Comparison of likelihood penalization and variance decomposition approaches for clinical prediction models: A simulation study.

Logistic regression is one of the most commonly used approaches to develop clinical risk prediction models. Developers of such models often rely on approaches that aim to minimize the risk of overfitting and improve predictive performance of the logistic model, such as through likelihood penalization and variance decomposition techniques. We present an extensive simulation study that compares the out-of-sample predictive performance of risk prediction models derived using the elastic net, with Lasso and ridge as special cases, and variance decomposition techniques, namely, incomplete principal component regression and incomplete partial least squares regression. We varied the expected events per variable, event fraction, number of candidate predictors, presence of noise predictors, and the presence of sparse predictors in a full-factorial design. Predictive performance was compared on measures of discrimination, calibration, and prediction error. Simulation metamodels were derived to explain the performance differences within model derivation approaches. Our results indicate that, on average, prediction models developed using penalization and variance decomposition approaches outperform models developed using ordinary maximum likelihood estimation, with penalization approaches being consistently superior over the variance decomposition approaches. Differences in performance were most pronounced on the calibration of the model. Performance differences regarding prediction error and concordance statistic outcomes were often small between approaches. The use of likelihood penalization and variance decomposition techniques methods was illustrated in the context of peripheral arterial disease.

Psychometric properties and measurement invariance of the Beck hopelessness scale (BHS): results from a German representative population sample.

BACKGROUND: The Beck Hopelessness Scale (BHS) has been the most frequently used instrument for the measurement of hopelessness in the past 40 years. Only recently has it officially been translated into German. The psychometric properties and factor structure of the BHS have been cause for intensive debate in the past. METHODS: Based on a representative sample of the German population (N = 2450) item analysis including item sensitivity, item-total correlation and item difficulty was performed. Confirmatory factor analyses (CFA) for several factor solutions from the literature were performed. Multiple group factor analysis was performed to assess measurement invariance. Construct validity was assessed via the replication of well-established correlations with concurrently assessed measures. RESULTS: Most items exhibited adequate properties. Items #4, #8 and #13 exhibited poor item characteristics- each of these items had previously received negative evaluations in international studies. A one-dimensional factor solution, favorable for the calculation and interpretation of a sum score, was regarded as adequate. A bi-factor model with one content factor and two method factors (defined by positive/negative item coding) resulted in an excellent model fit. Cronbach's alpha in the current sample was .87. Hopelessness, as measured by the BHS, significantly correlated in the expected direction with suicidal ideation (r = .36), depression (r = .53) and life satisfaction (r = -.53). Strict measurement invariance could be established regarding gender and depression status. Due to limited research regarding the interpretation of fit indices with dichotomous data, interpretation of CFA results needs to remain tentative. CONCLUSION: The BHS is a valid measure of hopelessness in various subgroups of the general population. Future research could aim at replicating these findings using item response theory and cross-cultural samples. A one-dimensional bi-factor model seems appropriate even in a non-clinical population.

German Beck Scale for Suicide Ideation (BSS): psychometric properties from a representative population survey.

BACKGROUND: Suicidal ideation has been identified as one of the major predictors of attempted or actual suicide. Routinely screening individuals for endorsing suicidal thoughts could save lives and protect many from severe psychological consequences following the suicide of loved ones. The aim of this study was to validate the German version of the Beck Scale for Suicide Ideation (BSS) in a sample representative for the Federal Republic of Germany. METHODS: All 2450 participants completed the first part of the Scale, the BSS-Screen. A risk group of n = 112 individuals (4.6%) with active or passive suicidal ideation was identified and subsequently completed the entire BSS. RESULTS: Satisfactory internal reliability (α = .97 for the BSS-Screen; α = .94 for the entire BSS) and excellent model fit indices for the one-dimensional factorial structure of the BSS-Screen (CFI = .998; TLI = .995; RMSEA = .045 [95%-CI: .030-.061]) were confirmed. Measurement invariance analyses supported strict invariance across gender, age, and depression status. We found correlations with related self-report measures in expected directions comparable to previous studies, indicating satisfactory construct validity. LIMITATIONS: Our study involved cross sectional data, hence neither predictive validity nor retest-reliability were examined. As only the risk group of n = 112 individuals completed the entire measure, confirmatory factor analyses could not be conducted for the full BSS. CONCLUSION: The German translation of the BSS is a reliable and valid instrument for assessing suicidal ideation in the general population. Using it as a screening device in general and specialized medical care could substantially advance suicide prevention.

The Effect of the Covid-19 Pandemic on Domestic Violence in Germany: A Comparison of Three Representative Population Surveys.

The COVID-19 pandemic has had profound societal and economic effects. Concerns were raised that domestic violence might increase because of the enacted infection control measures. Previous findings on this issue have been contradictory. Since existing studies mainly rely on official reports, administrative data, helpline calls, or retrospective measures, their findings are likely to prove unreliable. Few population-based surveys include pre-pandemic data, limiting their ability to test for causality regarding increasing violence. Therefore, the aim of this study was to compare findings from population-representative surveys on the prevalence of intimate partner violence (IPV) and violence against children (VAC) before and during the COVID-19 pandemic. Based on the data of N = 3,639 individuals living with a romantic partner and N = 1,313 parents living with at least one of their children from three German representative population surveys, we estimated average marginal effects for the temporal trends (i.e., pre vs. post infection control measures) of domestic violence separately for males and females. To minimize bias across survey waves, inverse probability weighting was used. Results show no statistically significant increase in either physical or psychological forms of IPV or VAC as a result of the implementation of COVID measures. On the contrary, the 1-year prevalence was decreasing for certain forms of violence. Our findings suggest that the assessment of the consequences of infection control measures needs an empirical basis. Further research should be conducted using high-quality data sources. Therefore, the present study should be considered a stepping stone for ongoing research efforts to examine the consequences of pandemic-related infection control measures on the general population.

It's time! Ten reasons to start replicating simulation studies.

The quantitative analysis of research data is a core element of empirical research. The performance of statistical methods that are used for analyzing empirical data can be evaluated and compared using computer simulations. A single simulation study can influence the analyses of thousands of empirical studies to follow. With great power comes great responsibility. Here, we argue that this responsibility includes replication of simulation studies to ensure a sound foundation for data analytical decisions. Furthermore, being designed, run, and reported by humans, simulation studies face challenges similar to other experimental empirical research and hence should not be exempt from replication attempts. We highlight that the potential replicability of simulation studies is an opportunity quantitative methodology as a field should pay more attention to.

Evaluation of the Universal Prevention Program Klasse2000 in Fourth Grade Primary School Children: Protocol for a Propensity Score-Matching Approach.

BACKGROUND: Klasse2000 is the most widely adopted school-based prevention program in Germany. It addresses health promotion, addiction, and violence prevention in primary schools. As a universal prevention program, it has reached more than 1.4 million German children in the past 25 years. OBJECTIVE: The effectiveness of Klasse2000 will be evaluated with a large representative survey among students. Students who have participated in the prevention program (intervention group) will be compared with students who did not participate (control group). The comparison will cover the following outcome domains: well-being, self-esteem, emotion regulation, food habits, behavioral problems, and school and classroom atmosphere. Furthermore, victimization and perpetration regarding bullying, alcohol consumption, smoking, and media consumption are assessed. METHODS: To control for potential group differences, treatment effects will be estimated using propensity score-matching, which matches students from the intervention and control groups based on an identical propensity score or a propensity score that does not differ by more than a previously defined distance. The treatment effect will then be estimated in the matched sample taking the matching process into account. RESULTS: Enrollment of schools began in March 2017. A total of 6376 students participated in the survey (n=4005 in control group; n=2371 in Klasse2000). The parent survey was returned by 52.13% (3324/6376) of parents. Results are expected in mid-2020. CONCLUSIONS: The results on the effectiveness of the Klasse2000 prevention program will form an empirical basis for legitimizing universal prevention programs and for planning future prevention approaches. TRIAL REGISTRATION: German Clinical Trials Register DRKS00014332; https://tinyurl.com/y2trvq4p. INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID): DERR1-10.2196/14371.

Factor Structure of the Inventory of Callous-Unemotional Traits in a Representative Sample of German 9th Grade Students.

Callous-unemotional (CU) traits represent the affective components of the psychopathy construct and show a strong relationship to violence and conduct-disorder in children. The Inventory of Callous-Unemotional Traits (ICU) is the current standard to assess CU traits. Despite the ICU having originally been constructed as a four-dimensional instrument, several studies found a three-factorial structure in combination with a general ICU-factor to be the best fitting factor-model. An imbalance in the number of positively and negatively worded items can be observed between the identified ICU dimensions. To investigate the effect of item keying on the factor structure we tested different factor-models in data stemming from a large sample (N = 3878) of German ninth grade students. Our findings suggest that the original four-factorial structure with a method factor shows good model fit. This solution is furthermore in line with the specifiers of the additional coding of Limited Prosocial Emotions for Conduct Disorder in the DSM-V.

Prediction models for diagnosis and prognosis of covid-19: systematic review and critical appraisal

OBJECTIVE: To review and appraise the validity and usefulness of published and preprint reports of prediction models for diagnosing coronavirus disease 2019 (covid-19) in patients with suspected infection, for prognosis of patients with covid-19, and for detecting people in the general population at increased risk of covid-19 infection or being admitted to hospital with the disease. DESIGN: Living systematic review and critical appraisal by the COVID-PRECISE (Precise Risk Estimation to optimise covid-19 Care for Infected or Suspected patients in diverse sEttings) group. DATA SOURCES: PubMed and Embase through Ovid, up to 1 July 2020, supplemented with arXiv, medRxiv, and bioRxiv up to 5 May 2020. STUDY SELECTION: Studies that developed or validated a multivariable covid-19 related prediction model. DATA EXTRACTION: At least two authors independently extracted data using the CHARMS (critical appraisal and data extraction for systematic reviews of prediction modelling studies) checklist; risk of bias was assessed using PROBAST (prediction model risk of bias assessment tool). RESULTS: 37 421 titles were screened, and 169 studies describing 232 prediction models were included. The review identified seven models for identifying people at risk in the general population; 118 diagnostic models for detecting covid-19 (75 were based on medical imaging, 10 to diagnose disease severity); and 107 prognostic models for predicting mortality risk, progression to severe disease, intensive care unit admission, ventilation, intubation, or length of hospital stay. The most frequent types of predictors included in the covid-19 prediction models are vital signs, age, comorbidities, and image features. Flu-like symptoms are frequently predictive in diagnostic models, while sex, C reactive protein, and lymphocyte counts are frequent prognostic factors. Reported C index estimates from the strongest form of validation available per model ranged from 0.71 to 0.99 in prediction models for the general population, from 0.65 to more than 0.99 in diagnostic models, and from 0.54 to 0.99 in prognostic models. All models were rated at high or unclear risk of bias, mostly because of non-representative selection of control patients, exclusion of patients who had not experienced the event of interest by the end of the study, high risk of model overfitting, and unclear reporting. Many models did not include a description of the target population (n=27, 12%) or care setting (n=75, 32%), and only 11 (5%) were externally validated by a calibration plot. The Jehi diagnostic model and the 4C mortality score were identified as promising models. CONCLUSION: Prediction models for covid-19 are quickly entering the academic literature to support medical decision making at a time when they are urgently needed. This review indicates that almost all pubished prediction models are poorly reported, and at high risk of bias such that their reported predictive performance is probably optimistic. However, we have identified two (one diagnostic and one prognostic) promising models that should soon be validated in multiple cohorts, preferably through collaborative efforts and data sharing to also allow an investigation of the stability and heterogeneity in their performance across populations and settings. Details on all reviewed models are publicly available at https://www.covprecise.org/. Methodological guidance as provided in this paper should be followed because unreliable predictions could cause more harm than benefit in guiding clinical decisions. Finally, prediction model authors should adhere to the TRIPOD (transparent reporting of a multivariable prediction model for individual prognosis or diagnosis) reporting guideline. SYSTEMATIC REVIEW REGISTRATION: Protocol https://osf.io/ehc47/, registration https://osf.io/wy245. READERS' NOTE: This article is a living systematic review that will be updated to reflect emerging evidence. Updates may occur for up to two years from the date of original publication. This version is update 3 of the original article published on 7 April 2020 (BMJ 2020;369:m1328). Previous updates can be found as data supplements (https://www.bmj.com/content/369/bmj.m1328/related#datasupp). When citing this paper please consider adding the update number and date of access for clarity.

Numerical investigation of complete mandibular dentures stabilized by conventional or mini implants in patient individual models.

Poor stability of a complete denture is a common problem due to bone atrophy of the edentulous ridge. The aim of the present study was to analyze denture stability after receiving implants and to study the biomechanical properties of denture implants and the bone bed using conventional or mini implants. Five models based on computed tomography (CT) data of edentulous patients were created. The overdentures' connection to the implants was assured by means of ball head abutments and rubber rings. In three models, the denture was supported by two to four conventional implants and in two models, the overdenture was supported by three to five mini implants. The dentures were loaded according to the individual biting forces which was clinically measured by means of pressure sheets. After implantation, the biting forces and displacements of overdentures increased in comparison to complete dentures. Displacements and stresses were higher with mini implants than with conventional ones. Stress in the implants was markedly below the yield stress of titanium grade 5 (880 MPa). An increase in the stress in the bone around the implants was noticed as compared to the situation with complete dentures which was below the physiological range of bone loading (<4 MPa).

The latent nature of prolonged grief - A taxometric analysis: Results from a representative population sample.

Individuals suffering from prolonged grief disorder (PGD) show severe grief reactions after the death of a significant other, even beyond a period of grieving that is within a person's cultural and religious context. In addition to this core element, PGD can manifest in various ways. Symptoms may include persistent preoccupation, intense emotional pain, or impairment in important life domains. The symptoms, furthermore, have to be of culturally or religiously inappropriate extent or severity, taking into account different norms of grieving. PGD is discussed as a distinct diagnostic category in the revision of the International Classification of Diseases (ICD-11). Nosology of PGD has been highly debated surrounding the suggested inclusion in the DSM-5, which had been declined due to insufficient evidence. This paper addresses the latent nature of PGD. Using a short form of the Inventory of Complicated Grief-Revised (ICG-R), we applied three popular taxometric methods: MAXEIG, MAMBAC and L-Mode. Data stemmed from a subsample of N = 1445 bereaved individuals that participated in a large representative German population survey (N = 2520). The analysis strongly indicated a dimensional latent structure of PGD. Implications of the conceptualization of PGD on a continuum are discussed, regarding measurement, diagnosis, etiology and future research.

Dimensional Latent Structure of Early Disruptive Behavior Disorders: A Taxometric Analysis in Preschoolers.

Although disruptive behavior disorders (DBDs) are used as a distinct categorical diagnosis in clinical practice, they have repeatedly been described as having a dimensional structure in taxometric analyses. In the current study the authors analyzed the latent status of disruptive behaviors (DB) in a large sample (N = 2,808) of German preschool children (2-6 years old, mean age 53.7 months, SD = 13.5, 48.4% girls). The Eyberg Child Behavior Inventory (ECBI) as well as the Strengths and Difficulties Questionnaire (SDQ) were used to compile indicators of the DB core dimensions (Temper Loss, Aggression, Noncompliance, and Low Concern for others). Three widely used taxometric methods (a) MAXEIG, (b) MAMBAC, and (c) L-Mode were applied. Simulation data were created to evaluate the Comparison Curve Fit Index values (CCFIs), which were below 0.45, supporting a dimensional solution. Hence, in the current study the latent structure of DB in preschoolers encompassed differences in degree rather than kind. Researchers and practitioners should be mindful of the dimensional latent status of DB in theory building, assessment, classification, and labeling.

Follow-up study regarding the medium-term effectiveness of the home-visiting program "Pro Kind" at age 7 years: study protocol for a randomized controlled trial.

BACKGROUND: Pro Kind is a German adaptation of the US Nurse Family Partnership program. It is an intervention based on home visits targeting first-time mothers from disadvantaged populations. Pro Kind was implemented as a randomized control trial from 2006 to 2012 with N = 755 first-time mothers (TG n = 394, CG n = 391). The 7-8-year follow-up aims to assess the mid-term effects of the program. METHODS/DESIGN: Mid-term outcomes are being assessed by trained assessors. In a multimethod approach telephone interviews, on-site interviews, observations and developmental tests will be held in order to assess children's and mothers' life satisfaction, mental health, cognitive and social development, parenting behavior, signs of child abuse or neglect as well as the family's socio-economic status. Furthermore, administrative data will be accessed to obtain information regarding the mother's usage of pediatric health care, welfare usage and employment history. DISCUSSION: Results regarding the mid-term effects of the intervention from the Pro Kind Follow-up will provide a scientific basis for future primary prevention programs as well as help stakeholders legitimizing early childhood investments. TRIAL REGISTRATION: German Clinical Trial Registration DRKS-ID, ID: DRKS00007554 . Registered on 11 June 2015, updated on 6 October 2017.

Differential Influences of Parenting Dimensions and Parental Physical Abuse during Childhood on Overweight and Obesity in Adolescents.

Besides other explanatory variables, parenting styles and parental violence might also be responsible for setting a path towards overweight/obesity in childhood. While this association has consistently been observed for adults, findings for adolescents still remain scarce and inconsistent. Therefore, the goal of this study is to add evidence on this topic for children and adolescents. Analyses are based on a sample of 1729 German, ninth-grade students. To analyze associations between parenting dimensions and weight status, non-parametric conditional inference trees were applied. Three gender-specific pathways for a heightened risk of overweight/obesity were observed: (1) female adolescents who report having experienced severe parental physical abuse and medium/high parental warmth in childhood; (2) male adolescents who report having experienced low or medium parental monitoring in childhood; and (3) this second pathway for male adolescents is more pronounced if the families receive welfare. The importance of promoting parenting styles characterized by warmth and a lack of physical abuse is also discussed. This is one of only a few studies examining the association of parenting dimensions/parental physical abuse and weight status in adolescence. Future studies should include even more parenting dimensions, as well as parental physical abuse levels, in order to detect and untangle gender-specific effects on weight status.

Brief assessment of subjective health complaints: Development, validation and population norms of a brief form of the Giessen Subjective Complaints List (GBB-8).

OBJECTIVE: Although there is no causal relationship to medical morbidity, routine clinical assessment of somatic symptoms aids medical diagnosis and assessment of treatment effectiveness. Regardless of their causes, somatic symptoms indicate suffering, distress, and help-seeking behavior. The aim of the present study was to develop and validate a brief self-report questionnaire to assess somatic symptom strain. METHODS: A brief form of the Giessen Subjective Complaints List (GBB-8) was developed and validated in a large population sample representative of the Federal Republic of Germany (N=2008). Psychometric analyses included confirmation of factor structure, classical item analysis, and measurement invariance tests. The sample furthermore served as a norm group. As indicators of construct validity, correlations with measures of anxiety, depression, alexithymia, and primary care contact were computed. RESULTS: Psychometric analyses yielded excellent scale properties regarding item characteristics, factor structure, and measurement invariance tests (Cronbach's alpha=0.88; CFI=0.980, TLI=0.965, RMSEA=0.049) for the second-order four-factor model; strict invariance was confirmed for gender, depression status, and physician contacts; strong invariance was confirmed regarding age and age×gender. CONCLUSIONS: The GBB-8 with its four subscales exhaustion, gastrointestinal complaints, musculoskeletal complaints, and cardiovascular complaints proves to be an economic measure of subjective symptom strain. Psychometric analyses deem it suitable for epidemiological research. The availability of norms makes it a potential everyday tool for general practitioners and psychosomatic clinics.