Mosai

Medalogix merged with Forcura and the combined company now sells as Mosai, so the models are no longer the whole purchase. Mosai's clinical management products are built on them: the Muse page says it uses machine learning on structured and unstructured record data and historical hospice cases to update each patient's risk after every visit, and Pulse and Transitions score hospitalisation risk and the timing of a move to hospice. That module does not exist without the models.

The platform around it does. Mosai also sells referral intake with AI data extraction, OASIS documentation review, physician order management and e-signature, the workflow side Forcura sold before the merger, and those carry value on their own. Remove the models and a buyer still has a care coordination platform, which is the B band rather than A.

The models rank and recommend; the care team decides. Muse offers a risk ranked census, guidance on visit frequency, staffing mix and care intensity, and support for the interdisciplinary group meeting that hospice rules already require, focusing it on patients whose care plan may need updating. Nothing on the Muse, home health or home pages says anything is ordered or changed automatically.

Held at B because one part is missing: nothing published says what a clinician is shown about the confidence of a prediction or what happens when their judgement disagrees, and a team told a patient is likely near death will act on it. A boundary worth watching is now inside the product: Transitions identifies home health patients who may benefit from moving to hospice sooner. Hospice eligibility is certified by physicians and closely enforced, so ask how a Transitions signal reaches the certifying physician and whether it is ever used in admission decisions.

The company has stated the number of data points compared per hospice patient at over 800 and the scale behind the home health product at more than 11 million visits, and names the input types: the federally mandated assessment, vital signs, medications and clinician narrative notes. The Muse page adds that it uses machine learning on structured and unstructured record data alongside historical hospice cases. Naming narrative notes matters, because it says the model reads free text rather than only structured fields.

No model version, release history or update practice is published, and the Muse page carries no accuracy figure: no confusion matrix, false positive rate or definition of the prediction window. On a prediction of death within a short window the base rate matters enormously, so any accuracy figure a buyer is given needs all three.

The architecture is described and no model provider is named. The products use machine learning on the agency's structured and unstructured record data, compared against historical hospice cases; nothing says whether any third party model is involved. The parties that are visible are the record system partners data moves through (Homecare Homebase, MatrixCare, Axxess) and the unnamed third parties the privacy notice says receive personal information, most storing it in the United States and some elsewhere. No subprocessor list is published.

The merger adds a party question: two companies' platforms now sit under one name, and nothing says whether patient data from the clinical management products and from the workflow products is held together. Ask for the subprocessor list, whether any model is sourced from a third party, and the retention and training terms that apply once data arrives.

The figures on the estate are vendor reported. The Muse page reports a 36 percent improvement in Service Intensity Add-on visits over baseline, 47 to 50 percent more visits in the last week of life and 82 percent more visits than the industry average, with a named BAYADA hospice administrator quoted; the home page says the top ten US home healthcare providers trust Mosai and reports over 90 percent user satisfaction. A Compassus executive is quoted on fewer hospitalisations with no figure. None has a method, denominator, period or comparison group.

The headline claim is that patients receive more visits at the end of life, and the visit figures speak to it. What nothing published answers is whether the predictions are accurate, or whether the added visits reach the patients who were in fact dying. Ask for the accuracy of the prediction against observed deaths, and for the visit figures with baseline, period and agency count.

The privacy notice covers the Mosai services, including the clinical management products, and says de identified information is not treated as personal information and that Mosai does not attempt to re identify it. The Muse page calls the software fully HIPAA compliant. Neither says how long patient records are kept, whether they are de identified before modelling, or whether they train the models.

The training question is the material one. Muse compares each patient against historical hospice cases, and that comparison base is built from agencies' records; whether one agency's data improves the models sold to its competitors is not published. The data describes the last weeks of people's lives, gathered in their homes. How an erroneous prediction is caught before it changes a visit schedule is not described either. Ask for retention by data type, the training position on agency data and whether it can be excluded, and de identification practice.

A status is stated but not supported. The Muse page calls the software fully HIPAA compliant and the home health page says it adheres to HIPAA, but no business associate statement or agreement is published. The privacy notice covers the Mosai services, including the clinical management products, yet addresses the personal information of users and never mentions protected health information or a business associate role.

Distribution adds a question: the products are also delivered through record system partners, so an agency may contract through its record vendor rather than with Mosai directly. Ask for the business associate agreement, which Mosai entity signs it, and who holds the agreement when the product is bought through a partner.

Mosai's own pages claim no certification: the home page says it aligns with HIPAA, AICPA SOC and HITRUST standards, and the Muse page says it adheres to HITRUST security standards. The attestation behind the platform comes from the Forcura side of the merger. Forcura announced HITRUST r2 certification in October 2024 for a second two year term, listing MyForcura, Forcura IQ, Forcura Circle and Forcura Analytics; Circle is still sold under the Mosai name.

That is a recognised certification with two questions a buyer has to raise. Scope: the listing predates the merger and names only Forcura's products, so it does not show that Pulse, Muse or Transitions are covered, or how the renamed products map to the listed ones. Renewal: a two year term from October 2024 runs out around October 2026. Ask which Mosai products sit inside the current HITRUST r2 scope, whether it has been renewed under the Mosai name, and whether any SOC 2 report exists.

No device authorisation is claimed and no regulatory position is stated. What the products do is rank and recommend: Pulse scores hospitalisation risk and utilisation, Transitions signals when a move to hospice may be needed, and Muse updates each hospice patient's risk after every visit (the home page lists mortality among the predictions) and recommends visit frequency, staffing mix and care intensity. Those are predictions of utilisation, eligibility and decline feeding decisions about visits and staffing, which is the population health shape graded C.

FLAG: two phrases sit near the edge and a buyer should ask where they stop: Pulse adapts care plans, and Muse guides care intensity. If either changes treatment rather than visit scheduling, the absence of any stated regulatory position becomes the issue. Hospice eligibility and length of stay are an area of active federal enforcement, which bears on how Transitions is used.

Nothing published describes how the models are evaluated, monitored or checked for uneven performance. The Muse page speaks of transparent, evidence based care planning; that describes the care plan, not the model.

The absence matters because the output allocates people. A hospice has a fixed number of nurses and social workers, and Muse's risk ranking and visit recommendations help decide whose last week gets the extra visits. A model learns the common course of decline, so the patients most likely to be missed are those whose course is atypical, such as non cancer diagnoses like dementia or heart failure, or whose records are thin because they were admitted late. Those groups are identifiable from the model's own inputs and no result for them is published. Transitions raises the same question at the other end, since a prediction that a home health patient belongs in hospice can move them out of curative care. Ask for performance by diagnosis group and length of stay, how the models are monitored after release, and what families are told about a prediction.

Nothing published says what happens when a prediction is wrong: no warranty, no remedy, and no error rate a buyer could hold the company to. The Muse product page publishes visit figures but no accuracy figure, false positive rate or definition of the prediction window.

The errors run both ways and neither is recoverable. A false alarm tells a team, and perhaps a family, to prepare for a death that does not come; a miss means the extra visits are not scheduled for one that does. Transitions adds a third: a home health patient flagged for hospice too early. Ask for the false positive and false negative rates at the operating threshold, the prediction window, what a clinician is shown about the confidence of an individual prediction, and what the contract says when a prediction misleads.

Mosai states integrations with Homecare Homebase, MatrixCare and Axxess, keeps partner pages for the first two, and its hospice page offers syncing critical data to the agency's record in one click. Named systems with data moving between them is the B band.

Held below A because neither the home page nor the hospice page sets out the direction and depth of each connection: whether scores are written back into the record, which fields are read, and whether every module runs on all three systems. Ask for the integration specification per record system and which modules run in production on each.

The privacy notice says the services, including the clinical management products, are controlled, operated and administered entirely within the United States, and that most third parties Mosai shares personal information with store it in the United States, though some may store it elsewhere. No hosting provider, region or isolation model is named, and whether patient data is ever held by one of those third parties outside the country is not stated.

Delivery also runs through the record systems, with data moving between Mosai and Homecare Homebase, MatrixCare or Axxess. Ask where patient data rests, which third parties hold it, and whether any of it sits outside the United States.

No price is published, but how the product is sold is discoverable. Mosai sells three product families (referrals and admissions, clinical management, networked care) to home health agencies, hospices and physician practices, integrated with Homecare Homebase, MatrixCare and Axxess; Pulse, Muse and Transitions sit in clinical management. The unit of charge is not stated.

The unit matters in a sector paid largely per episode or per day. Pulse is sold on fewer hospitalisations and tighter episode management, while Muse drives more visits at the end of life and is marketed partly on Service Intensity Add-on reimbursement for them. Ask how each module is priced (per patient, per census, per agency), whether pricing differs when bought through a record system partner, and whether the vendor shares in either direction.

Confined to home based care and complete within it. Pulse covers home health episodes and hospitalisation risk, Transitions the move from home health to hospice, and Muse hospice care through the final days of life. Since the merger Mosai also sells order management and e-signature to physician practices, but the predictive products stay within home health and hospice.

Reach is national and spans agency sizes, with the company stating that the top ten US home healthcare providers trust it, and availability to smaller agencies through the record platforms. Specialty coverage follows the population rather than a discipline, since home health and hospice patients arrive with every diagnosis. Nothing extends to facility based care or the hospital.