Olga Demler

The World Health Organization adult ADHD self-report scale (ASRS): a short screening scale for use in the general population

BACKGROUND A self-report screening scale of adult attention-deficit/hyperactivity disorder (ADHD), the World Health Organization (WHO) Adult ADHD Self-Report Scale (ASRS) was developed in conjunction with revision of the WHO Composite International Diagnostic Interview (CIDI). The current report presents data on concordance of the ASRS and of a short-form ASRS screener with blind clinical diagnoses in a community sample. METHOD The ASRS includes 18 questions about frequency of recent DSM-IV Criterion A symptoms of adult ADHD. The ASRS screener consists of six out of these 18 questions that were selected based on stepwise logistic regression to optimize concordance with the clinical classification. ASRS responses were compared to blind clinical ratings of DSM-IV adult ADHD in a sample of 154 respondents who previously participated in the US National Comorbidity Survey Replication (NCS-R), oversampling those who reported childhood ADHD and adult persistence. RESULTS Each ASRS symptom measure was significantly related to the comparable clinical symptom rating, but varied substantially in concordance (Cohens kappa in the range 0.16-0.81). Optimal scoring to predict clinical syndrome classifications was to sum unweighted dichotomous responses across all 18 ASRS questions. However, because of the wide variation in symptom-level concordance, the unweighted six-question ASRS screener outperformed the unweighted 18-question ASRS in sensitivity (68.7% v. 56.3%), specificity (99.5% v. 98.3%), total classification accuracy (97.9% v. 96.2%), and kappa (0.76 v. 0.58). CONCLUSIONS Clinical calibration in larger samples might show that a weighted version of the 18-question ASRS outperforms the six-question ASRS screener. Until that time, however, the unweighted screener should be preferred to the full ASRS, both in community surveys and in clinical outreach and case-finding initiatives.

Adequacy of treatment for serious mental illness in the United States.

OBJECTIVES The purpose of this study was to assess the prevalence and correlates of treatment for serious mental illness. METHODS Data were derived from the National Comorbidity Survey, a cross-sectional, nationally representative household survey assessing the presence and correlates of mental disorders and treatments. Crude and adjusted likelihoods of receiving treatment for serious mental illness in the previous 12 months were calculated. RESULTS Forty percent of respondents with serious mental illness had received treatment in the previous year. Of those receiving treatment, 38.9% received care that could be considered at least minimally adequate, resulting in 15.3% of all respondents with serious mental illness receiving minimally adequate treatment. Predictors of not receiving minimally adequate treatment included being a young adult or an African American, residing in the South, being diagnosed as having a psychotic disorder, and being treated in the general medical sector. CONCLUSIONS Inadequate treatment of serious mental illness is an enormous public health problem. Public policies and cost-effective interventions are needed to improve both access to treatment and quality of treatment.

Comorbid Mental Disorders Account for the Role Impairment of Commonly Occurring Chronic Physical Disorders: Results From the National Comorbidity Survey

Specify the effects of four common chronic physical disorders on role impairment (sickness absence days plus work cut-back days), and their association with comorbid mental disorders.Describe the individual and joint effects of mental disorder and physical disorder on role impairment.Evaluate competing explanations of the apparent connection between mental comorbidity and role impairment. Most health and work productivity studies have focused on individual conditions without considering comorbidity. We illustrate the implication of this neglect by examining the effects of comorbid mental disorders on role impairment (number of sickness absence and work cut-back days in the past month) among people with chronic physical disorders. A nationally representative household survey of 5877 respondents assessed current mental and physical disorders and role impairments. Four physical disorders were sufficiently common to be studied: hypertension, arthritis, asthma, and ulcers. All 4 physical disorders were associated with significant role impairments in bivariate analyses. However, further analysis showed that these impairments were almost entirely confined to cases with comorbid mental disorders. Effectiveness trials in workplace samples are needed to evaluate the cost-effectiveness of treating comorbid mental disorders among workers with chronic physical disorders.

The Prevalence and Correlates of Nonaffective Psychosis in the National Comorbidity Survey Replication (NCS-R)

BACKGROUND To estimate the prevalence and correlates of clinician-diagnosed DSM-IV nonaffective psychosis (NAP) in a national household survey. METHODS Data came from the United States National Comorbidity Survey Replication (NCS-R). A screen for NAP was followed by blinded sub-sample clinical reappraisal interviews. Logistic regression was used to impute clinical diagnoses to respondents who were not re-interviewed. The method of Multiple Imputation (MI) was used to estimate prevalence and correlates. RESULTS Clinician-diagnosed NAP was well predicted by the screen (area under the curve [AUC] = .80). The MI prevalence estimate of NAP (standard error in parentheses) is 5.0 (2.6) per 1000 population lifetime and 3.0 (2.2) per 1000 past 12 months. The vast majority (79.4%) of lifetime and 12-month (63.7%) cases met criteria for other DSM-IV hierarchy-free disorders. Fifty-eight percent of 12-month cases were in treatment, most in the mental health specialty sector. CONCLUSIONS The screen for NAP in the NCS-R greatly improved on previous epidemiological surveys in reducing false positives, but coding of open-ended screening scale responses was still needed to achieve accurate prediction. The lower prevalence estimate than in total-population incidence studies raises concerns that systematic nonresponse bias causes downward bias in survey prevalence estimates of NAP.

Novel metrics for evaluating improvement in discrimination: net reclassification and integrated discrimination improvement for normal variables and nested models

Net reclassification and integrated discrimination improvements have been proposed as alternatives to the increase in the area under the curve for evaluating improvement in the performance of risk assessment algorithms introduced by the addition of new phenotypic or genetic markers. In this paper, we demonstrate that in the setting of linear discriminant analysis, under the assumptions of multivariate normality, all three measures can be presented as functions of the squared Mahalanobis distance. This relationship affords an interpretation of the magnitude of these measures in the familiar language of effect size for uncorrelated variables. Furthermore, it allows us to conclude that net reclassification improvement can be viewed as a universal measure of effect size. Our theoretical developments are illustrated with an example based on the Framingham Heart Study risk assessment model for high-risk men in primary prevention of cardiovascular disease.

Misuse of DeLong test to compare AUCs for nested models

The area under the receiver operating characteristics curve (AUC of ROC) is a widely used measure of discrimination in risk prediction models. Routinely, the Mann-Whitney statistics is used as an estimator of AUC, while the change in AUC is tested by the DeLong test. However, very often, in settings where the model is developed and tested on the same dataset, the added predictor is statistically significantly associated with the outcome but fails to produce a significant improvement in the AUC. No conclusive resolution exists to explain this finding. In this paper, we will show that the reason lies in the inappropriate application of the DeLong test in the setting of nested models. Using numerical simulations and a theoretical argument based on generalized U-statistics, we show that if the added predictor is not statistically significantly associated with the outcome, the null distribution is non-normal, contrary to the assumption of DeLong test. Our simulations of different scenarios show that the loss of power because of such a misuse of the DeLong test leads to a conservative test for small and moderate effect sizes. This problem does not exist in cases of predictors that are associated with the outcome and for non-nested models. We suggest that for nested models, only the test of association be performed for the new predictors, and if the result is significant, change in AUC be estimated with an appropriate confidence interval, which can be based on the DeLong approach.

The effects of co-morbidity on the onset and persistence of generalized anxiety disorder in the ICPE surveys

BACKGROUND Although it is well known that generalized anxiety disorder (GAD) is highly co-morbid with other mental disorders, little is known about the extent to which earlier disorders predict the subsequent first onset and persistence of GAD. These associations are examined in the current report using data from four community surveys in the World Health Organization (WHO) International Consortium in Psychiatric Epidemiology (ICPE). METHOD The surveys come from Brazil, Canada, the Netherlands and the United States. The Composite International Diagnostic Interview (CIDI) was used to assess DSM-III-R anxiety, mood and substance use disorders in these surveys. Discrete-time survival analysis was used to examine the associations of retrospectively reported earlier disorders with first onset of GAD. Logistic regression analysis was used to examine the associations of the disorders with persistence of GAD. RESULTS Six disorders predict first onset of GAD in all four surveys: agoraphobia, panic disorder, simple phobia, dysthymia, major depression and mania. With the exception of simple phobia, only respondents with active disorders have elevated risk of GAD. In the case of simple phobia, in comparison, respondents with a history of remitted disorder also have consistently elevated risk of GAD. Simple phobia is also the only disorder that predicts the persistence of GAD. CONCLUSIONS The causal processes linking temporally primary disorders to onset of GAD are likely to be state-dependent. History of simple phobia might be a GAD risk marker. Further research is needed to explore the mechanisms involved in the relationship between simple phobia and subsequent GAD.

Equivalence of improvement in area under ROC curve and linear discriminant analysis coefficient under assumption of normality.

In this paper we investigate the addition of new variables to an existing risk prediction model and the subsequent impact on discrimination quantified by the area under the receiver operating characteristics curve (AUC of ROC). Based on practical experience, concerns have emerged that the significance of association of the variable under study with the outcome in the risk model does not correspond to the significance of the change in AUC: that is, often the variable is significant, but the change in AUC is not. This paper demonstrates that under the assumption of multivariate normality and employing linear discriminant analysis (LDA) to construct the risk prediction tool, statistical significance of the new predictor(s) is equivalent to the statistical significance of the increase in AUC. Under these assumptions the result extends asymptotically to logistic regression. We further show that equality of variance-covariance matrices of predictors within cases and non-cases is not necessary when LDA is used. However, our practical example from the Framingham Heart Study data suggests that the finding might be sensitive to the assumption of normality.

Cholesterol Efflux Capacity, HDL Particle Number, and Incident Cardiovascular Events. An Analysis from the JUPITER Trial (Justification for the Use of Statins in Prevention: An Intervention Trial Evaluating Rosuvastatin)

Background: Recent failures of drugs that raised high-density lipoprotein (HDL) cholesterol levels to reduce cardiovascular events in clinical trials have led to increased interest in alternative indices of HDL quality, such as cholesterol efflux capacity, and HDL quantity, such as HDL particle number. However, no studies have directly compared these metrics in a contemporary population that includes potent statin therapy and low low-density lipoprotein cholesterol. Methods: HDL cholesterol levels, apolipoprotein A-I, cholesterol efflux capacity, and HDL particle number were assessed at baseline and 12 months in a nested case-control study of the JUPITER trial (Justification for the Use of Statins in Prevention: An Intervention Trial Evaluating Rosuvastatin), a randomized primary prevention trial that compared rosuvastatin treatment to placebo in individuals with normal low-density lipoprotein cholesterol but increased C-reactive protein levels. In total, 314 cases of incident cardiovascular disease (CVD) (myocardial infarction, unstable angina, arterial revascularization, stroke, or cardiovascular death) were compared to age- and gender-matched controls. Conditional logistic regression models adjusting for risk factors evaluated associations between HDL-related biomarkers and incident CVD. Results: Cholesterol efflux capacity was moderately correlated with HDL cholesterol, apolipoprotein A-I, and HDL particle number (Spearman r= 0.39, 0.48, and 0.39 respectively; P<0.001). Baseline HDL particle number was inversely associated with incident CVD (adjusted odds ratio per SD increment [OR/SD], 0.69; 95% confidence interval [CI], 0.56–0.86; P<0.001), whereas no significant association was found for baseline cholesterol efflux capacity (OR/SD, 0.89; 95% CI, 0.72–1.10; P=0.28), HDL cholesterol (OR/SD, 0.82; 95% CI, 0.66–1.02; P=0.08), or apolipoprotein A-I (OR/SD, 0.83; 95% CI, 0.67–1.03; P=0.08). Twelve months of rosuvastatin (20 mg/day) did not change cholesterol efflux capacity (average percentage change −1.5%, 95% CI, −13.3 to +10.2; P=0.80), but increased HDL cholesterol (+7.7%), apolipoprotein A-I (+4.3%), and HDL particle number (+5.2%). On-statin cholesterol efflux capacity was inversely associated with incident CVD (OR/SD, 0.62; 95% CI, 0.42–0.92; P=0.02), although HDL particle number again emerged as the strongest predictor (OR/SD, 0.51; 95% CI, 0.33–0.77; P<0.001). Conclusions: In JUPITER, cholesterol efflux capacity was associated with incident CVD in individuals on potent statin therapy but not at baseline. For both baseline and on-statin analyses, HDL particle number was the strongest of 4 HDL-related biomarkers as an inverse predictor of incident events and biomarker of residual risk. Clinical Trial Registration: URL: http://www.clinicaltrials.gov. Unique identifier: NCT00239681.

Identifying an Optimal Cutpoint for the Diagnosis of Hypertriglyceridemia in the Nonfasting State

BACKGROUND Nonfasting triglycerides are similar or superior to fasting triglycerides at predicting cardiovascular events. However, diagnostic cutpoints are based on fasting triglycerides. We examined the optimal cutpoint for increased nonfasting triglycerides. METHODS We obtained baseline nonfasting (<8 h since last meal) samples from 6391 participants in the Womens Health Study who were followed prospectively for ≤17 years. The optimal diagnostic threshold for nonfasting triglycerides, determined by logistic regression models by use of c-statistics and the Youden index (sum of sensitivity and specificity minus 1), was used to calculate hazard ratios (HRs) for incident cardiovascular events. Performance was compared to thresholds recommended by the American Heart Association (AHA) and European guidelines. RESULTS The optimal threshold was 175 mg/dL (1.98 mmol/L), with a c-statistic of 0.656, statistically better than the AHA cutpoint of 200 mg/dL (c-statistic 0.628). For nonfasting triglycerides above and below 175 mg/dL, after adjusting for age, hypertension, smoking, hormone use, and menopausal status, the HR for cardiovascular events was 1.88 (95% CI 1.52-2.33, P < 0.001), and for triglycerides measured at 0-4 and 4-8 h since the last meal, 2.05 (1.54- 2.74) and 1.68 (1.21-2.32), respectively. We validated performance of this optimal cutpoint by use of 10-fold cross-validation and bootstrapping of multivariable models that included standard risk factors plus total and HDL cholesterol, diabetes, body mass index, and C-reactive protein. CONCLUSIONS In this study of middle-aged and older apparently healthy women, we identified a diagnostic threshold for nonfasting hypertriglyceridemia of 175 mg/dL (1.98 mmol/L), with the potential to more accurately identify cases than the currently recommended AHA cutpoint.