Delphine S. Courvoisier

Performance of logistic regression modeling: beyond the number of events per variable, the role of data structure

OBJECTIVE Logistic regression is commonly used in health research, and it is important to be sure that the parameter estimates can be trusted. A common problem occurs when the outcome has few events; in such a case, parameter estimates may be biased or unreliable. This study examined the relation between correctness of estimation and several data characteristics: number of events per variable (EPV), number of predictors, percentage of predictors that are highly correlated, percentage of predictors that were non-null, size of regression coefficients, and size of correlations. STUDY DESIGN Simulation studies. RESULTS In many situations, logistic regression modeling may pose substantial problems even if the number of EPV exceeds 10. Moreover, the number of EPV is not the only element that impacts on the correctness of parameter estimation. High regression coefficients and high correlations between the predictors may cause large problems in the estimation process. Finally, power is generally very low, even at 20 EPV. CONCLUSION There is no single rule based on EPV that would guarantee an accurate estimation of logistic regression parameters. Instead, the number of predictors, probable size of the regression coefficients based on previous literature, and correlations among the predictors must be taken into account as guidelines to determine the necessary sample size.

Risk prediction models for mortality in ambulatory patients with heart failure: a systematic review

Background—Optimal management of heart failure requires accurate assessment of prognosis. Many prognostic models are available. Our objective was to identify studies that evaluate the use of risk prediction models for mortality in ambulatory patients with heart failure and describe their performance and clinical applicability. Methods and Results—We searched for studies in Medline, Embase, and CINAHL in May 2012. Two reviewers selected citations including patients with heart failure and reporting on model performance in derivation or validation cohorts. We abstracted data related to population, outcomes, study quality, model discrimination, and calibration. Of the 9952 studies reviewed, we included 34 studies testing 20 models. Only 5 models were validated in independent cohorts: the Heart Failure Survival Score, the Seattle Heart Failure Model, the PACE (incorporating peripheral vascular disease, age, creatinine, and ejection fraction) risk score, a model by Frankenstein et al, and the SHOCKED predictors. The Heart Failure Survival Score was validated in 8 cohorts (2240 patients), showing poor-to-modest discrimination (c-statistic, 0.56–0.79), being lower in more recent cohorts. The Seattle Heart Failure Model was validated in 14 cohorts (16 057 patients), describing poor-to-acceptable discrimination (0.63–0.81), remaining relatively stable over time. Both models reported adequate calibration, although overestimating survival in specific populations. The other 3 models were validated in a cohort each, reporting poor-to-modest discrimination (0.66–0.74). Among the remaining 15 models, 6 were validated by bootstrapping (c-statistic, 0.74–0.85); the rest were not validated. Conclusions—Externally validated heart failure models showed inconsistent performance. The Heart Failure Survival Score and Seattle Heart Failure Model demonstrated modest discrimination and questionable calibration. A new model derived from contemporary patient cohorts may be required for improved prognostic performance.

Comparing the predictive validity of five cigarette dependence questionnaires

BACKGROUND To determine whether the long and short versions of the Cigarette Dependence Scale (CDS), the Fagerström Test for Nicotine Dependence (FTND), the Heaviness of Smoking Index (HSI) and the Nicotine Dependence Syndrome Scale (NDSS) predicted smoking abstinence. METHOD We collected data via Internet in 2004-2007 in 2343 current smokers. Follow-up surveys were conducted after 8 days (N=456) and 31 days (N=486). We assessed the ability of dependence ratings to predict smoking abstinence, controlling for confidence and intention to quit smoking. RESULTS All dependence scales were significant predictors of smoking abstinence at the 8-day follow-up, with the CDS being the best predictor (OR: 3.98 per SD unit, 15% of variance explained). Results were similar but stronger among respondents intending to quit (OR: 6.68, 26% of variance explained). At the 31-day follow-up, only the FTND, HSI and NDSS Continuity subscale were still significant, albeit weak, predictors of abstinence (OR: 1.56, 2% of variance explained). CONCLUSION Among the scales under scrutiny, the CDS had the best predictive validity after 8 days. None of the scales explained more than 2% of the variance in abstinence after 31 days.

Analyzing True Change in Longitudinal Multitrait-Multimethod Studies: Application of a Multimethod Change Model to Depression and Anxiety in Children

The authors show how structural equation modeling can be applied to analyze change in longitudinal multitrait-multimethod (MTMM) studies. For this purpose, an extension of latent difference models (McArdle, 1988; Steyer, Eid, & Schwenkmezger, 1997) to multiple constructs and multiple methods is presented. The model allows investigators to separate true change from measurement error and to analyze change simultaneously for different methods. The authors also show how Campbell and Fiskes (1959) guidelines for analyzing convergent and discriminant validity can be applied to the measurement of latent change. The practical application of the multimethod change model is illustrated in a reanalysis of child depression and anxiety scores (N = 906 American children) that were assessed by self- and parent reports on three measurement occasions. The analyses revealed that (a) the convergent validity of change was low for both constructs and (b) sex was a significant predictor of self-reported, but not of parent reported, anxiety states. Finally, the authors discuss advantages and limitations and compare the model with other approaches for analyzing longitudinal MTMM data.

Expression of mitofusin 2R94Q in a transgenic mouse leads to Charcot–Marie–Tooth neuropathy type 2A

Charcot-Marie-Tooth disease type 2A is an autosomal dominant axonal form of peripheral neuropathy caused by mutations in the mitofusin 2 gene. Mitofusin 2 encodes a mitochondrial outer membrane protein that participates in mitochondrial fusion in mammalian cells. How mutations in this protein lead to Charcot-Marie-Tooth disease type 2A pathophysiology remains unclear. We have generated a transgenic mouse expressing either a mutated (R94Q) or wild-type form of human mitofusin 2 in neurons to evaluate whether the R94Q mutation was sufficient for inducing a Charcot-Marie-Tooth disease type 2A phenotype. Only mice expressing mitofusin 2(R94Q) developed locomotor impairments and gait defects thus mimicking the Charcot-Marie-Tooth disease type 2A neuropathy. In these animals, the number of mitochondria per axon was significantly increased in the distal part of the sciatic nerve axons with a diameter smaller than 3.5 microm. Importantly, the analysis of R94Q transgenic animals also revealed an age-related shift in the size of myelinated axons leading to an over-representation of axons smaller than 3.5 microm. Together these data suggest a link between an increased number of mitochondria in axons and a shift in axonal size distribution in mitofusin 2(R94Q) transgenic animals that may contribute to their neurological phenotype.

Compliance to a cell phone-based ecological momentary assessment study: the effect of time and personality characteristics.

Ecological momentary assessment (EMA) is a method that is now widely used to study behavior and mood in the settings in which they naturally occur. It maximizes ecological validity and avoids the limitations of retrospective self-reports. Compliance patterns across time have not been studied. Consistent compliance patterns could lead to data not missing at random and bias the results of subsequent analyses. In order to use modern statistical approaches for handling missing data, it is important to include variables predicting missing values into the statistical analysis. Therefore, these predictors have to be known and measured. The authors collected data on 3 four-item mood scales measuring well-being, wakefulness, and nervousness on 6 occasions per day for 7 days (N = 305) and examined compliance rate across time, within day, and within week. Results show good global compliance (mean compliance: 74.9% of calls answered). Compliance varied more within day than within week. Within day, it was lower for the first call of the day between 9 p.m. and 11 p.m. and higher for the call between 5 p.m. and 7 p.m. Within week, calls were equally answered across days of the week, but, as the study progressed, there was a slight drop in compliance with a progressive decrease that was stronger for the first 2 calls. Compliance on the person level did not depend on personality or on satisfaction with life. Practical consequences of the results for conducting ambulatory assessment studies are discussed, and some recommendations are given.

Analyzing the convergent and discriminant validity of states and traits: Development and applications of multimethod latent state-trait models.

The analysis of convergent and discriminant validity is an integral part of the construct validation process. Models for analyzing the convergent and discriminant validity have typically been developed for cross-sectional data. There exist, however, only a few approaches for longitudinal data that can be applied for analyzing the construct validity of fluctuating states. In this article, the authors show how models of latent state-trait theory can be combined with models of multitrait-multimethod analysis to develop a model that allows for analyzing convergent and discriminant validity in time: the multimethod latent state-trait model. The model allows for identifying different sources of variance (trait consistency, trait-method specificity, occasion-specific consistency, occasion-specific method specificity, and unreliability). It is applied to the repeated measurement of depression and anxiety in children, which was assessed by self and teacher reports (N = 375). The application shows that the proposed models fit the data well and allow a deeper understanding of method effects in clinical assessment.

Risk Prediction Models for Mortality in Ambulatory Patients With Heart Failure

Background—Optimal management of heart failure requires accurate assessment of prognosis. Many prognostic models are available. Our objective was to identify studies that evaluate the use of risk prediction models for mortality in ambulatory patients with heart failure and describe their performance and clinical applicability. Methods and Results—We searched for studies in Medline, Embase, and CINAHL in May 2012. Two reviewers selected citations including patients with heart failure and reporting on model performance in derivation or validation cohorts. We abstracted data related to population, outcomes, study quality, model discrimination, and calibration. Of the 9952 studies reviewed, we included 34 studies testing 20 models. Only 5 models were validated in independent cohorts: the Heart Failure Survival Score, the Seattle Heart Failure Model, the PACE (incorporating peripheral vascular disease, age, creatinine, and ejection fraction) risk score, a model by Frankenstein et al, and the SHOCKED predictors. The Heart Failure Survival Score was validated in 8 cohorts (2240 patients), showing poor-to-modest discrimination (c-statistic, 0.56–0.79), being lower in more recent cohorts. The Seattle Heart Failure Model was validated in 14 cohorts (16 057 patients), describing poor-to-acceptable discrimination (0.63–0.81), remaining relatively stable over time. Both models reported adequate calibration, although overestimating survival in specific populations. The other 3 models were validated in a cohort each, reporting poor-to-modest discrimination (0.66–0.74). Among the remaining 15 models, 6 were validated by bootstrapping (c-statistic, 0.74–0.85); the rest were not validated. Conclusions—Externally validated heart failure models showed inconsistent performance. The Heart Failure Survival Score and Seattle Heart Failure Model demonstrated modest discrimination and questionable calibration. A new model derived from contemporary patient cohorts may be required for improved prognostic performance.

High density of nicotinic receptors in the cingulo-insular network.

The nicotinic system plays an important role in ordinary cognition, particularly in attention. The main nicotinic receptor in the human brain is the heteromeric α4β2 neuronal nicotinic acetylcholine receptor (nAChR), which is distributed throughout the brain, with an especially high density in the thalamus and brainstem. Despite the important role of α4β2 nAChRs in various physiological functions and pathological conditions, their distribution in the human cortex remains poorly characterized. We assessed the in vivo distribution of α4β2 nAChRs in the human cortex in a group of seven non-smoking healthy subjects, using 2-[(18)F]F-A-85380 PET and a volume-of-interest-based analysis. We showed that cortical nAChR density was highest in the insular and anterior cingulate cortices. In functional magnetic resonance imaging studies, these two cortical regions and the thalamus have been shown to be highly correlated during the resting state and various tasks. Here, we also directly assessed nAChR density in this cingulo-insular network as defined in an independent dataset using resting-state functional connectivity, and compared it to other control-related networks, to the default mode network as well as to sensory and motor networks. Receptor density was significantly higher in the cingulo-insular network. This network has been suggested to maintain a variety of foundational capacities fundamental to cognitive function. The demonstration of a high nAChR density in the insular and anterior cingulate cortices reflects a particular neurochemical organization of the cingulo-insular network, and suggests an important role of the nicotinic receptors in its functions.

Validation of alternative formulations of job strain.

Validation of Alternative Formulations of Job Strain: Delphine S. Courvoisier, et al. Clinical Epidemiology Division, Geneva University Hospitals and Faculty of Medicine, University of Geneva, Switzerland