A smart medical patch is the smallest package in medtech that still has to satisfy every discipline at once. Three to seven days of continuous wear on human skin, a sensor accurate enough to drive a clinical decision, a radio that reaches a phone through a shirt, a battery thin enough to be forgotten, and an adhesive that neither falls off in the shower nor strips the epidermis on removal. Any one of those is a manageable engineering problem. Together, inside a two millimeter stack, they trade against each other constantly.
Continuous glucose monitors, cardiac event patches, temperature and respiration monitors, and post-surgical wound sensors all live in this form factor. The pattern below is what actually decides whether a patch program reaches a submission or stalls in wear trials.
Skin Contact Sets the Materials, Not the Other Way Around
Every material touching the patient falls under ISO 10993 biocompatibility testing. A patch is classified as surface-contacting with intact skin, and the duration category matters: under 24 hours is limited exposure, 24 hours to 30 days is prolonged, and beyond that is permanent. Prolonged contact typically drives cytotoxicity, sensitization, and irritation testing, and each additional material in the stack is another line item.
The practical consequence is that you should pick materials from what is already tested. Medical-grade silicones, polyurethane films, and acrylic adhesives from the major suppliers come with master files and prior biocompatibility data that dramatically shorten the testing burden.
Adhesive Is the Component That Kills Programs
Adhesive selection is where most patch projects discover reality. Acrylic adhesives build tack over hours and hold for a week but can cause irritation and take skin with them. Silicone adhesives are gentle, repositionable, and appropriate for fragile or elderly skin, but they lose grip under sweat and shear. Hydrocolloids handle moisture well and are the standard for anything that stays on through showering.
Real answers come from wear trials, not datasheets. Plan on at least two rounds of twenty to thirty subjects wearing the mechanical mock-up through normal life, scoring adhesion daily and skin condition at removal. Perimeter geometry matters more than founders expect: rounded corners, a border zone with a softer adhesive, and a stiffness gradient from rigid electronics to compliant edge all reduce the peel-initiation that starts every failure.
Building Electronics That Bend
A patch that resists the body's motion gets pushed off. The usual architecture is a rigid island for the microcontroller, radio, and sensor front end, joined by flexible traces to the electrode or sensor sites, which is exactly what flex and rigid-flex circuits exist to do. Keep the rigid island small and place it away from anatomy that flexes hard, such as the ribcage edge or a joint.
Signal integrity is the other half. Biopotential measurements on skin sit in the microvolt range against a body acting as an antenna, so electrode placement, guard traces, and input impedance decide whether motion artifact swamps the signal. Integrated analog front ends built for wearable biosignals solve most of this if you follow the reference layout rather than improvising, and the general reasoning behind that choice is covered in choosing a microcontroller for your product.
The Power Budget Decides the Product
Wear duration is a battery arithmetic problem. A thin coin cell or a flexible printed cell gives you tens of milliamp-hours in a package thin enough to wear, which means the average current draw for a seven-day patch has to land in the tens of microamps. That is only reachable if the radio is off most of the time. Continuous streaming over Bluetooth burns the budget in a day; buffering samples in on-chip memory and connecting for a burst every few minutes can stretch the same cell across the full wear period, which is the design pattern described in low-power firmware and sleep modes.
Battery safety in a device worn against skin is not optional. Lithium cells in a patch need protection against short circuit and thermal runaway, and shipping them requires UN 38.3 transport testing. Many disposable patches deliberately use primary cells with no charging circuit at all, removing an entire class of hazard and simplifying the risk file.
Where the FDA Puts It
Classification follows the claim, not the hardware. A patch that displays a wellness trend and makes no medical claim may fall outside device regulation entirely. A patch that measures a physiological parameter used for diagnosis or monitoring is a device, and most land in Class II with a 510(k) pathway if a predicate exists. Cardiac event monitors and continuous glucose systems have well-populated predicate lists; a genuinely novel measurement may need the De Novo route instead.
Start by working through FDA device classes and then hunting for a predicate with a claim close to yours. The claim you write on day one determines the pathway, the clinical evidence, and roughly half of your budget. Founders routinely write an ambitious claim into marketing material before anyone checks what proving it costs.
Manufacturing a Disposable
Patches are made on roll-to-roll converting lines: laminate the adhesive stack, die-cut the outline, place the electronics, seal, and pouch. Setup tooling for die-cutting is inexpensive compared with injection molds, but the assembly step where rigid electronics meet a soft laminate is where yield problems live. Design so the electronics module can be tested before it is committed to the adhesive stack, because scrapping a finished patch scraps the expensive part along with the cheap one.
Sterility is usually unnecessary for intact-skin patches, which spares you a validation program, but shelf life is not. Adhesives age, and accelerated aging under ASTM F1980 is how you justify the expiration date printed on the pouch.
What a Realistic Program Looks Like
From concept to a cleared Class II patch, plan on eighteen to thirty months and a budget that starts in the high six figures once biocompatibility, clinical data, and verification testing are counted. The single best cost control is sequencing: prove the sensor works in a bulky benchtop form, prove the adhesive survives real wear with a dummy load, and only then combine them. Teams that miniaturize first almost always rebuild, a pattern documented in shrinking a bulky prototype into a compact product.
Start With a Feasibility Review
Projects House takes wearable patch concepts through sensor feasibility, adhesive and materials selection, flexible electronics design, and the documentation an FDA submission requires, working with US clients through a global engineering and manufacturing network. Describe what you want the patch to measure and how long it must stay on through our contact form.