In our latest publication, we used Medicare administrative data to examine the relationship between neighbourhood deprivation and survival outcomes following coronary artery bypass grafting (CABG). The results showed that while patients from less deprived neighbourhoods consistently had better survival rates compared to those from highly deprived areas, this survival advantage varied significantly across racial and ethnic groups.

Three Questions to Alessandro

What was the most interesting part of this project for you?

In this project, we used retrospective Medicare data from the US to study mortality rates in beneficiaries undergoing coronary artery bypass grafting (CABG), the most common cardiac surgery performed today worldwide.

The most interesting part of this project was the opportunity to work with large-scale, nationwide data to study the interplay between race/ethnicity and a neighbourhood-level measure of socio-economic deprivation and to understand how these factors jointly influence mortality rates.

You reported standardised survival probabilities as the main outcome measure in the manuscript. Could you briefly explain why you chose to focus over a more commonly used hazard ratio?

The results of survival analyses (e.g., using the commonly used Cox model) are often reported using hazard ratios. Despite their popularity and broad use, hazard ratios are often misinterpreted as relative risks. Survival probabilities provide an alternative measure that has a more intuitive and easier interpretation in terms of real-world probabilities. Moreover, since for this study we used non-randomised data, extensive covariate adjustment was required to reduce and eliminate bias due to confounding factors. Traditionally, adjusted models are used to predict conditional survival probabilities, i.e., a single survival curve per covariates profile. However, when a model includes many covariates, interpreting and comparing all these survival curves becomes difficult.

We therefore used regression standardisation to estimate standardised (marginal) survival probabilities. These can be interpreted as the survival probability that would be observed if the entire study population was exposed to a certain combination of deprivation status and race/ethnicity. Instead of predicting multiple survival curves, this approach allows us to estimate the average survival probability among exposed and unexposed while adjusting for the covariates included in the model. By comparing standardised survival probabilities between exposure levels, we obtain fair comparisons since the distribution of all covariates adjusted for in the survival model is the same across exposure levels.

What is the key takeaway from this publication?

The main take-home message is that the interaction between socio-economic status, race/ethnicity, and survival after CABG is complex. These findings suggest that policies aimed at reducing deprivation may not confer equitable benefits in terms of survival, and that more highly targeted interventions are needed.

Publication details

Schaffer JM, Pickering T, Squiers JJ, Banwait JK, Gasparini A, Mack MJ, Yancy CW, DiMaio JM. Variable Association of Neighborhood Deprivation and Race With Postoperative Survival After Coronary Artery Bypass Grafting. Journal of the American Heart Association (2025).