In April 2013, percutaneous vertebroplasty (PVP) and balloon kyphoplasty (BKP) without stenting—two vertebral augmentation procedures—were recommended by NICE to treat vertebral compression fractures (VCFs) due to osteoporosis (TA279). Although all-cause mortality was assessed as a secondary outcome, evidence from included RCTs did not achieve statistical significance, even when pooled, comparing operated patients (PVP or BKP) to patients receiving only optimal pain management (OPM). The Evidence Review Group stated that the effect of vertebral augmentation on mortality was an important, yet inadequately understood issue, despite evidence of improved survival from recently published large-scale registry studies from Germany and the USA. OBJECTIVES: To estimate the mortality differences between treatments for  osteoporotic VCFs by pooling randomised and observational data using Cox regression, propensity score matching, as well as, Thompson et al.’s (2010) and Welton et al.’s (2009) bias adjustment methods. METHODS: We extended the random effects meta-analysis from NICE’s TA279 to include observational data extracted from German and US (Medicare) insurance claims databases to estimate the mortality effect of PVP versus OPM and BKP versus OPM. All adjustment methods were compared and evaluated using a simple cost-effectiveness model. PRELIMINARY RESULTS: Survival hazard ratios were statistically significant, using all methods, in favour of either vertebral augmentation procedures versus OPM. The uncertainty in resulting estimates was artificially inflated to assess the level of uncertainty required to reach <50% probability of cost-effectiveness at common threshold values. Mortality benefit was shown to be a key driver of cost-effectiveness, particularly for BKP. CONCLUSIONS: Cox regression and propensity scoring adjustments are reliant on covariate information and thus may not capture all sources of bias; other proposed adjustment methods may play a pivotal role in assessing real-word evidence. An application of these methods to network meta-analysis is currently being undertaken to simultaneously compare operated patients with patients receiving OPM.