Bayesian methods in palliative care research: cancer-induced bone pain

Area of Science:

• Biostatistics
• Oncology
• Clinical Epidemiology

Background:

• Patient recruitment and retention pose significant challenges in clinical research, particularly with frail populations.
• Accurate diagnostic and predictive tools are crucial for optimizing treatment strategies in cancer-induced bone pain (CIBP).

Purpose of the Study:

• To demonstrate the utility of a simple Bayesian analysis for strengthening the evidence base in challenging patient populations.
• To evaluate the clinical utility of a bedside tool for identifying patients likely to benefit from palliative radiotherapy (XRT) for CIBP.

Main Methods:

• A Bayesian conjugate analysis was applied to binary data from the Thermal testing in Bone Pain (TiBoP) study.
• The study prospectively assessed diagnostic accuracy and predictive value in patients with CIBP.
• Prior information was integrated with TiBoP data to maximize the informativeness of the analysis.

Main Results:

• Only 27 patients were available for primary analysis due to recruitment and retention challenges.
• The Bayesian method improved the precision of positive and negative predictive values.
• Direct probability statements facilitated result interpretation, indicating a 7% probability that the true positive predictive value exceeded 80%.

Conclusions:

• Bayesian analysis offers advantages for small diagnostic or predictive studies, especially when prior information is available.
• The method increased confidence in interpreting results from a small sample size by incorporating data from a previous study.
• This approach is recommended for enhancing evidence in studies with limited patient recruitment and retention.