Propensity weighting plus adjustment in proportional hazards model is not doubly robust.

Recently, it has become common for applied works to combine commonly used survival analysis modeling methods, such as the multivariable Cox model and propensity score weighting, with the intention of forming a doubly robust estimator of an exposure effect hazard ratio that is unbiased in large samples when either the Cox model or the propensity score model is correctly specified. This combination does not, in general, produce a doubly robust estimator, even after regression standardization, when there is truly a causal effect. We demonstrate via simulation this lack of double robustness for the semiparametric Cox model, the Weibull proportional hazards model, and a simple proportional hazards flexible parametric model, with both the latter models fit via maximum likelihood. We provide a novel proof that the combination of propensity score weighting and a proportional hazards survival model, fit either via full or partial likelihood, is consistent under the null of no causal effect of the exposure on the outcome under particular censoring mechanisms if either the propensity score or the outcome model is correctly specified and contains all confounders. Given our results suggesting that double robustness only exists under the null, we outline 2 simple alternative estimators that are doubly robust for the survival difference at a given time point (in the above sense), provided the censoring mechanism can be correctly modeled, and one doubly robust method of estimation for the full survival curve. We provide R code to use these estimators for estimation and inference in the supporting information.

Transplantation of initially rejected donor lungs using ex vivo lung perfusion: A 5-year experience.

BACKGROUND: Ex vivo lung perfusion (EVLP) is a method for the evaluation and reconditioning of high-risk donor lungs to increase the pool of potential donor lungs. METHODS: We reviewed all consecutive patients who received lung transplants from May 2012 to May 2017 with follow-up until July 2021. EVLP was used in lungs initially rejected due to inadequate oxygenation but without other contraindications. Lungs with improved oxygenation levels above the threshold were transplanted. The primary endpoint was the time to graft failure, which was defined as the time from surgery to death or re-transplantation, whichever occurred first. The secondary outcome was freedom from chronic lung allograft dysfunction. RESULTS: A total of 157 patients underwent transplantation during the study period. Thirty-nine patients received EVLP-treated donor lungs. Restricted mean graft survival time up to 7 years is 5.14 years for non-EVLP and 4.19 for EVLP, the difference being -0.95 (confidence interval [CI]-1.93 to 0.04, p = .059). The hazard ratio is 1.66 (CI 1.00-2.75, p = .046). Chronic lung allograft dysfunction was the highest contributor to mortality in both groups. There were significant differences in freedom from chronic lung allograft dysfunction at 12 and 24 months of follow-up (p = .005 and p = .030, respectively). Subgroup analyses revealed that the first patients who received EVLP in 2012-2013 had a substantially worse 5-year graft survival than those who received EVLP more recently in 2016-2017 (14.3% vs. 60.0%). For the latter, the 5-year graft survival was observed to be remarkably close to the non-EVLP group (60.8%). CONCLUSION: Long-term survival was significantly lower, and lung function was poorer among recipients in the EVLP group than in the non-EVLP group. However, the outcome of patients who received EVLP-treated lungs was observed to improve steadily after the first 2 years after EVLP was introduced in Denmark.