TORTUS Health, Nolla Health and Doctronic illustrate different routes to AI-assisted care, not three interchangeable versions of an autonomous GP. In the October 2026 materials, TORTUS Health describes clinician-authorised virtual care, Nolla a bounded initial acne-treatment pathway in Utah, and Doctronic a restricted Utah medication-renewal pilot. The delegated decision is different in each case.
This comparison is published by iatroX and includes iatroX as a separate clinician-reference and learning tool. It does not rank the services by clinical superiority, and it does not treat iatroX as a prescribing alternative.
Compare the clinical task before the interface
Imagine three fictional people. One seeks an assessment for a new concern. Another wants an initial treatment for a problem that may fit a defined acne pathway. A third requests continuation of an established medicine. Each could begin by speaking to software, but the information required and the authority needed are not identical.
The TORTUS Health terms, recorded in the 9 October 2026 research, describe AI-generated proposed care plans that require clinician approval. Nolla's AI-prescription page, checked on that date, describes its Utah acne pilot. Utah's Doctronic description limits the pilot to renewal rather than a new treatment plan.
Starting treatment requires deciding that the presentation fits the pathway and selecting an appropriate initial approach. Renewal requires establishing that continuation remains appropriate, rather than assuming an earlier prescription settles every question. A general virtual assessment can encounter a wider range of uncertainty while retaining clinician authorisation.
The task-and-authority matrix
This table concerns the specific propositions described in the October 2026 sources, not every service that each company may market.
| Question | TORTUS Health | Nolla's Utah acne pilot | Doctronic's Utah renewal pilot |
|---|---|---|---|
| Main task | Virtual assessment and clinician-authorised care | Initial treatment within a bounded acne pathway | Continuation of an existing prescribed medicine |
| Starting point | Patient account requiring assessment | Eligibility for the specified condition and pathway | Existing prescription requiring renewal assessment |
| Human decision boundary | Clinician approval of proposed care | Staged oversight, not a claim of no doctors involved | Restricted renewal with physician oversight and escalation |
| Geography in scope | California-first staged service | Utah pilot | Utah pilot |
| What must not be inferred | Unrestricted autonomous medical practice | General autonomous primary care | Authority to initiate or change treatment under the renewal pilot |
| Key evaluation question | Is the combined intake-and-review service appropriate? | Is treatment appropriate and follow-up effective? | Is continuing the established plan appropriate? |
Utah's authorised-pilots page provides the official starting point for checking the two pilots. A permission attached to one arrangement should not be extended to another product, state or clinical task.
When the doctor reviews is part of the product
Prospective approval, retrospective review and sampled audit serve different purposes. Approval before a consequential decision creates an opportunity to alter that decision before it takes effect. Review afterwards can identify problems, trigger corrective contact and inform system improvement, but it cannot guarantee prevention of the original action.
The October 2026 descriptions of Nolla indicate a staged approach to oversight. Detailed review frequencies should not be treated as settled in this comparison without checking the applicable signed agreement and current phase. For TORTUS Health, the published terms describe clinician approval of care decisions. These are materially different statements.
Doctronic's permitted renewal also remains a bounded service, not a general replacement for clinical judgement. Utah's description states that new prescriptions or changes to treatment require a licensed physician. Readers should therefore ask about the current review and escalation pathway rather than reduce all three products to a yes-or-no human-oversight column.
A narrower condition is not necessarily a simpler patient
Limiting a pathway can make its boundaries easier to define. It does not eliminate the need to establish whether the person fits those boundaries. Concurrent illness, previous treatment, uncertainty in the history and an inability to complete follow-up can all change what the service should do.
A renewal request can also be clinically consequential even when the medicine is familiar. The useful question is not merely whether a prescription exists, but whether anything relevant has changed. Equally, an initial-treatment pathway may be tightly constrained without being universally safer or riskier than every renewal encounter.
These are reasoning frameworks, not allegations that any provider omits a specific check. A fair comparison would examine eligibility decisions and escalations as well as successful treatment transactions. Excluding difficult encounters from the reported denominator would hide part of the service's clinical work.
What happens when the request falls outside scope?
The route back to a person is central. Does the patient receive an explanation, an accessible clinical option and a clear account of what has not been decided? A generic instruction to seek care elsewhere may be appropriate in some circumstances, but it should not be mistaken for an accepted clinical handover.
Consider a fictional renewal request where the person also describes a new problem that may affect continuation. The safe service response could involve further assessment rather than immediate renewal. The outcome should be judged by whether the concern was appropriately addressed, not whether the software completed the original transaction.
For UK clinicians reviewing these developments, the GMC's prescribing guidance, checked on 9 October 2026, is a useful professional reference for the need for sufficient information. It is not the law governing the Utah pilots or a verdict on the US services.
Which evidence would permit a meaningful comparison?
A common benchmark score would not be enough because the tasks and patient populations differ. Useful comparison requires a defined eligible population, comparable clinical objectives and clear reporting of review, escalation and follow-up. Treatment initiation, renewal and virtual assessment should retain their separate denominators.
Relevant outcomes include appropriate decisions, consequential errors, completed escalation, patient understanding and total episode cost. The service should also account for people who abandon intake or cannot obtain the recommended next step. A completed AI exchange is not equivalent to a completed episode of care.
Utah's independent-evaluation arrangements, checked on 9 October 2026, describe a framework for external review. The existence of that framework is not evidence that an independent evaluation of every pilot has been completed successfully.
Verdicts for patients, clinicians and buyers
For a person seeking a new assessment, the appropriate comparison is between eligible clinical services, including their scope and access to a professional. For somebody considering a bounded acne pathway, the relevant questions concern suitability, treatment and follow-up. For an established medicine renewal, examine what the service can continue and what triggers clinical reassessment.
For clinicians and buyers, the strongest organising question is: what decision has been delegated, under which conditions, and how does the patient return to a clinician? That is more informative than choosing an overall AI-doctor winner.
Under iatroX's September 2026 specification, Ask-iatroX and question-based learning serve a different purpose. A useful fictional exercise is to identify the information that would make a clinician decline initiation or renewal, then check the relevant guidance. This is professional learning, not a substitute for the service's own clinical assessment.
Frequently asked questions
Does Doctronic's Utah pilot issue new treatment plans?
Utah's description, checked on 9 October 2026, concerns renewals of existing prescriptions. New prescriptions or treatment changes require a licensed physician under that description.
How is TORTUS Health different from Nolla's acne pilot?
TORTUS Health's October 2026 terms describe clinician approval of proposed care plans. Nolla's Utah pilot concerns bounded AI-issued acne treatment with staged oversight, which is a different decision pathway.
Can these services be ranked by one accuracy figure?
Not fairly without comparable populations, tasks and evaluation methods. A useful comparison should preserve the distinction between assessment, treatment initiation and renewal.