DoseMeRx

The engine is Bayesian dose optimisation over clinically validated population pharmacokinetic models, updated against a patient's measured drug concentrations and, where available, their genotype. That is quantitative pharmacology rather than machine learning, and the grade describes the mechanism rather than the quality of the software.

The candour is worth crediting explicitly, because it runs against the incentive in this market: the company calls itself a Bayesian dosing platform throughout its own material and does not reach for an artificial intelligence label, where a direct competitor references machine learning on its technology page for what amounts to model cluster assignment.

A buyer evaluating this category should understand that the dose is computed by statistical inference with a long clinical pedigree, and that a vendor saying so plainly is being accurate rather than modest.

This is the strongest oversight design in the category and the reason is what the clinician is shown alongside the answer. The interface displays how well the model actually fits this patient's observed levels rather than presenting a dose as a bare number, tracks serum creatinine on the same page so renal trajectory is visible while the decision is made, and compares the individualised result against guideline recommendations.

Most importantly it presents an Alternative Dosing Matrix showing multiple viable options across both dose amount and interval, any of which can be selected, rather than issuing a single recommendation to accept or reject. A clinician shown a fit diagnostic, an option space and a guideline reference is positioned to disagree with the software, which is the practical form of meaningful oversight and is materially more than a review step in a workflow. The remaining gap is the same one the whole category has: no published figure for how often clinicians adopt, modify or reject the recommendation.

The models are described as clinically validated pharmacokinetic drug models covering 42 drugs, and the platform exposes model fit to the user at the point of care, which is a genuine transparency artifact rather than a document: a clinician can see whether the model describes this patient's data well before acting on what it predicts.

Held below the top grade because the individual population models in the library are not named the way a direct competitor names its Thomson, Goti and Carreno models, so a pharmacist cannot check which published cohort underlies the prior for a given drug without asking.

No calibration data, threshold or validation methodology is published for the library as a whole, and the genotype input, which is the most interesting technical feature here, has no published description of which gene drug pairs are supported or how genotype is weighted against measured concentrations.

The independent evidence is thin and the vendor's strongest claim is uncited. DoseMe was included in the one independent head to head comparison in this segment, published in Pharmacotherapy in 2018, where it was not the least biased of the five programs assessed but was one of two whose accuracy improved when a second serum level was added, which is a clinically useful property.

Beyond that, retrieval located no peer reviewed outcome study, no prospective validation and no published performance data for the model library. Against that absence, acquisition era press repeatedly described the software as producing significant improvements to mortality, risk and patient outcomes, a mortality claim that carries no citation anywhere retrievable and that no located study supports. Deployment scale is real, with more than a thousand clinicians across more than a hundred locations, but scale is not evidence of benefit and this index grades it that way consistently.

This vendor publishes the answer to the question this index has had to record as unanswered for every other dosing vendor in the category. A public data security page states that all patient data entered by customers is owned by those customers, and where applicable by the patient, and then discloses precisely what the vendor takes in return: a licence back over de identified data, with the permitted purposes named as delivering, monitoring and improving the service, improving the models in use, and generating reports and research exports as determined by the data's owner.

Secondary use is therefore disclosed, bounded and framed as a licence from an owner rather than assumed. Set that against the clause elsewhere in this index granting a vendor exclusive ownership of de identified derivatives of a customer's own patient data, and the difference in posture is stark. Supporting controls are described rather than asserted, including encryption, firewalls, full user audit logging in the platform, and encryption of temporarily cached data on mobile devices.

The company also requires that storage facilities be independently audited, naming ISO 27001, the payment card industry data security standard, a service organisation controls report and the HIPAA security standards as the benchmarks it expects them to meet.

The published HIPAA position is the most developed in this category. A Notice of Health Information Practices is published as a HIPAA privacy statement describing how and when protected health information is used or disclosed and the patient rights that attach, and the company states that by implementing HITRUST certification requirements it has addressed every standard and implementation specification in the HIPAA Security Rule, which is a specific and checkable claim rather than a generic assertion of compliance.

A separate page addresses the General Data Protection Regulation for European operations. Held at B for the same reason as most vendors here: the business associate agreement itself is not published, so its breach notification timelines, subcontractor flow down terms and data return obligations on termination are not readable before contact.

HITRUST certification is held and the company states it is the only Bayesian dosing platform to carry it, a competitive claim that is checkable and that none of the directly comparable vendors contradicts. A published data security page describes multi layered controls including physical security, firewalls, encryption, full user audit logging and encrypted mobile caching, and commits to storing data only in facilities that have been independently audited, naming ISO 27001, the payment card industry data security standard, a service organisation controls report and the HIPAA security standards as the benchmarks it expects.

That passes the audit obligation through to the infrastructure layer explicitly, which few vendors bother to state. Held at B rather than higher because the company holds no service organisation controls report of its own, operates no trust centre where documents can be requested or reviewed, and publishes no penetration testing cadence and no vulnerability disclosure programme. The certification is strong and the surrounding assurance apparatus is thinner than the certification implies.

No United States clearance was located and none is claimed, consistent with the position taken across therapeutic drug monitoring software generally and supported by the pharmacist mediated workflow. The open question is broader here than for the American vendors in this category, because this product is sold across the United States, Europe, Asia Pacific, Africa and South America, with a named hospital customer in Portugal, and two retrieval passes located no CE mark, no United Kingdom conformity marking, no Australian Therapeutic Goods Administration listing and no other national designation.

A direct competitor selling the same function into European hospitals holds a CE mark, so the divergence is real rather than a documentation gap in the index. A buyer outside the United States should establish the regulatory basis on which the software is supplied in their own jurisdiction, since that answer is not published.

Retrieval located no performance reporting by age, sex, race, ethnicity, renal function band or body habitus, no bias assessment, no model governance documentation and no post deployment monitoring statement. The exposure is structural to Bayesian dosing rather than particular to this vendor: the prior comes from a published derivation cohort, so a patient unlike that cohort begins from a worse estimate, and the correction depends entirely on how many concentration measurements are drawn and when.

That interacts with access to care, because sicker and better resourced settings draw more levels. This vendor is better placed than most to say something useful about it, since its published licence back over de identified data explicitly names improving the models in use as a permitted purpose, which means the data to characterise model performance across populations is already being collected.

Integration with major hospital systems including Epic and Cerner is long established and predates the current ownership, and the company describes integration with the widest range of hospital systems in its segment. The most consequential integration is with a competitor's product: in October 2025 Inovalon announced a collaboration bringing this precision dosing capability into its pharmacy surveillance workflows, and Inovalon owns VigiLanz, so the software now reaches clinicians inside a clinical surveillance platform as well as directly.

Held at B because no FHIR conformance statement was located, a third party product directory reports that no public API is offered, and the specific electronic health record marketplace certifications that competitors publish were not found.

This is the most specific data residency disclosure in the category and one of the clearer ones in this index. The published HIPAA privacy statement names both the cloud provider and the physical location: United States healthcare data is stored in Microsoft Azure in an Illinois based facility.

The privacy policy adds a general commitment to taking all reasonable steps to store data on shore, and states what happens when an overseas transfer is required to deliver a service, including an obligation on overseas recipients under the Australian Privacy Principles, which reflects the product's Australian origin and is unusual to see spelled out. The platform is cloud delivered with no on premise option described. A privacy office evaluating this vendor can answer its residency question from the website, which is not true of any other dosing vendor graded here.

The price is undisclosed and two concrete commercial terms are published, which puts this vendor alongside the better disclosers in a category that mostly publishes nothing. A fourteen day free trial covers the full platform including all drug packages, so an evaluating pharmacy is not restricted to a limited module set.

More useful is a partnership with the American Society of Health System Pharmacists aimed at small and rural hospitals, which bundles that society's competency management subscription with unlimited use of the platform for up to ten users at a single site, and which is stated to be open to non members. That publishes a packaging unit, the site and the user count, even without publishing what it costs.

Graded C rather than higher because no price, contract term or renewal structure is disclosed for either the bundle or the standard product, and it is not stated whether the trial converts automatically.

Geographic breadth is the widest in this category by a clear margin, with more than a thousand clinicians across more than a hundred locations spanning the United States, Europe, Asia Pacific, Africa and South America, and a named academic hospital customer in Portugal alongside American health systems.

Setting breadth is also genuine and structurally varied: acute hospitals, infusion service centres, a programme of all inclusive care for the elderly, and small and rural hospitals reached through the pharmacists society bundle, which are four different staffing and purchasing models rather than four logos. Drug scope is stated at 42 agents, concentrated on high risk parenteral medications with therapeutic drug monitoring, so the function remains narrow even where the footprint is wide. Held at B because no paediatric validation or indication was located and the drug library is not individually enumerated in public material.