Stratified Proportional Hazards Model Research Articles

Abstract PO1-04-10: Body Mass Index and Treatment Efficacy in advanced Luminal Breast Cancer: Insights from the GEICAM/2013-02 (PEARL) trial

Abstract Background: Endocrine therapy (ET) and treatment directed against molecular targets (CDK 4/6, PI3K, mTOR) constitute the basis of initial treatments for patients with advanced luminal breast cancer. Most of these treatments are oral and have been developed at a fixed dose, without weight adjusted dosing. There is little evidence on the influence that patients’ weight might have on clinical outcomes of these oral drugs in the context of a randomized clinical trial. The present study analyzes the potential impact of the weight of patients treated in the two cohorts of the PEARL trial (NCT02028507) on Progression-Free Survival (PFS), Overall Survival (OS) and the incidence of side effects in each treatment arm. Methods: GEICAM/2013-02 (PEARL) is a multicenter, phase III randomized clinical trial that enrolled patients with hormone receptor-positive/human epidermal growth factor receptor 2-negative metastatic breast cancer (MBC) who progressed to a prior aromatase inhibitor. Patients were randomized 1:1 to receive ET + palbociclib or capecitabine. Patient weight was assessed based on baseline body mass index (BMI) categories: normal (BMI ≤25 Kg/m2), overweight (BMI 25-29.9 Kg/m2) and obesity (BMI ≥30 Kg/m2). Cox's stratified proportional hazard model was used for survival analysis. Adverse events were assessed and graded according to National Cancer Institute common terminology criteria for adverse events (NCI-CTCAE) version 4.0. Results: A total of 600 patients were included in the study and all were evaluable for the weight-based analysis, including 301 in the ET +palbociclib arm and 299 in the capecitabine arm. The distribution of patient’s weights is shown in Table 1. For the Intention-to-treat (ITT) population, no significant difference in the median OS was found between the different weight categories (Normal: 32.8 months vs overweight: 31.9 months, [p=0.467]; obese: 32 months [p=0.407]). Specifically, for patients with normal weight, the median PFS was 9.3 months in the ET + palbociclib group compared to 9.0 months in the capecitabine group (hazard ratio (HR) 0.83, 95% confidence interval [CI]: 0.61-1.13; p=0.23). In the overweight group, the PFS was 7.5 months in the ET+palbociclib group and 11.1 months in the capecitabine group (HR 1.17, 95% CI: 0.87-1.58; p=0.30) while for the obese group, the PFS was 5.7 months vs 11.7 months (HR 1.14, 95% CI: 0.82-1.59; p=0.43), respectively. No substantial differences in the incidence of grade ≥ 3 events were observed when considering weight categories in the different treatment groups (capecitabine arm (Normal: 60%, Overweight: 70.4%, Obese: 61.6%) and ET+palbociclib arm (Normal: 74.5%, Overweight: 76.1%, Obese: 51.6%). Conclusions: Our findings, despite not finding statistically significant differences, suggest a potentially clinically relevant trend towards improved PFS with capecitabine in overweight and obese patients compared to those receiving palbociclib plus ET. The impact of weight on the outcomes with fixed-dose oral drugs warrants further investigation in future studies. Table 1. Citation Format: Manuel Alva, Sara López-Tarruella, Christoph Zielinski, Manuel Ruíz - Borrego, Nicholas Turner, Eva Ciruelos, Montserrat Muñoz, Begoña Bermejo, Mireia Margelí, Antonio Antón, Zsuzsanna Kahan, Tibor Csöszi, Maribel Casas, Laura Murillo, Serafin Morales Murillo, Emilio Alba, Einav Nili-Gal Yam, Angel Guerrero, Lourdes Calvo, Juan de la Haba-Rodríguez, Manuel Ramos, Isabel Álvarez, Andrés García-Palomo, María Koehler, José Ángel García-Sáenz, J. Ignacio Chacón, Claire Swift, Christiane Thallinger, Miguel Gil-Gil, Miguel Martín. Body Mass Index and Treatment Efficacy in advanced Luminal Breast Cancer: Insights from the GEICAM/2013-02 (PEARL) trial [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO1-04-10.

Analysis and asymptotic theory for nested case-control designs under highly stratified proportional hazards models.

Nested case-control sampled event time data under a highly stratified proportional hazards model, in which the number of strata increases proportional to sample size, is described and analyzed. The data can be characterized as stratified sampling from the event time risk sets and the analysis approach of Borgan et al.(Ann Stat 23:1749-1778, 1995) is adapted to accommodate both the stratification and case-control sampling from the stratified risk sets. Conditions for the consistency and asymptotic normality of the maximum partial likelihood estimator are provided and the results are used to compare the efficiency of the stratified analysis to an unstratified analysis when the baseline hazards can be semi-parametrically modeled in two special cases. Using the stratified sampling representation of the stratified analysis, methods for absolute risk estimation described by Borgan et al.(1995) for nested case-control data are used to develop methods for absolute risk estimation under the stratified model. The methods are illustrated by a year of birth stratified analysis of radon exposure and lung cancer mortality in a cohort of uranium miners from the Colorado Plateau.

Marginal proportional hazards models for clustered interval-censored data with time-dependent covariates.

The Botswana Combination Prevention Project was a cluster-randomized HIV prevention trial whose follow-up period coincided with Botswana's national adoption of a universal test and treat strategy for HIV management. Of interest is whether, and to what extent, this change in policy modified the preventative effects of the study intervention. To address such questions, we adopt a stratified proportional hazards model for clustered interval-censored data with time-dependent covariates and develop a composite expectation maximization algorithm that facilitates estimation of model parameters without placing parametric assumptions on either the baseline hazard functions or the within-cluster dependence structure. We show that the resulting estimators for the regression parameters are consistent and asymptotically normal. We also propose and provide theoretical justification for the use of the profile composite likelihood function to construct a robust sandwich estimator for the variance. We characterize the finite-sample performance and robustness of these estimators through extensive simulation studies. Finally, we conclude by applying this stratified proportional hazards model to a re-analysis of the Botswana Combination Prevention Project, with the national adoption of a universal test and treat strategy now modeled as a time-dependentcovariate.

Effectiveness of eribulin as first-line or second-line chemotherapy for HER2-negative hormone-resistant advanced or metastatic breast cancer: findings from the multi-institutional, prospective, observational KBCRN A001: E-SPEC study.

The optimal positioning of eribulin treatment remains unclear. This study aimed to investigate the effectiveness of eribulin administration as first- and second-line chemotherapy in patients with endocrine-resistant advanced or metastatic breast cancer (AMBC) in the real-world clinical setting. This multi-institutional prospective cohort study enrolled patients with triple-negative AMBC or estrogen receptor-positive AMBC refractory to at least one previous endocrine therapy. The overall survival (OS) from the start of first-line (OS1) and second-line chemotherapy (OS2) was assessed. Data analysis included real-world chemotherapy sequences of first- to third-line chemotherapy regimens. The adjusted hazard ratio (HR) with 95% confidence interval (CI) for treatment regimen comparison was calculated using a stratified proportional hazards model. Among 201 patients enrolled, 180 were included in the final analysis. Eribulin was administered as first- and second-line chemotherapy to 46 (26.6%) and 70 (47.9%) patients, respectively. Median OS1 and OS2 were 2.25 (95% CI 1.07-2.68) and 1.75 (95% CI, 1.28-2.45) years for first- and second-line eribulin, respectively. Oral 5-FU followed by eribulin had a numerically longer OS1 (2.84years) than the other sequences. Among patients who proceeded to second-line or later chemotherapy, the median OS1 for those treated with anthracycline or taxane as first- or second-line (n = 98) was 2.56years (95% CI 2.27-2.74), while it was 2.87years (95% CI 2.20-4.32) for those who avoided anthracycline and taxane as first- and second-line (n = 48) (adjusted HR, 1.20; 95% CI 0.70-2.06). In the exploratory analysis, OS1 was 2.55 (95% CI 2.14-2.75) and 2.91years (95% CI 2.61-4.32) for those aged < 65 and ≥ 65years, respectively (adjusted HR of ≥ 65, 0.91; 95% CI 0.56-1.46). Eribulin or oral 5-FU administration in first- and second-line chemotherapy without anthracycline/taxane was acceptable in the real-world setting. This study is registered with Clinical Trials.gov (NCT 02,551,263).

A calibrated Bayesian method for the stratified proportional hazards model with missing covariates.

Missing covariates are commonly encountered when evaluating covariate effects on survival outcomes. Excluding missing data from the analysis may lead to biased parameter estimation and a misleading conclusion. The inverse probability weighting method is widely used to handle missing covariates. However, obtaining asymptotic variance in frequentist inference is complicated because it involves estimating parameters for propensity scores. In this paper, we propose a new approach based on an approximate Bayesian method without using Taylor expansion to handle missing covariates for survival data. We consider a stratified proportional hazards model so that it can be used for the non-proportional hazards structure. Two cases for missing pattern are studied: a single missing pattern and multiple missing patterns. The proposed estimators are shown to be consistent and asymptotically normal, which matches the frequentist asymptotic properties. Simulation studies show that our proposed estimators are asymptotically unbiased and the credible region obtained from posterior distribution is close to the frequentist confidence interval. The algorithm is straightforward and computationally efficient. We apply the proposed method to a stem cell transplantation data set.

Competing risks regression models with covariates-adjusted censoring weight under the generalized case-cohort design.

A generalized case-cohort design has been used when measuring exposures is expensive and events are not rare in the full cohort. This design collects expensive exposure information from a (stratified) randomly selected subset from the full cohort, called the subcohort, and a fraction of cases outside the subcohort. For the full cohort study with competing risks, He et al. (Scand J Stat 43:103-122, 2016) studied the non-stratified proportional subdistribution hazards model with covariate-dependent censoring to directly evaluate covariate effects on the cumulative incidence function. In this paper, we propose a stratified proportional subdistribution hazards model with covariate-adjusted censoring weights for competing risks data under the generalized case-cohort design. We consider a general class of weight functions to account for the generalized case-cohort design. Then, we derive the optimal weight function which minimizes the asymptotic variance of parameter estimates within the general class of weight functions. The proposed estimator is shown to be consistent and asymptotically normally distributed. The simulation studies show (i) the proposed estimator with covariate-adjusted weight is unbiased when the censoring distribution depends on covariates; and (ii) the proposed estimator with the optimal weight function gains parameter estimation efficiency. We apply the proposed method to stem cell transplantation and diabetes data sets.

Study of factors caused diabetes mellitus and stroke diseases toward gender

Gender consists of female and male. Diabetes mellitus is a chronic disease characterized by high levels of sugar in the blood. Glucose is controlled by the hormone insulin which is produced by the pancreas. In diabetics, the pancreas is unable to produce insulin according to the body’s needs. Blood sugar tests can be done to find out someone has diabetes. This method by a time blood sugar test and fasting blood sugar test. Stroke is a condition that occurs when the blood supply to the brain is interrupted or reduced due to a blockage (ischemic stroke) or rupture of a blood vessel (hemorrhagic stroke). We can use the proportional hazards regression and the stratified proportional hazards model to determine the factors causing diabetes mellitus and stroke. The significant covariate could be seen as a factor causing disease. Significant covariate also yields hazard ratio. If the event time assumed has loglogistics distribution. Some of the functions discussed in this paper, namely the cumulative distribution function, survival function, hazard rate function, and cumulative hazard function. These functions are useful for prediction tools provided the event time is known. The purpose of this study was to find the factors that cause diabetes mellitus and stroke in the gender strata and making functions of random variables with the logistic distribution. The results obtained from this study are the factors that influence a person at risk of stroke namely the type of stroke, hypertension, systolic blood pressure, diastolic blood pressure, and diabetes mellitus. The factors that affect a person at risk of stroke according to gender are the type of stroke, hypertension, systolic blood pressure. If the sample includes all women who have had a stroke, it is primarily affected by the type of stroke. If the sample consists of all men who have had a stroke, it is concluded that the significant covariates are type of stroke, hypertension, systolic blood pressure, and diabetes mellitus. Men have a higher rate of stroke than women.

Optimal stratification of survival data via Bayesian nonparametric mixtures

The stratified proportional hazards model represents a simple solution to take into account heterogeneity within the data while keeping the multiplicative effect of the predictors on the hazard function. Strata are typically defined a priori by resorting to the values of a categorical covariate. A general framework is proposed, which allows the stratification of a generic accelerated lifetime model, including, as a special case, the Weibull proportional hazards model. The stratification is determined a posteriori, taking into account that strata might be characterized by different baseline survivals, and also by different effects of the predictors. This is achieved by considering a Bayesian nonparametric mixture model and the posterior distribution it induces on the space of data partitions. An optimal stratification is then identified following a decision theoretic approach. In turn, stratum-specific inference is carried out. The performance of this method and its robustness to the presence of right-censored observations are investigated through an extensive simulation study. Further illustration is provided analysing a data set from the University of Massachusetts AIDS Research Unit IMPACT Study.

Recipient Obesity and Kidney Transplant Outcomes: A Mate-Kidney Analysis

The impact of extreme recipient obesity on long-term kidney transplant outcomes has been controversial. This study sought to evaluate the association of various levels of recipient obesity on kidney transplantation outcomes by comparing mate-kidney recipient pairs to address possible confounding effects of donor characteristics on posttransplant outcomes. Nationwide observational cohort study using mate-kidney models. In analysis based on the Organ Procurement and Transplant Network/United Network of Organ Sharing database, 44,560 adult recipients of first-time deceased-donor kidney transplants from 2001 through 2016 were paired by donor. Recipient body mass index (BMI) categorized as 18-25 (n= 12,446), >25-30 (n= 15,477), >30-35 (n= 11,144; obese), and >35 (n= 5,493; extreme obesity) kg/m2. Outcomes included patient survival, graft survival, death-censored graft survival, delayed graft function (DGF), and hospital length of stay. Conditional logistic regression and stratified proportional hazards models were used to compare outcomes as odds ratios and hazard ratios (HRs), adjusted for recipient and transplant factors, using recipients with a BMI >35 kg/m2 as a reference. At a median follow-up of 3.9 years, adjusted odds ratios for DGF were 0.42 (95%CI, 0.36-0.48), 0.55 (95% CI, 0.48-0.62), and 0.73 (95% CI, 0.64-0.83) for BMI 18-25, >25-30, and >30-35 kg/m2, respectively (P< 0.001 for all). Death-censored graft failure was less frequent for BMI≤25 and >25-30 kg/m2 (HRs of 0.66 [95% CI, 0.59-0.74] and 0.79 [95% CI, 0.70-0.88], respectively; P< 0.001 for both), but not for BMI >30-35 kg/m2 (HR, 0.91 [95% CI, 0.81-1.02]; P= 0.09). Length of stayand patient survival did not differ by recipient BMI. Observational study with limited detail regarding potential confounders. Despite an increased risk of DGFlikely unrelated to donor organ quality, long-term transplant outcomes among recipients with a BMI >35 kg/m2 are similar to those among recipients with a BMI >30-35 kg/m2, supporting a flexible approach to kidney transplantation candidacy in candidates with extreme obesity.

Stratified proportional subdistribution hazards model with covariate‐adjusted censoring weight for case‐cohort studies

AbstractThe case‐cohort study design is widely used to reduce cost when collecting expensive covariates in large cohort studies with survival or competing risks outcomes. A case‐cohort study dataset consists of two parts: (a) a random sample and (b) all cases or failures from a specific cause of interest. Clinicians often assess covariate effects on competing risks outcomes. The proportional subdistribution hazards model directly evaluates the effect of a covariate on the cumulative incidence function under the non‐covariate‐dependent censoring assumption for the full cohort study. However, the non‐covariate‐dependent censoring assumption is often violated in many biomedical studies. In this article, we propose a proportional subdistribution hazards model for case‐cohort studies with stratified data with covariate‐adjusted censoring weight. We further propose an efficient estimator when extra information from the other causes is available under case‐cohort studies. The proposed estimators are shown to be consistent and asymptotically normal. Simulation studies show (a) the proposed estimator is unbiased when the censoring distribution depends on covariates and (b) the proposed efficient estimator gains estimation efficiency when using extra information from the other causes. We analyze a bone marrow transplant dataset and a coronary heart disease dataset using the proposed method.

SurvBoost: An R Package for High-Dimensional Variable Selection in the Stratified Proportional Hazards Model via Gradient Boosting.

High-dimensional variable selection in the proportional hazards (PH) model has many successful applications in different areas. In practice, data may involve confounding variables that do not satisfy the PH assumption, in which case the stratified proportional hazards (SPH) model can be adopted to control the confounding effects by stratification without directly modeling the confounding effects. However, there is a lack of computationally efficient statistical software for high-dimensional variable selection in the SPH model. In this work an R package, SurvBoost, is developed to implement the gradient boosting algorithm for fitting the SPH model with high-dimensional covariate variables. Simulation studies demonstrate that in many scenarios SurvBoost can achieve better selection accuracy and reduce computational time substantially compared to the existing R package that implements boosting algorithms without stratification. The proposed R package is also illustrated by an analysis of gene expression data with survival outcome in The Cancer Genome Atlas study. In addition, a detailed hands-on tutorial for SurvBoost is provided.

Fertility in adult men born with hypospadias: A nationwide register-based cohort study on birthrates, the use of assisted reproductive technologies and infertility.

Fertility in men with hypospadias may be affected due to anatomical, surgical, or etiological factors and associated conditions. Fertility is further influenced by psychosocial and genetic factors, often shared within families. To evaluate fertility in men born with hypospadias and assess confounding by familial factors. A population-based cohort of 1.2 million men born in Sweden 1964-1998, identified through national demographic and healthcare registers. Associations between hypospadias and (a) being a biological father, (b) conceiving through ART, and (c) diagnosis of male infertility were investigated in the full cohort with logistic regression models and Cox proportional hazard models, expressed as odds ratios (ORs) and hazard ratios (HRs), respectively, with 95% confidence intervals (CIs). A stratified proportional hazard model, conditional on sibling group, was used to control for shared familial confounding. Men with hypospadias, as a whole group, had a lower probability of having biological children (adjusted HR 0.87, 95% CI 0.83-0.92). A significant association was present in both distal (adjusted HR 0.90, 95% CI 0.85-0.96) and proximal hypospadias (HR 0.59, 95% CI 0.42-0.81). Men with hypospadias more often became fathers through ART, regardless of concomitant cryptorchidism. The initial association between hypospadias and the diagnosis of infertility disappeared in sensitivity analyses excluding cryptorchidism. Men with hypospadias displayed lower birthrates as compared to their brothers and the general population. Mere birthrates may, however, be a questionable measure of fertility in a population using family planning. However, men with hypospadias were also at higher risk of reproducing through ART and did more often receive a diagnosis of male infertility. Altogether, these findings indicate impaired fertility in men with hypospadias. Fertility in men with hypospadias is impaired, as shown by lower birthrates, increased use of ART and higher risk of receiving a diagnosis of male infertility.

A goodness-of-fit test for the stratified proportional hazards model for survival data

Goodness-of-fit testing is addressed in the stratified proportional hazards model for survival data. A test statistic based on within-strata cumulative sums of martingale residuals over covariates is proposed and its asymptotic distribution is derived under the null hypothesis of model adequacy. A Monte Carlo procedure is proposed to approximate the critical value of the test. Simulation studies are conducted to examine finite-sample performance of the proposed statistic.

Final analysis of phase III LATITUDE study in patients (pts) with newly diagnosed high-risk metastatic castration-naïve prostate cancer (NDx-HR mCNPC) treated with abiraterone acetate + prednisone (AA+P) added to androgen deprivation therapy (ADT).

141 Background: The first and second interim analyses (IA1 and IA2) of the LATITUDE study in NDx-HR mCNPC pts with addition of AA+P to ADT vs placebo (PBO)+ADT showed significant benefit in the coprimary endpoints of overall survival (OS) and radiographic progression free survival (rPFS), and secondary endpoints. Final OS analysis and updated safety results are presented. Methods: 1,199 pts were randomized (1:1) to AA (1 g QD) + P (5 mg QD) + ADT or PBO + ADT. Primary endpoint of OS and secondary endpoints were assessed (stratified proportional hazards model), and adverse events (AEs) were summarized from final analysis, which was planned for around 852 total death events. Results: Final analysis was conducted after a median follow-up of 51.8 mo (~22 mo from IA1) and 618 deaths were observed (275 [46%] in AA+P and 343 [57%] in PBO). Treatment was ongoing for 157 (26.3%) pts in AA+P; 72 (12.0%) crossed over from PBO to AA+P, of which 59 (81.9%) remained on treatment. AA+P treatment continued to show significant OS benefits vs PBO (HR: 0.7, 95% CI: 0.6-0.8; p&lt;0.0001), along with significant improvements in secondary endpoints (table). Serious AEs were reported in 32.2% pts in AA+P and 25.1% in PBO. Grade 3/4 AEs of special interest (AA+P vs PBO; %) were hypertension (21.9 vs 10.5), hypokalemia (11.7 vs 1.7), hepatotoxicity (8.9 vs 3.5), cardiac disorders (3.9 vs 1) and fluid retention (0.8 vs 1). Conclusions: The final analysis continues to demonstrate a significant OS advantage of combining AA+P to ADT in NDx-HR mCNPC pts. These efficacy findings and AE profiles are consistent with IA1 and IA2 and reinforce AA+P as a standard of care in NDx-HR mCNPC pts. Clinical trial information: NCT01715285. [Table: see text]

Agricultural Suitability and Colonization Choices

Historical evidence indicates that the quest for new territories by the Western European colonizers was driven by a strong incentive to produce sugar, tea, timber, and other commodities. Once arrived, settlers also had a clear interest in becoming self-sufficient in food. This paper establishes that the timing and duration of colonization by the Western European powers were strongly influenced by regional agro-ecological factors. In particular, the potential average crop yield is shown to have a robust effect on both the timing and duration of colonization. Areas with higher agricultural productivity tended to be colonized earlier and for a more extended duration. The findings are robust to controlling for multiple indicators of early development, geographical and ecological conditions, selection issues, and to the utilization of the Cox stratified proportional hazard model. Finally, we provide initial evidence that via its effect on the duration of colonization, agricultural productivity influences contemporary levels of democracy, income inequality, and human capital.

On hazard ratio estimators by proportional hazards models in matched-pair cohort studies

BackgroundIn matched-pair cohort studies with censored events, the hazard ratio (HR) may be of main interest. However, it is lesser known in epidemiologic literature that the partial maximum likelihood estimator of a common HR conditional on matched pairs is written in a simple form, namely, the ratio of the numbers of two pair-types. Moreover, because HR is a noncollapsible measure and its constancy across matched pairs is a restrictive assumption, marginal HR as “average” HR may be targeted more than conditional HR in analysis.MethodsBased on its simple expression, we provided an alternative interpretation of the common HR estimator as the odds of the matched-pair analog of C-statistic for censored time-to-event data. Through simulations assuming proportional hazards within matched pairs, the influence of various censoring patterns on the marginal and common HR estimators of unstratified and stratified proportional hazards models, respectively, was evaluated. The methods were applied to a real propensity-score matched dataset from the Rotterdam tumor bank of primary breast cancer.ResultsWe showed that stratified models unbiasedly estimated a common HR under the proportional hazards within matched pairs. However, the marginal HR estimator with robust variance estimator lacks interpretation as an “average” marginal HR even if censoring is unconditionally independent to event, unless no censoring occurs or no exposure effect is present. Furthermore, the exposure-dependent censoring biased the marginal HR estimator away from both conditional HR and an “average” marginal HR irrespective of whether exposure effect is present. From the matched Rotterdam dataset, we estimated HR for relapse-free survival of absence versus presence of chemotherapy; estimates (95% confidence interval) were 1.47 (1.18–1.83) for common HR and 1.33 (1.13–1.57) for marginal HR.ConclusionThe simple expression of the common HR estimator would be a useful summary of exposure effect, which is less sensitive to censoring patterns than the marginal HR estimator. The common and the marginal HR estimators, both relying on distinct assumptions and interpretations, are complementary alternatives for each other.

The effect of a conditional cash transfer on HIV incidence in young women in rural South Africa (HPTN 068): a phase 3, randomised controlled trial

Cash transfers have been proposed as an intervention to reduce HIV-infection risk for young women in sub-Saharan Africa. However, scarce evidence is available about their effect on reducing HIV acquisition. We aimed to assess the effect of a conditional cash transfer on HIV incidence among young women in rural South Africa. We did a phase 3, randomised controlled trial (HPTN 068) in the rural Bushbuckridge subdistrict in Mpumalanga province, South Africa. We included girls aged 13-20 years if they were enrolled in school grades 8-11, not married or pregnant, able to read, they and their parent or guardian both had the necessary documentation necessary to open a bank account, and were residing in the study area and intending to remain until trial completion. Young women (and their parents or guardians) were randomly assigned (1:1), by use of numbered sealed envelopes containing a randomisation assignment card which were numerically ordered with block randomisation, to receive a monthly cash transfer conditional on school attendance (≥80% of school days per month) versus no cash transfer. Participants completed an Audio Computer-Assisted Self-Interview (ACASI), before test HIV counselling, HIV and herpes simplex virus (HSV)-2 testing, and post-test counselling at baseline, then at annual follow-up visits at 12, 24, and 36 months. Parents or guardians completed a Computer-Assisted Personal Interview at baseline and each follow-up visit. A stratified proportional hazards model was used in an intention-to-treat analysis of the primary outcome, HIV incidence, to compare the intervention and control groups. This study is registered at ClinicalTrials.gov (NCT01233531). Between March 5, 2011, and Dec 17, 2012, we recruited 10 134 young women and enrolled 2537 and their parents or guardians to receive a cash transfer programme (n=1225) or not (control group; n=1223). At baseline, the median age of girls was 15 years (IQR 14-17) and 672 (27%) had reported to have ever had sex. 107 incident HIV infections were recorded during the study: 59 cases in 3048 person-years in the intervention group and 48 cases in 2830 person-years in the control group. HIV incidence was not significantly different between those who received a cash transfer (1·94% per person-years) and those who did not (1·70% per person-years; hazard ratio 1·17, 95% CI 0·80-1·72, p=0·42). Cash transfers conditional on school attendance did not reduce HIV incidence in young women. School attendance significantly reduced risk of HIV acquisition, irrespective of study group. Keeping girls in school is important to reduce their HIV-infection risk. National Institute of Allergy and Infectious Diseases, National Institute of Mental Health of the National Institutes of Health.

Abstract P1-14-02: Disease-free (DFS) and overall survival (OS) data from ACOSOG Z1041 (Alliance) a randomized neoadjuvant trial comparing FEC followed by paclitaxel plus trastuzumab (FEC → P+T) with paclitaxel plus trastuzumab followed by FEC plus trastuzumab (P+T → FEC+T) in HER2-positive operable breast cancer

Abstract Background ACOSOG Z1041 (Alliance) found pathological complete response rates in women with operable HER2-positive breast cancer were similar with FEC → P+T (Arm 1) and P+T → FEC+T (Arm 2) where treatment was administered as 5-FU 500 mg/m2, epirubicin 75 mg/m2 and cyclophosphamide 500 mg/m2 day 1 of a 21-day cycle x 4; paclitaxel 80 mg/m2 weekly x 12 and trastuzumab 4 mg/kg once then 2 mg/kg weekly x 11. We now report DFS and OS results with a median follow up of 4.4 years (range: 26 days to 6.2 years). Methods All patients who began study treatment were included in the analyses. Stratified log rank tests and stratified proportional hazard models were used to assess differences in DFS and OS from randomization between treatment arms. Results From September 15, 2007 to December 15, 2011, 282 women with HER2-positive breast cancer were enrolled. Two patients randomized to arm 1 withdrew consent prior to treatment and are excluded from these analyses. Patient and disease characteristics of the remaining 280 women (arm 1, n=138; arm 2, n=142) were similar between treatment arms. Recurrences and deaths are shown in the table. DFS were not found to differ with respect to treatment arm (stratified logrank p=0.6870; HR stratified (Arm2/Arm1)=0.88; 95%CI: 0.48-1.61). Also, OS was not found to differ with respect to treatment arm (stratified logrank p=0.8790; HR stratified (Arm2/Arm1)=1.07; 95%CI: 0.43-2.69). Conclusions Concurrent administration of trastuzumab with anthracyclines offers no additional benefit in terms of achieving pathologic complete response or improving survival and is not needed. FEC → P+TP+T → FEC+T Arm 1Arm 2n138142Recurrences2022Deaths811 Grant Support: U10 CA76001 The study is registered with ClinicalTrials.gov, number NCT00513292. Citation Format: Buzdar AU, Suman VJ, Meric-Bernstam F, Leitch AM, Ellis MJ, Boughey JC, Unzeitig GW, Royce ME, Ewer MS, Hunt KK. Disease-free (DFS) and overall survival (OS) data from ACOSOG Z1041 (Alliance) a randomized neoadjuvant trial comparing FEC followed by paclitaxel plus trastuzumab (FEC → P+T) with paclitaxel plus trastuzumab followed by FEC plus trastuzumab (P+T → FEC+T) in HER2-positive operable breast cancer. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P1-14-02.

Antibiotics in Fetal and Early Life and Subsequent Childhood Asthma

Objective To investigate the association between exposure to antibiotics in fetal and early life and asthma in childhood, with adjustment for confounding factors. Design Nationwide prospective population based cohort study, including sibling control design. Setting Swedish population identified from national demographic and health registers. Participants 493 785 children born 2006-10; 180 894 of these were eligible for sibling analyses. Main outcome measure Asthma defined as having both an asthma diagnosis and dispensed asthma drugs. The association between antibiotic exposure and asthma was investigated in the whole cohort with Cox proportional hazard regression. A stratified proportional hazards model conditional on sibling group was used to adjust for shared factors within families. Confounding by respiratory infections was assessed by investigating whether specific groups of antibiotics were associated with asthma. Results Antibiotic exposure in fetal life was associated with an increased risk of asthma in cohort analyses (hazard ratio 1.28, 95% confidence interval 1.25 to 1.32), but not in sibling analyses (0.99, 0.92 to 1.07). In cohort analyses, antibiotics used to treat respiratory infections in childhood were associated with a more pronounced increased risk of asthma (4.12, 3.78 to 4.50) than antibiotics used for urinary tract and skin infections (1.54, 1.24 to 1.92). In sibling analyses, the excess risks after exposure to antibiotics for respiratory infections decreased (2.36, 1.78 to 3.13) and disappeared for antibiotics for urinary tract and skin (0.85, 0.47 to 1.55). Conclusions Previous positive associations between exposure to antibiotics in fetal and early life and subsequent childhood asthma could have been caused by confounding by shared familial factors, in addition to confounding by respiratory infections.

Nonparametric Regression-Based Failure Rate Model for Electric Power Equipment Using Lifecycle Data

In order to analyze the fault trends more accurately, a failure rate model appropriate for general electric power equipment is established based on a nonparametric regression method, improved from stratified proportional hazards model (PHM), which can make maximum use of equipment lifecycle data as the covariates, including manufacturer, service age, location, maintainer, health index, etc. All of covariates are represented in the hierarchy process of equipment health condition, which is beneficial for processing and classifying the lifecycle data into multitype recurrent events quantitatively. Meanwhile, based on new definitions of single health cycle and time between events, recurrent inspecting events distributed with martingale process can correspond with event-specific failure function during equipment lifecycle. On this occasion, more inspecting events can be utilized in a complete cycle to predict potential risk and assess equipment health condition. Then, stratified nonparametric PHM is employed to build the multitype recurrent events-specific failure model appropriate for competing risk problem toward interval censored. Lastly, the example in terms of transformers demonstrates the modeling procedure. Results show the well asymptotic property and goodness-of-fit tested by both of graphical and analytical methods. Compared with existing failure models, such as age-based or CBF model, this improved nonparametric regression model can mine lifecycle data acquisition from asset management system, depict the failure trend accurately considering both individual and group features, and lay the foundation for health prognosis, maintenance optimization, and asset management in power grid.