Competing approaches to analysis of failure times with competing risks

How Much Do Side Effects Contribute to Discontinuation? A Longitudinal Study of IUD and Implant Users in Senegal

Introduction

In Senegal, discontinuation due to sides effects of long-acting, reversible contraceptives (LARCs) is relatively low; 5% of new implant acceptors and 11% of new IUD acceptors stop using in their first year because of health or side effect concerns. This study investigated factors associated with LARC discontinuation in the first 12 months of use in Senegal and explored how LARC users cope with side effects.

Methods

This mixed-method study involved quantitative interviews at five time points with LARC acceptors recruited from three service channels between February 2018 and March 2019. Qualitative interviews were conducted in August 2018 with a subset of those who experienced side effects. Logistic regression models identified factors associated with discontinuation due to side effects and discontinuation for any reason. Twelve-month discontinuation rates due to side effects were also estimated using a cumulative incidence function (CIF) approach to account for time to discontinuation.

Results

In logistic models, method choice (IUD or implant) [OR = 3.15 (95% CI: 1.91–5.22)] and parity [OR = 0.81 (95% CI: 0.7–0.94)] were associated with discontinuation due to side effects; IUD users and women with fewer children were more likely to discontinue. Results for all-cause discontinuation were similar: method choice [OR = 2.39 (95% CI: 1.6–3.58)] and parity [OR = 0.86 (95% CI: 0.77–0.96)] were significant predictors. The 12-month side effect CIF discontinuation rate was 11.2% (95% CI: 7.9–15.0%) for IUDs and 4.9% (95% CI: 3.5–6.6%) for implants. Side effect experiences varied, but most women considered menstrual changes the least acceptable. No statistically significant differences across services channels were observed.

Conclusions

In this study in Senegal, the choice between implants and IUDs had a significant impact on continuation, and women with more children continued LARC methods longer, despite side effects.

Competing causes of death in the head and neck cancer population

PURPOSE/OBJECTIVES

The increasing survivorship of head and neck squamous cell carcinoma (HNSCC) comes with a risk of death from other causes, known as competing causes. The demographics of HNSCC are also evolving with increasing incidence of Human Papillomavirus (HPV) associated tumors. This study describes competing causes of death for the HNSCC population compared to the general population and identifies associated risk factors.

METHODS

Adult patients with first mucosal HNSCC (2004-2011) were identified from the Surveillance, Epidemiology and End Result database. Competing causes of death were compared to reference populations using proportion of deaths and Standardized Mortality Ratios (SMR). A multivariable competing risk survival analysis yielded subdistribution hazard ratios (HR) for competing mortality.

RESULTS

Of 64,598 HNSCC patients, 24,602 (38.1%) were deceased including 7142 deaths (29.0%) from competing causes. The most common were cardiovascular disease, lung cancer, and other cancers. All relative mortality rates were elevated, especially liver disease (SMR 38.7; 95% CI: 29.4-49.3), suicide (SMR 37.1; 95% CI: 26.1-48.6), and subsequent primary cancers (SMR 7.5; 95% CI: 6.78-8.32). Demographic and tumor factors independently increased risk of competing mortality, including age (HR per 5years 1.24; 95% CI: 1.22-1.25), sex (male HR 1.23; 95% CI: 1.16-1.32), race (Black HR 1.17; 95% CI: 1.09-1.26), insurance (uninsured HR 1.28; 95% CI: 1.09-1.50), and marital status (single HR 1.29; 95% CI: 1.21-1.37).

CONCLUSION

Nearly one in three HNSCC patients died from competing causes. When developing long term survivorship regimens for HNSCC patients, clinicians should be familiar with this population's specific risks.

Estimation of Mortality Rates for Disease Simulation Models Using Bayesian Evidence Synthesis

Purpose. The authors propose a Bayesian approach for estimating competing risks for inputs to disease simulation models. This approach is suggested when modeling a disease that causes a large proportion of all-cause mortality, particularly when mortality from the disease of interest and other-cause mortality are both affected by the same risk factor.

Methods. The authors demonstrate a Bayesian evidence synthesis by estimating other-cause mortality, stratified by smoking status, for use in a simulation model of lung cancer. National (US) survey data linked to death registries (National Health Interview Survey [NHIS]—Multiple Cause of Death files) were used to fit cause-specific hazard models for 3 causes of death (lung cancer, heart disease, and all other causes), controlling for age, sex, race, and smoking status. Synthesis of NHIS data with national vital statistics data on numbers and causes of deaths was performed in WinBUGS (version 1.4.1, MRC Biostatistics Unit, UK). Correction for inconsistencies between the NHIS and vital statistics data is described. A published cohort study was a source of prior information for smoking-related mortality.

Results. Marginal posterior densities of annual mortality rates for lung cancer and other-cause death (further divided into heart disease and all other causes), stratified by 5-year age interval, race (white and black), gender, and smoking status (current, former, never), were estimated, specific to a time period (1987-1995). Overall, black current smokers experienced the highest mortality rates.

Conclusions. Bayesian evidence synthesis is an effective method for estimation of cause-specific mortality rates, stratified by demographic factors.

Competing events in patients with malignant disease who are at risk for recurrent venous thromboembolism

Patients with malignant disease enrolled in trials of thrombotic disorders may experience competing events such as death. The occurrence of a competing event may prevent the thrombotic event from being observed. Standard survival analysis techniques ignore competing risks, resulting in possible bias and distorted inferences. To assess the impact of competing events on the results of a previously reported trial comparing low molecular weight heparin (LMWH) with oral anticoagulant (OAC) therapy for the prevention of recurrent venous thromboembolism (VTE) in patients with advanced cancer, we compare the results from standard survival analysis with those from competing risk techniques which are based on the cumulative incidence function (CIF) and Gray's test. The Kaplan-Meier method overestimates the risk of recurrent VTE (17.2% in the OAC group and 8.7% in the LMWH group). Risk of recurrence using the CIF is 12.0% and 6.0% in the OAC and LMWH groups, respectively. Both the log-rank test (p=0.002) and Gray's test (p=0.006) suggest evidence in favor of LMWH. The overestimation of risk is 30% in each treatment group, resulting in a similar relative treatment effect; using the Cox model the hazard ratio (HR) is 0.48 (95% confidence interval [CI], 0.30 to 0.78) and HR=0.47 (95% CI, 0.29 to 0.74) using the CIF model. Failing to account for competing risks may lead to incorrect interpretations of the probability of recurrent VTE. However, when the distribution of competing risks is similar within each treatment group, standard and competing risk methods yield comparable relative treatment effects.

Estimated epidemiologic parameters and morbidity associated with pandemic H1N1 influenza

BACKGROUND

In the face of an influenza pandemic, accurate estimates of epidemiologic parameters are required to help guide decision-making. We sought to estimate epidemiologic parameters for pandemic H1N1 influenza using data from initial reports of laboratory-confirmed cases.

METHODS

We obtained data on laboratory-confirmed cases of pandemic H1N1 influenza reported in the province of Ontario, Canada, with dates of symptom onset between Apr. 13 and June 20, 2009. Incubation periods and duration of symptoms were estimated and fit to parametric distributions. We used competing-risk models to estimate risk of hospital admission and case-fatality rates. We used a Markov Chain Monte Carlo model to simulate disease transmission.

RESULTS

The median incubation period was 4 days and the duration of symptoms was 7 days. Recovery was faster among patients less than 18 years old than among older patients (hazard ratio 1.23, 95% confidence interval 1.06-1.44). The risk of hospital admission was 4.5% (95% CI 3.8%-5.2%) and the case-fatality rate was 0.3% (95% CI 0.1%-0.5%). The risk of hospital admission was highest among patients less than 1 year old and those 65 years or older. Adults more than 50 years old comprised 7% of cases but accounted for 7 of 10 initial deaths (odds ratio 28.6, 95% confidence interval 7.3-111.2). From the simulation models, we estimated the following values (and 95% credible intervals): a mean basic reproductive number (R0, the number of new cases created by a single primary case in a susceptible population) of 1.31 (1.25-1.38), a mean latent period of 2.62 (2.28-3.12) days and a mean duration of infectiousness of 3.38 (2.06-4.69) days. From these values we estimated a serial interval (the average time from onset of infectiousness in a case to the onset of infectiousness in a person infected by that case) of 4-5 days.

INTERPRETATION

The low estimates for R0 indicate that effective mitigation strategies may reduce the final epidemic impact of pandemic H1N1 influenza.

Evaluating cardiovascular mortality in type 2 diabetes patients: an analysis based on competing risks Markov chains and additive regression models

RATIONALE, AIMS AND OBJECTIVES

Type 2 diabetes represents a condition significantly associated with increased cardiovascular mortality. The aims of the study are: (i) to estimate the cumulative incidence function for cause-specific mortality using Cox and Aalen model; (ii) to describe how the prediction of cardiovascular or other causes mortality changes for patients with different pattern of covariates; (iii) to show if different statistical methods may give different results.

METHODS

Cox and Aalen additive regression model through the Markov chain approach, are used to estimate the cause-specific hazard for cardiovascular or other causes mortality in a cohort of 2865 type 2 diabetic patients without insulin treatment. The models are compared in the estimation of the risk of death for patients of different severity.

RESULTS

For younger patients with a better covariates profile, the Cumulative Incidence Function estimated by Cox and Aalen model was almost the same; for patients with the worst covariates profile, models gave different results: at the end of follow-up cardiovascular mortality rate estimated by Cox and Aalen model was 0.26 [95% confidence interval (CI) = 0.21-0.31] and 0.14 (95% CI = 0.09-0.18).

CONCLUSIONS

Standard Cox and Aalen model capture the risk process for patients equally well with average profiles of co-morbidities. The Aalen model, in addition, is shown to be better at identifying cause-specific risk of death for patients with more severe clinical profiles. This result is relevant in the development of analytic tools for research and resource management within diabetes care.

Comparing the small sample performance of several variance estimators under competing risks

We examine several variance estimators for cumulative incidence estimators that have been proposed over time, some of which are derived from asymptotic martingale or counting process theory, and some of which are developed from the moments of the multinomial distribution. There is little published work comparing these variance estimators, largely because the variance estimators are algebraically complex and difficult to interpret and all but one have yet to be programmed for a standard statistical package. Through simulation and application to real data, we compare the performance of six variance estimators in relation to each other and the bootstrap in order to confirm earlier reports of their performance and to provide future direction toward their application. We find that the multinomial-moment-based estimators have performance close to that of the bootstrap, and are quite accurate for estimating the variance, even in samples of 20 subjects. All but one of the martingale theory-based estimators tend to perform poorly in small samples, tending to either overestimate or underestimate the empirical variance in samples of fewer than 100 subjects.

Estimation of distribution function in bivariate competing risk models

We consider lifetime data involving pairs of study individuals with more than one possible cause of failure for each individual. Non-parametric estimation of cause-specific distribution functions is considered under independent censoring. Properties of the estimators are discussed and an illustration of their application is given.

Biliary atresia: clinical profiles, risk factors, and outcomes of 755 patients listed for liver transplantation

OBJECTIVES

To test the hypothesis that risk analysis from the time of listing for liver transplantation (LT) focuses attention on areas where outcomes can be improved.

STUDY DESIGN

Competing outcomes and multivariate models were used to determine significant risk factors for pretransplantation and posttransplantation mortality and graft failure in patients with biliary atresia (BA) listed for LT and enrolled in the Studies of Pediatric Liver Transplantation (SPLIT) registry.

RESULTS

Of 755 patients, most were infants (age < 1 year). Significant waiting list mortality risk factors included infancy and pediatric end-stage liver disease (PELD) score > or = 20, whose components were also continuous risk factors. Survival posttransplantation (n=567) was 88% at 3 years. Most deaths were from infection (37%). Posttransplantation mortality risk factors included infant recipients, height/weight < -2 standard deviations (SD), use of cyclosporine versus tacrolimus and retransplantation. Graft failure risks included height/weight < -2 SD, cadaveric partial donors, donor age < or = 5 months, use of cyclosporine versus tacrolimus, and rejection.

CONCLUSIONS

Referral for LT should be anticipatory for infants with BA with failed portoenterostomies. Failing nutrition should prompt aggressive support. Post-LT risk factors are mainly nonsurgical, including nutrition, the relative risk of infection over rejection, and the choice of immunosuppression.

Factors affecting the outcome of stem cell transplantation from unrelated donors for childhood acute lymphoblastic leukemia in third remission

Between July 1990 and March 2002, 35 consecutive children with ALL in third complete remission (CR3) underwent stem cell transplantation (SCT) from unrelated donors (UD). All patients received CAMPATH-1M 5-20 mg daily for 5 days. Grafts were T-cell depleted in 30 patients, 29 by CAMPATH antibodies and one by CD34 selection. Median follow-up was 3.8 years (0.3-9.3). Event-free survival (EFS) at 3 years was 35% (SE 8%); relapse rate and transplant-related mortality (TRM) at 3 years was 42 and 23%. Short first complete remission (CR1) <2.5 years was associated with lower EFS (P=0.001), higher TRM (P=0.019) and higher relapse rate (P=0.023). Short second complete remission (CR2) <2.5 years was associated with lower EFS (P=0.003) and higher TRM (0.009). Higher relapse rate and lower EFS were associated with isolated first extramedullary relapse (P=0.019, 0.012). There was no significant difference in outcome between mismatched unrelated donor stem cell transplantation (MMUD-SCT) and matched unrelated donor stem cell transplantation (UD-SCT). We conclude that UD-SCT is an effective treatment of ALL in CR3. The outcome remains limited by TRM and a high relapse rate. Short duration of CR1 and of CR2 and extramedullary site at first relapse are particularly adverse. MMUD should also be considered in high-risk patients, since the outcome of MMUD appears similar to MUD.

Frameless versus classical intrauterine device for contraception

BACKGROUND

The frameless intrauterine device (IUD) dispenses with the frame in the classical IUD and holds the device in the uterus by anchoring one end of a nylon thread in the fundal myometrium, to which copper sleeves are attached.

OBJECTIVES

This review examines the hypothesis that the frameless IUD Gynefix reduces risk of expulsion and pregnancy, and the problems of bleeding and pain necessitating early removal.

SEARCH STRATEGY

We searched the Cochrane Controlled Trial Register (Cochrane Library Issue 2, 2004), MEDLINE, and Popline from 1980 to March 2004, and reference list of articles.

SELECTION CRITERIA

We selected for the review randomised trials that compared the frameless device to a classical framed device for contraception.

DATA COLLECTION AND ANALYSIS

Both authors extracted data independently. We contacted study author for additional data. We calculated rate ratios and rate differences for cumulative rates for each outcome at yearly intervals. We used the inverse variance-based method to combine trials, and tested the results for heterogeneity.

MAIN RESULTS

Four trials were included in the review involving 5,939 women randomised to either a frameless device or TCu380, with data up to eight years for the largest, and with a total experience of 23,180 years. Apart from one small trial, nulliparous women were excluded from the trials. The two earlier trials used a prototype introducer and there was a higher expulsion rate at one year (relative risk 2.48, 95% confidence intervals 1.89 to 3.26). However, between two and six years in the large WHO trial the risk of pregnancy was lower with the frameless device (relative risk 0.53, 95% confidence intervals 0.32 to 0.91). In a recent trial using GyneFix with a new introducer early expulsions and pregnancies were not statistically different from the control device. Removals rates for excessive bleeding and/or pain were no different between the devices (relative risks 0.92, 95% confidence intervals 0.74 to 1.14, at one year and 1.13, 0.93 to 1.37, at six years). There was a tendency towards fewer removals for pain in early years but no difference at six years (relative risk 1.13, 95% confidence intervals 0.93 to 1.37).

AUTHORS' CONCLUSIONS

There is insufficient data to show that problems of early expulsions have been overcome with the modified introducer used in GyneFix. Apart from that, the frameless device performs similarly to TCu380, and appears to have a lower pregnancy rate in later years, although the absolute difference is small.

The graft-versus-leukaemia effect after allogeneic bone-marrow transplantation: assessment through competing risks approaches

In failure time studies involving a chronic disease such as cancer, data often focus on one or more non-fatal events, in addition to survival, to describe the course of the disease. In the example of allogeneic bone marrow transplantation in leukaemic patients, acute graft-versus host disease (aGvHD), relapse and death were taken as the reported events, and we focused on testing the existence of graft-versus-leukaemia (GvL) effect, i.e. that the occurrence of aGvHD modifies the probability of relapse. One of the weaknesses of the standard competing risks models is their inability to model secondary relapses. We thus derived, from two competing risks models, two estimators of cumulative incidence functions of primary and secondary relapses, as well as statistics to test the GvL effect. The approach is illustrated by application to a data set from a cohort of 442 children with acute leukaemia who received an unrelated transplant.