What you’re forecasting from matters as much as how you’re forecasting. Reading de-identified patient-level claims data throughout the quarter changes the calculus for healthcare investors.
Every healthcare investor is forecasting. The question is what they’re forecasting from.
Most are working from the same inputs: prescription volume trends, earnings guidance, analyst consensus. These inputs are late. They’re also shared – every investor is looking at the same numbers at the same time. By the time volume confirms a thesis in healthcare, the move is largely priced in. The investors who consistently get healthcare right are not just running better models. They are starting from better observations.
Alternative data in healthcare is not new. What’s still underutilized is the depth of what de-identified patient-level longitudinal data can reveal: the details of what comes after prescriptions are made, across millions of patient journeys, with medical and pharmacy benefits stitched together, linked across the full care pathway and built to be timely and nationally representative.
Volume Is a Conclusion, Not a Signal
Prescription volume is the output of a chain of events that happens well before a script is written: a patient gets diagnosed, a physician decides on a treatment path, a payer approves or denies coverage, the patient fills, or doesn’t, stays on therapy, or doesn’t. Each step leaves a trace in patient-level claims data, visible earlier in the chain than the volume numbers that eventually show up in earnings or syndicated feeds.
That logic holds beyond prescriptions, too. A procedure volume, a device’s share of cases, a health plan’s MLR: Each is the downstream sum of upstream patient events. The claims trail sits upstream of all of them, most of it currently invisible to anyone not built to trace it end to end.
Investors who are looking at volume are reading the last chapter. The signal shows up earlier: intra-quarter shifts in treatment initiation, changes in persistence and adherence, movement in line of therapy. These are the mechanical inputs to the volume number, more so than soft indicators.
The Inputs That Actually Build Conviction
The questions that determine whether a healthcare thesis holds up are answered at the patient level – across the full patient journey, over time. Final prescription volume alone does not answer them.
- Is uptake inflecting – and can you see it before the print? The single most time-sensitive signal in claims data is the uptake curve: how fast an asset is being adopted – a drug into its eligible population, a device into its addressable procedures, a diagnostic into its screening funnel – into its eligible population, month by month, and whether the slope is accelerating, plateauing, or breaking. This is intra-quarter visibility into the trend that becomes next quarter’s volume number – an early read on inflection points before they surface in consensus or syndicated feeds.
- Is this patient population staying on therapy? Persistence and adherence patterns in longitudinal claims data give you an early read on whether a drug’s real-world performance matches trial expectations and whether volume is durable or about to erode
- Who’s winning share – and where in the treatment sequence? Share dynamics tell you whether an asset is gaining ground, and whether the category itself is expanding or two regimens are simply trading patients. Line-of-therapy data tells you the why beneath the share number – first-line capture, refractory pickup, or displacement to later lines as new entrants move in. Together, they set the ceiling on volume growth and reveal whether a position is durable or exposed to the next launch.
- How big is the treatable population, really? Diagnosed prevalence built from actual claims – not just modeled from epidemiology – gives you the real addressable population for an asset, how fast it’s growing as diagnosis expands, and how much of it a drug has yet to reach. Adjusted to a national denominator, it turns a fuzzy TAM assumption into a countable patient base you can build a penetration curve against.
- What are the ripple effects? Healthcare assets do not operate in isolation. A drug that displaces a device category, reduces a procedure volume, or shifts the comorbidity treatment landscape creates second- and third-order effects that move through the market before they surface in any single asset’s numbers or sell side models.
These are the inputs for a defensible healthcare thesis that most investors are not yet systematically building from today.
What Deeper and Earlier Observation Actually Looks Like
Consider what happened when tirzepatide received FDA approval for obstructive sleep apnea in late 2024. The investment question was what tirzepatide’s growth would mean for adjacent device markets — specifically CPAP manufacturers.
Patient-level claims data from the Komodo Healthcare Map® surfaced a specific pattern in the months following approval: Among OSA patients with no prior CPAP use, only 4% of those who initiated tirzepatide went on to start CPAP therapy within six months. Among comparable patients without GLP-1 exposure, it was 22% — an 83% relative difference in observed CPAP initiation rates between cohorts.
That is a treatment-behavior signal with direct implications for device category demand, and the kind of cross-market dynamic that surfaces in claims data as it unfolds, rather than only in retrospective studies published after the quarter closes.
From Observation to Your Own Thesis
This is where Marmot™, Komodo’s healthcare analytics engine, comes in. The Healthcare Map (including both observed events and projected volume) is the asset; and Marmot is how you put it to work,
Ask a question in plain terms, like, “Among newly diagnosed MASH patients in the last two quarters, what share of new drug starts is going to incumbent Rezdiffra versus semaglutide entrant, and how do they split by prescriber specialty and payer channel?” — then build the cohort, the funnel, and the answer yourself. That’s the difference between reading a real-world data brief and producing the one that informs your position. Komodo Signals shows what the data can reveal. The Healthcare Map and Marmot let you reveal what matters to you, before it shows up in volume.
Komodo Health publishes Komodo Signals, a series of real-world data briefs surfacing patient behavior and treatment pattern observations relevant to investment research. The GLP-1 and CPAP analysis referenced above is available here.
Coming Next
In upcoming Komodo Signals briefs and research, we’ll be looking at biosimilar adoption curves across Humira and Stelara, and what the pattern implies for Enbrel ahead. We’ll map second- and third-order effects of GLP-1 drugs on procedures, other drug categories, and portfolio targets. We’ll trace which insurance cost trends are driving Medicare Advantage expenses and when they show up in claims before carriers’ earnings.
To see more articles like this, follow Komodo Health on LinkedIn, YouTube, or X, and visit our Resources Hub.
