There is no "best" outcome or "gold standard" in the assessment of inhaled drug delivery. All levels of inquiry are important in development of drug formulations and delivery devices, and culminate in optimal therapy for the patient. Some advantages and disadvantages of each approach are summarized in Table 3. Each level forms a framework for the next step in the process, with in vitro leading to in vivo, followed by clinical trials. For optimal design, predictions and measurements of inhaled drug delivery should be done prior to large, expensive clinical trials. Bench testing with in vitro models is essential to document the performance of the delivery device/drug combination, whereas in vivo studies examine the behavior of the aerosol in human subjects. Both scintigraphy and pharmacokinetic/pharmacodynamic studies can add insights into the relationship between drug delivery and clinical efficacy and toxicity. There is a need for continued investigation on the subject of inhaled drug delivery, including more trials that bridge the levels of testing. We need more trials comparing in vitro with in vivo outcomes, as well as trials that relate in vivo assessments to clinical outcomes. In the final analysis, however, there is no substitute for clinical trials in patients. There is a lack of published, well-designed, randomized, controlled clinical trials comparing delivery devices or drug formulations. For these, attention must be given to the definition of the population, the dose-response relationships, and the expression of the physiologic responses. It is also important that clinical trials monitor both objective and subjective outcomes, as symptoms and quality of life measures help quantify the impact of disease and therapy on daily life.