence: optimizing cost management strategies for payers (Curafi; Serif)

Theme 3: Supercharged Operational Workflows with AI 

While clinical decisioning is high stakes and will always need a high percentage of experts in the loop, operational administrative workflows can be 10x more efficient. 

Automating repetitive manual documentation could save hours of skilled worker time: think of the nurses working overtime on OASIS documentation or pharma teams spending hours collating toxicology reports and protocols. 

AI-powered workflows make care cheaper, more accessible, and more efficient

Care isn’t just about the quality of provider, but backend steps prior to and after an episode of care such as prior authorization, proper reimbursement, and access to physicians in-network. Places where there’s need for specialized solutions include: 

• Network Design and Optimization: building data-driven provider networks to improve compliance, cost, and adequacy (J2 Health).
• Documentation Management: generating specialized documentation such as home health scribe and OASIS (Enzo Health; Andy AI) or prior authorization submissions for payers and providers (Basys.ai; Latent).
• Care Coordination and Patient Engagement: handling critical voice workflows such as referral management and pharmacy ordering for virtual care (Parakeet Health; Kairo Health)
• Services Infrastructure: providing an infrastructure backbone to support extended care such as pharmacy-as-a-service through API (Foundation Health).

AI-powered workflows accelerate Pharma R&D and commercialization

Scientific discovery is a complex interaction of many processes, some of which include scientists working efficiently together, well operated clinical trials, and regulatory submissions completed accurately. Yet many of these areas are mainly manual, highly prone to error, and detail intensive with regulatory scrutiny. These startups already make several processes faster and more accurate, but there’s room for more:

• Manufacturing and Supply Chain: accelerating data analysis for biopharma manufacturing (Fathom).
• Clinical Trial Management and Operations: monitoring trial performance and providing real-time transparency in clinical trials through companies (StudyOS; Miracle)
• Regulatory Submissions: learning from precise regulatory language to create and manage AI-Powered documents (Artos AI)
• Commercial Operations: performing data analysis for drug launches (Cellbyte)

We’re looking for startups creating ambitious software in these spaces. If that’s you, reach out to Nicole or Mark at Costanoa via LinkedIn. We can share more about our belief that a new moment in healthcare has begun.