Semi-Competing Risks Data Analysis

Abstract

Hospital readmission is a key marker of quality of health care. Notwithstanding its widespread use, however, it remains controversial in part because statistical methods used to analyze readmission, primarily logistic regression and related models, may not appropriately account for patients who die prior to experiencing a readmission event within the timeframe of interest. Towards resolving this, we describe and illustrate the semi-competing risks framework, which refers to the general setting where scientific interest lies with some non-terminal event (e.g. readmission), the occurrence of which is subject to a terminal event (e.g. death). Although a number of statistical analysis methods have been proposed for semi-competing risks data, we describe in detail the use of illness-death models primarily because of their relation to well-known methods for survival analysis and the availability of software. We also describe and consider in detail a number of existing approaches that could, in principle, be used to analyze semi-competing risks data including composite endpoint and competing risks analyses. Throughout we illustrate the ideas and methods using data on N=49,763 Medicare beneficiaries hospitalized between 2011–2013 with a principle discharge diagnosis of heart failure.

Variation in Acute Care Rehabilitation and 30-Day Hospital Readmission or Mortality in Adult Patients With Pneumonia

“…We chose this composite outcome to address the competing risk of death with hospital readmission. 27 Death outside the health system may have also been secondary to readmission to another hospital outside the system.…”

Survival Associated With Sirolimus Plus Tacrolimus Maintenance Without Induction Therapy Compared With Standard Immunosuppression After Lung Transplant

“…For the 3 most common mortality-causing events (chronic rejection, infection, and malignant disease), semicompeting risks Cox proportional hazards regression analyses compared event risks and postevent death risks between drugs. The semicompeting risks framework, intended to address possible informative censoring of nonfatal events by death due to other causes, is explained elsewhere by Alvares et al 21 and Haneuse and Lee, 22 authors of the SemiCompRisks R package. A reference manual is available at https://cran.r-project.org, and these methods have been previously applied by Jazić et al 23…”

Association of Infectious Diseases Consultation With Long-term Postdischarge Outcomes Among Patients With Staphylococcus aureus Bacteremia

“…We also examined time to SAB recurrence and mortality as secondary outcomes, taking semicompeting risks of death and recurrence into account (eFigure 1 in the Supplement). 33 We first examined stratified cumulative incidence function plots for recurrence and mortality without SAB recurrence, using the Gray test for equality to test differences between patients with and without ID consultation. 34 We then used subdistribution hazard Cox regression models with similar interaction terms between time and ID consultation to estimate time-dependent HRs for recurrence and death without recurrence, while accounting for competing risks (ie, death and SAB recurrence) simultaneously (h 1 and h 2 in eFigure 1 in the Supplement).…”