Limitations of hazard ratios in clinical trials

Stensrud, Mats Julius; Aalen, John M.; Aalen, Odd O.; Valberg, Morten

doi:10.1093/eurheartj/ehy770

Cited by 71 publications

(46 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…On the other hand, the ratio or difference in cumulative incidence functions will not be affected by this bias since it is inherently a function of follow‐up time. Nevertheless, both measures over time can be affected by selection bias from the changing profile of patients at risk over time, an issue that entails particular attention 44‐46 …”

Section: Discussionmentioning

confidence: 99%

Time‐related biases in pharmacoepidemiology

Suissa

Dell’Aniello

2020

Pharmacoepidemiology and Drug

111

View full text Add to dashboard Cite

Purpose Observational studies using computerized healthcare databases have become popular to investigate the potential effectiveness of old drugs for new indications. Many of these studies reporting remarkable effectiveness were shown to be affected by different time‐related biases. We describe these biases and illustrate their effects using a cohort of patients treated for chronic obstructive pulmonary disease (COPD). Methods The Quebec healthcare databases were used to form a cohort of 124 030 patients with COPD, 50 years or older, treated between 2000 and 2015. Inhaled corticosteroids (ICS) and long‐acting bronchodilators were used as exposures, with diverse outcomes, including lung cancer, acute myocardial infarction and death, to illustrate protopathic, latency time, immortal time, time‐window, depletion of susceptibles, and immeasurable time biases. Results Protopathic bias affected bronchodilator‐defined cohort entry with an incident rate of lung cancer of 23.9 per 1000 in the first year, compared with around 12.0 in the subsequent years. When latency and immortal times were misclassified, ICS were associated with decreased incidence of lung cancer (hazard ratio [HR] 0.32; 95% CI: 0.30‐0.34), compared with 0.50 (95% CI: 0.48‐0.53) after correcting for immortal time bias and 0.96 (95% CI: 0.91‐1.02) after also correcting for latency time bias. Time‐window, depletion of susceptibles and immeasurable time biases also affected the findings similarly. Conclusions Many observational studies of new indications for older drugs reporting unrealistic effectiveness were affected by avoidable time‐related biases. The apparent effectiveness often disappears with proper design and analysis. Future studies should consider these time‐related issues to avoid severely biased results.

show abstract

Section: Discussionmentioning

confidence: 99%

Time‐related biases in pharmacoepidemiology

Suissa

Dell’Aniello

2020

Pharmacoepidemiology and Drug

111

View full text Add to dashboard Cite

show abstract

“…We therefore see little practical reason to use alternative analysis strategies as the pre-specified primary analysis when deviation from PH is not expected, despite recent critiques of the hazard ratio for estimating treatment effects. 13 However, when major deviations are anticipated, it may be possible to adapt the design.…”

Section: Discussionmentioning

confidence: 99%

Nonproportional Hazards for Time-to-Event Outcomes in Clinical Trials

Gregson

Sharples

Stone

et al. 2019

Journal of the American College of Cardiology

View full text Add to dashboard Cite

show abstract

“…In roughly, 19 % of oncology clinical trials using time-to-event endpoint there is evidence of non-proportional hazards, and in this case RMST offer a more parsimonious estimates while keeping the conclusions (Rulli et al, 2018). Additionally, RMST is a viable alternative to the HR if causality is of interest (Stensrud et al, 2018). Just as HR (or any other regression estimands) RMST can be adjusted either by using pseudo-observations and generalized estimating equations (Andersen et al, 2004) or by integrating an adjusted Kaplan-Meier estimator with inverse probability weighting (Conner et al, 2019).…”

Section: Discussionmentioning

confidence: 99%

Hazard Rates and Restricted Mean Survival Time

Nemes

2019

IJSA

View full text Add to dashboard Cite

Restricted Mean Survival Time (RMST) is well-established, but underutilized measure that can be interpreted as the average event-free survival time up to a pre-specified time point. In the last decade RMST received substantial attention and was advocated as an alternative for the Hazard Rate when the proportionality assumption is not met. Currently studies with time-to-evet outcomes routinely report survival curves and hazard rates. Research planning assumes extraction of comparative effect measures and variances that facilitates sample size calculations. Here we assessed the possibility of extracting clinically meaningful effect size estimates for RMST based research plans from studies that report survival curves and hazard rates. This assessment was based on simulations using Exponential and Weibull distributions. The simulations suggest that under certain conditions meaningful RMST effect size estimates can be extrapolated form published hazard rates. However, in cases when the proportionality assumption is in doubt (i.e. when RMST have most utility) extraction of meaningful estimates is not feasible.

show abstract

Limitations of hazard ratios in clinical trials

Abstract: Abstract

Cited by 71 publications

References 13 publications

Time‐related biases in pharmacoepidemiology

Time‐related biases in pharmacoepidemiology

Nonproportional Hazards for Time-to-Event Outcomes in Clinical Trials

Hazard Rates and Restricted Mean Survival Time

Contact Info

Product

Resources

About