Uncontrolled Bias Research Articles

Improving the Reliability of, and Confidence in, DFT Functional Benchmarking through Active Learning.

Validating the performance of exchange-correlation functionals is vital to ensure the reliability of density functional theory (DFT) calculations. Typically, these validations involve benchmarking data sets. Currently, such data sets are usually assembled in an unprincipled manner, suffering from uncontrolled chemical bias, and limiting the transferability of benchmarking results to a broader chemical space. In this work, a data-efficient solution based on active learning is explored to address this issue. Focusing─as a proof of principle─on pericyclic reactions, we start from the BH9 benchmarking data set and design a chemical reaction space around this initial data set by combinatorially combining reaction templates and substituents. Next, a surrogate model is trained to predict the standard deviation of the activation energies computed across a selection of 20 distinct DFT functionals. With this model, the designed chemical reaction space is explored, enabling the identification of challenging regions, i.e., regions with large DFT functional divergence, for which representative reactions are subsequently acquired as additional training points. Remarkably, it turns out that the function mapping the molecular structure to functional divergence is readily learnable; convergence is reached upon the acquisition of fewer than 100 reactions. With our final updated model, a more challenging─and arguably more representative─pericyclic benchmarking data set is curated, and we demonstrate that the functional performance has changed significantly compared to the original BH9 subset.

Comprehensive evaluation of pure and hybrid collaborative filtering in drug repurposing

Drug development is known to be a costly and time-consuming process, which is prone to high failure rates. Drug repurposing allows drug discovery by reusing already approved compounds. The outcomes of past clinical trials can be used to predict novel drug-disease associations by leveraging drug- and disease-related similarities. To tackle this classification problem, collaborative filtering with implicit feedback (and potentially additional data on drugs and diseases) has become popular. It can handle large imbalances between negative and positive known associations and known and unknown associations. However, properly evaluating the improvement over the state of the art is challenging, as there is no consensus approach to compare models. We propose a reproducible methodology for comparing collaborative filtering-based drug repurposing. We illustrate this method by comparing 11 models from the literature on eight diverse drug repurposing datasets. Based on this benchmark, we derive guidelines to ensure a fair and comprehensive evaluation of the performance of those models. In particular, an uncontrolled bias on unknown associations might lead to severe data leakage and a misestimation of the model’s true performance. Moreover, in drug repurposing, the ability of a model to extrapolate beyond its training distribution is crucial and should also be assessed. Finally, we identified a subcategory of collaborative filtering that seems efficient and robust to distribution shifts. Benchmarks constitute an essential step towards increased reproducibility and more accessible development of competitive drug repurposing methods.

O-106 Bringing Transparency to Oocyte Assessment: the importance of including confounders when building Artificial Intelligence (AI) based support tools to quantify oocyte viability

Abstract Study question Which confounders (sperm quality, oocyte dysmorphism, culture time, images pre or post-ICSI, age) affect the ability of AI to predict blastulation based on oocyte images? Summary answer Sperm quality, oocyte dysmorphism, pre or post-ICSI image should be controlled for when building AI algorithms to predict blastulation based on oocyte images. What is known already Previous studies reporting on the use of AI to predict blastulation based on oocyte images have: (i) not accounted for confounders affecting blastulation (i.e. sperm quality, culture time), and (ii) used post-ICSI images; without assessing whether the ICSI procedure affects the oocyte image as assessed by AI. Therefore, there is a risk of mislabeling viable oocytes as non-viable due to external factors, which could cause uncontrolled bias and failure to generalize when used in clinical practice. The objective was to assess how these confounders affect efficacy of prediction of blastulation from oocyte images by an AI-based oocyte assessment tool: CHLOE-OQ(Fairtility). Study design, size, duration Cohort study. Images of 1281 oocytes (February to June 2022) were taken pre and post ICSI using the Embryoscope, and the embryos cultured until day 7. Oocyte donor source and age, oocyte dysmorphias and sperm quality were documented. CHLOE-OQ algorithm was trained, validated and tested in a diverse data set, accounting for pre and post ICSI image datasets, quality of oocytes, quality of sperm and patient age. Participants/materials, setting, methods The primary endpoint was blastulation. Sperm quality data was classified into 4 groups: (A)All (n = 1281), (B)donor sperm only (n = 51), (C)donor sperm and normospermic samples from men not diagnosed with male factor infertility (n = 557), (D)abnormal sperm samples and other diagnosed male factor cycles (n = 747). Eggs were classified by source (own/donor), and by dysmorphisms: enlarged perivitelline space, abnormal Zona pellucida, cytoplasmic abnormalities, dark, enlarged oocytes. Main results and the role of chance Post-ICSI images had higher mean CHLOE-OQ score than pre ICSI images (0.28±0.1 vs 0.33±0.1, p < 0.001, paired t-test). Discrepancies were particularly identified in oocytes that degenerated following ICSI, and scored 0 by CHLOE-OQ despite having higher scores pre-ICSI. Using Post-ICSI images (AUC=0.66, 95% confidence interval, CI: 0.63-0.69, n = 1281) improved the efficacy of prediction of blastulation compared to pre-ICSI images (AUC=0.57: 0.53-0.60, n = 1281, p < 0.001), suggesting that ICSI affected the quality morphology of the oocyte, and how an oocyte responds to ICSI, as assessed by AI, contributes to prediction of blastulation. Efficacy of prediction (AUC) was not affected by the quality of the sperm: (A-OVERALL 0.658 [CI(95%): 0.626-0.687]; B-Donor 0.586 [CI(95%): 0.449-0.728]; C-normospermic 0.645 [CI(95%): 0.600-0.688], D male factor 0.678 [CI(95%): 0.639-0.715]). Oocyte features associated with low CHLOE-OQ scores were: enlarged perivitelline space, dysmorphic oocytes, abnormal Zona pellucida, cytoplasmic abnormalities and dark and enlarged oocytes. Whilst spherical oocytes with normal zona and perivitelline space were characterized as being more likely to form a blastocyst. Limitations, reasons for caution This single-clinic study is retrospective. A multi-center study is underway. External factors affecting blastulation must be accounted for to avoid mislabeling of good oocytes as non-viable. There is also a need to understand oocyte dysmorphias identified by the AI algorithm to ensure biological transparency in clinical decision making. Wider implications of the findings Taking into account clinical and gamete confounders when building AI algorithms is a necessary strategy to ensure AI algorithms are generalized when incorporated into clinical practice, whilst reducing bias and promoting transparency in clinical decision making. The risk of not considering confounders leads to mislabeling, bias and inaccurate predictions. Trial registration number not applicable

P-310 Bringing Transparency to Oocyte Assessment: the importance of including confounders when building Artificial Intelligence (AI) based support tools to quantify oocyte viability

Abstract Study question Which confounders (sperm quality, oocyte dysmorphism, culture time, images pre or post-ICSI, age) affect the ability of AI to predict blastulation based on oocyte images? Summary answer Sperm quality, oocyte dysmorphism, pre or post-ICSI image should be controlled for when building AI algorithms to predict blastulation based on oocyte images. What is known already Previous studies reporting on the use of AI to predict blastulation based on oocyte images have: (i) not accounted for confounders affecting blastulation (i.e. sperm quality, culture time), and (ii) used post-ICSI images; without assessing whether the ICSI procedure affects the oocyte image as assessed by AI. Therefore, there is a risk of mislabeling viable oocytes as non-viable due to external factors, which could cause uncontrolled bias and failure to generalize when used in clinical practice. The objective was to assess how these confounders affect efficacy of prediction of blastulation from oocyte images by an AI-based oocyte assessment tool: CHLOE-OQ(Fairtility). Study design, size, duration Cohort study. Images of 1281 oocytes (February to June 2022) were taken pre and post ICSI using the Embryoscope, and the embryos cultured until day 7. Oocyte donor source and age, oocyte dysmorphias and sperm quality were documented. CHLOE-OQ algorithm was trained, validated and tested in a diverse data set, accounting for pre and post ICSI image datasets, quality of oocytes, quality of sperm and patient age. Participants/materials, setting, methods The primary endpoint was blastulation. Sperm quality data was classified into 4 groups: (A)All (n = 1281), (B)donor sperm only (n = 51), (C)donor sperm and normospermic samples from men not diagnosed with male factor infertility (n = 557), (D)abnormal sperm samples and other diagnosed male factor cycles (n = 747). Eggs were classified by source (own/donor), and by dysmorphisms: enlarged perivitelline space, abnormal Zona pellucida, cytoplasmic abnormalities, dark, enlarged oocytes. Main results and the role of chance Post-ICSI images had higher mean CHLOE-OQ score than pre ICSI images (0.28 ± 0.1 vs 0.33 ± 0.1, p < 0.001, paired t-test). Discrepancies were particularly identified in oocytes that degenerated following ICSI, and scored 0 by CHLOE-OQ despite having higher scores pre-ICSI. Using Post-ICSI images (AUC = 0.66, 95% confidence interval, CI: 0.63-0.69, n = 1281) improved the efficacy of prediction of blastulation compared to pre-ICSI images (AUC = 0.57: 0.53-0.60, n = 1281, p < 0.001), suggesting that ICSI affected the quality of the oocyte, and how an oocyte responds to ICSI, as assessed by AI, contributes to prediction of blastulation. Efficacy of prediction (AUC) was not affected by the quality of the sperm: (A-OVERALL 0.658 [CI(95%): 0.626-0.687]; B-Donor 0.586 [CI(95%): 0.449-0.728]; C-normospermic 0.645 [CI(95%): 0.600-0.688], D male factor 0.678 [CI(95%): 0.639-0.715]). Oocyte features associated with low CHLOE-OQ scores were: enlarged perivitelline space, dysmorphic oocytes, abnormal Zona pellucida, cytoplasmic abnormalities and dark and enlarged oocytes. Whilst spherical oocytes with normal zona and perivitelline space were characterized as being more likely to form a blastocyst. Limitations, reasons for caution This single-clinic study is retrospective. A multi-center study is underway. External factors affecting blastulation must be accounted for to avoid mislabeling of good oocytes as non-viable. There is also a need to understand oocyte dysmorphias identified by the AI algorithm to ensure biological transparency in clinical decision making. Wider implications of the findings Taking into account clinical and gamete confounders when building AI algorithms is a necessary strategy to ensure AI algorithms are generalized when incorporated into clinical practice, whilst reducing bias and promoting transparency in clinical decision making. The risk of not considering confounders leads to mislabeling, bias and inaccurate predictions. Trial registration number not applicable

Gradient-Based Simulation Optimization Algorithms via Multi-Resolution System Approximations

We propose gradient-based simulation-optimization algorithms to optimize systems that have complicated stochastic structure. The presence of complicated stochastic structure, such as the involvement of infinite-dimensional continuous-time stochastic processes, may cause the exact simulation of the system to be costly or even impossible. On the other hand, for a complicated system, one can sometimes construct a sequence of approximations at different resolutions, where the sequence has finer and finer approximation resolution but higher and higher cost to simulate. With the goal of optimizing the complicated system, we propose algorithms that strategically use the approximations with increasing resolution and higher simulation cost to construct stochastic gradients and perform gradient search in the decision space. To accommodate scenarios where approximations cause discontinuities and lead path-wise gradient estimators to have an uncontrollable bias, stochastic gradients for the proposed algorithms are constructed through finite difference. As a theory support, we prove algorithm convergence, convergence rate, and optimality of algorithm design under the assumption that the objective function for the complicated system is strongly convex, whereas no such assumptions are imposed on the approximations of the complicated system. We then present a multilevel version of the proposed algorithms to further improve convergence rates, when in addition the sequence of approximations can be naturally coupled. History: Accepted by Bruno Tuffin, Area Editor for Simulation. Supplemental Material: The software that supports the findings of this study is available within the paper and its Supplemental Information ( https://pubsonline.informs.org/doi/suppl/10.1287/ijoc.2023.1279 ) as well as from the IJOC GitHub software repository ( https://github.com/INFORMSJoC/2021.0289 ) at ( http://dx.doi.org/10.5281/zenodo.7485443 ).

Cardiovascular events and all-cause mortality in patients with chronic obstructive pulmonary disease using olodaterol and other long-acting beta2-agonists.

PurposeWe examined the effect of olodaterol on the risk of myocardial ischaemia, cardiac arrhythmia, and all‐cause mortality compared with use of other long‐acting beta2‐agonists (LABAs). Channelling bias was also explored.MethodsThis Danish population‐based cohort study used data linked from registries of hospital diagnoses, outpatient dispensings, and deaths. It included patients (aged ≥40 years) with a diagnosis of chronic obstructive pulmonary disease (COPD) who initiated olodaterol or another LABA. Using matching and propensity score (PS) stratification, we calculated adjusted incidence rate ratios (IRRs) using Poisson regression, followed by several additional analyses to evaluate and control channelling bias.ResultsThe IRRs of cardiac arrhythmias or myocardial ischaemia among users of olodaterol (n = 14 239) compared to users of other LABAs (n = 51 167) ranged from 0.96 to 1.65 in various analyses, although some estimates had low precision. Initial analysis suggested an increased risk for death with olodaterol compared with other LABAs (IRR, 1.63; 95% CI, 1.44–1.84). Because olodaterol prescribing was associated with COPD severity, the mortality association was attenuated by using different methods of tighter confounding control: the IRRs were 1.26 (95% CI, 0.97–1.64) among LABA‐naïve LABA/LAMA users without recent COPD hospitalisation; 1.27 (95% CI, 1.03–1.57) in a population with additional trimming from the tails of the PS distribution; and 1.32 (95% CI, 1.19–1.48) after applying overlap‐weights analysis.ConclusionsOlodaterol users had a similar risk for cardiac arrhythmias or myocardial ischaemia as other LABA users. The observed excess all‐cause mortality associated with olodaterol use could be due to uncontrolled channelling bias.

Collective Housing of Mice of Different Age Groups before Maturity Affects Mouse Behavior.

Background Although population housing is recommended by many animal management and ethical guidelines, the effect of collective housing of mice of different age groups on mouse behavior has not been clarified. Since the development of the central nervous system continues to occur before sexual maturation, the stress of social ranking formation among male individuals in mixed housing conditions can affect postmaturation behavior. To assess these effects, sexually immature mice of different ages were housed in the same cage and a series of behavioral tests were performed after maturation. Results The findings for three groups of mice—junior mice housed with older mice, senior mice housed with younger mice, and mice housed with other mice of the same age—were compared. Junior mice showed higher body weight and activity as well as lower grip strength and anxiety-like behaviors than other mice. In contrast, senior mice showed lower body temperature and increased aggression, antinociceptive effect, and home-cage activity in the dark period in comparison with other mice. Conclusions Thus, combined housing of immature mice of different age groups affects mouse behavior after maturation. Appropriate prematuration housing conditions are crucial to eliminate the uncontrollable bias caused by age-related social stratification.

2-D nanometer thickness mapping applying a reduced bias soft X-ray NEXAFS approach.

We present a 2-D mapping of a sample thickness with nanometer accuracy employing a compact arrangement of near-edge X-ray absorption fine structure (NEXAFS) technique. A NEXAFS spectrum coupled with a scanning system was used to generate a 2-D thickness map of the TiO2 sample (anatase form) deposited on the top of a SiN membrane. The thickness values were retrieved from the experimental data by applying different methods of data processing. In the paper, the detailed analysis of the data processing methods and the identified sources of the errors show that the proposed procedure based on averaging two imperfect estimates reduces the error caused by the uncontrolled bias of the measured signals. This procedure was termed as the average one. The estimates from the proposed average approach and the standard absorption-jump ratio in the absorption edge vicinity were compared with the direct results obtained by applying scanning electron microscopy (SEM). The experimental arrangement of the NEXAFS spectroscopy system, the data acquisition method, as well as the possible error sources, are presented and discussed in detail.

Grid-based diffusion Monte Carlo for fermions without the fixed-node approximation.

A diffusion Monte Carlo algorithm is introduced that can determine the correct nodal structure of the wave function of a few-fermion system and its ground-state energy without an uncontrolled bias. This is achieved by confining signed random walkers to the points of a uniform infinite spatial grid, allowing them to meet and annihilate one another to establish the nodal structure without the fixed-node approximation. An imaginary-time propagator is derived rigorously from a discretized Hamiltonian, governing a non-Gaussian, sign-flipping, branching, and mutually annihilating random walk of particles. The accuracy of the resulting stochastic representations of a fermion wave function is limited only by the grid and imaginary-time resolutions and can be improved in a controlled manner. The method is tested for a series of model problems including fermions in a harmonic trap as well as the He atom in its singlet or triplet ground state. For the latter case, the energies approach from above with increasing grid resolution and converge within 0.015E_{h} of the exact basis-set-limit value for the grid spacing of 0.08 a.u. with a statistical uncertainty of 10^{-5}E_{h} without an importance sampling or Jastrow factor.

Cancer immunology and radiobiology: Oliver Scott's struggle for the perfect tumour model in translational research.

Oliver Scott is best known for his research into the role of tumour hypoxia in radiation oncology. Yet no less important were Oliver's activities in the development of concepts and methods for performing translational research on the effect of ionising radiation on tumour in experimental animals, stressing the importance of using strictly inbred animals for transplantation of tumours which had arisen in exactly the identical mouse strain. Otherwise residual immunity would lead to uncontrollable bias in the results of cure experiments, invalidating conclusions. These pioneering views are no less valid in today's cancer research.

Should benzodiazepines be avoided?

Benzodiazepines (BZDs) have been on the market primarily as anxiolytics and hypnotics since 1960 where they – being much safer – gradually replaced the far more toxic drugs such as barbiturates, meprobamate, and chloral hydrate. However, after some time, it was realized that some patients developed addiction and tolerance to BZD, and during the last decades, the regulation of BZD has been rather tight. This has been followed by a marked decrease in prescriptions of BZD to patients. At the same time, studies on potential health hazards associated with BZD emerged. Thus, BZD has been associated with increased mortality and risk of traffic accidents, falls, hip fractures, cancer, cognitive decline, and dementia, but most studies have been hampered by inappropriate adjustment for factors associated with the selection of patient into treatment 1. Due to the cognitive disturbances associated with BZD, additional studies of the potential long-term effect on cognition and dementia seem relevant. In the present issue of Acta, a timely, well-performed, register-based nested case–control study finds that BZD use was associated with modestly increased odds of Alzheimer's disease 2 with no clear dose–response relationship or difference between different types of BZD. The authors do take precautions though by assuming there was ‘no uncontrolled bias or confounding’. Thus, they introduce a time lag of 5 years to reduce the possibility of reverse causality (protopathic bias). Also, they reduce indication bias by accounting for ‘combination BZD that are indicated for a larger variety of symptoms (such as gastrointestinal cramps, dizziness and muscle tension) than traditional BZDs’. Furthermore, the authors adjust for socioeconomic position, several comorbidities, and other medication. However, to limit residual confounding, it is important that covariates are included with sufficient detail 3 and it is puzzling why the authors lump all psychiatric diagnosis in one category instead of a more detailed adjustment for diseases strongly associated with BZD and dementia such as depression and anxiety 4, 5. The lack of dose–response relation is also surprising given the large range of exposure from those who once purchased BZD to heavy long-term users. This broad definition of exposure is also reflected by a very high number of exposed among both cases (40%) and controls (37%). More than 50 years ago, Hill proposed nine criteria for causation, which are still considered to be relevant and applicable to pharmacoepidemiology 6. The criteria include strength, consistency, specificity, temporality, biological gradient, plausibility, coherence, experimental evidence, and analogy. If these criteria are applied on the existing evidence of an effect of BZD on most of the above health outcomes including dementia, none of them seems to be fulfilled. Thus, one may wonder what leads to the final statement in the paper: ‘BZDRs should be avoided if possible and threshold for prescribing should be high enough due to overall adverse effect profile of the BZDRs’. While this may be well advised if given to users and their doctors, it will not be due to risk of dementia but probably more due to risk of habituation. And if the advice is directed to the health authorities, as a further encouragement to chase BZD using patients and their doctors, they might be barking up the wrong tree. Most guidelines recommend use of BZD drugs for only a few weeks. However, most anxiety disorders, sleep disorders, and psychiatric disorders with symptom-anxiety (or as residual symptom following first-line treatments) last longer. Comforting then to know that, by far, most patients who are prescribed BZD are short-term or intermittent users 7. Also, even though habituation can be a problem in some patients and real misuse in much fewer, not all patients experience withdrawal when quitting. Further, the statement that BZD does not have long-term effect is not based on long-term studies. Rather it is based on lack of them. In fact, when studied long term, they seem to work long term 8, 9. Even if the relation between BZD and dementia shown in the present paper is causal and we use the risk estimates to calculate the population attributable risk, for example, for the Danish population, it will show that <0.1% of dementia cases would be prevented if BZD was eliminated, whereas the corresponding risks for low education, hearing loss, smoking, and depression have been estimated to be 9, 5, 8, and 4% respectively 4. None.

Double blinded: the uncontrolled bias.

Double blinded: the uncontrolled bias.

Comparison of soft tissue balancing, femoral component rotation, and joint line change between the gap balancing and measured resection techniques in primary total knee arthroplasty: A meta-analysis.

This meta-analysis was designed to compare the accuracy of soft tissue balancing and femoral component rotation as well as change in joint line positions, between the measured resection and gap balancing techniques in primary total knee arthroplasty. Studies were included in the meta-analysis if they compared soft tissue balancing and/or radiologic outcomes in patients who underwent total knee arthroplasty with the gap balancing and measured resection techniques. Comparisons included differences in flexion/extension, medial/lateral flexion, and medial/lateral extension gaps (LEGs), femoral component rotation, and change in joint line positions. Finally, 8 studies identified via electronic (MEDLINE, EMBASE, and the Cochrane Library) and manual searches were included. All 8 studies showed a low risk of selection bias and provided detailed demographic data. There was some inherent heterogeneity due to uncontrolled bias, because all included studies were observational comparison studies. The pooled mean difference in gap differences between the gap balancing and measured resection techniques did not differ significantly (-0.09 mm, 95% confidence interval [CI]: -0.40 to +0.21 mm; P = 0.55), except that the medial/LEG difference was 0.58 mm greater for measured resection than gap balancing (95% CI: -1.01 to -0.15 mm; P = 0.008). Conversely, the pooled mean difference in femoral component external rotation (0.77°, 95% CI: 0.18° to 1.35°; P = 0.01) and joint line change (1.17 mm, 95% CI: 0.82 to 1.52 mm; P < 0.001) were significantly greater for the gap balancing than the measured resection technique. The gap balancing and measured resection techniques showed similar soft tissue balancing, except for medial/LEG difference. However, the femoral component was more externally rotated and the joint line was more elevated with gap balancing than measured resection. These differences were minimal (around 1 mm or 1°) and therefore may have little effect on the biomechanics of the knee joint. This suggests that the gap balancing and measured resection techniques are not mutually exclusive.

FTA Effects on Agricultural Trade with Matching Approaches

Abstract While the trade effect of free trade agreements (FTAs) is a global issue, little research has examined the economic effects of trade liberalization on agricultural products with robust empirical methods. In this study, propensity score matching for controlling selection bias is used to examine and analyze the effect of FTAs on the trade of South Korea’s agricultural products. To enhance the robustness of the estimated results, differences between the FTA treatment effects in 2010 and 2012 are analyzed. The results reveal that the effect of FTAs on agricultural trade varies slightly, depending on the matching approach used; however, the signs of all estimated average treatment effects on the treated (ATT) values are positive. Analysis of the difference between selection bias controlled through matching and uncontrolled selection bias shows that the value of the average treatment effect (ATE) with uncontrolled bias is greater than the ATT estimate calculated through matching. This implies that controlled versus uncontrolled selection bias can result in different ATE and ATT estimates, and that previous studies on FTA trade effects have overestimated the effect, because selection bias was not fully addressed.

PRM277 - Moving Beyond the Picos: Appropriate Comparative Observational Data Selection Can Facilitate Meta-Analysis of Relative Treatment Effects

Evidence synthesis based on the quantitative analysis (indirect treatment comparisons ITC) of data from randomized controlled trials (RCTs) is the gold standard in healthcare decision making. However, when ITC of RCTs may not be practical or possible, comparative observational data may provide a valuable alternative despite potential challenges with internal validity. The present project aims to develop a conceptual framework from which valid quantitative assessment of real-world evidence (RWE) can provide necessary insights into treatment effects outside the RCT settings. Considerations of quantitative evidence synthesis in RWE were identified using published best-practice guidelines for quantitative analysis of RCT data and applied to two test cases, atrial fibrillation (AF) treated with novel oral anticoagulant (NOAC) and type II diabetes treated with antihyperglycemic agents. Data sources of RWE, sample selection, outcome definitions, and statistical method variation were assessed. Significant variations across all key factors were observed in studies for both indications. In the case of NOAC treated AF, homogeneity of the data source and geographic location did not guarantee appropriate quantitative comparison. Several factors unique to assessing observational database study comparability for meta-analyses were identified including a need for a greater focus on selection bias inherent to the type and location of the database and increased sensitivity of estimates to the index dates particularly in relation to local and international product launches, as well various outcome definition and reporting complexities. Valid ITC could be performed using RWE from comparative observational studies. Careful selection of studies is required to mitigate factors that would otherwise compromise results through uncontrolled bias. Researchers need to go beyond the obvious individual study reporting biases and fully assess the complexities of comparative data sources to be able to appropriately select studies for ITC.

Childhood traumatic experiences of patients with bipolar disorder type I and type II

BackgroundChildhood trauma is an important environmental stressor associated with bipolar disorders (BD). It is still not clear if it is differently distributed between BD I and BD II. Therefore, the aim of this research was to investigate the distribution patterns of childhood trauma in BD I and BD II. In this perspective, we also studied the relationship between childhood trauma and suicidality. MethodsWe assessed 104 outpatients diagnosed with BD I (n=58) or BD II (n=46) according to DSM-IV-TR criteria and 103 healthy controls (HC) matched for age, sex and education level. History of childhood trauma was obtained using the Childhood Trauma Questionnaire (CTQ). ResultsAll patients with BD had had more severe traumatic childhood experiences than HC. Both BD I and BD II patients differed significantly from HC for trauma summary score and emotional abuse. BD I patients differed significantly from HC for sexual abuse, and BD II differed from HC for emotional neglect. BD I and BD II did not significantly differ for any type of trauma. Suicide attempts were linked to both emotional and sexual abuse in BD I and only to emotional abuse in BD II. Emotional abuse was an independent predictor of lifetime suicide attempts in BD patients. LimitationsThe reliability of the retrospective assessment of childhood trauma experiences with the CTQ during adulthood may be influenced by uncontrolled recall bias. ConclusionsThe assessment of childhood trauma, which has great clinical importance because of its strong link with suicidality, can unveil slight differences between BD subtypes and HC.

ARID1A Is Recurrently and Significantly Mutated in Follicular Lymphoma (FL) and Impairs DNA Repair Efficiency

FL is the most common indolent nodal lymphoma and considered incurable for the majority of patients. Virtually all patients eventually develop resistant disease, and up to 45% histologic transformation (tFL) with poor outcome. Alterations associated with progression to tFL include increased copy number variations (e.g. CDKN2A deletions, MYC amplification), acquisition of additional translocations (e.g. MYC or BCL6), and higher mutation load (including TP53). However the underlying molecular mechanisms promoting this genomic instability remain elusive. Interestingly, histologic transformation occurs at a remarkably constant rate of ~3% per year implying a rather stochastic process.We previously described convergent evolution for disruptive ARID1A mutations in FLs of a donor and the recipient patient following an allogeneic bone marrow transplantation (Weigert, Canc Discovery 2012). The independent acquisition of ARID1Aloss in two separate hosts after transplantation of an identical FL ancestor clone suggested (i) a later acquired event in the molecular ontogeny and (ii) functional relevance.ARID1A promotes the formation of SWI/SNF nucleosome remodeling complexes containing BRG1 or BRM, which catalyze disruption of DNA-histone contacts, thereby controlling chromatin condensation and DNA accessibility. Nucleosome-remodelling has been linked to DNA damage response, including nucleotide and base excision repair, homologous recombination and non-homologous end-joining (NHEJ). We thus hypothesized that ARID1Ahaplodeficiency might contribute to genetic and genomic instability in FL.In a cohort of 305 FL specimens, we identified 44 ARID1A mutations in 43 patients (14%). The majority of these mutations were disruptive (66%; 19 nonsense and 10 frame shift mutations), 6 cases harbored splice site mutations. MutSigCV analysis (Lawrence, Nature 2013) indicated that ARID1A was significantly mutated in FL, i.e. more often than expected by chance given background mutation processes. Consistent with our hypothesis, the number of non-synonymous mutations in 104 targeted genes was higher in ARID1A mutated FLs (5.6+0.34, mean+sem) compared to ARID1A wild-type (wt) FLs (4.7+0.12; p=0.0099), and similar compared to 17 FLs with TP53 mutations (5.6+0.12). Of note, ARID1A and TP53 mutations were mutually exclusive.For functional studies, we confirmed deleterious ARID1A alterations in 6 lymphoma cell lines (Namalwa, Karpas422, SUPHD-1, SU-DHL5, WSU-FSCCL, U-OH1). Immunoblotting showed decreased ARID1A protein expression in all 6 lines compared to 5 ARID1A wt lymphomas. Tet-induced re-expression of ARID1A in two tested ARID1A mutant cell lines (Namalwa and SUP-HD1) reduced cell proliferation compared to tet-induced control (LacZ), but had no effect in ARID1A wt OCI-Ly1.As a proof-of-principle experiment we assayed the impact of ARID1A loss on DNA double strand break (DSB) repair efficiency in genetically well defined mouse embryo fibroblasts (MEFs), avoiding uncontrolled bias introduced by the genetic and genomic complexity in lymphoma cell lines. Early passage MEFs from either wt or conditional ARID1A knock-out mice (ARID1Af/f; Gao, PNAS 2008) were transiently transfected or retrovirally transduced with self-excising Cre-recombinase (Silver&Livingston, Mol Cell 2001) or control, and irradiated with 2 Gy after 36 hrs. Cre-induced ARID1A loss was confirmed by Western blot. To assay NHEJ, the predominate DSB-repair pathway used by cells in G1 phase, we stained cells for γH2A.X and 53BP1. In 3 independent and blinded experiments, evaluation of >100 cells per condition demonstrated significantly more double-positive cells (>6 foci/cell) in ARID1A-/- MEFs (43+14%) compared to both ARID1Af/fMEFs (9+1%) and Cre-exposed wt MEFs (15+5%) at 16 hrs post irradiation, whereas there was no significant difference for ARID1Af/for wt MEFs with or without Cre at baseline (all <10%), and 1 hr post irradiation (all >50%). This data indicates that ARID1A loss slows the repair kinetics of NHEJ and delays the clearance of DSB.We conclude that ARID1A is recurrently and significantly mutated in FL. Most mutations are disruptive leading to functionally relevant ARID1A protein haplodeficiency. ARID1A loss impairs NHEJ, which might sensitize tumors to DNA damaging agents and irradiation, but also contribute to genetic and genomic instability promoting histologic transformation. DisclosuresNo relevant conflicts of interest to declare.

A critical assessment of the HGCA grain sampling guide

Author Summary: HGCA’s grain sampling guide is assessed with respect to the principles for representative sampling as set forward in the Theory of Sampling (TOS). Sampling correctness, which requires the elimination of all Incorrect Sampling Errors (ISE), constitutes the only guarantee for valid, representative grain quality control; presence of ISEs causes a varying, uncontrollable sampling bias that cannot be corrected for. Contrary to a first superficial observation (“grain is grain”), many different species and varieties, as well as differences caused by soil types, availability of local nutrients, make “grain” a significantly heterogeneous commodity, which requires special attention when sampled at various process locations (from harvesting, storage until commercial intake). The present appraisal shows that most of the respected HGCA grain guide’s recommendations do not comply with TOS principles of sampling correctness. The suggested sampling procedures constitute major error potentials, which strongly compromise sample representativity.

Developing templates for uniform data documentation and reporting in critical care using a modified nominal group technique.

BackgroundClinical practice in trauma and critical care is predominantly derived from quantitative observational cohort studies based on data retrospectively collected from medical records. Such data create uncontrolled bias and influence external and internal validity, thereby hindering systematic reviews. Templates or standards for uniform documenting and scientific reporting may result in high quality and internationally standardised data being collected on a regular basis, enhance large international multi-centre studies, and increase the quality of evidence. Templates or standards may be developed using multidisciplinary expert panel consensus methods.We present three consensus processes aimed at developing templates for documenting and scientific reporting. We discuss the advantages, limitations, and possible future improvements of our method.MethodsThe template preparation was based on expert panel consensus derived through a modified nominal group technique (NGT) method that combined the traditional Delphi method with the traditional NGT method in a four-step process.ResultsStandard templates for documenting and scientific reporting were developed for major trauma, pre-hospital advanced airway handling, and physician-staffed pre-hospital EMS. All templates were published in scientific journals.ConclusionOur modified NGT consensus method can successfully be used to establish templates for reporting trauma and critical care data. When used in a structured manner, the method uses recognised experts to achieve consensus, but based on our experiences, we recommend the consensus process to be followed by feasibility, reliability, and validity testing.

Eliminating experimental bias in anisotropic-flow measurements of high-energy nuclear collisions

We argue that the traditional event-plane method, which is still widely used to analyze anisotropic flow in ultrarelativistic heavy-ion collisions, should be abandoned because flow fluctuations introduce an uncontrolled bias in the measurement. Instead, one should use an alternative, such as the scalar-product method or cumulant method, which always measures an unambiguous property of the underlying anisotropic flow and therefore eliminates this bias, and does so without any disadvantages. It is known that this correction is important for precision comparisons of traditional v_n measurements requiring better than a few percent accuracy. However, we show that it is absolutely essential for correlations between different harmonics, such as those that have been recently measured by the ATLAS Collaboration, which can differ from the nominally-measured quantity by a factor two or more. We also describe how, using the corrected analysis method, the information from different subevents can be combined in order to optimize the precision of analyses.