Quantifying the bias due to observed individual confounders in causal treatment effect estimates

Effectiveness of Immediate Video-Assisted Thoracoscopic Surgery for Empyema: A Multicentre, Retrospective Cohort Study

BACKGROUND

Because of limitations in previous randomised controlled trials and observational studies, the effectiveness of immediate video-assisted thoracoscopic surgery (VATS) for patients with empyema in real-world settings remains unclear.

OBJECTIVE

This study aimed to evaluate whether immediate VATS improves clinical outcomes in patients with empyema.

METHODS

This multicentre retrospective cohort study included 744 patients with physician-diagnosed empyema from six hospitals between 2006 and 2021. The exposure was VATS performed within 3 days of empyema diagnosis, the primary outcome was 30-day mortality, and secondary outcomes were 90-day mortality, length of hospital stay, and time from diagnosis to discharge. We used propensity score weighting to account for potential confounders. For outcome analyses, we used logistic regression for mortality outcomes and gamma regression for the number of days.

RESULTS

Among the 744 patients, 53 (7.1%) underwent VATS within 3 days, and 691 (92.9%) initially received conservative treatment. After propensity score weighting, the differences in 30- and 90-day mortalities between the immediate VATS and initial conservative treatment groups were 1.18% (95% confidence interval [CI], -10.7 to 13.0%) and -0.08% (95% CI, -10.3 to 10.2%), respectively. The differences in length of hospital stay and time from diagnosis to discharge were -3.22 (95% CI, -6.19 to -0.25 days) and -5.04 days (95% CI, -8.19 to -1.90 days), respectively.

CONCLUSIONS

Our real-world study showed that immediate VATS reduced the length of hospital stay and the time from diagnosis to discharge. Considering the small sample and differences in protocols between countries, further large-scale studies are warranted.

Causal analyses with target trial emulation for real-world evidence removed large self-inflicted biases: systematic bias assessment of ovarian cancer treatment effectiveness

BACKGROUND AND OBJECTIVES

Drawing causal conclusions from real-world data (RWD) poses methodological challenges and risk of bias. We aimed to systematically assess the type and impact of potential biases that may occur when analyzing RWD using the case of progressive ovarian cancer.

METHODS

We retrospectively compared overall survival with and without second-line chemotherapy (LOT2) using electronic medical records. Potential biases were determined using directed acyclic graphs. We followed a stepwise analytic approach ranging from crude analysis and multivariable-adjusted Cox model up to a full causal analysis using a marginal structural Cox model with replicates emulating a reference randomized controlled trial (RCT). To assess biases, we compared effect estimates (hazard ratios [HRs]) of each approach to the HR of the reference trial.

RESULTS

The reference trial showed an HR for second line vs. delayed therapy of 1.01 (95% confidence interval [95% CI]: 0.82-1.25). The corresponding HRs from the RWD analysis ranged from 0.51 for simple baseline adjustments to 1.41 (95% CI: 1.22-1.64) accounting for immortal time bias with time-varying covariates. Causal trial emulation yielded an HR of 1.12 (95% CI: 0.96-1.28).

CONCLUSION

Our study, using ovarian cancer as an example, shows the importance of a thorough causal design and analysis if one is expecting RWD to emulate clinical trial results.

Can inverse probability treatment weighting (IPTW) be used to assess differences of CRBSI rates between non-tunneled femoral and jugular CVCs in PICU patients?

BACKGROUND

In children in the ICU, catheter-related bloodstream infections (CRBSI) have also been linked to mortality, morbidity, and healthcare costs. Although CRBSI poses many potential risks, including the need to avoid femoral access, there is debate regarding whether jugular access is preferable to femoral access in adults. Study reports support both perspectives. There is no consensus in meta-analyses. Children have yet to be examined in depth. Based on compliance with the central line bundle check lists, we aim to determine CRBSI risk in pediatric intensive care units for patients with non-tunneled femoral and internal jugular venous access.

METHODS

A retrospective cohort study was conducted on patients with central venous catheters in the pediatric ICU of King Abdulaziz University Hospital between January 1st, 2017 and January 30th, 2018. For the post-match balance, we use a standardized mean difference of less than 0.1 after inverse probability treatment weighting for all baseline covariates, and then we draw causal conclusions. As a final step, the Rosenbaum sensitivity test was applied to see if any bias influenced the results.

RESULTS

We recorded 145 central lines and 1463 central line days with 49 femoral accesses (33.79%) and 96 internal jugular accesses (66.21%). CRBSI per 1000 central line days are 4.10, along with standardized infections of 3.16. CRBSI risk differed between non-tunneled femoral vein access and internal jugular vein access by 0.074 (- 0.021, 0.167), P-value 0.06, and relative risk was 4.67 (0.87-25.05). Using our model, the actual probability was 4.14% (0.01-0.074) and the counterfactual probability was 2.79% (- 0.006, 0.062). An unobserved confounding factor was not identified in the sensitivity analysis.

CONCLUSIONS

So long as the central line bundle is maintained, a femoral line does not increase the risk of CRBSI. Causation can be determined through propensity score weighting, as this is a trustworthy method of estimating causality. There is no better way to gain further insight in this regard than through the use of randomized, double-blinded, multicenter studies.