Lightbeam Health Solutions

Genuine predictive and prescriptive modelling sits inside this platform, and the company bought it rather than built it.

The acquired asset is the model centric part. Jvion's technology combines clinical, socioeconomic, environmental and behavioural data to identify patients on a trajectory toward becoming high risk and to indicate which interventions would alter that trajectory. Moving from prediction to prescription is a real step up in difficulty, because ranking who is at risk requires only a forecast while recommending what to do requires a view about what changes the outcome, and few vendors in this lane attempt it. Deviceless monitoring adds predictive models over patient reported responses collected by text and telephone.

The founding platform is a different kind of work. Aggregating sources, normalising and representing data, mining it for care gaps, stratifying risk and returning findings to the point of care is data engineering, rules and workflow, and it is what the company was before any acquisition. It remains the substrate everything else runs on.

Graded C on the same basis as the other platform records in this lane. That consistency is itself a finding worth stating: across the value based care intelligence category the recurring shape is a data platform with models layered on top, and the differentiating asset is usually the aggregation, the content or the contract knowledge rather than the inference.

What would move this grade is evidence that the prescriptive layer is doing work no rules engine could. Ask what the prescriptive models recommend, how those recommendations are validated, and how the acquired models have been maintained since the purchase.

The autonomy sits at the recommendation layer rather than the action layer, which is the less obvious and more interesting place for it.

The analytics core is advisory in the conventional way. Care gaps, risk stratification and quality findings surface to clinicians and care teams at the point of care, and a person decides what to do.

The prescriptive layer behaves differently in kind. Technology that identifies a patient trajectory and then indicates which interventions would change it is not presenting information for a clinician to interpret, it is proposing a clinical course of action. That is a meaningful step, and the oversight question it raises is whether the recommending logic is inspectable. A care manager told that a particular patient should receive a particular intervention needs to know what drove that conclusion in order to agree or disagree with it, and nothing located describes whether any explanation accompanies a recommendation.

The monitoring product operates with real autonomy at the contact layer, sending scheduled text and telephone check ins and collecting responses without a person initiating each one, which is the design and the source of its efficiency.

What is absent across all three is the escalation account. Nothing located describes what happens when a monitoring response indicates deterioration, who is notified, within what time, or what a clinician can override in the prescriptive layer and whether disagreement is recorded.

Graded C for a clearly bounded advisory core with the recommending and contacting layers unaccounted for.

Ask what explanation accompanies an intervention recommendation, and what the escalation path and response time are for a concerning monitoring response.

Function is described throughout and model behaviour is described nowhere.

The functional account is reasonably clear. The platform aggregates, normalises and represents data from many sources, mines it for care gaps and stratifies risk. The prescriptive layer combines four categories of data to identify trajectories and indicate interventions. The monitoring product uses predictive models over patient reported responses. A reader can follow what each component is for.

Everything about how well any of it works is closed. A dedicated pass located no model card, no accuracy, calibration or validation figure for any predictive or prescriptive function, no description of algorithms, no retraining cadence and no drift monitoring account.

The prescriptive layer is where the omission matters most, and it is a harder question than the usual one. A risk prediction can be validated against what subsequently happened. A recommendation that a particular intervention will change a trajectory is a causal claim, and validating it requires evidence that the intervention worked, not merely that the prediction was accurate. Nothing published describes how such recommendations were derived or whether they have ever been tested against outcomes.

Acquisition history adds a maintenance question. Models bought as operating assets in 2022 were trained on data from before that point, and nothing describes whether they have been retrained since, on whose data, or how performance has been monitored across the transition.

Graded D.

Ask for accuracy figures on risk stratification, the evidence basis for intervention recommendations, and how the acquired models have been maintained.

One infrastructure link is identifiable and the rest of the chain is closed, with an acquisition history that makes the closure harder to accept.

The visible part is the cloud relationship. Marketplace availability and a partner of the year award identify the infrastructure provider with reasonable confidence, so a buyer can reason about where data sits and what concentration risk follows.

Everything else is undisclosed. No sub processor register was located, no subcontractor list, no named third party component, and no identification of the messaging or telephony providers that carry patient outreach. That last omission is concrete rather than theoretical, since text messages and calls to patients pass through carriers and communication platform providers, and a buyer cannot determine who handles clinical content in transit or what those intermediaries retain.

The acquisition history is the distinctive supply chain question here. Three businesses were absorbed over four years, each built by a different company with its own third party components, model dependencies and data licences. Whether those inventories have been consolidated, or even collected, is not described. Models acquired as operating assets carry a further question about what data they were trained on, whether the training data transferred with them, and on what legal basis it may continue to be used.

Graded D.

Ask for the sub processor register, the messaging and telephony providers behind patient outreach, and what data and licences transferred with the acquired models.

The strongest evidence base in this lane so far, and the reason is that most of it was produced by people other than the vendor.

The category ranking is the substantive item. A healthcare research firm ranked the remote monitoring product first in its category in its 2024 annual report, and that firm's rankings are derived from structured interviews with customers rather than from vendor submissions. A first place finish measures delivered experience across a client base, which is a different and harder thing to obtain than an analyst placement based on briefings. Recognition as a major cloud vendor's healthcare and life sciences partner of the year in 2025 and inclusion in a trade publication's annual list of notable companies in 2024 add corroboration from two further independent sources.

Scale is stated in the right units. The company reports that in the 2022 performance year its accountable care organisation clients managed care for 1.1 million patients, which is a lives under management figure attached to a named performance year rather than a cumulative total. A public health department is named as a customer of the monitoring product, which is a verifiable public sector reference.

What holds this below the top grade is the absence of clinical outcome evidence at the same standard. No peer reviewed publication was located. Outcome claims attached to the monitoring programmes, including reductions in avoidable readmissions, appear in acquisition announcements without a denominator, period or comparison group. And the 1.1 million patient figure describes the size of the client base rather than what happened to those patients.

Ask for outcome data by programme with denominators, and whether any result has been externally evaluated.

Credible custodial credentials, an unusually sensitive data set, and no statement about what the models are permitted to learn.

The custodial position is the better half. Certification at the top assurance tier covers the platform, aggregation and normalisation are described as core functions rather than incidental ones, and the company has a named chief data scientist in a public leadership role, which puts identifiable accountability on the data side rather than only the security side.

The data set is what raises the stakes. The prescriptive layer combines clinical, socioeconomic, environmental and behavioural information about individuals, which is a broader and more intimate profile than a claims history. Environmental and behavioural inference about a person who never consented to being profiled that way, and who will never see the resulting score, is a stewardship question distinct from ordinary protected health information handling, and no published material addresses what is inferred, how long it is retained or whether a patient can see or contest it.

The model governance question is unanswered in the usual way. Nothing located states whether customer data trains or tunes models, whether learning is confined within a customer tenancy, or whether models built from one client's population are applied to another. That question has extra force here because the predictive assets were acquired as operating assets from another company, and how data and models transferred in those transactions is not described.

The monitoring channel adds retention questions covering message and call content that no published statement addresses.

Graded C.

Ask what is inferred beyond clinical data, whether learning crosses customer boundaries, and what transferred with the acquired models.

Certification at the top assurance tier, and no contractual detail behind it.

The control evidence is real. The population health platform holds certification under the healthcare assurance framework at its risk based two year level, which maps federal and state regulation, recognised standards and industry requirements onto an externally validated control set tailored to the organisation's risk profile. Business associate status is the correct posture for a processor holding identified clinical, claims and social data on behalf of providers and payers.

What is missing is the agreement. No template business associate agreement, breach notification window, liability cap position, audit rights statement, subcontractor flow down provision or data return clause was located.

The monitoring product carries obligations the platform alone would not. Patients are contacted by text message and telephone and respond with clinical information, which means protected health information travels through consumer channels in both directions. The specific questions are what clinical content appears in an unencrypted message, how consent to that channel is captured and recorded, and how a patient withdraws it. A public health department deployment raises a further question, since population level programmes may reach people who are not established patients of the contracting organisation at all.

Graded C for a strong certification posture with the contractual and channel specifics undisclosed.

Ask for the template agreement, the breach notification window, and how consent for text and telephone contact is captured, recorded and revoked.

Top tier certification with a visible renewal history, and no current attestation surface.

The substance is strong. The platform earned certification under the healthcare assurance framework in January 2022 and renewed at the risk based two year level in February 2024. The risk based tier is the demanding one, assessing a control set tailored to the organisation's own risk profile rather than a fixed baseline, and it is the tier large healthcare buyers normally ask for. A named infrastructure and security executive speaks for the programme publicly.

The renewal history is worth more than either certification alone. A single certification shows a point in time achievement. Progressing from an initial certification to a higher assurance tier two years later shows a maintained programme, and it answers the question a buyer would otherwise have to ask about whether the credential was a one off exercise.

Three things hold this below the top. No audited service organisation control report of either type was located, and that report is the standard artefact a health system procurement process requests alongside certification. The two year certification located dates from February 2024, so it would require renewal by early 2026, and no statement of current standing was found. And no trust centre or portal was located where a buyer could retrieve current attestations, scope statements or testing summaries without asking.

One scope question is specific to this company. The certification is described as covering the population health management platform, and three acquired products now sit alongside it. Nothing states whether they are in scope.

Ask for current certification standing, whether the acquired products are in scope, and any service organisation control report.

Outside device regulation by function, with one branded category a buyer should examine closely for a different reason.

Population analytics, risk stratification, care gap identification, quality reporting and contract modelling are administrative and financial activities rather than diagnosis or treatment, so no clearance is required and none is claimed.

The item worth attention is reimbursement rather than device classification. The monitoring product is branded as deviceless remote patient monitoring. Federal remote physiologic monitoring reimbursement generally requires physiologic data transmitted automatically from a medical device, and a programme built on text messages and telephone calls collecting patient reported responses does not obviously meet that description. The branding is arguably candid, since deviceless states the distinction openly rather than hiding it, and there are other billing pathways for care management and communication based services. The risk is a buyer assuming the familiar reimbursement applies because the familiar phrase appears in the product name. This index records that as a question for the buyer to resolve directly rather than as a finding against the vendor.

Quality reporting carries genuine regulatory weight, since electronic measure submissions determine payment and public reporting under federal programmes, and supporting that migration is a compliance sensitive function even though it is not device regulated.

Graded C.

Ask which billing codes each monitoring programme is designed to support and what documentation the platform produces for them.

Health equity is positioned as a product capability, and no evidence that the models delivering it have been tested for the bias they are meant to address.

The positioning is genuine and worth crediting. The acquired prescriptive analytics were marketed explicitly on social determinants and health equity, and incorporating socioeconomic, environmental and behavioural factors into risk prediction is a recognised approach to catching need that a claims only model misses. A vendor that has made equity a named capability has at least framed the problem.

Framing it is not measuring it. A dedicated pass located no published bias testing, no subgroup performance figures, no fairness review process, no model documentation and no statement of what any model takes as its target variable.

The exposure runs in two directions here and the second is specific to this product. The first is the documented one for the category, where historical cost used as a proxy for health need understates illness in populations who have historically received less care. The second belongs to models that use socioeconomic, environmental and behavioural features directly. Those features are close proxies for race and class, and a model built on them can encode disadvantage as risk in ways that are difficult to distinguish from encoding need. Whether such a model directs additional support toward a deprived population or flags that population as a poor investment depends entirely on the objective it optimises, and that objective is not published.

A model marketed as an equity intervention carries a higher obligation to show subgroup performance than one that makes no such claim, not a lower one.

Ask what the prescriptive models optimise, how socioeconomic features are weighted, and whether performance has been measured across race, language and deprivation groups.

Recourse is undefined at every point where a buyer would need it, and this product generates more distinct failure modes than most records in this lane.

A dedicated pass located no indemnification position, no warranty covering model output, no accuracy guarantee, no service credit regime tied to data quality, availability or outreach delivery, and no described route to dispute an output.

The failure modes divide into three kinds. Financial: a quality measure computed on incompletely ingested data moves a reported score and the shared savings attached to it, and for an accountable care organisation a single performance year can decide continued participation. Clinical: a monitoring programme that fails to escalate a patient reporting deterioration, or a prescriptive recommendation that directs care toward the wrong patients, produces harm to people rather than to budgets. Actuarial: contract modelling and incentive design shape the financial terms an organisation commits to for years, and an error there is not a bad report but a bad contract, discovered long after signature and effectively irreversible.

That third category is unusual in this index and deserves emphasis. Advice that determines the structure of a risk arrangement is closer to professional services than to software output, and professional services carry different liability conventions from software licences. Which convention applies to actuarial and incentive design work delivered through a technology platform is not described anywhere located.

Graded D.

Ask for the indemnification position on quality submissions, whether any warranty attaches to actuarial and contract modelling work, and how liability for an automated patient interaction is allocated.

Aggregation is the founding competence and one concrete piece of standards work backs it up.

The core description has been consistent since the company's earliest public account of itself: pull data from many different sources, normalise it, represent it properly, mine it for care gaps, then return those findings to providers at the point of care. Returning findings into the clinical workflow rather than into a separate portal is the part most platforms in this lane describe least well, and it is stated here as a defining function.

The standards work is the strongest single item. Building capability for accountable care organisations migrating to fully electronic clinical quality measure reporting is demanding interoperability engineering, because electronic measure submission requires structured clinical data in specified formats with defined value sets and it fails on data that merely looks correct. A vendor that has done that work has demonstrated something about data fidelity that no integration count establishes.

Cloud marketplace availability indicates the platform is packaged for deployment into an existing enterprise cloud estate rather than requiring a bespoke arrangement.

What holds it below the top is unstated mechanics. No count of natively supported electronic health record systems, no description of how patient identity is resolved across sources, no account of whether integration is bidirectional, and nothing describing how the acquired products connect to the core platform or whether each carries its own integration path.

Ask how many source systems are supported natively, how identity is resolved, and whether the acquired products share the platform's integrations.

The infrastructure provider is identifiable, and nothing beyond that is stated.

What is established is useful. The care orchestration suite is transactable through a major cloud vendor's marketplace and the company was named that vendor's healthcare and life sciences partner of the year in 2025, which together identify the hosting platform with reasonable confidence and tell a buyer that procurement can run through an existing cloud agreement against committed spend.

The security review questions remain open. Whether customer data occupies a dedicated tenancy or a shared environment with logical separation is not stated. Geographic residency is not addressed and no region selection is described. No recovery objective, availability commitment or failover description was located, for a platform that carries scheduled patient outreach as well as analytics, which means an outage is not only an unavailable dashboard but missed contacts with patients enrolled in a monitoring programme.

The acquired estate raises a question the material does not answer. Three businesses purchased over four years may well retain separately architected infrastructure, and nothing describes whether a customer buying the platform and the monitoring product is buying one deployment or two.

Graded C: infrastructure identifiable, everything a security reviewer would ask about it undisclosed.

Ask whether tenancy is dedicated, where data resides, the recovery objective, and whether the acquired products share infrastructure with the core platform.

Cost is absent from every published surface. A dedicated pass located no pricing page, no unit of charge, no range, no tiering, no implementation fee position and no minimum commitment.

One procurement route is established without a price attached. The care orchestration suite is transactable through a major cloud marketplace, which means an organisation can buy through an existing cloud agreement and draw down committed spend. That is a practical benefit and it discloses the channel rather than the cost.

The unit question is unusually wide here because the product is four businesses. A population health platform plausibly prices per attributed life or per member per month. A monitoring programme plausibly prices per enrolled patient per month or per programme. Contract modelling and actuarial work plausibly price per engagement or per contract analysed. Nothing indicates whether these are sold together, separately, or as tiers, and an organisation buying the platform cannot tell what adding monitoring or actuarial support would cost.

The services element compounds it. The company describes itself as delivering technology and services and the most recent acquisition explicitly added actuarial services, so labour is part of the offering, and no ratio of services to technology in a typical engagement is published.

Ask for the unit of charge per component, how the acquired products price alongside the platform, the implementation fee, and the contract term.

Coverage extends across buyer types and across the lifecycle of a risk contract, which is a less common form of breadth than coverage across clinical conditions.

Buyer types named include accountable care organisations, payers, provider groups, health systems and, through the monitoring product, a county public health department. Public health is a materially different customer from a health system, with population level obligations, no attributed panel and different funding, and serving it indicates the monitoring product travels beyond the value based care setting it was bought for.

The lifecycle coverage is the distinctive part. With the December 2025 acquisition of contract modelling, incentive design and actuarial capability, the company now spans designing a risk arrangement, forecasting its performance, operating against it, monitoring the population inside it and reporting quality out of it. Most competitors in this lane begin after the contract is signed.

Clinical coverage runs through condition specific monitoring programmes and through prescriptive analytics that incorporate socioeconomic, environmental and behavioural factors alongside clinical data. Quality reporting capability supports accountable care organisations moving to fully electronic measure submission.

What holds it at B is depth evidence per area. Breadth assembled through three acquisitions in four years raises a question the published material does not answer about how completely the parts function as one product, and no figure establishes installed base by setting beyond the accountable care population.

Ask how the acquired products integrate with the core platform, and what the installed base looks like outside accountable care.