Prognostics

Both patients and clinicians confronted with rheumatic disease need prognostic information to direct their management strategies. Prognosis refers to the prediction of outcome of disease. Clinicians confronted with prognostic problems in individual patients usually have to rely on evidence from prognostic studies on prognostic factors or clinical prediction models. Prognostic factors are patient characteristics that can be used to predict the outcome of disease. Clinical prediction models are sets of prognostic factors developed by logistic regression modelling. These models estimate the probabilities of the various forms of outcome of a disease and they can be valuable prognostic tools in daily clinical practice. To judge whether an article on prognosis fulfils the necessary methodological standards and contains relevant information, clinicians can use a number of guidelines that are described in this chapter. To find out whether the results of a prognostic study are applicable to individual patients in daily practice, the clinician should obtain clarity on three aspects: (i) generalizability, (ii) the possibility of extrapolating the results to the particular clinical situation, and (iii) feasibility. Decision analysis values the information provided by prognostic testing insofar as it alters treatment strategies. The treatment threshold is the probability of a particular disease outcome for which the net risks of treatment and no treatment are equal. Treatment is indicated if the probability of the disease outcome is higher than the treatment threshold.