Understanding random resampling techniques for class imbalance correction and their consequences on calibration and discrimination of clinical risk prediction models.

OBJECTIVE: Class imbalance is sometimes considered a problem when developing clinical prediction models and assessing their performance. To address it, correction strategies involving manipulations of the training dataset, such as random undersampling or oversampling, are frequently used. The aim of this article is to illustrate the consequences of these class imbalance correction strategies on clinical prediction models' internal validity in terms of calibration and discrimination performances. METHODS: We used both heuristic intuition and formal mathematical reasoning to characterize the relations between conditional probabilities of interest and probabilities targeted when using random undersampling or oversampling. We propose a plug-in estimator that represents a natural correction for predictions obtained from models that have been trained on artificially balanced datasets ("naïve" models). We conducted a Monte Carlo simulation with two different data generation processes and present a real-world example using data from the International Stroke Trial database to empirically demonstrate the consequences of applying random resampling techniques for class imbalance correction on calibration and discrimination (in terms of Area Under the ROC, AUC) for logistic regression and tree-based prediction models. RESULTS: Across our simulations and in the real-world example, calibration of the naïve models was very poor. The models using the plug-in estimator generally outperformed the models relying on class imbalance correction in terms of calibration while achieving the same discrimination performance. CONCLUSION: Random resampling techniques for class imbalance correction do not generally improve discrimination performance (i.e., AUC), and their use is hard to justify when aiming at providing calibrated predictions. Improper use of such class imbalance correction techniques can lead to suboptimal data usage and less valid risk prediction models.

Predicting the SARS-CoV-2 effective reproduction number using bulk contact data from mobile phones.

Over the last months, cases of SARS-CoV-2 surged repeatedly in many countries but could often be controlled with nonpharmaceutical interventions including social distancing. We analyzed deidentified Global Positioning System (GPS) tracking data from 1.15 to 1.4 million cell phones in Germany per day between March and November 2020 to identify encounters between individuals and statistically evaluate contact behavior. Using graph sampling theory, we estimated the contact index (CX), a metric for number and heterogeneity of contacts. We found that CX, and not the total number of contacts, is an accurate predictor for the effective reproduction number R derived from case numbers. A high correlation between CX and R recorded more than 2 wk later allows assessment of social behavior well before changes in case numbers become detectable. By construction, the CX quantifies the role of superspreading and permits assigning risks to specific contact behavior. We provide a critical CX value beyond which R is expected to rise above 1 and propose to use that value to leverage the social-distancing interventions for the coming months.

COVID-19 Vaccine Effectiveness and Digital Pandemic Surveillance in Germany (eCOV Study): Web Application-Based Prospective Observational Cohort Study.

BACKGROUND: The COVID-19 pandemic posed significant challenges to global health systems. Efficient public health responses required a rapid and secure collection of health data to improve the understanding of SARS-CoV-2 and examine the vaccine effectiveness (VE) and drug safety of the novel COVID-19 vaccines. OBJECTIVE: This study (COVID-19 study on vaccinated and unvaccinated subjects over 16 years; eCOV study) aims to (1) evaluate the real-world effectiveness of COVID-19 vaccines through a digital participatory surveillance tool and (2) assess the potential of self-reported data for monitoring key parameters of the COVID-19 pandemic in Germany. METHODS: Using a digital study web application, we collected self-reported data between May 1, 2021, and August 1, 2022, to assess VE, test positivity rates, COVID-19 incidence rates, and adverse events after COVID-19 vaccination. Our primary outcome measure was the VE of SARS-CoV-2 vaccines against laboratory-confirmed SARS-CoV-2 infection. The secondary outcome measures included VE against hospitalization and across different SARS-CoV-2 variants, adverse events after vaccination, and symptoms during infection. Logistic regression models adjusted for confounders were used to estimate VE 4 to 48 weeks after the primary vaccination series and after third-dose vaccination. Unvaccinated participants were compared with age- and gender-matched participants who had received 2 doses of BNT162b2 (Pfizer-BioNTech) and those who had received 3 doses of BNT162b2 and were not infected before the last vaccination. To assess the potential of self-reported digital data, the data were compared with official data from public health authorities. RESULTS: We enrolled 10,077 participants (aged ≥16 y) who contributed 44,786 tests and 5530 symptoms. In this young, primarily female, and digital-literate cohort, VE against infections of any severity waned from 91.2% (95% CI 70.4%-97.4%) at week 4 to 37.2% (95% CI 23.5%-48.5%) at week 48 after the second dose of BNT162b2. A third dose of BNT162b2 increased VE to 67.6% (95% CI 50.3%-78.8%) after 4 weeks. The low number of reported hospitalizations limited our ability to calculate VE against hospitalization. Adverse events after vaccination were consistent with previously published research. Seven-day incidences and test positivity rates reflected the course of the pandemic in Germany when compared with official numbers from the national infectious disease surveillance system. CONCLUSIONS: Our data indicate that COVID-19 vaccinations are safe and effective, and third-dose vaccinations partially restore protection against SARS-CoV-2 infection. The study showcased the successful use of a digital study web application for COVID-19 surveillance and continuous monitoring of VE in Germany, highlighting its potential to accelerate public health decision-making. Addressing biases in digital data collection is vital to ensure the accuracy and reliability of digital solutions as public health tools.

transferGWAS: GWAS of images using deep transfer learning.

MOTIVATION: Medical images can provide rich information about diseases and their biology. However, investigating their association with genetic variation requires non-standard methods. We propose transferGWAS, a novel approach to perform genome-wide association studies directly on full medical images. First, we learn semantically meaningful representations of the images based on a transfer learning task, during which a deep neural network is trained on independent but similar data. Then, we perform genetic association tests with these representations. RESULTS: We validate the type I error rates and power of transferGWAS in simulation studies of synthetic images. Then we apply transferGWAS in a genome-wide association study of retinal fundus images from the UK Biobank. This first-of-a-kind GWAS of full imaging data yielded 60 genomic regions associated with retinal fundus images, of which 7 are novel candidate loci for eye-related traits and diseases. AVAILABILITY AND IMPLEMENTATION: Our method is implemented in Python and available at https://github.com/mkirchler/transferGWAS/. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Assessing the transportability of clinical prediction models for cognitive impairment using causal models.

BACKGROUND: Machine learning models promise to support diagnostic predictions, but may not perform well in new settings. Selecting the best model for a new setting without available data is challenging. We aimed to investigate the transportability by calibration and discrimination of prediction models for cognitive impairment in simulated external settings with different distributions of demographic and clinical characteristics. METHODS: We mapped and quantified relationships between variables associated with cognitive impairment using causal graphs, structural equation models, and data from the ADNI study. These estimates were then used to generate datasets and evaluate prediction models with different sets of predictors. We measured transportability to external settings under guided interventions on age, APOE ε4, and tau-protein, using performance differences between internal and external settings measured by calibration metrics and area under the receiver operating curve (AUC). RESULTS: Calibration differences indicated that models predicting with causes of the outcome were more transportable than those predicting with consequences. AUC differences indicated inconsistent trends of transportability between the different external settings. Models predicting with consequences tended to show higher AUC in the external settings compared to internal settings, while models predicting with parents or all variables showed similar AUC. CONCLUSIONS: We demonstrated with a practical prediction task example that predicting with causes of the outcome results in better transportability compared to anti-causal predictions when considering calibration differences. We conclude that calibration performance is crucial when assessing model transportability to external settings.

Comparison of Bayesian Networks, G-estimation and linear models to estimate causal treatment effects in aggregated N-of-1 trials with carry-over effects.

BACKGROUND: The aggregation of a series of N-of-1 trials presents an innovative and efficient study design, as an alternative to traditional randomized clinical trials. Challenges for the statistical analysis arise when there is carry-over or complex dependencies of the treatment effect of interest. METHODS: In this study, we evaluate and compare methods for the analysis of aggregated N-of-1 trials in different scenarios with carry-over and complex dependencies of treatment effects on covariates. For this, we simulate data of a series of N-of-1 trials for Chronic Nonspecific Low Back Pain based on assumed causal relationships parameterized by directed acyclic graphs. In addition to existing statistical methods such as regression models, Bayesian Networks, and G-estimation, we introduce a carry-over adjusted parametric model (COAPM). RESULTS: The results show that all evaluated existing models have a good performance when there is no carry-over and no treatment dependence. When there is carry-over, COAPM yields unbiased and more efficient estimates while all other methods show some bias in the estimation. When there is known treatment dependence, all approaches that are capable to model it yield unbiased estimates. Finally, the efficiency of all methods decreases slightly when there are missing values, and the bias in the estimates can also increase. CONCLUSIONS: This study presents a systematic evaluation of existing and novel approaches for the statistical analysis of a series of N-of-1 trials. We derive practical recommendations which methods may be best in which scenarios.

Study protocol: combined N-of-1 trials to assess open-label placebo treatment for antidepressant discontinuation symptoms [FAB-study].

BACKGROUND: Antidepressant discontinuation is associated with a broad range of adverse effects. Debilitating discontinuation symptoms can impede the discontinuation process and contribute to unnecessary long-term use of antidepressants. Antidepressant trials reveal large placebo effects, indicating a potential use of open-label placebo (OLP) treatment to facilitate the discontinuation process. We aim to determine the effect of OLP treatment in reducing antidepressant discontinuation symptoms using a series of N-of-1 trials. METHODS: A series of randomized, single-blinded N-of-1 trials will be conducted in 20 patients with fully remitted DSM-V major depressive disorder, experiencing moderate to severe discontinuation symptoms following antidepressant discontinuation. Each N-of-1 trial consists of two cycles, each comprising two-week alternating periods of OLP treatment and of no treatment in a random order, for a total of eight weeks. Our primary outcome will be self-reported discontinuation symptoms rated twice daily via the smartphone application 'StudyU'. Secondary outcomes include expectations about discontinuation symptoms and (depressed) mood. Statistical analyses will be based on a Bayesian multi-level random effects model, reporting posterior estimates of the overall and individual treatment effects. DISCUSSION: Results of this trial will provide insight into the clinical application of OLP in treating antidepressant discontinuation symptoms, potentially offering a new cost-effective therapeutic tool. This trial will also determine the feasibility and applicability of a series of N-of-1 trials in a clinical discontinuation trial. TRIAL REGISTRATION: ClinicalTrials.gov: NCT05051995, first registered September 20, 2021.

Diagnostic accuracy of cerebrospinal fluid biomarkers for the differential diagnosis of sporadic Creutzfeldt-Jakob disease: a (network) meta-analysis.

BACKGROUND: There are no systematic reviews of cerebrospinal fluid and blood biomarkers for sporadic Creutzfeldt-Jakob disease (sCJD) in specialized care settings that compare diagnostic accuracies in a network meta-analysis (NMA). METHODS: We searched Medline, Embase, and Cochrane Library for diagnostic studies of sCJD biomarkers. Studies had to use established diagnostic criteria for sCJD and for diseases in the non-CJD groups, which had to represent a consecutive population of patients suspected as a CJD case, as reference standard. Risk of bias was assessed with QUADAS-2. We conducted individual biomarker meta-analyses with generalized bivariate models. To investigate heterogeneity, we performed subgroup analyses based on QUADAS-2 quality and clinical criteria. For the NMA, we applied a Bayesian beta-binomial ANOVA model. The study protocol was registered at PROSPERO (CRD42019118830). RESULTS: Of 2976 publications screened, we included 16 studies, which investigated 14-3-3ß (n = 13), 14-3-3γ (n = 3), neurofilament light chain (NfL, n = 1), neuron-specific enolase (n = 1), p-tau181/t-tau ratio (n = 2), RT-QuIC (n = 7), S100B (n = 3), t-tau (n = 12), and t-tau/Aß42 ratio (n = 1). Excluded diagnostic studies had strong limitations in study design. In the NMA, RT-QuIC (0.91; 95% CI [0.83, 0.95]) and NfL (0.93 [0.78, 0.99]) were the most sensitive biomarkers for the diagnosis of definite, probable, and possible sCJD cases. RT-QuIC was the most specific biomarker (0.97 [0.89, 1.00]). Heterogeneity in accuracy estimates was high between studies. CONCLUSIONS: We identified RT-QuIC as the most accurate biomarker, partially confirming currently applied diagnostic criteria. The shortcomings identified in many diagnostic studies for sCJD biomarkers need to be addressed in future studies.

Participatory disease surveillance for a mass gathering - a prospective cohort study on COVID-19, Germany 2021.

BACKGROUND: Mass gatherings (MGs) such as music festivals and sports events have been associated with a high risk of SARS-CoV-2 transmission. On-site research can foster knowledge of risk factors for infections and improve risk assessments and precautionary measures at future MGs. We tested a web-based participatory disease surveillance tool to detect COVID-19 infections at and after an outdoor MG by collecting self-reported COVID-19 symptoms and tests. METHODS: We conducted a digital prospective observational cohort study among fully immunized attendees of a sports festival that took place from September 2 to 5, 2021 in Saxony-Anhalt, Germany. Participants used our study app to report demographic data, COVID-19 tests, symptoms, and their contact behavior. This self-reported data was used to define probable and confirmed COVID-19 cases for the full "study period" (08/12/2021 - 10/31/2021) and within the 14-day "surveillance period" during and after the MG, with the highest likelihood of an MG-related COVID-19 outbreak (09/04/2021 - 09/17/2021). RESULTS: A total of 2,808 of 9,242 (30.4%) event attendees participated in the study. Within the study period, 776 individual symptoms and 5,255 COVID-19 tests were reported. During the 14-day surveillance period around and after the MG, seven probable and seven PCR-confirmed COVID-19 cases were detected. The confirmed cases translated to an estimated seven-day incidence of 125 per 100,000 participants (95% CI [67.7/100,000, 223/100,000]), which was comparable to the average age-matched incidence in Germany during this time. Overall, weekly numbers of COVID-19 cases were fluctuating over the study period, with another increase at the end of the study period. CONCLUSION: COVID-19 cases attributable to the mass gathering were comparable to the Germany-wide age-matched incidence, implicating that our active participatory disease surveillance tool was able to detect MG-related infections. Further studies are needed to evaluate and apply our participatory disease surveillance tool in other mass gathering settings.

Digital Assessment Tools Using Animation Features to Quantify Alcohol Consumption: Systematic App Store and Literature Review.

BACKGROUND: Accurate and user-friendly assessment tools for quantifying alcohol consumption are a prerequisite for effective interventions to reduce alcohol-related harm. Digital assessment tools (DATs) that allow the description of consumed alcoholic drinks through animation features may facilitate more accurate reporting than conventional approaches. OBJECTIVE: This review aims to identify and characterize freely available DATs in English or Russian that use animation features to support the quantitative assessment of alcohol consumption (alcohol DATs) and determine the extent to which such tools have been scientifically evaluated in terms of feasibility, acceptability, and validity. METHODS: Systematic English and Russian searches were conducted in iOS and Android app stores and via the Google search engine. Information on the background and content of eligible DATs was obtained from app store descriptions, websites, and test completions. A systematic literature review was conducted in Embase, MEDLINE, PsycINFO, and Web of Science to identify English-language studies reporting the feasibility, acceptability, and validity of animation-using alcohol DATs. Where possible, the evaluated DATs were accessed and assessed. Owing to the high heterogeneity of study designs, results were synthesized narratively. RESULTS: We identified 22 eligible alcohol DATs in English, 3 (14%) of which were also available in Russian. More than 95% (21/22) of tools allowed the choice of a beverage type from a visually displayed selection. In addition, 36% (8/22) of tools enabled the choice of a drinking vessel. Only 9% (2/22) of tools allowed the simulated interactive pouring of a drink. For none of the tools published evaluation studies were identified in the literature review. The systematic literature review identified 5 exploratory studies evaluating the feasibility, acceptability, and validity of 4 animation-using alcohol DATs, 1 (25%) of which was available in the searched app stores. The evaluated tools reached moderate to high scores on user rating scales and showed fair to high convergent validity when compared with established assessment methods. CONCLUSIONS: Animation-using alcohol DATs are available in app stores and on the web. However, they often use nondynamic features and lack scientific background information. Explorative study data suggest that such tools might enable the user-friendly and valid assessment of alcohol consumption and could thus serve as a building block in the reduction of alcohol-attributable health burden worldwide. TRIAL REGISTRATION: PROSPERO International Prospective Register of Systematic Reviews CRD42020172825; https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42020172825.

StudyU: A Platform for Designing and Conducting Innovative Digital N-of-1 Trials.

N-of-1 trials are the gold standard study design to evaluate individual treatment effects and derive personalized treatment strategies. Digital tools have the potential to initiate a new era of N-of-1 trials in terms of scale and scope, but fully functional platforms are not yet available. Here, we present the open source StudyU platform, which includes the StudyU Designer and StudyU app. With the StudyU Designer, scientists are given a collaborative web application to digitally specify, publish, and conduct N-of-1 trials. The StudyU app is a smartphone app with innovative user-centric elements for participants to partake in trials published through the StudyU Designer to assess the effects of different interventions on their health. Thereby, the StudyU platform allows clinicians and researchers worldwide to easily design and conduct digital N-of-1 trials in a safe manner. We envision that StudyU can change the landscape of personalized treatments both for patients and healthy individuals, democratize and personalize evidence generation for self-optimization and medicine, and can be integrated in clinical practice.

Powerful rare variant association testing in a copula-based joint analysis of multiple phenotypes.

In genetic association studies of rare variants, the low power of association tests is one of the main challenges. In this study, we propose a new single-marker association test called C-JAMP (Copula-based Joint Analysis of Multiple Phenotypes), which is based on a joint model of multiple phenotypes given genetic markers and other covariates. We evaluated its performance and compared its empirical type I error and power with existing univariate and multivariate single-marker and multi-marker rare-variant tests in extensive simulation studies. C-JAMP yielded unbiased genetic effect estimates and valid type I errors with an adjusted test statistic. When strongly dependent traits were jointly analyzed, C-JAMP had the highest power in all scenarios except when a high percentage of variants were causal with moderate/small effect sizes. When traits with weak or moderate dependence were analyzed, whether C-JAMP or competing approaches had higher power depended on the effect size. When C-JAMP was applied with a misspecified copula function, it still achieved high power in some of the scenarios considered. In a real-data application, we analyzed sequencing data using C-JAMP and performed the first genome-wide association studies of high-molecular-weight and medium-molecular-weight adiponectin plasma concentrations. C-JAMP identified 20 rare variants with p-values smaller than 10-5 , while all other tests resulted in the identification of fewer variants with higher p-values. In summary, the results indicate that C-JAMP is a powerful, flexible, and robust method for association studies, and we identified novel candidate markers for adiponectin. C-JAMP is implemented as an R package and freely available from https://cran.r-project.org/package=CJAMP.

Directed acyclic graphs and causal thinking in clinical risk prediction modeling.

BACKGROUND: In epidemiology, causal inference and prediction modeling methodologies have been historically distinct. Directed Acyclic Graphs (DAGs) are used to model a priori causal assumptions and inform variable selection strategies for causal questions. Although tools originally designed for prediction are finding applications in causal inference, the counterpart has remained largely unexplored. The aim of this theoretical and simulation-based study is to assess the potential benefit of using DAGs in clinical risk prediction modeling. METHODS: We explore how incorporating knowledge about the underlying causal structure can provide insights about the transportability of diagnostic clinical risk prediction models to different settings. We further probe whether causal knowledge can be used to improve predictor selection in clinical risk prediction models. RESULTS: A single-predictor model in the causal direction is likely to have better transportability than one in the anticausal direction in some scenarios. We empirically show that the Markov Blanket, the set of variables including the parents, children, and parents of the children of the outcome node in a DAG, is the optimal set of predictors for that outcome. CONCLUSIONS: Our findings provide a theoretical basis for the intuition that a diagnostic clinical risk prediction model including causes as predictors is likely to be more transportable. Furthermore, using DAGs to identify Markov Blanket variables may be a useful, efficient strategy to select predictors in clinical risk prediction models if strong knowledge of the underlying causal structure exists or can be learned.

Association of dietary intake of milk and dairy products with blood concentrations of insulin-like growth factor 1 (IGF-1) in Bavarian adults.

PURPOSE: Circulating IGF-1 concentrations have been associated with higher cancer risk, particularly prostate, breast and colorectal cancer. There is evidence from observational and intervention studies that milk and dairy products intake is associated with higher IGF-1 concentrations, but results were not always consistent. The purpose of this study was to examine the relationship between dairy intake and circulating IGF-1 concentrations in participants of the Second Bavarian Food Consumption Survey, thereby providing data for a German population for the first time. METHODS: In this cross-sectional study of 526 men and women aged 18-80 years, in contrast to most previous investigations, dietary intake was assessed with a more detailed instrument than food frequency questionnaires (FFQs), i.e., by three 24-h dietary recalls conducted on random days close in time to the blood collection. Circulating IGF-1 concentrations were measured in blood samples. Multivariable linear regression models were used to examine the association of dairy intake with IGF-1 concentrations. RESULTS: Each 400 g increment in daily dairy intake was associated with 16.8 µg/L (95% CI 6.9, 26.7) higher IGF-1 concentrations. Each 200 g increment in milk per day was associated with 10.0 µg/L (95% CI 4.2, 15.8) higher IGF-1. In contrast, we observed no association between cheese or yogurt intake and IGF-1 concentrations. CONCLUSIONS: Our findings are in line with most previous investigations and support the hypothesis that dairy and milk intake are associated with higher IGF-1 concentrations.

Estimating and testing direct genetic effects in directed acyclic graphs using estimating equations.

In genetic association studies, it is important to distinguish direct and indirect genetic effects in order to build truly functional models. For this purpose, we consider a directed acyclic graph setting with genetic variants, primary and intermediate phenotypes, and confounding factors. In order to make valid statistical inference on direct genetic effects on the primary phenotype, it is necessary to consider all potential effects in the graph, and we propose to use the estimating equations method with robust Huber-White sandwich standard errors. We evaluate the proposed causal inference based on estimating equations (CIEE) method and compare it with traditional multiple regression methods, the structural equation modeling method, and sequential G-estimation methods through a simulation study for the analysis of (completely observed) quantitative traits and time-to-event traits subject to censoring as primary phenotypes. The results show that CIEE provides valid estimators and inference by successfully removing the effect of intermediate phenotypes from the primary phenotype and is robust against measured and unmeasured confounding of the indirect effect through observed factors. All other methods except the sequential G-estimation method for quantitative traits fail in some scenarios where their test statistics yield inflated type I errors. In the analysis of the Genetic Analysis Workshop 19 dataset, we estimate and test genetic effects on blood pressure accounting for intermediate gene expression phenotypes. The results show that CIEE can identify genetic variants that would be missed by traditional regression analyses. CIEE is computationally fast, widely applicable to different fields, and available as an R package.

Variability and reliability study of overall physical activity and activity intensity levels using 24 h-accelerometry-assessed data.

BACKGROUND: 24 h-accelerometry is now used to objectively assess physical activity (PA) in many observational studies like the German National Cohort; however, PA variability, observational time needed to estimate habitual PA, and reliability are unclear. METHODS: We assessed 24 h-PA of 50 participants using triaxial accelerometers (ActiGraph GT3X+) over 2 weeks. Variability of overall PA and different PA intensities (time in inactivity and in low intensity, moderate, vigorous, and very vigorous PA) between days of assessment or days of the week was quantified using linear mixed-effects and random effects models. We calculated the required number of days to estimate PA, and calculated PA reliability using intraclass correlation coefficients. RESULTS: Between- and within-person variance accounted for 34.4-45.5% and 54.5-65.6%, respectively, of total variance in overall PA and PA intensities over the 2 weeks. Overall PA and times in low intensity, moderate, and vigorous PA decreased slightly over the first 3 days of assessment. Overall PA (p = 0.03), time in inactivity (p = 0.003), in low intensity PA (p = 0.001), in moderate PA (p = 0.02), and in vigorous PA (p = 0.04) slightly differed between days of the week, being highest on Wednesday and Friday and lowest on Sunday and Monday, with apparent differences between Saturday and Sunday. In nested random models, the day of the week accounted for < 19% of total variance in the PA parameters. On average, the required number of days to estimate habitual PA was around 1 week, being 7 for overall PA and ranging from 6 to 9 for the PA intensities. Week-to-week reliability was good (intraclass correlation coefficients, range, 0.68-0.82). CONCLUSIONS: Individual PA, as assessed using 24 h-accelerometry, is highly variable between days, but the day of assessment or the day of the week explain only small parts of this variance. Our data indicate that 1 week of assessment is necessary for reliable estimation of habitual PA.

Joint analysis of multiple phenotypes: summary of results and discussions from the Genetic Analysis Workshop 19.

For Genetic Analysis Workshop 19, 2 extensive data sets were provided, including whole genome and whole exome sequence data, gene expression data, and longitudinal blood pressure outcomes, together with nongenetic covariates. These data sets gave researchers the chance to investigate different aspects of more complex relationships within the data, and the contributions in our working group focused on statistical methods for the joint analysis of multiple phenotypes, which is part of the research field of data integration. The analysis of data from different sources poses challenges to researchers but provides the opportunity to model the real-life situation more realistically.Our 4 contributions all used the provided real data to identify genetic predictors for blood pressure. In the contributions, novel multivariate rare variant tests, copula models, structural equation models and a sparse matrix representation variable selection approach were applied. Each of these statistical models can be used to investigate specific hypothesized relationships, which are described together with their biological assumptions.The results showed that all methods are ready for application on a genome-wide scale and can be used or extended to include multiple omics data sets. The results provide potentially interesting genetic targets for future investigation and replication. Furthermore, all contributions demonstrated that the analysis of complex data sets could benefit from modeling correlated phenotypes jointly as well as by adding further bioinformatics information.

Analyzing population-level trials as N-of-1 trials: An application to gait.

Studying individual causal effects of health interventions is important whenever intervention effects are heterogeneous between study participants. Conducting N-of-1 trials, which are single-person randomized controlled trials, is the gold standard for their analysis. As an alternative method, we propose to re-analyze existing population-level studies as N-of-1 trials, and use gait as a use case for illustration. Gait data were collected from 16 young and healthy participants under fatigued and non-fatigued, as well as under single-task (only walking) and dual-task (walking while performing a cognitive task) conditions. As a reference to the N-of-1 trials approach, we first computed standard population-level ANOVA models to evaluate differences in gait parameters (stride length and stride time) across conditions. Then, we estimated the effect of the interventions on gait parameters on the individual level through Bayesian repeated-measures models, viewing each participant as their own trial, and compared the results. The results illustrated that while few overall population-level effects were visible, individual-level analyses revealed differences between participants. Baseline values of the gait parameters varied largely among all participants, and the effects of fatigue and cognitive task were also heterogeneous, with some individuals showing effects in opposite directions. These differences between population-level and individual-level analyses were more pronounced for the fatigue intervention compared to the cognitive task intervention. Following our empirical analysis, we discuss re-analyzing population studies through the lens of N-of-1 trials more generally and highlight important considerations and requirements. Our work encourages future studies to investigate individual effects using population-level data.

FASDetect as a machine learning-based screening app for FASD in youth with ADHD.

Fetal alcohol-spectrum disorder (FASD) is underdiagnosed and often misdiagnosed as attention-deficit/hyperactivity disorder (ADHD). Here, we develop a screening tool for FASD in youth with ADHD symptoms. To develop the prediction model, medical record data from a German University outpatient unit are assessed including 275 patients aged 0-19 years old with FASD with or without ADHD and 170 patients with ADHD without FASD aged 0-19 years old. We train 6 machine learning models based on 13 selected variables and evaluate their performance. Random forest models yield the best prediction models with a cross-validated AUC of 0.92 (95% confidence interval [0.84, 0.99]). Follow-up analyses indicate that a random forest model with 6 variables - body length and head circumference at birth, IQ, socially intrusive behaviour, poor memory and sleep disturbance - yields equivalent predictive accuracy. We implement the prediction model in a web-based app called FASDetect - a user-friendly, clinically scalable FASD risk calculator that is freely available at https://fasdetect.dhc-lab.hpi.de .