How SFDA Medical Device Classification Works

SFDA classification is not a letter selected from A to D by looking at a product name or a foreign certificate. The starting point is the manufacturer's intended purpose, followed by a test of the device's functions, design and risks against Annex 5 of MDS-REQ 1. Two products that look similar in a catalogue can therefore follow different rules when the user, contact duration or consequence of a wrong result differs.

This article explains the decision framework that should precede an MDMA file. It does not classify a particular product. A defensible conclusion requires the device description, claims, components, use environment and consequences of failure. The practical benefit is to expose missing facts before unsupported assumptions enter the technical documentation, labeling and commercial plan.

1 Classification starts with intended purpose

MDS-REQ 1 derives intended purpose from the information supplied by the manufacturer on the label, instructions for use and promotional or sales materials. A broad phrase such as monitoring device or diagnostic platform is not enough. The record must state what the product monitors, for whom, in which setting and what action a user may take from the output. An added medical claim can change the candidate rule.

The first regulatory question is whether the item is a medical device, IVD, independent software or an accessory. Once the regulated unit is identified, the rules are applied to each function. A commercial system may contain a principal device, accessory and software without automatically sharing one class. MDS-REQ 1 applies classification separately to devices used in combination and treats independent software in its own right.

2 The facts that change the rule

For non-IVD devices, decision-critical facts include invasiveness, contact site, duration, implantability, active functions and the energy or substance administered or removed. Special rules also address medicinal constituents, tissues or cells, nanomaterials, disinfection and contraception. The description non-invasive does not settle risk when an active diagnostic function or clinically consequential output triggers another rule.

To turn this analysis into a product-specific plan, review Medical Device Registration and MDMA Support and the related regulatory insight.

For an IVD, the record needs the specimen, analyte, purpose of the result, user, test setting and whether the result is the sole determinant or one input to a wider clinical process. The effect of a false positive and false negative on the person and public health also matters. For software, the focus is the information produced and the diagnostic or therapeutic decision it informs, not the complexity of the algorithm or the phone on which it runs.

3 Test every plausible rule

A sound method lists each candidate rule and records why it applies or does not. Annex 5 states that intended purpose governs application, combined use does not remove separate classification, and the strictest applicable rule or sub-rule resulting in the higher class controls. This prevents a team from selecting a convenient rule while overlooking a higher trigger.

A rationale should not stop at Class B. It should connect a documented product fact to a specific rule, show the exclusions reviewed and identify the fact that could change the outcome. For example, software that supplies information for a therapeutic decision requires a Rule 11 analysis. The clinical consequence if that decision is wrong separates the B, C and D scenarios. The product claim and workflow are needed before choosing among them.

4 Pre-classification sorting matrix

5 Class is separate from grouping and evidence sufficiency

Classification answers the regulatory risk question. It does not by itself determine the number of MDMA applications or prove that the evidence package is sufficient. Models may share a class and still fail the family criteria because their intended purpose, design or manufacturing process differs. A device can also have a plausible provisional class while its clinical, performance or risk evidence remains inadequate for the selected claims.

Maintain three separate records: the classification rationale, the bundling rationale and the evidence gap assessment. Collapsing them into one conclusion creates poor commercial assumptions, such as pricing registration work on one application or printing a catalogue before the authorized model scope is settled. Classification starts the pathway; it does not close the other file decisions.

6 What changes operationally after class is set

Class affects the depth of documentation and lifecycle controls, but it works with other factors such as sterile or measuring status, the intended user and local establishment roles. MDS-REQ 1 requires the manufacturer to prepare, retain and update technical documentation and maintain a quality system, and links Saudi market availability to MDMA. Those principles still need device-specific mapping.

Commercially, the justified class should be reflected in model scope, technical materials, Saudi labeling, distributor training and the evidence preparation plan. When the classification rationale is fixed late, these activities often need rework. A better sequence is to close the product fact record first, then approve the rationale with the technical-file, marketing and supply-chain owners.

The rationale also needs version control. Record the decision date, intended-purpose wording, rules tested, sources used and internal approver. When a claim, accessory or use setting changes, the same record shows whether the facts supporting the class remain valid. This prevents an old class from surviving in ERP or catalogue data after the regulated product has changed.

References: [1]