When a nasal spray or inhaler works well in a patient, a lot of things must go right. The formulation must stay stable. The device must be atomised correctly. The resulting aerosol must travel to the right place in the airway. And all of that must happen under the unpredictable conditions of real-world use, including variable actuation, variable shaking, variable angle, a nose that might be congested, and lungs that breathe at different rates.
Performance is not a property of the formulation nor device. It emerges from the interaction between formulation, device, and how the patient uses it. That is a complex system, and complex systems can reveal their weaknesses late unless they are interrogated early.
This is what working inside the spray means: building development around the mechanisms that drive performance, not just the measurements that confirm a specification has been met.
Measurement and Understanding Are Not the Same Thing
The orally inhaled and nasal drug product (OINDP) field knows how to generate data. Spray pattern, plume geometry, aerodynamic particle size distribution, droplet size distribution, and delivered dose uniformity are described in general pharmacopeial chapters and supported by regulatory guidance, although the chapters themselves
do not prescribe detailed step-by-step methods in the way some other tests do. The harder question is whether the data being generated is the data needed to understand performance. All data provides information, but it only becomes useful when it captures the variables that matter for how the product behaves in use. A product can pass every required test yet still fail to perform as intended in a real patient. The gap between passing a test and working clinically is where programmes get into trouble, often quietly, and often late.
The scale of that gap is quantifiable. Orally inhaled drugs carry a clinical attrition rate of approximately 70% – seven out of ten candidates that enter trials do not reach the market. The cumulative probability of a respiratory drug achieving approval is just 3%, compared to 6–14% for drugs in other therapeutic areas.6 These are products that have already passed preclinical and in vitro testing. The failure is not happening before the data is generated. It is happening after.
