Publicatie

Publication date

Methodological issues of clinical prediction models for shoulder pain in general practice.

Vergouw, D. Methodological issues of clinical prediction models for shoulder pain in general practice. Amsterdam: Vrije Universiteit, 2013. 177 p. Proefschrift Vrije Universiteit Amsterdam
In primary care and general practice especially, musculoskeletal complaints are frequently occurring and pose a major burden on health care and society. Managing musculoskeletal disorders in primary care poses difficulties since identifying the exact cause of musculoskeletal complaints in individual patients proves to be problematic.

Because of the lack of good diagnostic criteria, research has therefore focussed on exploring the determinants of an unfavourable course of musculoskeletal complaints, rather than trying to find a precise cause. Combining such determinants in a clinical prediction model facilitates a quantitative estimate of the likely future outcome which may subsequently be used by physicians or informing and advising patients. However, finding a parsimonious set of determinants, i.e., predictors, to form a simple yet good model that can consistently be applied in a broad patient population proves to be difficult. Some methodological issues such as missing data, or model stability might hinder the development of a prediction model and therefore remain to be resolved.

The principle aim of this thesis was to address several methodological issues of clinical prediction models by applying various modelling techniques in several musculoskeletal datasets in order to contribute to the identification of optimal methods for the development and validation of prediction models.