Search

The prevalence of pollen-induced allergies stagnates on a high level, ranging from 15% to 25% worldwide. Since allergy is a chronic health condition, it requires long-term therapy. The efficiency of every pharmacological treatment or supporting health behavior is limited by the patient’s cooperation, especially, if it requires self-administration and is performed outside of healthcare institutions. It is well-known, that most noncompliance is intentional. Thus, interventions focusing on improvement of health behavior of allergic individuals, should address the root causes of inappropriate health behavior. Anti-allergic medications are symptomatic and have to be taken as needed. Allergen avoidance strategies make sense only if performed at the moment, when airborne pollen concentration is high. Pollen information provided to allergic individuals via a pollen application, might become an important aid in avoiding exposure to allergenic pollen, as well as planning medication and outdoor activities. However, little is known about factors motivating sustained pollen application use. Depending on the phenological and meteorological factors, airborne pollen concentration shows considerable fluctuation in its amount during the main pollen season. Robust forecasting techniques providing prediction of airborne pollen levels on a diurnal scale are of paramount importance to support a proper allergy management. The doctoral thesis contains four contributions to scientific literature. Contribution 1 examines the current situation regarding the impairment caused by allergic symptoms and frequently performed health behavior. Contribution 2 investigates the influencing factors explaining the health behavior of allergic individuals. Contribution 3 focuses on influencing factors facilitating the acceptance and utilization of pollen applications as a supporting tool in allergy management. Contribution 4 is devoted to development of predictive models of airborne pollen concentrations on a 3-hourly scales of pollen data using time series analysis and machine learning techniques. The contribution 1 explores health behavior of allergic individuals by means of a cross-sectional study. It confirms that pollen allergy remains a serious health-related problem with a profound effect on the health-related quality of life of allergic individuals, with negative implications in social life, everyday activities, and significant decline of work productivity. Despite perceived symptoms, a considerably small proportion of the allergic individuals seek medical support or undergo specific immunotherapy. Allergen avoidance strategies and pollen information services are moderately used by allergic individuals. The biggest share of allergic individuals self-manages allergic symptoms using over-the-counter medication. The contribution 2 investigates the determinants of utilization of various allergy management measures using Protection Motivation Theory. It shows that the threat appraisal, consisting of perceived severity of the symptoms and perceived seriousness of allergy, is the most relevant motivator of allergy management efforts performed by allergic individuals. However, educational interventions aiming at promoting appropriate allergy management and raising awareness of health risks associated with inadequate allergy management should be accompanied by measures increasing self-efficacy of allergic individuals. The contribution 3 explores motivational factors facilitating the acceptance and utilization of pollen applications by allergic individuals. Empirical data collected in the course of an online experiment shows that the IT-driven factors have substantially greater influence on the acceptance of pollen applications, than allergy characteristics. Therefore, to assure sustained use, pollen applications have to focus on providing high quality health content and pollen information in order to be perceived as a useful supporting tool in allergy management. The contribution 4 tests ability of four forecasting techniques, namely ARIMA, dynamic regression, artificial neural network and neural network autoregression to accurately predict airborne pollen concentrations of Betula and Poaceae on a 3-hourly scale of data. In general, forecasting techniques explicitly using autoregression and considering external meteorological variables performed well in forecasting airborne pollen levels. However, seasonal ARIMA, being the simplest among tested forecasting methods, was superior in predicting Poaceae airborne pollen concentration. A possible extension of the present scientific work, is to test the utility of the provided recommendation using experimental and longitudinal study designs. These research questions are especially interesting in context of the third contribution presenting a pollen application as a supporting tool in allergy management.

Additional file 1. Operationalization of the PMT constructs.