On the role of Volterra integral equations in self-consistent, product-limit, inverse probability of censoring weighted, and redistribution-to-the-right estimators for the survival function

This paper reconsiders several results of historical and current importance to nonparametric estimation of the survival distribution for failure in the presence of right-censored observation times, demonstrating in particular how Volterra integral equations help inter-connect the resulting estimators. The paper begins by considering Efron's self-consistency equation, introduced in a seminal 1967 Berkeley symposium paper. Novel insights provided in the current work include the observations that (i) the self-consistency equation leads directly to an anticipating Volterra integral equation whose solution is given by a product-limit estimator for the censoring survival function; (ii) a definition used in this argument immediately establishes the familiar product-limit estimator for the failure survival function; (iii) the usual Volterra integral equation for the product-limit estimator of the failure survival function leads to an immediate and simple proof that it can be represented as an inverse probability of censoring weighted estimator; (iv) a simple identity characterizes the relationship between natural inverse probability of censoring weighted estimators for the survival and distribution functions of failure; (v) the resulting inverse probability of censoring weighted estimators, attributed to a highly influential 1992 paper of Robins and Rotnitzky, were implicitly introduced in Efron's 1967 paper in its development of the redistribution-to-the-right algorithm. All results developed herein allow for ties between failure and/or censored observations.

Affordability of Forensic Assertive Community Treatment Programs: A Return-on-Investment Analysis

OBJECTIVE

In this study, the authors assessed return on investment (ROI) associated with a forensic assertive community treatment (FACT) program.

METHODS

A retrospective secondary data analysis of a randomized controlled trial comprising 70 legal-involved patients with severe mental illness was conducted in Rochester, New York. Patients were randomly assigned to receive either FACT or outpatient psychiatric treatment including intensive case management. Unit of service costs associated with psychiatric emergency department visits, psychiatric inpatient days, and days in jail were obtained from records of New York State Medicaid and the Department of Corrections. The total dollar value difference between the two trial arms calculated on a per-patient-per-year (PPPY) basis constituted the return from the FACT intervention. The FACT investment cost was defined by the total additional PPPY cost associated with FACT implementation relative to the control group. ROI was calculated by dividing the return by the investment cost.

RESULTS

The estimated return from FACT was $27,588 PPPY (in 2019 dollars; 95% confidence interval [CI]=$3,262-$51,913), which was driven largely by reductions in psychiatric inpatient days, and the estimated investment cost was $18,440 PPPY (95% CI=$15,215-$21,665), implying an ROI of 1.50 (95% CI=0.35-2.97) for FACT.

CONCLUSIONS

The Rochester FACT program was associated with approximately $1.50 return for every $1 spent on its implementation, even without considering potential returns from other sources, including reductions in acute medical care, crime-related damages, and public safety costs. ROI estimates were highly dependent on context-specific factors, particularly Medicaid reimbursement rates for assertive community treatment and hospital stays.

Insertable cardiac monitor-guided early intervention to reduce atrial fibrillation burden following catheter ablation: Study design and clinical protocol (ICM-REDUCE-AF trial)

BACKGROUND

Percutaneous catheter ablation (CA) to achieve pulmonary vein isolation is an effective treatment for drug-refractory paroxysmal and persistent atrial fibrillation (AF). However, recurrence rates after a single AF ablation procedure remain elevated. Conventional management after CA ablation has mostly been based on clinical AF recurrence. However, continuous recordings with insertable cardiac monitors (ICMs) and patient-triggered mobile app transmissions post-CA can now be used to detect early recurrences of subclinical AF (SCAF). We hypothesize that early intervention following CA based on personalized ICM data can prevent the substrate progression that promotes the onset and maintenance of atrial arrhythmias.

METHODS

This is a randomized, double-blind (to SCAF data), single-tertiary center clinical trial in which 120 patients with drug-refractory paroxysmal or persistent AF are planned to undergo CA with an ICM. Randomization will be to an intervention arm (n = 60) consisting of ICM-guided early intervention based on SCAF and patient-triggered mobile app transmissions versus a control arm (n = 60) consisting of a standard intervention protocol based on clinical AF recurrence validated by the ICM. Primary endpoint is AF burden, which will be assessed from ICMs at 15 months post-AF ablation. Secondary endpoints include healthcare utilization, functional capacity, and quality of life.

CONCLUSION

We believe that ICM-guided early intervention will provide a novel, personalized approach to post-AF ablation management that will result in a significant reduction in AF burden, healthcare utilization, and improvements in functional capacity and quality of life.

Regression trees and ensembles for cumulative incidence functions

The use of cumulative incidence functions for characterizing the risk of one type of event in the presence of others has become increasingly popular over the past two decades. The problems of modeling, estimation and inference have been treated using parametric, nonparametric and semi-parametric methods. Efforts to develop suitable extensions of machine learning methods, such as regression trees and ensemble methods, have begun comparatively recently. In this paper, we propose a novel approach to estimating cumulative incidence curves in a competing risks setting using regression trees and associated ensemble estimators. The proposed methods use augmented estimators of the Brier score risk as the primary basis for building and pruning trees, and lead to methods that are easily implemented using existing R packages. Data from the Radiation Therapy Oncology Group (trial 9410) is used to illustrate these new methods.

Screening for chronic diseases: optimizing lead time through balancing prescribed frequency and individual adherence

Screening for chronic diseases, such as cancer, is an important public health priority, but traditionally only the frequency or rate of screening has received attention. In this work, we study the importance of adhering to recommended screening policies and develop new methodology to better optimize screening policies when adherence is imperfect. We consider a progressive disease model with four states (healthy, undetectable preclinical, detectable preclinical, clinical), and overlay this with a stochastic screening-behavior model using the theory of renewal processes that allows us to capture imperfect adherence to screening programs in a transparent way. We show that decreased adherence leads to reduced efficacy of screening programs, quantified here using elements of the lead time distribution (i.e., the time between screening diagnosis and when diagnosis would have occurred clinically in the absence of screening). Under the assumption of an inverse relationship between prescribed screening frequency and individual adherence, we show that the optimal screening frequency generally decreases with increasing levels of non-adherence. We apply this model to an example in breast cancer screening, demonstrating how accounting for imperfect adherence affects the recommended screening frequency.

Intratumoral heterogeneity of second-harmonic generation scattering from tumor collagen and its effects on metastatic risk prediction

Background

Metastases are the leading cause of breast cancer-related deaths. The tumor microenvironment impacts cancer progression and metastatic ability. Fibrillar collagen, a major extracellular matrix component, can be studied using the light scattering phenomenon known as second-harmonic generation (SHG). The ratio of forward- to backward-scattered SHG photons (F/B) is sensitive to collagen fiber internal structure and has been shown to be an independent prognostic indicator of metastasis-free survival time (MFS). Here we assess the effects of heterogeneity in the tumor matrix on the possible use of F/B as a prognostic tool.

Methods

SHG imaging was performed on sectioned primary tumor excisions from 95 untreated, estrogen receptor-positive, lymph node negative invasive ductal carcinoma patients. We identified two distinct regions whose collagen displayed different average F/B values, indicative of spatial heterogeneity: the cellular tumor bulk and surrounding tumor-stroma interface. To evaluate the impact of heterogeneity on F/B’s prognostic ability, we performed SHG imaging in the tumor bulk and tumor-stroma interface, calculated a 21-gene recurrence score (surrogate for OncotypeDX®, or S-ODX) for each patient and evaluated their combined prognostic ability.

Results

We found that F/B measured in tumor-stroma interface, but not tumor bulk, is prognostic of MFS using three methods to select pixels for analysis: an intensity threshold selected by a blinded observer, a histogram-based thresholding method, and an adaptive thresholding method. Using both regression trees and Random Survival Forests for MFS outcome, we obtained data-driven prediction rules that show F/B from tumor-stroma interface, but not tumor bulk, and S-ODX both contribute to predicting MFS in this patient cohort. We also separated patients into low-intermediate (S-ODX < 26) and high risk (S-ODX ≥26) groups. In the low-intermediate risk group, comprised of patients not typically recommended for adjuvant chemotherapy, we find that F/B from the tumor-stroma interface is prognostic of MFS and can identify a patient cohort with poor outcomes.

Conclusions

These data demonstrate that intratumoral heterogeneity in F/B values can play an important role in its possible use as a prognostic marker, and that F/B from tumor-stroma interface of primary tumor excisions may provide useful information to stratify patients by metastatic risk.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12885-020-07713-4.

Robust Q-learning

Q-learning is a regression-based approach that is widely used to formalize the development of an optimal dynamic treatment strategy. Finite dimensional working models are typically used to estimate certain nuisance parameters, and misspecification of these working models can result in residual confounding and/or efficiency loss. We propose a robust Q-learning approach which allows estimating such nuisance parameters using data-adaptive techniques. We study the asymptotic behavior of our estimators and provide simulation studies that highlight the need for and usefulness of the proposed method in practice. We use the data from the "Extending Treatment Effectiveness of Naltrexone" multi-stage randomized trial to illustrate our proposed methods.

Genomics models in radiotherapy: From mechanistic to machine learning

Machine learning (ML) provides a broad framework for addressing high-dimensional prediction problems in classification and regression. While ML is often applied for imaging problems in medical physics, there are many efforts to apply these principles to biological data toward questions of radiation biology. Here, we provide a review of radiogenomics modeling frameworks and efforts toward genomically guided radiotherapy. We first discuss medical oncology efforts to develop precision biomarkers. We next discuss similar efforts to create clinical assays for normal tissue or tumor radiosensitivity. We then discuss modeling frameworks for radiosensitivity and the evolution of ML to create predictive models for radiogenomics.

Modeling the rate of HIV testing from repeated binary data amidst potential never-testers

Many longitudinal studies with a binary outcome measure involve a fraction of subjects with a homogeneous response profile. In our motivating data set, a study on the rate of human immunodeficiency virus (HIV) self-testing in a population of men who have sex with men (MSM), a substantial proportion of the subjects did not self-test during the follow-up study. The observed data in this context consist of a binary sequence for each subject indicating whether or not that subject experienced any events between consecutive observation time points, so subjects who never self-tested were observed to have a response vector consisting entirely of zeros. Conventional longitudinal analysis is not equipped to handle questions regarding the rate of events (as opposed to the odds, as in the classical logistic regression model). With the exception of discrete mixture models, such methods are also not equipped to handle settings in which there may exist a group of subjects for whom no events will ever occur, i.e. a so-called "never-responder" group. In this article, we model the observed data assuming that events occur according to some unobserved continuous-time stochastic process. In particular, we consider the underlying subject-specific processes to be Poisson conditional on some unobserved frailty, leading to a natural focus on modeling event rates. Specifically, we propose to use the power variance function (PVF) family of frailty distributions, which contains both the gamma and inverse Gaussian distributions as special cases and allows for the existence of a class of subjects having zero frailty. We generalize a computational algorithm developed for a log-gamma random intercept model (Conaway, 1990. A random effects model for binary data. Biometrics46, 317-328) to compute the exact marginal likelihood, which is then maximized to obtain estimates of model parameters. We conduct simulation studies, exploring the performance of the proposed method in comparison with competitors. Applying the PVF as well as a Gaussian random intercept model and a corresponding discrete mixture model to our motivating data set, we conclude that the group assigned to receive follow-up messages via SMS was self-testing at a significantly lower rate than the control group, but that there is no evidence to support the existence of a group of never-testers.

Readmission Patterns During Long-Term Follow-Up After Left Ventricular Assist Device Implantation

As more patients are supported for longer periods by a left ventricular assist device (LVAD), hospital readmission is becoming a growing problem. However, data about temporal changes in readmission rates and causes for patients with prolonged LVAD support are limited. We aimed to evaluate rates, causes, and predictors of any and long-term readmission after LVAD placement at our institution. We followed 177 HeartMate II LVAD patients for a mean of 1.90 ± 1.33 years post initial discharge after implantation. A marginal rate model was used to evaluate readmission rates, accounting for mortality. During the first year, the readmission rate was 1.79 (95% confidence interval 1.51 to 2.10) readmissions per year. The readmission rate was 1.54 (95% confidence interval 1.07 to 1.93) 2 to 3 years after initial discharge. There was a further decrease in readmission rate in the 3- to 4-year interval. The most common causes of readmission during the first year and even after 3 to 4 years of LVAD support were bleeding (excluding intracranial bleeding) and infection. Female gender was associated with an increased risk of readmission in multivariable analyses, while blood urea nitrogen was predictive of long-term readmissions. In conclusion, readmission after LVAD implantation is common, but readmission rates decrease during long-term follow-up. Bleeding and infection remain leading causes of readmission during longer follow-up and strategies to decrease these complications may reduce readmission rates. Female patients and patients with renal dysfunction have increased risk of readmission and further studies are needed to improve outcomes in these groups.

Censoring Unbiased Regression Trees and Ensembles

This paper proposes a novel paradigm for building regression trees and ensemble learning in survival analysis. Generalizations of the CART and Random Forests algorithms for general loss functions, and in the latter case more general bootstrap procedures, are both introduced. These results, in combination with an extension of the theory of censoring unbiased transformations applicable to loss functions, underpin the development of two new classes of algorithms for constructing survival trees and survival forests: Censoring Unbiased Regression Trees and Censoring Unbiased Regression Ensembles. For a certain "doubly robust" censoring unbiased transformation of squared error loss, we further show how these new algorithms can be implemented using existing software (e.g., CART, random forests). Comparisons of these methods to existing ensemble procedures for predicting survival probabilities are provided in both simulated settings and through applications to four datasets. It is shown that these new methods either improve upon, or remain competitive with, existing implementations of random survival forests, conditional inference forests, and recursively imputed survival trees.

Regularity of a renewal process estimated from binary data

Assessment of the regularity of a sequence of events over time is important for clinical decision-making as well as informing public health policy. Our motivating example involves determining the effect of an intervention on the regularity of HIV self-testing behavior among high-risk individuals when exact self-testing times are not recorded. Assuming that these unobserved testing times follow a renewal process, the goals of this work are to develop suitable methods for estimating its distributional parameters when only the presence or absence of at least one event per subject in each of several observation windows is recorded. We propose two approaches to estimation and inference: a likelihood-based discrete survival model using only time to first event; and a potentially more efficient quasi-likelihood approach based on the forward recurrence time distribution using all available data. Regularity is quantified and estimated by the coefficient of variation (CV) of the interevent time distribution. Focusing on the gamma renewal process, where the shape parameter of the corresponding interevent time distribution has a monotone relationship with its CV, we conduct simulation studies to evaluate the performance of the proposed methods. We then apply them to our motivating example, concluding that the use of text message reminders significantly improves the regularity of self-testing, but not its frequency. A discussion on interesting directions for further research is provided.

A Randomized Controlled Trial of the Rochester Forensic Assertive Community Treatment Model

OBJECTIVE

Forensic assertive community treatment (FACT) is an adaptation of the assertive community treatment model and is designed to serve justice-involved adults with serious mental illness. This study compared the effectiveness of a standardized FACT model and enhanced treatment as usual in reducing jail and hospital use and in promoting engagement in outpatient mental health services.

METHODS

Seventy adults with psychotic disorders who were arrested for misdemeanor crimes and who were eligible for conditional discharge were recruited from the Monroe County, New York, court system. Participants were randomly assigned to receive either FACT (N=35) or enhanced treatment as usual (N=35) for one year. Criminal justice and mental health service utilization outcomes were measured by using state and county databases.

RESULTS

Forty-nine participants (70%) completed the full one-year intervention period. Nineteen (27%) were removed early by judicial order, one was removed by county health authorities, and one died of a medical illness. Intent-to-treat analysis for all 70 participants showed that those receiving the FACT intervention had fewer mean±SD convictions (.4±.7 versus .9±1.3, p=.023), fewer mean days in jail (21.5±25.9 versus 43.5±59.2, p=.025), fewer mean days in the hospital (4.4±15.1 versus 23.8±64.2, p=.025), and more mean days in outpatient mental health treatment (305.5±92.1 versus 169.4±139.6, p<.001) compared with participants who received treatment as usual.

CONCLUSIONS

The Rochester FACT model was associated with fewer convictions for new crimes, less time in jail and hospitals, and more time in outpatient treatment among justice-involved adults with psychotic disorders compared with treatment as usual.

Estimation in the semiparametric accelerated failure time model with missing covariates: improving efficiency through augmentation

This paper considers linear regression with missing covariates and a right censored outcome. We first consider a general two-phase outcome sampling design, where full covariate information is only ascertained for subjects in phase two and sampling occurs under an independent Bernoulli sampling scheme with known subject-specific sampling probabilities that depend on phase one information (e.g., survival time, failure status and covariates). The semiparametric information bound is derived for estimating the regression parameter in this setting. We also introduce a more practical class of augmented estimators that is shown to improve asymptotic efficiency over simple but inefficient inverse probability of sampling weighted estimators. Estimation for known sampling weights and extensions to the case of estimated sampling weights are both considered. The allowance for estimated sampling weights permits covariates to be missing at random according to a monotone but unknown mechanism. The asymptotic properties of the augmented estimators are derived and simulation results demonstrate substantial efficiency improvements over simpler inverse probability of sampling weighted estimators in the indicated settings. With suitable modification, the proposed methodology can also be used to improve augmented estimators previously used for missing covariates in a Cox regression model.

Doubly robust survival trees

Estimating a patient's mortality risk is important in making treatment decisions. Survival trees are a useful tool and employ recursive partitioning to separate patients into different risk groups. Existing 'loss based' recursive partitioning procedures that would be used in the absence of censoring have previously been extended to the setting of right censored outcomes using inverse probability censoring weighted estimators of loss functions. In this paper, we propose new 'doubly robust' extensions of these loss estimators motivated by semiparametric efficiency theory for missing data that better utilize available data. Simulations and a data analysis demonstrate strong performance of the doubly robust survival trees compared with previously used methods. Copyright © 2016 John Wiley & Sons, Ltd.

The choice of prior distribution for a covariance matrix in multivariate meta-analysis: a simulation study

Bayesian meta-analysis is an increasingly important component of clinical research, with multivariate meta-analysis a promising tool for studies with multiple endpoints. Model assumptions, including the choice of priors, are crucial aspects of multivariate Bayesian meta-analysis (MBMA) models. In a given model, two different prior distributions can lead to different inferences about a particular parameter. A simulation study was performed in which the impact of families of prior distributions for the covariance matrix of a multivariate normal random effects MBMA model was analyzed. Inferences about effect sizes were not particularly sensitive to prior choice, but the related covariance estimates were. A few families of prior distributions with small relative biases, tight mean squared errors, and close to nominal coverage for the effect size estimates were identified. Our results demonstrate the need for sensitivity analysis and suggest some guidelines for choosing prior distributions in this class of problems. The MBMA models proposed here are illustrated in a small meta-analysis example from the periodontal field and a medium meta-analysis from the study of stroke. Copyright © 2015 John Wiley & Sons, Ltd.

The Care Span: Lower Hispanic participation in Medicare Part D may reflect program barriers

Despite the successes of Medicare's Part D prescription drug program, an estimated 12.5 percent of Americans ages sixty-five and older do not have prescription drug coverage. It is possible that some who remain without coverage do so for rational economic reasons. However, barriers to insurance uptake, such as the program's complexity, may exist for certain elderly people. Racial and ethnic minorities may be particularly susceptible to these barriers. To investigate the role that race and ethnicity may play in Medicare Part D participation, we analyzed data from the 2011 National Health and Aging Trends Study. We found that Hispanics were 35 percent less likely than non-Hispanic whites to have coverage, after individual predictors of prescription drug demand were controlled for. There was no statistically significant difference in Part D coverage between non-Hispanic blacks and non-Hispanic whites. Results of a stratified analysis suggest that the difference between Hispanics and non-Hispanic whites in Part D coverage may be driven by ethnic disparities among those eligible for the low-income Part D subsidy but not automatically enrolled in it. Further research is needed to identify both the exact mechanisms underlying the observed differential uptake in the rapidly growing elderly Hispanic population and potential policy-based solutions.

Implementation of a mortality prediction rule for real-time decision making: feasibility and validity

BACKGROUND

A previously published, retrospectively derived prediction rule for death within 30 days of hospital admission has the potential to launch parallel interdisciplinary team activities. Whether or not patient care improves will depend on the validity of prospectively generated predictions, and the feasibility of generating them on demand for a critical proportion of inpatients.

OBJECTIVE

To determine the feasibility of generating mortality predictions on admission and to validate their accuracy using the scoring weights of the retrospective rule.

DESIGN

Prospective, sequential cohort.

SETTING

Large, tertiary care, community hospital in the Midwestern United States

PATIENTS

Adult patients admitted from the emergency department or scheduled for elective surgery

RESULTS

Mortality predictions were generated on demand at the beginning of the hospitalization for 9312 (92.9%) out of a possible 10,027 cases. The area under the receiver operating curve for 30-day mortality was 0.850 (95% confidence interval: 0.833-0.866), indicating very good to excellent discrimination. The prospectively generated 30-day mortality risk had a strong association with the receipt of palliative care by hospital discharge, in-hospital mortality, and 180-day mortality, a fair association with the risk for 30-day readmissions and unplanned transfers to intensive care, and weak associations with receipt of intensive unit care ever within the hospitalization or the development of a new diagnosis that was not present on admission (ie, complication).

CONCLUSIONS

Important prognostic information is feasible to obtain in a real-time, single-assessment process for a sizeable proportion of hospitalized patients.

Continuity of care and the risk of preventable hospitalization in older adults

IMPORTANCE

Preventable hospitalizations are common among older adults for reasons that are not well understood.

OBJECTIVE

To determine whether Medicare patients with ambulatory visit patterns indicating higher continuity of care have a lower risk of preventable hospitalization.

DESIGN

Retrospective cohort study.

SETTING

Ambulatory visits and hospital admissions.

PARTICIPANTS

Continuously enrolled fee-for-service Medicare beneficiaries older than 65 years with at least 4 ambulatory visits in 2008.

EXPOSURES

The concentration of patient visits with physicians measured for up to 24 months using the continuity of care score and usual provider continuity score on a scale from 0 to 1.

MAIN OUTCOMES AND MEASURES

Index occurrence of any 1 of 13 preventable hospital admissions, censoring patients at the end of their 24-month follow-up period if no preventable hospital admissions occurred, or if they died.

RESULTS

Of the 3,276,635 eligible patients, 12.6% had a preventable hospitalization during their 2-year observation period, most commonly for congestive heart failure (25%), bacterial pneumonia (22.7%), urinary infection (14.9%), or chronic obstructive pulmonary disease (12.5%). After adjustment for patient baseline characteristics and market-level factors, a 0.1 increase in continuity of care according to either continuity metric was associated with about a 2% lower rate of preventable hospitalization (continuity of care score hazard ratio [HR], 0.98 [95% CI, 0.98-0.99; usual provider continuity score HR, 0.98 [95% CI, 0.98-0.98). Continuity of care was not related to mortality rates.

CONCLUSIONS AND RELEVANCE

Among fee-for-service Medicare beneficiaries older than 65 years, higher continuity of ambulatory care is associated with a lower rate of preventable hospitalization.

Mortality predictions on admission as a context for organizing care activities

BACKGROUND

Favorable health outcomes are more likely to occur when the clinical team recognizes patients at risk and intervenes in consort. Prediction rules can identify high-risk subsets, but the availability of multiple rules for various conditions present implementation and assimilation challenges.

METHODS

A prediction rule for 30-day mortality at the beginning of the hospitalization was derived in a retrospective cohort of adult inpatients from a community hospital in the Midwestern United States from 2008 to 2009, using clinical laboratory values, past medical history, and diagnoses present on admission. It was validated using 2010 data from the same and from a different hospital. The calculated mortality risk was then used to predict unplanned transfers to intensive care units, resuscitation attempts for cardiopulmonary arrests, a condition not present on admission (complications), intensive care unit utilization, palliative care status, in-hospital death, rehospitalizations within 30 days, and 180-day mortality.

RESULTS

The predictions of 30-day mortality for the derivation and validation datasets had areas under the receiver operating characteristic curve of 0.88. The 30-day mortality risk was in turn a strong predictor for in-hospital death, palliative care status, 180-day mortality; a modest predictor for unplanned transfers and cardiopulmonary arrests; and a weaker predictor for the other events of interest.

CONCLUSIONS

The probability of 30-day mortality provides health systems with an array of prognostic information that may provide a common reference point for organizing the clinical activities of the many health professionals involved in the care of the patient.

Randomized controlled effectiveness trial of reciprocal peer support in heart failure

BACKGROUND

Although disease management programs for patients hospitalized with heart failure (HF) are effective, they are, however, often resource intensive, limiting their uptake. Peer support programs have led to improved outcomes among patients with other chronic conditions and may result in similar improvements for patients with HF.

METHODS AND RESULTS

In this randomized controlled trial, reciprocal peer support (RPS) arm patients participated in a HF nurse practitioner-led goal setting group session, received brief training in peer communication skills, and were paired with another participant in their cohort with whom they were encouraged to talk weekly using a telephone platform. Participants were also encouraged to attend 3 nurse practitioner-facilitated peer support group sessions. Patients in the nurse care management arm attended a nurse practitioner-led session to address their HF care questions and receive HF educational materials and information on how to access care management services. The median age of the patients was 69 years; 51% were female and 26% were racial/ethnic minorities. Only 55% of RPS patients participated in peer calls or group sessions. In intention-to-treat analyses, the RPS and nurse care management groups did not differ in time-to-first all-cause rehospitalization or death or in mean numbers of rehospitalizations or deaths. There were no differences in improvements in 6-month measures of HF-specific quality of life or social support. Conclusions- Among patients recently hospitalized for HF, more than half of RPS participants had no or minimal engagement with the RPS program, and the program did not improve outcomes compared with usual HF nurse care management.

CLINICAL TRIAL REGISTRATION

URL: http://www.clinicaltrials.gov. UNIQUE IDENTIFIER: NCT00508508.

Analyzing repeated measures semi-continuous data, with application to an alcohol dependence study

Two-part random effects models (Olsen and Schafer,(1) Tooze et al.(2)) have been applied to repeated measures of semi-continuous data, characterized by a mixture of a substantial proportion of zero values and a skewed distribution of positive values. In the original formulation of this model, the natural logarithm of the positive values is assumed to follow a normal distribution with a constant variance parameter. In this article, we review and consider three extensions of this model, allowing the positive values to follow (a) a generalized gamma distribution, (b) a log-skew-normal distribution, and (c) a normal distribution after the Box-Cox transformation. We allow for the possibility of heteroscedasticity. Maximum likelihood estimation is shown to be conveniently implemented in SAS Proc NLMIXED. The performance of the methods is compared through applications to daily drinking records in a secondary data analysis from a randomized controlled trial of topiramate for alcohol dependence treatment. We find that all three models provide a significantly better fit than the log-normal model, and there exists strong evidence for heteroscedasticity. We also compare the three models by the likelihood ratio tests for non-nested hypotheses (Vuong(3)). The results suggest that the generalized gamma distribution provides the best fit, though no statistically significant differences are found in pairwise model comparisons.

Classical conditioning through auditory stimuli in Drosophila: methods and models

The role of sound in Drosophila melanogaster courtship, along with its perception via the antennae, is well established, as is the ability of this fly to learn in classical conditioning protocols. Here, we demonstrate that a neutral acoustic stimulus paired with a sucrose reward can be used to condition the proboscis-extension reflex, part of normal feeding behavior. This appetitive conditioning produces results comparable to those obtained with chemical stimuli in aversive conditioning protocols. We applied a logistic model with general estimating equations to predict the dynamics of learning, which successfully predicts the outcome of training and provides a quantitative estimate of the rate of learning. Use of acoustic stimuli with appetitive conditioning provides both an alternative to models most commonly used in studies of learning and memory in Drosophila and a means of testing hearing in both sexes, independently of courtship responsiveness.

Effect of Johne's disease status on reproduction and culling in dairy cattle

Among the costs attributed to Mycobacterium avium ssp. paratuberculosis (MAP) infection in dairy cattle, the effects on reproduction and culling are the least documented. To estimate the cost of MAP infections and Johne's disease in a dairy herd, the rates of calving and culling were calculated for cows in each stage of MAP infection relative to uninfected cows. Data from 6 commercial dairy herds, consisting of 2,818 cows with 2,754 calvings and 1,483 cullings, were used for analysis. Every cow in each study herd was tested regularly for MAP, and herds were followed for between 4 and 7 yr. An ordinal categorical variable for Johne's disease status [test-negative, low-positive (low-shedding or ELISA-positive only), or high-shedding] was defined as a time-dependent variable for all cows with at least 1 positive test result or 2 negative test results. A Cox regression model, stratified on herd and controlling for the time-dependent infection variable, was used to analyze time to culling. Nonshedding animals were significantly less likely to be culled in comparison with animals in the low-shedding or ELISA-positive category, and high-shedding animals had nonsignificantly higher culling rates than low-shedding or ELISA-positive animals. Time to calving was analyzed using a proportional rates model, an analog to the Andersen-Gill regression model suitable for recurrent event data, stratifying on herd and weighted to adjust for the dependent censoring caused by the culling effects described above. High-shedding animals had lower calving rates in comparison with low-shedding or ELISA-positive animals, which tended to have higher calving rates than test-negative animals.

A flexible two-part random effects model for correlated medical costs

In this paper, we propose a flexible "two-part" random effects model (Olsen and Schafer, 2001; Tooze et al., 2002) for correlated medical cost data. Typically, medical cost data are right-skewed, involve a substantial proportion of zero values, and may exhibit heteroscedasticity. In many cases, such data are also obtained in hierarchical form, e.g., on patients served by the same physician. The proposed model specification therefore consists of two generalized linear mixed models (GLMM), linked together by correlated random effects. Respectively, and conditionally on the random effects and covariates, we model the odds of cost being positive (Part I) using a GLMM with a logistic link and the mean cost (Part II) given that costs were actually incurred using a generalized gamma regression model with random effects and a scale parameter that is allowed to depend on covariates (cf., Manning et al., 2005). The class of generalized gamma distributions is very flexible and includes the lognormal, gamma, inverse gamma and Weibull distributions as special cases. We demonstrate how to carry out estimation using the Gaussian quadrature techniques conveniently implemented in SAS Proc NLMIXED. The proposed model is used to analyze pharmacy cost data on 56,245 adult patients clustered within 239 physicians in a mid-western U.S. managed care organization.

Decomposing trends in nonmarital fertility among Latinas

CONTEXT

For Latinos, high rates of nonmarital fertility reinforce economic inequality and slow the pace of social and economic incorporation into American society.

METHODS

Changes in the nonmarital fertility ratio--nonmarital births as a percentage of all births (NMFR)--among women aged 15-44 over the period 1994-2005 were partitioned into three components: changes in marital and in nonmarital fertility, and in the proportion of women who were married. Annual birth data were drawn from the national Natality Detail File, and population estimates were drawn from the Current Population Surveys. Analyses were conducted for blacks, whites and Latinas, as well as for selected subgroups of Latinas; differences in NMFRs between racial and ethnic groups were also calculated.

RESULTS

NMFRs were largely unchanged between 1994 and 2002, and then began to rise; they averaged 43% for Latinas, 69% for blacks and 23% for whites over the study period. In 2005, 48% of births to Latinas were nonmarital. Most of the rise in Latinas' NMFR was linked to a decline in marriage. Among foreign-born Latinas, a six-percentage-point increase in the NMFR was due mostly to a rise in nonmarital fertility and a decline in marital fertility, which offset the beneficial effects of a rising marriage rate. The difference between Latinas' and whites' NMFRs was largely attributed to Latinas' higher nonmarital fertility, whereas the difference between blacks' and whites' NMFRs was driven mostly by lower marriage rates among blacks.

CONCLUSIONS

Efforts to reduce out-of-wedlock childbearing among Latinas are needed, and programs should promote healthy marriages, especially among foreign-born Latinas.

Induced smoothing for the semiparametric accelerated failure time model: asymptotics and extensions to clustered data

This paper extends the induced smoothing procedure of Brown & Wang (2006) for the semiparametric accelerated failure time model to the case of clustered failure time data. The resulting procedure permits fast and accurate computation of regression parameter estimates and standard errors using simple and widely available numerical methods, such as the Newton-Raphson algorithm. The regression parameter estimates are shown to be strongly consistent and asymptotically normal; in addition, we prove that the asymptotic distribution of the smoothed estimator coincides with that obtained without the use of smoothing. This establishes a key claim of Brown & Wang (2006) for the case of independent failure time data and also extends such results to the case of clustered data. Simulation results show that these smoothed estimates perform as well as those obtained using the best available methods at a fraction of the computational cost.

Conditional GEE for recurrent event gap times

This paper deals with the analysis of recurrent event data subject to censored observation. Using a suitable adaptation of generalized estimating equations for longitudinal data, we propose a straightforward methodology for estimating the parameters indexing the conditional means and variances of the process interevent (i.e. gap) times. The proposed methodology permits the use of both time-fixed and time-varying covariates, as well as transformations of the gap times, creating a flexible and useful class of methods for analyzing gap-time data. Censoring is dealt with by imposing a parametric assumption on the censored gap times, and extensive simulation results demonstrate the relative robustness of parameter estimates even when this parametric assumption is incorrect. A suitable large-sample theory is developed. Finally, we use our methods to analyze data from a randomized trial of asthma prevention in young children.

Predictors and outcomes of emergency department visits within 30 days following percutaneous coronary intervention

Our objective was to determine the frequency and predictive factors for cardiac-related emergency department (ED) encounters within 30 days after percutaneous coronary intervention (PCI). The data source was an electronic database of 2,731 patients who had PCI from 2002 to 2004. Almost all underwent stent placement. Risk factors for returning to the ED were identified from clinical, anatomic, and demographic candidate variables using multivariate logistic regression. Approximately 9% of the cohort (255 of 2,731 patients) returned to the ED for cardiac reasons within 30 days, peaking around 3 days. ED visits were more likely in those whose index PCI was emergent or urgent (odds ratio [OR] 2.0, 95% confidence interval [CI] 1.3 to 3.0), in women (OR 1.9, 95% CI 1.5 to 2.5), and in those who had previous encounters with the ED or hospital (OR 1.7, 95% CI 1.5 to 2.0). Patients receiving stents were somewhat less likely to return (OR 0.7, 95% CI 0.5 to 1.0). In conclusion, the clinical courses of the 255 returning patients were generally benign, but 12% had a subsequent myocardial infarction or repeat PCI within 30 days of the ED encounter.

Polymorphisms in cytoplasmic serine hydroxymethyltransferase and methylenetetrahydrofolate reductase affect the risk of cardiovascular disease in men

Genetic variation in folate-regulating enzymes contributes to the risk of cardiovascular disease (CVD). The cytoplasmic serine hydroxymethyltransferase (cSHMT) enzyme is proposed to regulate a key metabolic intersection in folate metabolism. We hypothesized that a variant in cSHMT (cSHMT 1420C-->T) affects CVD risk, and that the effect depends on a linked step in the metabolic pathway catalyzed by methylenetetrahydrofolate reductase (MTHFR). A nested case-control study of incident CVD was conducted within the all-male Normative Aging Study cohort. Of the incident CVD cases, 507 had DNA samples; 2 controls/case were selected by risk set sampling (matched on age and birth year). A significant gene-gene interaction (P-values 0.0013, 0.0064) was found between MTHFR and cSHMT, and there was little or no change in the coefficients in covariate-adjusted models. The effect of MTHFR 677C-->T genotype on CVD risk varied by cSHMT 1420C-->T genotype. Among men with cSHMT 1420C-->T TT genotype, the odds ratios (OR) for CVD risk for MTHFR 677C-->T CT and TT genotypes compared with the MTHFR 677C-->T CC genotype were 3.6 (95% CI, 1.7-7.8) and 10.6 (95% CI, 2.5-46.0), respectively. Among men with the cSHMT 1420C-->T CC/CT genotype, the corresponding ORs were 1.0 (95% CI, 0.8-1.2) and 1.3 (95% CI, 0.9-1.8). Plasma total homocysteine concentrations were highest in the subgroup of men with both polymorphisms, MTHFR 677C-->T TT and cSHMT 1420C-->T TT, consistent with a higher risk of CVD in this subgroup. A more complete understanding of the molecular mechanism awaits identification of the functional effect of the polymorphism.

Effects of diseases on reproductive performance in Swedish Red and White dairy cattle

The objectives of this study were to evaluate the effect of diseases on days open (DO), days to first breeding (DFB) and days from first breeding to conception (DFBC) using survival analysis models, and to assess the significance of the sire component and its possible confounding effect. The data consisted of a random sample of 20% of all herds enrolled in Swedish recording system and using 100% artificial insemination with at least 15 Swedish Red and White cows calving in 1991. The follow-up period was from 45 to 145 d after calving. After editing, the data sets had 23,927, 28,197, and 22,089 cows for days open, days to first breeding, and days from first breeding to conception, respectively. The Cox models included parity, calving season, cow milk production and age at first calving as fixed effects, and herd and sire as random effects. Ten disease groups were considered as possible risk factors for the reproductive traits. Disease groups were treated differently if they occurred before or after 45 d postpartum. Diseases occurring in the first 45 d after calving were treated as time-independent covariates and diseases occurring after day 45 were treated as time-dependent covariates for days open and days to first breeding. The percentages of censored cows were 35% for days open, 19% for days to first breeding, and 33% for days from first breeding to conception. Days open increased in cows with dystocia, stillbirth, retained placenta, metritis, or other diseases occurring in the first 45 d after calving, and in cows with metritis, mastitis, or other diseases occurring after 45 d. Days to first breeding increased in cows with stillbirth, retained placenta, milk fever, mastitis, foot and leg problems, or other diseases occurring before day 45, and in cows with metritis, mastitis, foot and leg problems, or other diseases occurring after 45 d. Days to first breeding decreased in cows treated for ovulatory dysfunctions either before or after 45 d. Days from first breeding to conception increased in cows with dystocia, stillbirth, retained placenta, metritis, or ovulatory dysfunctions occurring before first breeding, and in cows with mastitis occurring after first breeding. Although the additive genetic components were significant for all traits considered, the sires did not act as confounders because only a small amount of variability for the traits considered in this study was explained by the sires, with estimated heritabilities of 2% on the logarithmic scale and from 3 to 4% on the real scale.

Using a standardized donor ratio to assess the performance of organ procurement organizations

OBJECTIVE

To develop a Standardized Donor Ratio (SDR) as an outcome measure for evaluating the effectiveness of organ procurement organizations (OPOs).

DATA SOURCES/STUDY SETTING

All deaths by cause in the United States during 1993-1994 as reported in the Vital Mortality Statistics, Multiple Cause of Death files. The OPO-specific data were provided by the United Network for Organ Sharing (UNOS).

STUDY DESIGN

Each OPO's expected number of donors was calculated by applying national donation rates to deaths with potential for donation in 24 age, sex, and race cells. The SDR was calculated by dividing the observed number of donors by the expected number. The chi2 tests of the hypothesis that the OPO's performance differed from the national norm of 1.0 were performed. The SDR was compared to the existing performance standard based on the unadjusted number of donors per million live population in the OPO's service area. An ordinary least squares (OLS) regression assessed predictors of the SDR.

PRINCIPAL FINDINGS

The SDRs ranged from 0.41 to 1.99. Twenty-nine of 64 OPOs had SDRs significantly different than 1.0. The SDRs were positively associated with the percent of white living population and the number of organ types transplanted per transplant center served by the OPO.

CONCLUSIONS

The SDRs can be used by Centers for Medicare and Medicaid Services (CMS), UNOS, and OPOs to target quality improvement initiatives, present more accurate comparisons of OPO performance, and develop public policy on the evaluation of the effectiveness of organ procurement efforts.

Quantifying the physician contribution to managed care pharmacy expenses: a random effects approach

BACKGROUND

Despite the availability of more sophisticated techniques, few alternatives to ordinary least squares (OLS) regression have been utilized to profile physician prescribing in managed care. It is not known to what extent the modest R values derived from OLS models reflect incomplete risk adjustment or widely varying physician prescribing patterns.

OBJECTIVES

To quantify the role of interphysician variability relative to overall variability in managed care pharmacy expenses, and to examine the extent to which different statistical approaches generate meaningful differences in profile results.

RESEARCH DESIGN

Comparison of three basic statistical modeling approaches: OLS, fixed effects regression, and random effects (ie, hierarchical) regression models.

SETTING

Two managed care populations that differed more than 2-fold in per member pharmacy expenditures in 1999, one from the Midwestern United States, the other from three Western States.

MAIN OUTCOME MEASURES

The intraclass correlation coefficient (ICC, the proportion of variability in expenses attributable to differences among physicians) and the range of projected expenses attributed to each physician's prescribing style.

RESULTS

The ICCs were small for aggregated pharmacy expenditures, 0.04 or less in both populations. As determined by OLS, the most costly physician contributed 94,399 U.S. dollars in excess expenses to the organization whereas the most parsimonious saved 89,940 U.S. dollars. When derived from random effects models, the range in performance was 63% of that derived from OLS.

CONCLUSIONS

In the populations studied, systematic prescribing differences among physicians were small relative to the overall variability in pharmacy expenses, suggesting other factors were more likely driving these costs. Random effects models generated smaller estimates of the individual physicians' contribution to costs, sometimes considerably, relative to those derived from OLS and fixed effects approaches.

Flexible hazard regression modeling for medical cost data

The modeling of lifetime (i.e. cumulative) medical cost data in the presence of censored follow-up is complicated by induced informative censoring, rendering standard survival analysis tools invalid. With few exceptions, recently proposed nonparametric estimators for such data do not extend easily to handle covariate information. We propose to model the hazard function for lifetime cost endpoints using an adaptation of the HARE methodology (Kooperberg, Stone, and Truong, Journal of the American Statistical Association, 1995, 90, 78-94). Linear splines and their tensor products are used to adaptively build a model that incorporates covariates and covariate-by-cost interactions without restrictive parametric assumptions. The informative censoring problem is handled using inverse probability of censoring weighted estimating equations. The proposed method is illustrated using simulation and also with data on the cost of dialysis for patients with end-stage renal disease.

Clinical significance of histological classification of idiopathic interstitial pneumonia

Patients with idiopathic interstitial pneumonias (IIPs) can be subdivided into groups based on the histological appearance of lung tissue obtained by surgical biopsy. The quantitative impact of histological diagnosis, baseline factors and response to therapy on survival has not been evaluated. Surgical lung biopsy specimens from 168 patients with suspected IIP were reviewed according to the latest diagnostic criteria. The impact of baseline clinical, physiological, radiographic and histological features on survival was evaluated using Cox regression analysis. The predictive value of honeycombing on high-resolution computed tomography (HRCT) as a surrogate marker for usual interstitial pneumonia (UIP) was examined. The response to therapy and survival of 39 patients treated prospectively with high-dose prednisone was evaluated. The presence of UIP was the most important factor influencing mortality. The risk ratio of mortality when UIP was present was 28.46 (95% confidence interval (CI) 5.5-148.0; p=0.0001) after controlling for patient age, duration of symptoms, radiographic appearance, pulmonary physiology, smoking history and sex. Honeycombing on HRCT indicated the presence of UIP with a sensitivity of 90% and specificity of 86%. Patients with nonspecific interstitial pneumonia were more likely to respond or remain stable (9 of 10) compared to patients with UIP (14 of 29) after treatment with prednisone. Patients remaining stable had the best prognosis. The risk ratio of mortality for stable patients compared to nonresponders was 0.32 (95% CI 0.11-0.93; p=0.04) in all patients and 0.33 (95% CI 0.12-0.96; p=0.04) in patients with UIP. The histological diagnosis of usual interstitial pneumonia is the most important factor determining survival in patients with suspected idiopathic interstitial pneumonia. The presence of honeycombing on high-resolution computed tomography is a good surrogate for usual interstitial pneumonia and could be utilized in patients unable to undergo surgical lung biopsy. Patients with nonspecific interstitial pneumonia are more likely to respond or remain stable following a course of prednisone. Patients remaining stable following prednisone therapy have the best prognosis.