Pearl Health

AI powered patient prioritization is the platform engine: predictive insights identify the highest urgency patients before costs escalate. The product is a hybrid, predictive software wrapped in risk bearing value based care enablement services, and the AI platform cannot be purchased separately from the enablement relationship.

Advisory by design: predictions and insights surface to care teams who act on them, and workflow automation covers scheduling and post discharge follow up. No formal oversight or governance documentation retrieved beyond this implied posture.

The functions are clearly described and the models beneath them are not. The technology page says an AI powered algorithm pinpoints the highest urgency patients from claims, ADT feeds and record system data, that each signal carries a suggested next step such as scheduling a follow up or addressing a diagnosis gap, and that routine tasks are automated; actuarial and data science teams match provider cohorts to Medicare models and risk levels. A Signal-Action Framework Report describing the algorithm is offered on the technology page behind a request form.

No model, features, training data, validation design, performance figure or version practice is published on the home, technology, physicians, privacy or terms pages; the methodology report is the right artifact and is gated. A prioritized list is only actionable if the clinician can see why a patient is on it, and nothing published describes what explanation accompanies a prediction at the point of use.

The architecture is described and no model provider is named. The technology page says the platform continuously analyses claims, ADT feeds and record system data to surface signals, each with a suggested next step and automated routine tasks; the privacy policy names data hosting and storage partners, analytics and advertising providers, technology services and data security advisors as categories of recipient for website data, and says it does not apply to the secure platform customers log into. No model, cloud host or subprocessor for the platform is named on the home, technology, physicians, privacy or terms pages.

The data flowing through that chain is Medicare claims history at population scale, ADT events and record extracts from practices across more than 40 states, and the beneficiaries whose data it is were attributed to the arrangement through claims rather than choosing it. Who else processes that data, and under what terms, is not a question a practice can answer from the site.

Named customers and vendor figures with no method. The home page reports a 1 to 3 percent medical loss ratio improvement in year one, a 25 percent potential workload reduction, $3.6 billion of healthcare premiums managed and twofold year on year growth in lives covered, under a customer results heading; the technology page lists results themes including same practice utilization decreases and program quality improvement. Customers quoted by name include BayCare Health System, Holzer Health System, Santé Physicians, Hudson Headwaters, PraxisCare, MDX Hawaii, Virginia Care Partners, BlueSky MD and Odibo Medical Group.

The company is the claimed source, and no baseline, period, population or method accompanies the figures; the quotes describe the partnership rather than measured outcomes. Scale of use is not evidence of benefit. The Medicare programs Pearl participates in publish performance results by ACO, an independent source the savings claims can be checked against.

A general commitment to protecting patient data, with nothing on what the platform does with it. The company's HITRUST announcement says the certification validates the security and privacy measures protecting the patient data entrusted to it by customers and partners; the privacy policy says it may anonymize and aggregate information for analytics and reporting but covers only the website and excludes the secure platform. No retention period, training use statement or de identification position for platform data appears on the home, technology, physicians, privacy or terms pages.

The holdings are Medicare claims history at population scale, longitudinal and identifiable, for beneficiaries who did not choose this company: alignment to an accountable care arrangement happens through claims attribution rather than enrollment, so many will not know the relationship exists. Nothing states whether their data develops the prioritization models, whether improvements from one practice network serve another, or what happens when a practice leaves. No safety engineering is published either: how a wrongly ranked or missed patient is detected and handled.

HIPAA is acknowledged and nothing published supports a position on it. The terms of use, issued by Latitude Health, Inc. doing business as Pearl Health and last updated 21 October 2024, say protected health information provided in connection with the services is also governed by the applicable Notice of Privacy Practices; the privacy policy, issued by Pearl Health, Inc., covers the website and states that it does not apply to the secure platform customers log into. No business associate statement, agreement, subcontractor position or breach notification term is published. The HITRUST r2 certification of the Pearl Health Application assesses security controls.

The structure makes the instrument matter. The company provides administrative and management services to physician practices, which is business associate territory, and also bears risk alongside them in Medicare arrangements, where beneficiary data is used under program rules rather than a business associate agreement alone. Nothing published says which governs a given data flow, or which of the two named entities signs.

A demanding certification is named for the product, without the artifact. The company's March 2025 announcement on its own blog says the Pearl Health Application earned HITRUST r2 certification for information security, the framework's risk based tier, assessed by an authorized external assessor with quality review by HITRUST on a two year cycle. The announcement names the application as the certified scope, so what is covered is stated.

There is no trust portal, letter or downloadable report on the site, so the certification cannot be checked from the site itself. No penetration testing or vulnerability disclosure statement is published, and on its stated cycle the certification expires within 2027. The business spans the application, actuarial and data science teams and risk bearing operations, so whether data handled outside the application sits inside the certified scope is not stated.

No device position is needed for what is marketed, and the regime that governs is Medicare program law. The platform ranks patients by urgency for outreach and suggests next steps such as scheduling follow ups and addressing diagnosis gaps; the terms say the company is not a medical group, does not provide medical advice, care or treatment, and provides administrative and management services to physician owned practices. Ranking for outreach is population health work, and surfacing a possible diagnosis gap for the practice's clinician to confirm is documentation, not diagnosis. Answering to program rules is not a regulatory pathway.

Those rules are substantive: ACO REACH and MSSP participation terms define beneficiary alignment, quality reporting, benchmarks and the settlement of shared savings and losses; risk adjustment carries data validation audits of whether submitted diagnoses are supported; and because the company bears risk, program integrity rules on inducements and steering apply to how it engages practices and beneficiaries. The company's own blog argues against documentation driven risk adjustment, a position that bears on the diagnosis gap prompts.

No evaluation or subgroup analysis is published, and the models make two consequential kinds of decision. The first is clinical prioritization: an algorithm ranks which patients a care team reaches first across a reported 250,000 Medicare beneficiaries. If it learns from historical utilization it reflects who has been able to access care, so patients who used less care because of distance, cost, language or mistrust generate a thinner signal and can rank lower, reproducing the gap while appearing to optimize.

The second is financial. The technology page says the company's data science and actuarial teams match provider cohorts to the Medicare model and risk level that boosts savings and unlocks revenue, and the company bears risk alongside those practices, so it has a direct interest in the outcome of decisions its models inform. That is a legitimate structure, and nothing published says how patient interest is represented in either optimization. The company's blog carries serious methodological writing on risk adjustment and sepsis criteria. It is the nearest thing to governance of its own models that the company publishes, and it is not a governance framework.

Nothing published on what happens when a prediction is wrong, from the contract's side. The terms of use provide the platform as is, disclaim any warranty that it is error free or that defects will be corrected, disclaim any guarantee of the correctness or accuracy of its content, and exclude liability for damages of any kind arising from it. No accuracy commitment, remediation obligation, published error rate or route for a practice or patient to challenge a priority ranking appears on the home, technology, physicians, privacy or terms pages, and nothing describes what a clinician is shown to check why a patient was ranked.

The company's commercial structure does not change that. Pearl bears financial risk alongside its practices: it offers implementation advances, predictable monthly payments and incentives tied to preventive care, and pools practices into risk arrangements across ACO REACH, MSSP and Medicare Advantage. That downside runs to the Medicare program against a benchmark and is shared with the practices in the pool, so a practice pays part of the cost when Pearl's prioritization misses, and the position is rewarded on lower total spend rather than on a correct ranking. It answers for cost, not for the output.

Workflow fit is claimed and no integration is named. The technology page calls the platform EHR agnostic and says it fits visit preparation, care management and post discharge follow up; the home page says it connects with a practice's existing record data and workflows; the platform analyses claims, ADT feeds and record system data, and the physicians page says next steps surface directly in the practice's workflow. No record system, interoperability standard, interface documentation or ADT source is named on the home, technology, physicians, privacy or terms pages, and nothing says whether insights appear inside the chart or in Pearl's own application.

Claims and alignment files supplied through Medicare programs let a platform serve thousands of practices without an integration project at each, which is a real scaling advantage with a known limit: claims lag by weeks or months. ADT feeds narrow that gap for admissions and discharges, and how far they reach across the network matters.

Nothing published about where the platform runs or where its data rests. The privacy policy names data hosting and storage partners only as a category, for website data, and excludes the secure platform from its scope; the home page says the platform is designed to meet the highest data security standards. No hosting provider, region, tenancy model or residency commitment appears on the home, technology, physicians, privacy or terms pages. The HITRUST r2 certification of the Pearl Health Application covers infrastructure controls within its assessed scope, but a certification announcement is not a residency or tenancy statement.

Two questions are specific here. Tenancy separation matters because the platform serves practices and health systems that compete in the same markets, and it holds their panel cost and utilization performance. Retention matters because accountable care is settled retrospectively over years, with reconciliation and audit long after the care, so what happens to data when those obligations end, or when a practice leaves, needs to be written down.

The deal structure is set out in detail, with no price attached. The physicians page says Pearl designs a financial structure with each practice, through risk contracts, technology adoption or both: it provides implementation advances, predictable monthly payments and incentives tied to preventive care, pools practices with others that meet its performance standards to share risk, and matches each practice to the Medicare program that fits, across ACO REACH, MSSP and Medicare Advantage. From the site alone it is clear how the relationship works and that economics are tied to attributed populations and performance.

What drives the cost is the unstated part: what share of shared savings Pearl keeps, how losses are allocated within a risk pool, whether advances are repaid from savings, and what a technology only arrangement costs.

The core coverage is named and validated, and one line is newer than its evidence. The scope is Medicare primary care: health systems, physician networks and independent practices in ACO REACH, MSSP and Medicare Advantage, with the company reporting 250,000 beneficiaries, 10,000 providers and more than 40 states. The traditional Medicare side is supported by named customers quoted on the home and technology pages, among them BayCare Health System, Holzer Health System, Santé Physicians, Hudson Headwaters, MDX Hawaii and Virginia Care Partners, most describing ACO work.

Medicare Advantage is presented as extending that success across a practice's full Medicare population, and no customer or result on the home, technology, physicians, privacy or terms pages is specific to it. That is the part the validation does not yet reach.