The FDA takes a risk-based approach to mobile health applications under its medical-device framework. As a result, most apps are not regulated. The main question is what the software does and what could happen if it does that job incorrectly.
An app that performs a medical device function is regulated. This includes software that analyzes data to produce a diagnosis, calculates a dose, or controls a connected device. An app that logs symptoms, reminds a person to take medication, or displays information is generally subject to enforcement discretion.
That distinction does not answer every question about the quality of a digital care service. An app may be part of a health care transaction even when its clinical process has received little or no meaningful review. The function performed by the software and the care delivered through the service are separate matters.
The function matters more than the format
The same health information can be used in different ways. The relevant issue is what the software does with the information, not the format in which the information appears.
A symptom log records what a person reports. A reminder prompts the user at a particular time. An information display presents content. These functions are generally subject to enforcement discretion.
An app performs a regulated medical device function when it uses information to produce a diagnosis, calculate a dose, or control a connected device. The task does not have to seem technically difficult for the function to raise regulatory concerns. The question is whether the software performs a task where an error could directly harm a patient.
Dose calculators often cause confusion. They may look like simple arithmetic, but they are treated as device functions because a wrong result can directly affect a patient’s treatment. That gives them a different risk profile from software that records what a person reports or displays fixed information.
This distinction also limits what users can conclude from an app’s role in a care pathway. A tool that is generally subject to enforcement discretion does not necessarily provide clinical oversight. It may perform only a lower-risk software function.
Telehealth platforms are mostly a separate question
A telehealth platform that connects a patient with a clinician mostly falls outside this FDA framework, as discussed in software as a medical device and where telehealth sits. The platform acts as a channel rather than a device. Connecting a patient to a clinician does not, by itself, make the platform a regulated medical device.
A platform may use intake logic, escalation defaults, or side-effect triage before the patient reaches a clinician. The fact that a platform makes these product decisions does not mean anyone reviewed them before patients used the service. The decisions may differ greatly in the amount of clinical thought behind them.
The platform’s role and the quality of its care process cannot be determined from its appearance or its connection to a clinician alone. A service may use a polished interface while relying heavily on automated decisions. A connection with a clinician also does not show how much that clinician participates in the patient’s care.
One useful observable proxy is whether a real clinician is genuinely involved in a patient’s care rather than merely approving a form. This proxy does not make a telehealth platform a regulated device. It helps distinguish clinical involvement from a transaction that is largely automated.
Dose tools deserve extra scrutiny
Dose calculators show the practical effect of the FDA’s risk-based approach. To a user, a calculator may look like one feature among many. The relevant questions are what the feature does and what harm could result from an error.
When a calculation determines a dose, the software performs a medical device function. The output has a direct connection to a patient’s treatment. That connection makes accuracy consequential even when the calculation relies on a formula that appears simple.
This does not mean that every app referring to dosing performs the same function. Software that displays dosing information differs from software that calculates a patient-specific dose. Software that reminds a user to take something differs from software that determines how much the user should take.
This distinction matters to developers because a minor-looking feature may perform a regulated function. It also matters to users because the FDA’s device approval and clearance framework describes what the software does. It does not describe how much clinical thought went into the broader service.
Clinical involvement is a separate policy signal
For people considering a digital service connected with GLP-1 care, the FDA’s treatment of an app does not establish how meaningfully a clinician participates in the care process. That question depends on the service’s clinical involvement. One consideration is whether a real clinician evaluates the patient’s circumstances rather than simply approving a form.
Our clinic rankings weigh clinical involvement for that reason. They do not treat an app’s regulatory status as proof of quality. They focus on whether the service includes genuine participation by a clinician.
These are two separate policy questions. FDA oversight asks whether software performs a regulated medical device function. A person evaluating a digital care service also needs to know whether a real clinician is genuinely involved. The two signals measure different things.