Janez Stare

On Estimation in Relative Survival

Estimation of relative survival has become the first and the most basic step when reporting cancer survival statistics. Standard estimators are in routine use by all cancer registries. However, it has been recently noted that these estimators do not provide information on cancer mortality that is independent of the national general population mortality. Thus they are not suitable for comparison between countries. Furthermore, the commonly used interpretation of the relative survival curve is vague and misleading. The present article attempts to remedy these basic problems. The population quantities of the traditional estimators are carefully described and their interpretation discussed. We then propose a new estimator of net survival probability that enables the desired comparability between countries. The new estimator requires no modeling and is accompanied with a straightforward variance estimate. The methods are described on real as well as simulated data.

The Risk of Lower Mental Outcome in Infantile Spasms Increases after Three Weeks of Hypsarrhythmia Duration

Summary:  To assess the correlation between hypsarrhythmia duration and mental outcome in infantile spasms (IS) the medical records of 48 infants with IS were reviewed retrospectively and psychological assessments undertaken at follow‐up at the age of 3 to 13 years. We found 18 (38%) cryptogenic IS cases with typical hypsarrhythmia and 30 symptomatic with modified hypsarrhythmia—further classified into 15 cases as multifocal, 10 as pseudoperiodic and 5 as unilateral hypsarrhythmia. A short treatment lag (one to two weeks) occurred in 25, three to four weeks in 10 cases. Spasms ceased within one month after treatment in 23 infants. At follow‐up 15 children had normal mental outcome (borderline included). A correlation between hypsarrhythmia duration longer than three weeks and lower mental outcome was found using the logistic regression model. The duration of hypsarrhythmia represents a sensitive prognostic parameter in IS; the risk of mental retardation increases after three weeks of hypsarrhythmia.

An approach to estimation in relative survival regression.

The goal of relative survival methodology is to compare the survival experience of a cohort with that of the background population. Most often an additive excess hazard model is employed, which assumes that each persons hazard is a sum of 2 components--the population hazard obtained from life tables and an excess hazard attributable to the specific condition. Usually covariate effects on the excess hazard are assumed to have a proportional hazards structure with parametrically modelled baseline. In this paper, we introduce a new fitting procedure using the expectation-maximization algorithm, treating the cause of death as missing data. The method requires no assumptions about the baseline excess hazard thus reducing the risk of bias through misspecification. It accommodates the possibility of knowledge of cause of death for some patients, and as a side effect, the method yields an estimate of the ratio between the excess and the population hazard for each subject. More importantly, it estimates the baseline excess hazard flexibly with no additional degrees of freedom spent. Finally, it is a generalization of the Cox model, meaning that all the wealth of options in existing software for the Cox model can be used in relative survival. The method is applied to a data set on survival after myocardial infarction, where it shows how a particular form of the hazard function could be missed using the existing methods.

Rebuttal to editorial saying cancer survival statistics are misleading

Beral and Peto’s 2010 editorial on cancer survival statistics is unfounded, untenable, and inconsistent.1 Godlee reported in September 2010 that they were then too busy to defend it.2 The editorial is indefensible. It should be withdrawn. The editorial is unfounded. The provocative subtitle, “UK cancer survival statistics are misleading and make survival look worse than it is,” is pure conjecture. Conjecture becomes assertion, then …

BJ: an S-Plus program to fit linear regression models to censored data using the Buckley–James method

Most researchers are familiar with ordinary multiple regression models, most commonly fitted using the method of least squares. The method of Buckley and James (J. Buckley, I. James, Linear regression with censored data, Biometrika 66 (1979) 429-436.) is an extension of least squares for fitting multiple regression models when the response variable is right-censored as in the analysis of survival time data. The Buckley-James method has been shown to have good statistical properties under usual regularity conditions (T.L. Lai, Z. Ying, Large sample theory of a modified Buckley-James estimator for regression analysis with censored data, Ann. Stat. 19 (1991) 1370-1402.). Nevertheless, even after 20 years of its existence, it is almost never used in practice. We believe that this is mainly due to lack of software and we describe here an S-Plus program that through its inclusion in a public domain function library fully exploits the power of the S-Plus programming environment. This environment provides multiple facilities for model specification, diagnostics, statistical inference, and graphical depiction of the model fit.

A Measure of Explained Variation for Event History Data

There is no shortage of proposed measures of prognostic value of survival models in the statistical literature. They come under different names, including explained variation, correlation, explained randomness, and information gain, but their goal is common: to define something analogous to the coefficient of determination R(2)  in linear regression. None however have been uniformly accepted, none have been extended to general event history data, including recurrent events, and many cannot incorporate time-varying effects or covariates. We present here a measure specifically tailored for use with general dynamic event history regression models. The measure is applicable and interpretable in discrete or continuous time; with tied data or otherwise; with time-varying, time-fixed, or dynamic covariates; with time-varying or time-constant effects; with single or multiple event times; with parametric or semiparametric models; and under general independent censoring/observation. For single-event survival data with neither censoring nor time dependency it reduces to the concordance index. We give expressions for its population value and the variance of the estimator and explore its use in simulations and applications. A web link to R software is provided.

Mild Hyperhomocysteinemia and Fibrinolytic Factors in Patients with History of Venous Thromboembolism

Mild hyperhomocysteinemia is recognized as a risk factor for venous thromboembolism (VTE), though its role in the thrombogenic processes is not understood. Its possible association with impaired fibrinolysis was investigated in 157 patients (61 women, 96 men) below the age of 60 years (43+/-11, mean+/-SD) with a history of objectively confirmed VTE. Patients had significantly higher fasting total plasma homocysteine (tHcy) levels than 138 apparently healthy subjects (8.0, 6.6-9.9 micromol/L vs. 7.2, 5.9-8.6 micromol/L, P=0. 001; median, range between first and third quartile). In 17 of 157 patients (12%) hyperhomocysteinemia (tHcy>11.4 micromol/L for women and tHcy>12.6 micromol/L for men) was established. The adjusted odds ratio as an estimate of relative risk for VTE was 2.3 (0.8-7.0; 95% confidence interval). When patients with hyperhomocysteinemia were compared to patients without hyperhomocysteinemia, no significant differences in t-PA (antigen 9.2+/-5.5 microg/L and 9.7+/-4.7 microg/L, respectively; activity 1.3+/-0.5 IU/mL and 1.3+/-0.7 IU/mL, respectively) and PAI-1 (antigen 19.3+/-17.5 microg/L and 22.6+/-20. 4 microg/L, respectively; activity 15.0+/-12.6 and 15.8+/-13.3 IU/mL, respectively) were observed. In conclusion, this study showed an association between mild hyperhomocysteinemia and VTE, but provided no evidence for an independent association between hyperhomocysteinemia and alterations in fibrinolytic proteins.

A hospital outbreak of Legionella from a contaminated water supply.

The authors performed a cross-sectional epidemiological survey to investigate the source of a hospital Legionella outbreak originating in contaminated water. Water temperature and air humidity were measured around possible contamination sources. A deadend pipe was found to contain Legionella pneumophila serogroup 1. All individuals who acquired legionellosis had spent at least 30 min within 2 m of the contamination source. Among staff, 41 of 71 were exposed, and 31 of these fell ill. All 7 patients exposed to the contaminated water acquired legionellosis. None of the 94 bed-ridden patients from the same units developed the disease. An aerosol with 60% relative air humidity was formed near the suspect water faucets, but the humidity fell rapidly farther from the water source, suggesting that desiccation decreased the risk of infection. The healthy personnel and patients closest to the source acquired legionellosis, suggesting that risk was related less to compromised patients than to exposure.

Segmentation priors from local image properties: without using bias field correction, location-based templates, or registration.

We present a novel approach for generating information about a voxels tissue class membership based on its signature--a collection of local image textures estimated over a range of neighborhood sizes. The approach produces a form of tissue class priors that can be used to initialize and regularize image segmentation. The signature-based approach is a departure from current location-based methods, which derive tissue class likelihoods based on a voxels location in standard template space. To use location-based priors, one needs to register the volume in question to the template space, and estimate the image intensity bias field. Two optimizations, over more than a thousand parameters, are needed when high order nonlinear registration is employed. In contrast, the signature-based approach is independent of volume orientation, voxel position, and largely insensitive to bias fields. For these reasons, the approach does not require the use of population derived templates. The prior information is generated from variations in image texture statistics as a function of spatial scale, and an SVM approach is used to associate signatures with tissue types. With the signature-based approach, optimization is needed only during the training phase for the parameter estimation stages of the SVM hyperplanes, and associated PDFs; a training process separate from the segmentation step. We found that signature-based priors were superior to location-based ones aligned under favorable conditions, and that signature-based priors result in improved segmentation when replacing location-based ones in FAST (Zhang et al., 2001), a widely used segmentation program. The software implementation of this work is freely available as part of AFNI http://afni.nimh.nih.gov.

Effects of international collaboration and status of journal on impact of papers

This study examines the effect of international collaboration of Slovenian authors and the status of journals where papers are published (as determined by their impact factors) on the impact of papers as measured by the number of citations papers receive. Research programme groups working in Slovenia in the 2004–2008 period in the fields of physics, chemistry, biology, biotechnology, and medical science were used for analyses. The results of the analyses show that the effects of the two factors differ among the fields. We discuss possible reasons for this, including the possibility that differences are the result of Slovenia’s science policy.