General regression model for the subdistribution of a competing risk under left-truncation and right-censoring.

Left-truncation poses extra challenges for the analysis of complex time-to-event data. We propose a general semiparametric regression model for left-truncated and right-censored competing risks data that is based on a novel weighted conditional likelihood function. Targeting the subdistribution hazard, our parameter estimates are directly interpretable with regard to the cumulative incidence function. We compare different weights from recent literature and develop a heuristic interpretation from a cure model perspective that is based on pseudo risk sets. Our approach accommodates external time-dependent covariate effects on the subdistribution hazard. We establish consistency and asymptotic normality of the estimators and propose a sandwich estimator of the variance. In comprehensive simulation studies we demonstrate solid performance of the proposed method. Comparing the sandwich estimator with the inverse Fisher information matrix, we observe a bias for the inverse Fisher information matrix and diminished coverage probabilities in settings with a higher percentage of left-truncation. To illustrate the practical utility of the proposed method, we study its application to a large HIV vaccine efficacy trial dataset.

Transitions in pregnancy planning in women recruited for a large prospective cohort study.

STUDY QUESTION: Do the rates at which women transition among different intensities of pregnancy planning vary with age, marital status and race/ethnicity? SUMMARY ANSWER: Rates of transition from low or moderate pregnancy probability groups (PPGs) to higher PPGs vary by age, marital status and race/ethnicity. WHAT IS KNOWN ALREADY: The design of prospective studies of the effects of pre- and peri-conception exposures on fecundity, pregnancy and children's health is challenging because at any specific time only a small percentage of reproductive age women is attempting to conceive. To our knowledge, there has been no population-based, prospective study that repeatedly assessed pregnancy planning, which included women who were not already planning pregnancy at enrollment and whose ages spanned the female reproductive age range. STUDY DESIGN, SIZE, DURATION: A longitudinal study was carried out that repeatedly assessed pregnancy probability in 12 916 women for up to 21 months from January 2009 to September 2010. PARTICIPANTS/MATERIALS, SETTING, METHOD: We analyzed data from the National Children's Study Vanguard Study, a pilot study for a large-scale epidemiological birth cohort study of children and their parents. During the Vanguard Study, investigators followed population-based samples of reproductive age women in each of seven geographically dispersed and diverse study locations over time to identify when they sought to become pregnant, providing a unique opportunity to prospectively assess changes in pregnancy planning in a large sample of US women. At study entry and each follow-up contact, which occurred at 1, 3 or 6 month intervals depending on PPG, a questionnaire was used to assess behavior dimensions of pregnancy planning to assign women to low, moderate, high non-tryer and high tryer PPGs. MAIN RESULTS AND THE ROLE OF CHANCE: Crude rates of pregnancy increased with higher assigned PPG, validating the utility of the instrument. The initial PPG and probabilities of transitioning from low or moderate PPG to higher PPG or pregnancy varied with age, marital status and race/ethnicity. Women aged 25 to <35 years had shorter times to transition to higher PPGs or to pregnant compared with women <25 years. Women who were not currently married had longer times to transition from any initial PPG to pregnant, high tryer or high non-tryer status than currently married women. Non-Hispanic Black (NHB) and Hispanic women had shorter time to transition from low or moderate to high non-tryer than non-Hispanic White (NHW) women. NHB women also had shorter time to transition from low to high tryer than NHW women. High tryers are more likely to be aged 25 to <30 years, to be married, and to be Hispanic, NHB or other race/ethnicity than women in the low PPG. LIMITATIONS, REASONS FOR CAUTION: Loss to follow-up varied by age, marital status and race/ethnicity. Although weights were not developed for the Vanguard study, the self-weighting design minimizes the bias of unweighted analysis. Nonetheless, the SEs for some estimates may be under-estimated. WIDER IMPLICATIONS OF THE FINDINGS: Our results show that demographic characteristics are strong predictors of women's behaviors toward pregnancy. The results further show that frequent follow-up assessments of pregnancy planning behavior in large numbers of women are required to recruit an unbiased sample of preconception women. These findings will be useful to investigators designing prospective studies of fecundability, pregnancy outcomes and children's health. STUDY FUNDING/COMPETING INTERESTS: National Institutes of Health (contracts N01-HD53414, N01-HD63416, N01-HD53410, N01-HD53415, N01-HD53396, N01-HD53413 and N01-HD-53411; grant R21 ES016846) and by the University of California Irvine Center for Occupational and Environmental Health. No competing interests. TRIAL REGISTRATION NUMBER: None.

Standardization and validation of a novel and simple method to assess lumbar dural sac size.

AIM: To develop and validate a simple, reproducible method to assess dural sac size using standard imaging technology. MATERIALS AND METHODS: This study was institutional review board-approved. Two readers, blinded to the diagnoses, measured anterior-posterior (AP) and transverse (TR) dural sac diameter (DSD), and AP vertebral body diameter (VBD) of the lumbar vertebrae using MRI images from 53 control patients with pre-existing MRI examinations, 19 prospectively MRI-imaged healthy controls, and 24 patients with Marfan syndrome with prior MRI or CT lumbar spine imaging. Statistical analysis utilized linear and logistic regression, Pearson correlation, and receiver operating characteristic (ROC) curves. RESULTS: AP-DSD and TR-DSD measurements were reproducible between two readers (r = 0.91 and 0.87, respectively). DSD (L1-L5) was not different between male and female controls in the AP or TR plane (p = 0.43; p = 0.40, respectively), and did not vary by age (p = 0.62; p = 0.25) or height (p = 0.64; p = 0.32). AP-VBD was greater in males versus females (p = 1.5 × 10(-8)), resulting in a smaller dural sac ratio (DSR) (DSD/VBD) in males (p = 5.8 × 10(-6)). Marfan patients had larger AP-DSDs and TR-DSDs than controls (p = 5.9 × 10(-9); p = 6.5 × 10(-9), respectively). Compared to DSR, AP-DSD and TR-DSD better discriminate Marfan from control subjects based on area under the curve (AUC) values from unadjusted ROCs (AP-DSD p < 0.01; TR-DSD p = 0.04). CONCLUSION: Individual vertebrae and L1-L5 (average) AP-DSD and TR-DSD measurements are simple, reliable, and reproducible for quantitating dural sac size without needing to control for gender, age, or height.

On the robustness of the adaptive lasso to model misspecification.

Penalization methods have been shown to yield both consistent variable selection and oracle parameter estimation under correct model specification. In this article, we study such methods under model misspecification, where the assumed form of the regression function is incorrect, including generalized linear models for uncensored outcomes and the proportional hazards model for censored responses. Estimation with the adaptive least absolute shrinkage and selection operator, lasso, penalty is proven to achieve sparse estimation of regression coefficients under misspecification. The resulting estimators are selection consistent, asymptotically normal and oracle, where the selection is based on the limiting values of the parameter estimators obtained using the misspecified model without penalization. We further derive conditions under which the penalized estimators from the misspecified model may yield selection consistency under the true model. The robustness is explored numerically via simulation and an application to the Wisconsin Epidemiological Study of Diabetic Retinopathy.

Nonparametric tests for continuous covariate effects with multistate survival data.

SUMMARY: In clinical trials and observational studies, it is often of scientific interest to evaluate the effects of covariates on complex multistate event probabilities. With discrete covariates, nonparametric tests may be constructed using estimates of the relevant quantities. With continuous covariates, a common approach is to arbitrarily discretize the covariates, which may lead to substantial information loss. Another strategy is to formulate the covariate effects in a regression model. Model-based tests may have either low power or be biased under misspecification. We propose nonparametric tests not requiring arbitrary discretization. The tests involve integrals of estimates continuously indexed by dichotomizations of the covariates. General asymptotic results are derived under null and alternative hypotheses, and verified using empirical process theory in several special cases. The tests are consistent under stochastic ordering, which arises naturally with multistate data. A novel nonparametric measure of covariate effect is studied as a natural byproduct of the testing procedure. Simulation studies and two real data analyses demonstrate the gains of the new testing procedure over those based either on categorization or on regression models.

Classification and localization of acetabular labral tears.

OBJECTIVE: The purpose of this study was to compare the findings on hip MR arthrography (MRA) with the published MRA and arthroscopic classifications of hip labral tears and to evaluate a clock-face method for localizing hip labral tears. DESIGN/PATIENTS: We retrospectively reviewed 65 hip MRA studies with correlative hip arthroscopies. Each labrum was evaluated on MRA using the classification system of Czerny and an MRA modification of the Lage arthroscopic classification. In addition, each tear was localized on MRA by using a clock-face description where 6 o'clock was the transverse ligament and 3 o'clock was anterior. These MRA findings were then correlated with the arthroscopic findings using the clock-face method of localization and the Lage arthroscopic classification of labral tears. RESULTS: At MRA, there were 42 Czerny grade 2 and 23 grade 3 labral tears and 22 MRA Lage type 1, 11 type 2, 22 type 3 and 10 type 4 tears. At arthroscopy, there were 10 Lage type 1 flap tears, 20 Lage type 2 fibrillated tears, 18 Lage type 3 longitudinal peripheral tears and 17 Lage type 4 unstable tears. The Czerny MRA classification and the modified MRA Lage classification had borderline correlation with the arthroscopic Lage classification. Localization of the tears using a clock-face description was within 1 o'clock of the arthroscopic localization of the tears in 85% of the patients. CONCLUSIONS: The Lage classification, which is the only published arthroscopic classification system for hip labral tears, does not correlate well with the Czerny MRA or an MRA modification of the Lage classification. Using a clock-face description to localize tears provides a way to accurately localize a labral tear and define its extent.

Aortic aneurysm volume calculation: effect of operator experience.

BACKGROUND: We have successfully applied sequential volumetric analysis of abdominal aortic aneurysms to exclude endoleak in patients who have an aortic endostent. This study compared the effect of variable operator experience on volumetric calculation accuracy. METHODS: Four operators with different experience levels calculated abdominal aneurysm volumes in 10 patients at two different times (>/= 1 week apart). The four reviewers were ranked as having a high level of experience (one full-time laboratory worker specializing in three dimensions with 3 years of experience), a moderate level of experience (one part-time laboratory worker specializing in three dimensions/computed tomographic technician with 1 year of part-time experience), and a low level of experience (two individuals taught volumetric measurements for the purposes of this study: a fellow in abdominal imaging and a computed tomographic technician). All volumes were calculated with a GE Advantage 4.0 workstation (General Electric, Waukesha, WI, USA). RESULTS: Mean aneurysm volume and volume difference between two measurements were calculated for four operators. The average (standard deviation) percent volume differences were 1.2% (0.2%) for the experienced reader, 3.2% (0.3%) for the moderately experienced reader, and 6.0% (1.0%) and 5.8% (1.1%) for the two readers with light experience. Differences between averages were statistically significant (p < 0.005). CONCLUSION: We have defined a percent margin of error for aortic aneurysm volume measurement and have shown a direct correlate to level of experience. Diagnosis of endoleak based on aneurysm volume enlargement on serial scans needs to account for the level of operator experience.

Effect of influenza immunization on CYP3A4 activity.

A number of clinical reports of drug interactions with influenza vaccine have been made. We hypothesized that CYP3A4 activity would decrease following influenza immunization. Fifteen healthy subjects had erythromycin breath tests (ERMBT) and influenza antibody titer hemagglutinin inhibition assay (HIA) before and after receiving influenza vaccine. The mean age of the subjects was 31.9 years (S.D. 10.2). The change in ERMBT following influenza immunization was not significant (mean -4%, S.D. 17%, P=0.25). Influenza immunization does not significantly change CYP3A4 activity. Changes in serum drug concentrations noted previously after influenza immunization are either due to very small changes in CYP3A4 activity or other pharmacokinetic interactions.

Aggregation of refractive error and 5-year changes in refractive error among families in the Beaver Dam Eye Study.

OBJECTIVE: To examine aggregation of refraction, myopia, hyperopia, and astigmatism, as well as the 5-year change in each of these measures, among adult family members. DESIGN: Geographically defined, population-based cohort study in Beaver Dam, Wis. Participants were all 43 to 84 years of age in 1988. Family relationships among participants of the study were identified through interviews. The main outcome measures were noncycloplegic refractions. Aggregation was assessed by Pearson correlations and odds ratios (ORs) that both members of a pair were affected. RESULTS: Age-adjusted sibling correlation of refraction was 0.37 and the OR for a sibling to be myopic was 4.18, whereas the OR for being hyperopic was 2.87 (all statistically significant, P<.05). Correlations and ORs for parent-child and cousin relationships were smaller, and those for spousal relationships were not significant. Correlations and ORs for cylinder power and astigmatism were not statistically significant for most relationships considered. There were no statistically significant correlations or ORs for changes in any measure of refractive error. CONCLUSIONS: The strong aggregation of refractive error, including myopia and hyperopia, among siblings along with weaker associations among parent-child and cousin pairs and no associations among spouses suggest a potential genetic influence on refractive error. There is no such suggestion for a genetic influence on the changes in refraction or in cylinder power and astigmatism.

Statistical issues in the analysis of quantitative traits in combined crosses.

We consider some practical statistical issues in QTL analysis where several crosses originate in multiple inbred parents. Our results show that ignoring background polygenic variation in different crosses may lead to biased interval mapping estimates of QTL effects or loss of efficiency. Threshold and power approximations are derived by extending earlier results based on the Ornstein-Uhlenbeck diffusion process. The results are useful in the design and analysis of genome screen experiments. Several common designs are evaluated in terms of their power to detect QTL.

Joint regression analysis of survival and quality-adjusted survival.

In this paper, a semiparametric bivariate linear regression model for survival and quality-adjusted survival is investigated. Even with a parametric specification for the joint, distribution, maximum likelihood is not applicable because of induced informative censoring. We propose inference procedures based on estimating functions. The estimators are consistent and asymptotically normal. Hypothesis tests and confidence intervals may be constructed with easy-to-implement resampling techniques. Simultaneous regression modeling of survival and quality-adjusted survival has not been studied formally. Our methodology gives parameter estimates that are highly interpretable in the context of a cost-effectiveness analysis. The usefulness of the proposal is illustrated with a breast cancer dataset.

Regression modeling of competing crude failure probabilities.

In a randomized trial of tamoxifen therapy for breast cancer, women can experience tumor recurrence or die from competing causes. One goal of analysis is to describe the effect of tamoxifen on the probabilities of recurrence or death from other causes. To this end, we propose a semi-parametric transformation model for the crude failure probabilities of a competing risk, conditional on covariates. The model is developed as an extension of the standard approach to survival data with independent right censoring. Estimation of the regression coefficients is achieved with a rank-based least squares criterion. Simulations show that the procedure works well with practical sample sizes. A separate estimating function is developed for the baseline parameter. Prediction of covariate-adjusted failure probabilities is considered. The methodology is motivated and illustrated with data from the tamoxifen trial.

Testing for differences in survival with delayed ascertainment.

During the interim stages of most large-scale clinical trials, knowledge that a patient is alive or dead is usually not up-to-date. This is due to the pattern of patient visits to hospitals as well as the administrative set-up used by the study to obtain information on vital status. On a two-armed study, if the process of ascertaining vital status is not the same in both treatment groups, then the standard method of testing based on the logrank statistic may not be applicable. Instead, an ad hoc modification to the logrank test, which artificially truncates follow-up prior to the time of analysis, is often used. These approaches have not been formally addressed in the literature. In the early stages of a clinical trial, severe bias or loss of power may result. For this situation, we propose a class of test statistics that extends the usual class of U statistics. Asymptotic normality is derived by reformulating the statistics in terms of counting processes and employing the theory of U statistics along with martingale techniques. For early interim analyses, a numerical study indicates that the new tests can be more powerful than the current practice when differential ascertainment is present. To illustrate the potential loss of information when lagging follow-up to control for ascertainment delays, we reanalyze an AIDS clinical trial with the truncated logrank and the new statistics.

Prediction of cumulative incidence function under the proportional hazards model.

In the presence of dependent competing risks in survival analysis, the Cox model can be utilized to examine the covariate effects on the cause-specific hazard function for the failure type of interest. For this situation, the cumulative incidence function provides an intuitively appealing summary curve for marginal probabilities of this particular event. In this paper, we show how to construct confidence intervals and bands for such a function under the Cox model for future patients with certain covariates. Our proposals are illustrated with data from a prostate cancer trial.

The Rural Physician as Preceptor: How rural physicians can supervise residents in a rural practice.

Rural family practice offers a wealth of potential clinical teaching material for family practice residents. In order to exploit this potential, the Department of Family Practice at the University of British Columbia has developed a community-based program in which second-year residents in family practice spend the year in rural communities under the supervision of a family physician preceptor. This paper describes some of the problems and satisfactions facing the rural physician who is interested in becoming a preceptor and offers practical guidelines on how a resident can be incorporated into the practice.

Practice morbidity and activity analysis with the aid of a computer. A trainee project.

In a five doctor group practice a system was devised for collecting information about morbidity and practice activity. A structured 'day sheet' was used to record the doctors' case load and this information was then transferred on to a computer file. The computer was programmed to produce monthly morbidity returns and practice activity analyses, printed out in a format that allowed easy comparison to existing published data.