The correct use of inhaler devices is an inclusion criterion for all studies comparing inhaled treatments. However, in real life patients make many errors when inhaling their medication which may negate the benefits observed in clinical trials. A recently published observational study evaluated inhaler handling in 3811 patients for at least 1 month using the Aerolizer, Autohaler, Diskus, pressurised metered dose inhaler (pMDI) or Turbuhaler devices. Inhalation errors were considered critical if they could have substantially affected drug delivery to the lung. The two most common errors made by patients were device-independent errors and included not breathing out before actuation of the device (28.9%) and failure to breath-hold for a few seconds after inhalation (28.3%). These errors were observed in 40%-47% of patients. The number of patients making at least one error with breath-actuated inhalers was high; with less than 50% of patients inhaling correctly. Seventy-six per cent of patients made at least one error with pMDI compared to 49%-55% with breath-actuated inhalers. With respect to device-dependent errors, the pMDI fared worst with 69% of patients exhibiting at least one error, closely followed by the Turbuhaler (32%) and Autohaler (41%). Critical errors were made by only 11%-12% of patients treated with Aerolizer, Autohaler or Diskus compared to 28% and 32% of patients treated with pMDI and Turbuhaler, respectively. Over-estimation of good inhalation by GPs was maximal for Turbuhaler (24%) and lowest for Autohaler and pMDI (6%). Ninety per cent of GPs felt that participation in the study would improve error detection. Compliance may be improved by educating patients and physicians in the correct use of inhaler devices. Inhalers should be easy to use correctly, and have multiple feedback and control mechanisms which would reduce physician over-estimation of a correct inhalation, allow compliance to be monitored, facilitate patient self-education and give reassurance to patients in the real life setting.