Semiparametric bayes' proportional odds models for current status data with underreporting

Current status data are a type of interval-censored event time data in which all the individuals are either left or right censored. For example, our motivation is drawn from a cross-sectional study, which measured whether or not fibroid onset had occurred by the age of an ultrasound exam for each woman. We propose a semiparametric Bayesian proportional odds model in which the baseline event time distribution is estimated nonparametrically by using adaptive monotone splines in a logistic regression model and the potential risk factors are included in the parametric part of the mean structure. The proposed approach has the advantage of being straightforward to implement using a simple and efficient Gibbs sampler, whereas alternative semiparametric Bayes' event time models encounter problems for current status data. The model is generalized to allow systematic underreporting in a subset of the data, and the methods are applied to an epidemiologic study of uterine fibroids.     

Estimating and comparing the distribution of onset of disease from prevalence case-control data

Age at ascertainment from prevalence case-control data identifies the age-specific odds of disease. When age at onset is available from the cases, the conditional distribution of age at onset, given that disease occurs, is identifiable. Combining both kinds of information by introducing a multiplicative intercept allows identification of the marginal distribution of age at onset. Here, the approach is extended to the two-sample setting through a generalization of the multiplicative intercept model. The efficiency of the approach is explored and a test statistic based on the integrated difference between distribution function estimates is proposed. An approach to regularization of the likelihood is discussed. The methods are illustrated through an application to data on colorectal polyps obtained from a case-control study of individuals undergoing colonoscopy.     

Analyzing the relationship between age at onset and risk to relatives.

Correlations in age at onset between relatives affect risk to relatives of a given age. Either an increase or a decrease in risk may be observed for a relative of a proband, according to whether there is a causal relationship between liability to disease and age at onset. Likelihood formulas are given for pairs of relatives under a number of different sampling schemes, and it is shown how data collected from relatives enable maximum-likelihood estimation of parameters of a linear model relating disease liability and age at onset. A genotype-environment extension of this model was fitted to data on age at onset for schizophrenia that were obtained from the National Academy of Sciences-National Research Council Twin Registry. Age at onset is correlated between twins, but this correlation appears to be associated with factors that are separate from those which affect liability to disease. However, even this relatively large sample of twins is too small to draw firm conclusions about any causal relationship between disease liability and onset.     

Bayesian modeling of incidence and progression of disease from cross-sectional data

In the absence of longitudinal data, the current presence and severity of disease can be measured for a sample of individuals to investigate factors related to disease incidence and progression. In this article, Bayesian discrete-time stochastic models are developed for inference from cross-sectional data consisting of the age at first diagnosis, the current presence of disease, and one or more surrogates of disease severity. Semiparametric models are used for the age-specific hazards of onset and diagnosis, and a normal underlying variable approach is proposed for modeling of changes with latency time in disease severity. The model accommodates multiple surrogates of disease severity having different measurement scales and heterogeneity among individuals in disease progression. A Markov chain Monte Carlo algorithm is described for posterior computation, and the methods are applied to data from a study of uterine leiomyoma.     

Bayesian Estimation of the Time‐Varying Sensitivity of a Diagnostic Test with Application to Mother‐to‐Child Transmission of HIV

Summary We present a Bayesian model to estimate the time‐varying sensitivity of a diagnostic assay when the assay is given repeatedly over time, disease status is changing, and the gold standard is only partially observed. The model relies on parametric assumptions for the distribution of the latent time of disease onset and the time‐varying sensitivity. Additionally, we illustrate the incorporation of historical data for constructing prior distributions. We apply the new methods to data collected in a study of mother‐to‐child transmission of HIV and include a covariate for sensitivity to assess whether two different assays have different sensitivity profiles.     

Regression analysis of a disease onset distribution using diagnosis data

Summary .   We consider methods for estimating the effect of a covariate on a disease onset distribution when the observed data structure consists of right-censored data on diagnosis times and current status data on onset times amongst individuals who have not yet been diagnosed. Dunson and Baird (2001, Biometrics 57, 306–403) approached this problem using maximum likelihood, under the assumption that the ratio of the diagnosis and onset distributions is monotonic nondecreasing. As an alternative, we propose a two-step estimator, an extension of the approach of van der Laan, Jewell, and Petersen (1997, Biometrika 84, 539–554) in the single sample setting, which is computationally much simpler and requires no assumptions on this ratio. A simulation study is performed comparing estimates obtained from these two approaches, as well as that from a standard current status analysis that ignores diagnosis data. Results indicate that the Dunson and Baird estimator outperforms the two-step estimator when the monotonicity assumption holds, but the reverse is true when the assumption fails. The simple current status estimator loses only a small amount of precision in comparison to the two-step procedure but requires monitoring time information for all individuals. In the data that motivated this work, a study of uterine fibroids and chemical exposure to dioxin, the monotonicity assumption is seen to fail. Here, the two-step and current status estimators both show no significant association between the level of dioxin exposure and the hazard for onset of uterine fibroids; the two-step estimator of the relative hazard associated with increasing levels of exposure has the least estimated variance amongst the three estimators considered.     

Age of child, more than HPV type, is associated with clinical course in recurrent respiratory papillomatosis

BACKGROUND: RRP is a devastating disease in which papillomas in the airway cause hoarseness and breathing difficulty. The disease is caused by human papillomavirus (HPV) 6 or 11 and is very variable. Patients undergo multiple surgeries to maintain a patent airway and in order to communicate vocally. Several small studies have been published in which most have noted that HPV 11 is associated with a more aggressive course. METHODOLOGY/PRINCIPAL FINDINGS: Papilloma biopsies were taken from patients undergoing surgical treatment of RRP and were subjected to HPV typing. 118 patients with juvenile-onset RRP with at least 1 year of clinical data and infected with a single HPV type were analyzed. HPV 11 was encountered in 40% of the patients. By our definition, most of the patients in the sample (81%) had run an aggressive course. The odds of a patient with HPV 11 running an aggressive course were 3.9 times higher than that of patients with HPV 6 (Fisher's exact p = 0.017). However, clinical course was more closely associated with age of the patient (at diagnosis and at the time of the current surgery) than with HPV type. Patients with HPV 11 were diagnosed at a younger age (2.4y) than were those with HPV 6 (3.4y) (p = 0.014). Both by multiple linear regression and by multiple logistic regression HPV type was only weakly associated with metrics of disease course when simultaneously accounting for age. CONCLUSIONS/SIGNIFICANCE ABSTRACT: The course of RRP is variable and a quarter of the variability can be accounted for by the age of the patient. HPV 11 is more closely associated with a younger age at diagnosis than it is associated with an aggressive clinical course. These data suggest that there are factors other than HPV type and age of the patient that determine disease course.     

Risk prediction with imperfect survival outcome information from electronic health records

Readily available proxies for the time of disease onset such as the time of the first diagnostic code can lead to substantial risk prediction error if performing analyses based on poor proxies. Due to the lack of detailed documentation and labor intensiveness of manual annotation, it is often only feasible to ascertain for a small subset the current status of the disease by a follow-up time rather than the exact time. In this paper, we aim to develop risk prediction models for the onset time efficiently leveraging both a small number of labels on the current status and a large number of unlabeled observations on imperfect proxies. Under a semiparametric transformation model for onset and a highly flexible measurement error model for proxy onset time, we propose the semisupervised risk prediction method by combining information from proxies and limited labels efficiently. From an initially estimator solely based on the labeled subset, we perform a one-step correction with the full data augmenting against a mean zero rank correlation score derived from the proxies. We establish the consistency and asymptotic normality of the proposed semisupervised estimator and provide a resampling procedure for interval estimation. Simulation studies demonstrate that the proposed estimator performs well in a finite sample. We illustrate the proposed estimator by developing a genetic risk prediction model for obesity using data from Mass General Brigham Healthcare Biobank.     

Lexis Diagram and Illness-Death Model: Simulating Populations in Chronic Disease Epidemiology

Chronic diseases impose a tremendous global health problem of the 21st century. Epidemiological and public health models help to gain insight into the distribution and burden of chronic diseases. Moreover, the models may help to plan appropriate interventions against risk factors. To provide accurate results, models often need to take into account three different time-scales: calendar time, age, and duration since the onset of the disease. Incidence and mortality often change with age and calendar time. In many diseases such as, for example, diabetes and dementia, the mortality of the diseased persons additionally depends on the duration of the disease. The aim of this work is to describe an algorithm and a flexible software framework for the simulation of populations moving in an illness-death model that describes the epidemiology of a chronic disease in the face of the different times-scales. We set up a discrete event simulation in continuous time involving competing risks using the freely available statistical software R. Relevant events are birth, the onset (or diagnosis) of the disease and death with or without the disease. The Lexis diagram keeps track of the different time-scales. Input data are birth rates, incidence and mortality rates, which can be given as numerical values on a grid. The algorithm manages the complex interplay between the rates and the different time-scales. As a result, for each subject in the simulated population, the algorithm provides the calendar time of birth, the age of onset of the disease (if the subject contracts the disease) and the age at death. By this means, the impact of interventions may be estimated and compared.     

Assessing familial aggregation of age at onset, by using estimating equations, with application to breast cancer.

In genetic research of chronic diseases, age-at-onset outcomes within families are often correlated. The nature of correlation of age-at-onset outcomes is indicative of common genetic and/or shared environmental risk factors among family members. Understanding patterns of such correlation may shed light on the disease etiology and, hence, is an important step to take prior to further searching for the responsible genes via segregation and linkage studies. Age-at-onset outcomes are different from those familiar quantitative or qualitative traits for which many statistical methods have been developed. In comparison with the quantitative traits, age-at-onset outcomes are often censored, i.e., instead of actual age-at-onset outcomes, only the current ages or ages at death are observed. They are also different from qualitative traits because of their continuity. Because of the complexity of correlated censored outcomes, few methods have yet been developed. A traditional approach is to impose a parametric joint distribution for the correlated age-at-onset outcomes, which has been criticized for requiring a stringent assumption about the entire distribution of age at onset. The purpose of this paper is to describe a method for assessing familial aggregation of correlated age-at-onset outcomes semiparametrically, by use of estimating equations. This method does not require any parametric assumption for modeling the age at onset. The estimates of parameters, including those quantifying the correlation within families, are consistent and have an asymptotic normal distribution that can be used to make inferences. To illustrate this new method, we analyzed two age-at-onset data sets that were obtained from studies conducted in the States of Washington and Hawaii, with the objective of quantifying the familial aggregations of age at onset of breast cancer.     

Defining natural history: assessment of the ability of college students to aid in characterizing clinical progression of Niemann-Pick disease, type C

Niemann-Pick Disease, type C (NPC) is a fatal, neurodegenerative, lysosomal storage disorder. It is a rare disease with broad phenotypic spectrum and variable age of onset. These issues make it difficult to develop a universally accepted clinical outcome measure to assess urgently needed therapies. To this end, clinical investigators have defined emerging, disease severity scales. The average time from initial symptom to diagnosis is approximately 4 years. Further, some patients may not travel to specialized clinical centers even after diagnosis. We were therefore interested in investigating whether appropriately trained, community-based assessment of patient records could assist in defining disease progression using clinical severity scores. In this study we evolved a secure, step wise process to show that pre-existing medical records may be correctly assessed by non-clinical practitioners trained to quantify disease progression. Sixty-four undergraduate students at the University of Notre Dame were expertly trained in clinical disease assessment and recognition of major and minor symptoms of NPC. Seven clinical records, randomly selected from a total of thirty seven used to establish a leading clinical severity scale, were correctly assessed to show expected characteristics of linear disease progression. Student assessment of two new records donated by NPC families to our study also revealed linear progression of disease, but both showed accelerated disease progression, relative to the current severity scale, especially at the later stages. Together, these data suggest that college students may be trained in assessment of patient records, and thus provide insight into the natural history of a disease.     

Age at time of first intercourse v. chronologic age as a basis for Pap smear screening.

The Walton Report on cervical cancer screening programs recently recommended a new program for screening for cervical cancer based on chronologic age, calling for 3- and 5-year intervals between examinations. It recommended that such examinations be discontinued after 60 years of age. In a group of 232 routinely examined women (aged 18 to 47 years) in whom cervical intraepithelial neoplasia developed the timing of onset of the disease and the implications for screening were studied. The average age at the time of diagnosis was 30 years; in 20% of the patients the diagnosis had been made after age 35. The screening program recommended in the Walton Report would have been effective in diagnosing most cases (80%) in this sample by age 35 and all by age 60. However, when the patients were grouped according to age at the time of first intercourse, the diagnosis had been made after age 35 in only 13% of those who started having intercourse at age 15 to 17 years, 20% of those who started at age 18 to 19 years and 33% of those who started at age 20 years of later. When the times of diagnosis were expressed by number of years of intercourse the distributions became uniform in the same three groups; in 72% of all the patients the diagnosis had been made within the first 15 years of intercourse, in 88% it had been made within 20 years and in 100% it had been made by 30 years. These data suggest that a program based on number of years of intercourse may be more uniform and more efficient than one based on chronologic age, and that cytologic examinations should be concentrated during the time when most cases develop -- 6 to 20 years after the time of first intercourse.     

Nonparametric estimation of asymptomatic duration from a randomized prospective cancer screening trial

We propose a nonparametric estimation of preclinical duration distribution in cancer based on data from a randomized early detection trial. In cancer screening studies, the preclinical duration of a disease is of great interest for better understanding the natural history of the disease, and for developing optimal screening strategies. To estimate the sojourn time distribution nonparametrically, we first estimate the distribution of the age at onset of preclinical disease nonparametrically using data from the screening arm in a randomized screening trial, and the distribution for the age at onset of clinical disease from the control arm of the randomized screening trial. Finally, by using deconvolution the two estimated distributions lead to a nonparametric estimate of the distribution for the gap time between the onset of preclinical disease and the onset of clinical disease. We illustrate the methodology using data from a randomized breast cancer screening trial.     

Risk of schizophrenia in adults born after obstetric complications and their association with early onset of illness: a controlled study.

OBJECTIVE--To determine whether obstetric complications occur to excess in the early histories of individuals who go on to develop schizophrenia when compared with controls, and to seek clinical correlates of any such excess. DESIGN--Contemporaneous maternity hospital records were identified and extracted verbatim, and these extracts evaluated for obstetric complications by two independent assessors who were blind to subjects'' status. SUBJECTS--65 patients having an ICD-9 diagnosis of schizophrenia, the records of the previous same sex live birth being deemed to be those of a control subject. MAIN OUTCOME MEASURE--Presence of one or more obstetric complications recorded in maternity notes of patients and controls. RESULTS--When two recognised scales for specifying obstetric complications were used the patients with schizophrenia were significantly more likely than controls to have experienced at least one obstetric complication (odds ratio 2.44, 95% confidence interval 1.08 to 6.03). Patients also showed a greater number and severity of and total score for obstetric complications, fetal distress being the only complication to occur to significant individual excess (present in five (8%) patients, absent in controls). There was a marked sex effect, male patients being more vulnerable (odds ratio 4.24, 1.39 to 12.90) to such complications. Obstetric complications in patients were unrelated to family history or season of birth but were associated with a significantly younger age at onset of illness (mean difference--4.5 years,--1.2 to--7.8 years). CONCLUSIONS--Patients with schizophrenia, particularly males, have an excess of obstetric complications in their early developmental histories, and such complications are associated with a younger age at onset of their disease. Though the data are not conclusive, they also suggest that obstetric complications may be secondary to yet earlier events.     

Time since childbirth and prognosis in primary breast cancer: population based study.

OBJECTIVE: To investigate whether time since birth of last child was of prognostic importance in women with primary breast cancer. DESIGN: Retrospective cohort study based on a population based database of breast cancer diagnoses with detailed information on tumour characteristics, treatment regimens, reproductive factors, and vital status. SETTING: Denmark. SUBJECTS: 5652 women with primary breast cancer aged 45 years or less at the time of diagnosis. MAIN OUTCOME MEASURES: 5 and 10 year survival; relative risk of dying. RESULTS: Women diagnosed in the first 2 years after last childbirth had a crude 5 year survival of 58.7% and 10 year survival of 46.1% compared with 78.4% and 66.0% for women whose last childbirth was more than 2 years before their diagnosis. After adjustment for age, reproductive factors, and stage of disease (tumour size, axillary nodal status, and histological grading), a diagnosis sooner than 2 years since last childbirth was significantly associated with a poor survival (relative risk 1.58, 95% confidence interval 1.24 to 2.02) compared with women who gave birth more than 5 years previously. Further analyses showed that the effect was not modified by age at diagnosis, tumour size, and nodal status. CONCLUSIONS: A diagnosis of breast cancer less than 2 years after having given birth is associated with a particularly poor survival irrespective of the stage of disease at debut. Therefore, a recent pregnancy should be regarded as a negative prognostic factor and should be considered in counselling these patients and in the decisions regarding adjuvant treatment.     

Asthma,allergy and the risk of prostate cancer: Results from the Montreal PROtEuS study

BackgroundThe few previous studies examining the association between asthma or allergy and prostate cancer (PCa) risk were inconclusive. This study aimed to evaluate these associations, and to explore in details the possible influence of current versus former allergic condition, age at onset, time since onset, and duration of each allergic condition.MethodsDetailed information on self-reported asthma and allergy was collected in the context of a large population-based case–control study conducted in Montreal, Canada. Study subjects included 1936 cases, diagnosed between 2005 and 2009, and 1995 population controls. Unconditional multivariate logistic regression was used to estimate odds ratios (OR) and 95% confidence intervals (CI) adjusting for age, ancestry and familial history of prostate cancer.ResultsThe ORs were 1.11 (95% CI: 0.89–1.40) and 0.98 (95% CI: 0.84–1.14) for ever reporting of asthma and allergy, respectively. These ORs did not substantially vary according to status (former or current), age at onset, time since onset, and duration of each allergic condition. PCa screening was not associated with allergic diseases reporting.ConclusionsOverall, our findings are in line with the absence of an association between a history of asthma or allergy, and PCa risk.     

面肩肱型肌营养不良症的分子机制和治疗方法研究进展

面肩肱型肌营养不良症(facioscapulohumeral muscular dystrophy, FSHD)是世界范围内位列第三的肌营养不良症,呈常染色体显性遗传。FSHD以肌肉坏死为特征,在疾病表现、进展和发病年龄方面具有显著的家族内和家庭间变异性。在不同的年龄段,FSHD的症状表现程度各不相同,通常在青少年时期发病,20岁前的外显率达95%。本文综述了FSHD的发病机制、诊断和治疗的现状,并重点阐述了与治疗相关的研究进展,为今后该疾病的机制研究和临床治疗提供了参考。     

A model to estimate risk of infection with human herpesvirus 8 associated with transfusion from cross-sectional data

In cross-sectional studies of infectious diseases, the data typically consist of: age at the time of study, status (presence or absence) of infection, and a chronology of events possibly associated with the disease. Motivated by a study of how human herpesvirus 8 (HHV-8) is transmitted among children with sickle cell anemia in Uganda, we have developed a flexible parametric approach for combining current-status data with a history of blood transfusions. We model heterogeneity in transfusion-associated risk by a child-specific random effect. We present an extension of the model to accommodate the fact that there is no gold standard for HHV-8 infection and infection status was assessed by a serological assay. The parameters are estimated via maximum likelihood. Finally, we present results from applying various parameterizations of the model to the Ugandan study.     

Robust combination of multiple diagnostic tests for classifying censored event times

Recent advancement in technology promises to yield a multitude of tests for disease diagnosis and prognosis. When there are multiple sources of information available, it is often of interest to construct a composite score that can provide better classification accuracy than any individual measurement. In this paper, we consider robust procedures for optimally combining tests when test results are measured prior to disease onset and disease status evolves over time. To account for censoring of disease onset time, the most commonly used approach to combining tests to detect subsequent disease status is to fit a proportional hazards model (Cox, 1972) and use the estimated risk score. However, simulation studies suggested that such a risk score may have poor accuracy when the proportional hazards assumption fails. We propose the use of a nonparametric transformation model (Han, 1987) as a working model to derive an optimal composite score with theoretical justification. We demonstrate that the proposed score is the optimal score when the model holds and is optimal "on average" among linear scores even if the model fails. Time-dependent sensitivity, specificity, and receiver operating characteristic curve functions are used to quantify the accuracy of the resulting composite score. We provide consistent and asymptotically Gaussian estimators of these accuracy measures. A simple model-free resampling procedure is proposed to obtain all consistent variance estimators. We illustrate the new proposals with simulation studies and an analysis of a breast cancer gene expression data set.     

Localization of a novel melanoma susceptibility locus to 1p22

Over the past 20 years, the incidence of cutaneous malignant melanoma (CMM) has increased dramatically worldwide. A positive family history of the disease is among the most established risk factors for CMM; it is estimated that 10% of CMM cases result from an inherited predisposition. Although mutations in two genes, CDKN2A and CDK4, have been shown to confer an increased risk of CMM, they account for only 20%-25% of families with multiple cases of CMM. Therefore, to localize additional loci involved in melanoma susceptibility, we have performed a genomewide scan for linkage in 49 Australian pedigrees containing at least three CMM cases, in which CDKN2A and CDK4 involvement has been excluded. The highest two-point parametric LOD score (1.82; recombination fraction [theta] 0.2) was obtained at D1S2726, which maps to the short arm of chromosome 1 (1p22). A parametric LOD score of 4.65 (theta=0) and a nonparametric LOD score of 4.19 were found at D1S2779 in nine families selected for early age at onset. Additional typing yielded seven adjacent markers with LOD scores >3 in this subset, with the highest parametric LOD score, 4.95 (theta=0) (nonparametric LOD score 5.37), at D1S2776. Analysis of 33 additional multiplex families with CMM from several continents provided further evidence for linkage to the 1p22 region, again strongest in families with the earliest mean age at diagnosis. A nonparametric ordered sequential analysis was used, based on the average age at diagnosis in each family. The highest LOD score, 6.43, was obtained at D1S2779 and occurred when the 15 families with the earliest ages at onset were included. These data provide significant evidence of a novel susceptibility gene for CMM located within chromosome band 1p22.