Search

The hazard ratio (HR) remains the most frequently employed metric in assessing treatment effects on survival times. However, the difference in restricted mean survival time (RMST) has become a popular alternative to the HR when the proportional hazards assumption is considered untenable. Moreover, independent of the proportional hazards assumption, many comparative effectiveness studies aim to base contrasts on survival probability rather than on the hazard function. Causal effects based on RMST are often estimated via inverse probability of treatment weighting (IPTW). However, this approach generally results in biased results when the assumed propensity score model is misspecified. Motivated by the need for more robust techniques, we propose an empirical likelihood-based weighting approach that allows for specifying a set of propensity score models. The resulting estimator is consistent when the postulated model set contains a correct model; this property has been termed multiple robustness. In this report, we derive and evaluate a multiply robust estimator of the causal between-treatment difference in RMST. Simulation results confirm its robustness. Compared with the IPTW estimator from a correct model, the proposed estimator tends to be less biased and more efficient in finite samples. Additional simulations reveal biased results from a direct application of machine learning estimation of propensity scores. Finally, we apply the proposed method to evaluate the impact of intrapartum group B streptococcus antibiotic prophylaxis on the risk of childhood allergic disorders using data derived from electronic medical records from the Children's Hospital of Philadelphia and census data from the American Community Survey., Competing Interests: Declaration of conflicting interestsThe author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

The 30-day hospital readmission rate has been used in provider profiling for evaluating inter-provider care coordination, medical cost effectiveness, and patient quality of life. Current profiling analyzes use logistic regression to model 30-day readmission as a binary outcome, but one disadvantage of this approach is that this outcome is strongly affected by competing risks (e.g., death). Thus, one, perhaps unintended, consequence is that if two facilities have the same rates of readmission, the one with the higher rate of competing risks will have the lower 30-day readmission rate. We propose a discrete time competing risk model wherein the cause-specific readmission hazard is used to assess provider-level effects. This approach takes account of the timing of events and focuses on the readmission rates which are of primary interest. The quality measure, then is a standardized readmission ratio, akin to a standardized mortality ratio. This measure is not systematically affected by the rate of competing risks. To facilitate the estimation and inference of a large number of provider effects, we develop an efficient Blockwise Inversion Newton algorithm, and a stabilized robust score test that overcomes the conservative nature of the classical robust score test. An application to dialysis patients demonstrates improved profiling, model fitting, and outlier detection over existing methods.

The performance of health care facilities (e.g. hospitals, transplant centers, etc.) is often evaluated through time-to-event outcomes. In this paper, we consider the case where, for each subject, the failure event is due to one of several mutually exclusive causes (competing risks). Since the distribution of patient characteristics may differ greatly by the center, some form of covariate adjustment is generally necessary in order for center-specific outcomes to be accurately compared (to each other or to an overall average). We propose a weighting method for comparing facility-specific cumulative incidence functions to an overall average. The method directly standardizes each facility's non-parametric cumulative incidence function through a weight function constructed from a multivariate prognostic score. We formally define the center effects and derive large-sample properties of the proposed estimator. We evaluate the finite sample performance of the estimator through simulation. The proposed method is applied to the end-stage renal disease setting to evaluate the center-specific pre-transplant mortality and transplant cumulative incidence functions from the Scientific Registry of Transplant Recipients.

Objective: Although important enhancements to continuous ambulatory peritoneal dialysis (CAPD) have occurred since its inception, few studies have explicitly evaluated trends over time in CAPD technique failure rates. To assist in quantifying the net benefit of improvements to CAPD for patient outcomes, we examined trends in technique failure rates among Canadian CAPD patients., Patients: Patients initiating renal replacement therapy on CAPD (n = 7110) between 1981 and 1997., Main Outcome Measures: Technique failure (i.e., switch to hemodialysis)., Results: Total follow-up was 12,831 patient-years (pt-yr). There were 1976 technique failures, for a crude CAPD failure rate of 154.0/1000 pt-yr. Technique failure rate ratios (RR) estimated using Poisson regression and adjusted for age, gender, race, province, primary renal diagnosis, and follow-up time, were significantly reduced for the 1990-93 [RR = 0.75, 95% confidence interval (CI) = (0.68, 0.83)], 1994-95 [RR = 0.83, CI (0.75, 0.93)], and 1996-97 [RR = 0.78, CI (0.70, 0.87)] calendar periods relative to 1981-89 (RR = 1, reference). Among cause-specific technique failure rates, the greatest improvement was observed for peritonitis-attributable technique failure, with RR = 0.46, CI (0.41, 0.50) for 1990-97 relative to 1981-89. However, rates of technique failure due to inadequate dialysis were significantly elevated for the 1990-97 period [RR = 1.68, CI (1.44, 1.96)]., Conclusions: The collection of more detailed data on practice patterns would enable future studies to elucidate the cause-and-effect relationship between CAPD descriptors and technique failure, and hence assist in clinical decision-making.

Objective: Primarily, to determine whether peritoneal small solute clearance is related to patient and technique survival among anuric peritoneal dialysis [continuous ambulatory (CAPD) and automated peritoneal dialysis (APD)] patients. A secondary goal was to describe the ability to attain Dialysis Outcomes Quality Initiative (DOQI) targets among anuric patients on peritoneal dialysis., Design: Retrospective cohort study via chart reviews., Setting: Peritoneal Dialysis Unit of Toronto Hospital (Western Division)., Patients: The study included 122 CAPD and APD patients between January 1992 and September 1997, with 24-hour urine volume less than 100 mL, or renal creatinine clearance (CCr) less than 1 mL/minute. Adequacy data were available for 115 patients., Outcome Measures: Mortality and technique failure (TF). Regression analysis was used to estimate the mortality and TF rate ratios (RR) for peritoneal Kt/V urea (pKt/V) and pCCr, adjusting for age, gender, diabetes, months of follow-up prior to anuria, albumin, transport status, coronary artery disease, cardiovascular disease, and peripheral vascular disease., Results: Fifty seven per cent (51/89) of patients on CAPD and 81% (21/26) on APD had a weekly pKt/V > or = 2 and > or = 2.2, respectively (DOQI targets); whereas only 35% on CAPD (31/89) and 35% (9/26) on APD had a weekly pCCr > or = 60 U1.73 m2 and 66 L/1.73 m2, respectively. Median follow-up times among patients were 16.5 and 19.5 months pre- and postanuria, respectively. Patients with pKt/V > or = 1.85 experienced a strong decrease in patient mortality (RR = 0.54, p= 0.10); the effect was less pronounced for pCCr > or = 50 L/1.73 m2 (RR = 0.63, p = 0.25). No relationship was observed between pKt/V or pCCr and TF., Conclusion: Mortality was noticeably less frequent among patients with a pKt/V > or = 1.85 compared with those with a Kt/W < 1.85 (p = 0.10). Given the magnitude of the association, the failure to observe statistical significance relates to the size of the patient cohort. Our results imply that it is, in fact, possible to achieve DOQI targets among anuric patients on peritoneal dialysis.

Objective: To compare mortality rates on hemodialysis (HD) to rates on continuous ambulatory/cyclic peritoneal dialysis (CAPD/CCPD), to contrast our results with those of other recent investigations, and to discuss reasons for discrepancies., Data Sources: Patient-specific data obtained from the Canadian Organ Replacement Register on patients initiating renal replacement therapy (RRT) between 1 January 1990 and 31 December 1995 (n = 14 483). Recent mortality comparisons of CAPD and HD., Main Outcome Measures: Mortality rate ratio (RR) based on "as-treated" (AT) analysis incorporating treatment modality switches and adjusting for age, primary renal diagnosis, and comorbid conditions using Poisson regression. Hazard ratios (HR) were estimated using Cox regression and based on an "intent-to-treat" (ITT) analysis wherein patients were classified based on dialytic modality received on follow-up day 90., Results: Adjusted mortality rates were significantly decreased on CAPD/CCPD relative to HD [RR = 0.73, 95% confidence interval (CI) = (0.69, 0.77)] based on the AT analysis. Most of the protective effect of CAPD/CCPD was concentrated in the first 2 years of follow-up post-RRT initiation. Based on the ITT analysis, the estimated CAPD/ CCPD effect was greatly reduced, with HR = 0.93 (0.87, 0.99)., Conclusions: We provide further evidence that CAPD/CCPD is not an inferior dialytic modality to HD, particularly in the short term. Comparing mortality rates on CAPD/CCPD and HD is inherently difficult due to the potential for bias. Discrepancies between our results and those of previous investigations, and variability in findings among previous studies, relate to differences in clinical and demographic setting, patient populations, study design, statistical methods, and interaction between the dialytic modality effect and various other covariables.