BackgroundAllergic Rhinitis (AR) is a common disorder in Europe with Allergic Asthma (AA) as a frequent comorbidity. Allergy immunotherapy (AIT) is the only causal therapy of AR and AA, and can be administered as subcutaneous injections at the physician or as sublingual drops or tablets at home. The usual treatment duration is 3 years.ObjectiveThis study aimed to elicit patient preferences to identify the AIT administration mode preferred by patients.MethodsA discrete-choice-experiment (DCE) was developed to determine how people weight different treatment options using a paper-based questionnaire from June to September 2014, including 16 study centres. Main inclusion criteria: >18 years, grass, birch and/or house dust mite AR with moderate to severe symptoms, AIT-naïve and AIT-indicated. DCE-attributes were: Administration form, number and duration of physician visits, frequency of life-threatening anaphylactic shocks, local side-effects and co-payments.ResultsTwo-hundred thirty-nine subjects participated, resulting in analysable 1842 choices. All attributes were significant predictors for the treatment-choice. Ranked by importance, the following first three attributes are most preferred by patients:1st Number and duration of physician visits:Fewer visits with shorter duration preferred (0.658*)2nd Frequency of life-threatening anaphylactic shocks:Lower risk of shocks preferred (0.285*)3rd Local side-effects:Preference for rash/swelling on upper arm over itching/swelling under the tongue (0.210*)(*coefficient-size represents relative importance of the attributes)ConclusionThe most important attribute is the number and duration of visits to a physician. A lower risk of life-threatening anaphylactic shocks was ranked as the second whereas co-payments and administration form play a limited role.

correction. A simple Markov model was built to illustrate the impact of the halfcycle correction, and to demonstrate how a more accurate correction factor can be applied to models. RESULTS: Half-cycle corrections appear to be used routinely in Markov models. In nearly all cases, the so-called "correction" is applied without due consideration of the implications. Two major flaws were identified with the approach. The first, mathematical, flaw is that the half-cycle correction approach assumes that all events occur at the mid-point of each cycle. It can be demonstrated that, for one-directional events (such as death), events will be more likely to occur in the first half of the cycle since more patients will be exposed to the event at the start of the cycle, and the number of patients 'at risk' falls throughout the cycle. The second flaw is that, for many events, the implications of the event may not actually become apparent until the next cycle. For instance, in oncology, the increased costs associated with disease progression will not occur until progression is confirmed, which may only happen at regular routine follow-up visits. CONCLUSIONS: Half-cycle corrections are frequently applied inappropriately in modelling. This study has produced two key recommendations to generate more accurate outcomes and to avoid biases in decision analytic models.