Assessment in Health Psychology: Development, Validation and Applications of Self-Report Tools for Patients and Caregivers

This Ph.D. thesis aimed to improve the state of the art regarding the assessment in the context of health psychology. Specifically, it focused on creating and validating assessment tools designed for clinical constructs specific to both patients and informal caregivers facing illness. Chapter I provides a general introduction to the most relevant topics in health psychology, with particular attention to uncertainty in illness (UI) and caregiver burden in its multiple facets. These crucial constructs have been rather overlooked from both a psychometric and a clinical point of view. In addition, particular emphasis is placed on psychometric assessment in psychological science. Psychometrically sound measures are fundamental ways to investigate constructs. Importantly, accurate instruments lead to precise estimations of effects and relations between constructs, ultimately contributing to increased replicability in psychological science. Relying on this background, these constructs and their measurement are deepened throughout the following chapters, consisting of original research studies conducted in wide samples of patients and informal caregivers dealing with illness. Chapter II focuses on UI and the development, validation, and study of the psychometric properties of the Uncertainty in Illness Questionnaire (UIQ). The Exploratory Graph Analysis (EGA) identified 4 dimensions of UI, namely uncertainty about symptoms, uncertainty about treatments, uncertainty about future change, and uncertainty about relationships (Study 1). Then, a Confirmatory Factor Analysis (CFA) tested a hierarchical second-order model that provided a good fit to the data; also, the measurement invariance of the UIQ between patients and caregivers was reached until the highest level of latent means (Study 2). Chapter III focuses on the assessment of the discriminant validity of the UIQ and Intolerance of Uncertainty Scale-Revised (IUS-R) assessing intolerance of uncertainty (IU), a construct similar but conceptually distinct from UI. A Structural Equation Model (SEM) showed that the latent factors of UIQ and IUS-R were not too highly correlated, thus supporting discriminant validity between measures of these intertwined but different constructs. Chapter IV is on the definition of a short form of the Zarit Burden Interview (ZBI), one of the most used measures to assess caregiver burden. Its psychometric validation was never performed, as its Italian adaptation only included a principal component analysis (PCA). Thus, a first EGA detected 3 dimensions, but 5 items were removed due to instability across dimensions (Study 1). A second EGA still detected 3 subdimensions of caregiver burden, named impairment, emotions, and uncertainty, showing optimal item stability. Then, in a CFA, a hierarchical second-order model with 3 first-order factors provided a good fit to the data (Study 2). Moving to constructs relations, through a SEM with item parceling in Chapter V, the theoretically driven hypothesis of a sequential mediation from IU, through worry and UI to anxiety, was tested. The model provided a good fit to the data, showing that the path from IU to anxiety was completely mediated by worry and UI measured with the UIQ. Overall, these findings emphasize the relevance of developing psychometrically sound assessment tools by combining robust techniques and clinical knowledge which are useful for research and clinical practice in the context of health with patients and caregivers. From the applications of psychometrics in the field of health, valuable results can be derived. Hopefully, these findings will have a positive impact on promoting scientific research in this domain and improving the health conditions and psychological well-being of individuals. In conclusion, this research aimed to point out the need to rely on robust methodologies and statistics within psychological science, particularly in contexts of health psychology and illness.