Using data from studies of health-related quality of life to describe clinical issues examples from a longitudinal study of patients with advanced stages of cervical cancer

Abstract

The focus of the paper is to describe how to present data from studies on health-related quality of life (H-QoL) in a way that is simple and clinically relevant. Data from a longitudinal study of patients with advanced stages of cervix cancer are used. One hundred and eighteen patients filled out questionnaires (including EORTC QLQ-C30) 7 times over a period of 2 years. The following issues are considered: (1) The use of a panel for an initial overview of data. (2) The visual difference between using mean and median values. (3) Box-whisker plots to illustrate the variability of the data. (4) The effect of combining categorical data into fewer categories. (5) Individual patient profiles showing the wide variability among patients. (6) A table showing the change of scores over a one-year period. (7) "Prognostic plots" dividing the initial scores and the following scores. (8) Plotting changes over time. (9) Illustration of the impact of non-random dropout. (10) The effect of drop-out for the patients who fill out two sequential assessments. (11) The use of healthy controls to help answer the question "what is normal?".