Back to Bayesian: A strategy to enhance prognostication of metastatic spine disease

Abstract

Aims Clinicians must consider prognosis when offering treatment to patients with spine metastases. Although several prognostic indices have been developed and validated for this purpose, they may not be applicable in the current era of targeted systemic therapies. Even before the introduction of targeted therapies, these prognostic indices should not have been directly used for individual patient decision making without contextualising with other sources of data. By contextualising, we mean that prognostic estimates should not be based on these scores alone and formally incorporate clinically relevant factors not part of prognostic indices. Contextualisation requires the use of Bayesian statistics which may be unfamiliar to many readers. In this paper we show readers how to correctly apply prognostic scores to individual patients using Bayesian statistics. Through Bayesian analysis, we explore the impact of new targeted therapies on prognostic estimates obtained using the Tokuhashi score. Methods We provide a worked calculation for the probability of a patient surviving up to 6 months using dichotomous prognostication. We then demonstrate how to calculate a patient's expected survival using continuous prognostication. Sensitivity of the posterior distribution to prior assumptions is illustrated through effective sample size adjustment. Results When the predicted prognosis from the Tokuhashi score is contextualised with data on contemporary systemic treatments, patients previously deemed non‐surgical candidates may be eligible for surgery. Conclusions Bayesian prognostication generates intuitive results and allows multiple data points to be synthesised transparently. These techniques can extend the usefulness of existing prognostic scores in the era of targeted systemic therapies.

Clinical medicine and the legacy of the reverend Bayes

“…Given this growing use of Bayesian analyses, it is crucial that clinicians and researchers become as familiar with them as with the more traditional, “frequentist,” statistical methods. In this issue, Pahuta et al walk the reader through an example of predicting the outcome of an intervention using a Bayesian approach to take “contextualizing” factors into account. This should be fairly familiar territory to readers who interpret diagnostic tests taking prevalence into consideration.…”

Think Bayesian, Think Smarter!

“…Professor Streiner returns to the pages of the International Journal of Clinical Practice with an erudite editorial on the legacy of the Reverend Bayes . The purpose was to provide some context to an article that follows, that of using Bayesian statistics to better prognosticate metastatic spine disease . I won't lie; this article is not easy to casually skim and understand.…”

“…Bayesian statistics formalise this philosophy and provide a mathematical means of calculating probabilities based on prior information; current data and prior information are input into Bayes Theorem and the output is a posterior distribution that can perhaps lead to more meaningful inferences, especially under complex conditions . Pahuta et al provide an example that can also be adapted to other disease states . An important caveat is that faulty assessment of a priori probability can lead to error, and consequently a poor decision .…”