Trends and current spectrum of contact allergy in Central Europe: results of the Information Network of Departments of Dermatology (IVDK) 2007-2018.

BACKGROUND: Changing patterns of exposure lead to changes in the spectrum of contact allergy in the general population and in patients patch tested for suspected contact allergy. The main contact allergens (haptens) are assembled in the so-called 'baseline series'. OBJECTIVES: To present the current spectrum of contact allergy to baseline series allergens, and its temporal development over the last 12 years, based on data collected by the 56 active departments of the trinational Information Network of Departments of Dermatology (IVDK). METHODS: Patch-test data, along with core demographic and clinical information, subdivided into 4-year periods for all patients patch tested with the baseline series in the IVDK, were analysed descriptively. Moreover, log-binomial regression analysis addressed temporal trends of specific contact allergies, adjusted for potentially confounding factors. RESULTS: Among the 125 436 patients tested with the German baseline series, the most common allergens were nickel (14·7%), fragrance mix I (8·1%), Myroxylon pereirae resin (7·5%) and cobalt (5·2%), with no conclusive trend. The rise and fall of contact allergy to methylchloroisothiazolinone/methylisothiazolinone (MI), following (self-) regulation in the European Union, reflected the MI contact allergy epidemic. Propolis showed a marked upward trend with a prevalence of 3·94% during the period 2015-2018. CONCLUSIONS: Decreases in sensitization prevalence likely reflect reduced exposure, with some lag, as seen with hydroxyisohexyl 3-cyclohexene carboxaldehyde. If no (sufficient) decrease can be observed despite interventions, such as for nickel and chromium, affected subgroups should be identified and their causative exposures explored. Finally, increases such as that observed with propolis, certainly warrant targeted investigation of the exposures driving sensitization, and possibly intervention. What is already known about this topic? Contact allergy is prevalent in the general population. Surveillance based on clinical data offers timely information on trends concerning certain allergens or subgroups at risk. What does this study add? This analysis provides an update of the current contact allergy prevalence in Central Europe, including time trends. This study describes the demographic and clinical profile of patients tested from 2007 to 2018. This research identifies contact allergens that require (further) preventive efforts. Linked Comment: Schuttelaar. Br J Dermatol 2020; 183:800-801.

Memorandum "Open Metadata". Open Access to Documentation Forms and Item Catalogs in Healthcare.

At present, most documentation forms and item catalogs in healthcare are not accessible to the public. This applies to assessment forms of routine patient care as well as case report forms (CRFs) of clinical and epidemiological studies. On behalf of the German chairs for Medical Informatics, Biometry and Epidemiology six recommendations to developers and users of documentation forms in healthcare were developed. Open access to medical documentation forms could substantially improve information systems in healthcare and medical research networks. Therefore these forms should be made available to the scientific community, their use should not be unduly restricted, they should be published in a sustainable way using international standards and sources of documentation forms should be referenced in scientific publications.

The evolution of boosting algorithms. From machine learning to statistical modelling.

BACKGROUND: The concept of boosting emerged from the field of machine learning. The basic idea is to boost the accuracy of a weak classifying tool by combining various instances into a more accurate prediction. This general concept was later adapted to the field of statistical modelling. Nowadays, boosting algorithms are often applied to estimate and select predictor effects in statistical regression models. OBJECTIVES: This review article attempts to highlight the evolution of boosting algorithms from machine learning to statistical modelling. METHODS: We describe the AdaBoost algorithm for classification as well as the two most prominent statistical boosting approaches, gradient boosting and likelihood-based boosting for statistical modelling. We highlight the methodological background and present the most common software implementations. RESULTS: Although gradient boosting and likelihood-based boosting are typically treated separately in the literature, they share the same methodological roots and follow the same fundamental concepts. Compared to the initial machine learning algorithms, which must be seen as black-box prediction schemes, they result in statistical models with a straight-forward interpretation. CONCLUSIONS: Statistical boosting algorithms have gained substantial interest during the last decade and offer a variety of options to address important research questions in modern biomedicine.

Extending statistical boosting. An overview of recent methodological developments.

BACKGROUND: Boosting algorithms to simultaneously estimate and select predictor effects in statistical models have gained substantial interest during the last decade. OBJECTIVES: This review highlights recent methodological developments regarding boosting algorithms for statistical modelling especially focusing on topics relevant for biomedical research. METHODS: We suggest a unified framework for gradient boosting and likelihood-based boosting (statistical boosting) which have been addressed separately in the literature up to now. RESULTS: The methodological developments on statistical boosting during the last ten years can be grouped into three different lines of research: i) efforts to ensure variable selection leading to sparser models, ii) developments regarding different types of predictor effects and how to choose them, iii) approaches to extend the statistical boosting framework to new regression settings. CONCLUSIONS: Statistical boosting algorithms have been adapted to carry out unbiased variable selection and automated model choice during the fitting process and can nowadays be applied in almost any regression setting in combination with a large amount of different types of predictor effects.

Risk factors of polysensitization to contact allergens.

BACKGROUND: 'Polysensitization' (PS) is usually defined as contact sensitization to three or more unrelated haptens of the baseline patch test series. Despite PS being an important clinical phenotype indicating increased susceptibility to contact allergy, statistical approaches to analyse PS have hitherto been preliminary. OBJECTIVES: To apply an appropriate regression model for count data, namely, negative binomial hurdle regression, to a large set of clinical patch test data with the aim of estimating PS risk in more detail than previously achieved. METHODS: The detailed information provided by the hurdle model includes a separate estimation of an 'increment factor' quantifying the likelihood of further positive reactions, i.e. PS. Clinical data of 126 878 patients patch tested by departments comprising the IVDK (Information Network of Departments of Dermatology) network (www.ivdk.org) between 1995 and 2010 were included. RESULTS: Regarding anatomical sites as exposure (surrogate), the axillae and the feet were found to be strong PS risk factors. Moreover, age was a strong PS risk factor, and less so, female sex. In comparison, atopic eczema and occupational dermatitis were less important risk factors. Single allergens contributed to PS to a varying extent. CONCLUSIONS: The data presented point to some, very likely exposure-related, risk factors which need to be considered in future PS research, e.g. addressing the genetic basis for PS.

Assessing skin pigmentation in epidemiological studies: the reliability of measurements under different conditions.

BACKGROUND/PURPOSE: Different devices are used for the non-invasive measurement of (constitutive) skin pigmentation in (epidemiological) studies. Reproducibility of measurements with the Chromameter (CM, model CR 300, Minolta, Osaka) and the Reflectometer (RM, Courage & Khazaka, Cologne) has not yet been examined in detail and was addressed in a set of four experiments and studies respectively. METHODS: Regarding the CM, the Y value of the Yxy CIE 1931 colour system was utilized, representing lightness in this colour space, while the RM measured reflectance at 660 nm with a small bandwidth of 20 nm. Both devices measure reflectance on a scale from 0 to 100%, however, in different wavelength ranges. Between 3 and 20 repetitions were performed on standard grey scales and different sets of human volunteers, including RM measurements in a large epidemiological field study. RESULTS: While the coefficient of variation (V) increases and the intraclass correlation coefficient decreases from controlled laboratory to field conditions, reproducibility remained in a range considered acceptable, if adequate study conditions were maintained. In a direct comparison on human skin, V of the RM was significantly smaller than that of the CM. CONCLUSION: Both devices can be used confidently in field studies; however, based on considerations of skin optics and in view of slightly lesser variability, the RM may be preferable.

A comparison of patterns of sun protection during beach holidays and everyday outdoor activities in a population sample of young German children.

BACKGROUND: Reducing exposure to ultraviolet (UV) radiation is the main effective measure for preventing skin cancer. Educational campaigns targeting sun protection have been focused either on behaviour on the beach during the summer holiday alone, or during everyday outdoor activities of the children. Little is known about the comparison between these different settings. OBJECTIVES: To analyse whether parents apply similar protective measures to reduce UV exposure for their young children in different outdoor environments. METHODS: Families (n = 2619) with children aged 3-6 years (response: 64·7%) were enrolled in a population-based survey in the German city of Erlangen and its surrounding rural county. Using a self-administered standardized questionnaire parents gave information about demographic and photosensitivity data of their children, their knowledge about risk factors for skin cancer and their typical instructions given to their children when these played outside on a summer day in different outdoor environments. RESULTS: Significant discrepancies regarding the four UV protective measures (clothes, shade, sunhat, sunscreen) for children between an everyday outdoor setting and a holiday setting on the beach were observed. A high level of parental risk factor knowledge was significantly associated with a better protection for children in all four measures only on the beach. Photosensitivity and demographic characteristics had some impact on protective behaviour, too. Measures of sun protection were reduced with children's increasing age. CONCLUSIONS: Skin cancer prevention campaigns should target the encouragement of sun protection for children also in outdoor activities of daily living, not only during a summer holiday on the beach.

The birth and evolution of a discipline devoted to information in biomedicine and health care. As reflected in its longest running journal.

BACKGROUND: The journal Methods of Information in Medicine, founded in 1962, has now completed its 50th volume. Its publications during the last five decades reflect the formation of a discipline that deals with information in biomedicine and health care. OBJECTIVES: To report about 1) the journal's origin, 2) the individuals who have significantly contributed to it, 3) trends in the journal's aims and scope, 4) influential papers and 5) major topics published in Methods over the years. METHODS: Methods included analysing the correspondence and journal issues in the archives of the editorial office and of the publisher, citation analysis using the ISI and Scopus databases, and analysing the articles' Medical Subject Headings (MeSH) in MEDLINE. RESULTS: In the journal's first 50 years 208 editorial board members and/or editors contributed to the journal's development, with most individuals coming from Europe and North America. The median time of service was 11 years. At the time of analysis 2,456 articles had been indexed with MeSH. Topics included computerized systems of various types, informatics methodologies, and topics related to a specific medical domain. Some MeSH topic entries were heavily and regularly represented in each of the journal's five decades (e.g. information systems and medical records), while others were important in a particular decade, but not in other decades (e.g. punched-card systems and systems integration). Seven papers were cited more than 100 times and these also covered a broad range of themes such as knowledge representation, analysis of biomedical data and knowledge, clinical decision support and electronic patient records. CONCLUSIONS: Methods of Information in Medicine is the oldest international journal in biomedical informatics. The journal's development over the last 50 years correlates with the formation of this new discipline. It has and continues to stress the basic methodology and scientific fundamentals of organizing, representing and analysing data, information and knowledge in biomedicine and health care. It has and continues to stimulate multidisciplinary communication on research that is devoted to high-quality, efficient health care, to quality of life and to the progress of biomedicine and the health sciences.

Recombinant human erythropoietin delays loss of gray matter in chronic schizophrenia.

Neurodevelopmental abnormalities together with neurodegenerative processes contribute to schizophrenia, an etiologically heterogeneous, complex disease phenotype that has been difficult to model in animals. The neurodegenerative component of schizophrenia is best documented by magnetic resonance imaging (MRI), demonstrating progressive cortical gray matter loss over time. No treatment exists to counteract this slowly proceeding atrophy. The hematopoietic growth factor erythropoietin (EPO) is neuroprotective in animals. Here, we show by voxel-based morphometry in 32 human subjects in a placebo-controlled study that weekly high-dose EPO for as little as 3 months halts the progressive atrophy in brain areas typically affected in schizophrenia, including hippocampus, amygdala, nucleus accumbens, and several neocortical areas. Specifically, gray matter protection is highly associated with improvement in attention and memory functions. These findings suggest that a neuroprotective strategy is effective against common pathophysiological features of schizophrenic patients, and strongly encourage follow-up studies to optimize EPO treatment dose and duration.

The importance of context information in the diagnostic rating of digital images of patch test reactions.

BACKGROUND: Patch test (PT) guidelines recommend classifying PT reactions based on morphological criteria only, i.e. irrespective of context information such as substance, concentration and vehicle. OBJECTIVES: To analyse reclassification of PT reactions after revelation of context information subsequent to an initial blinded reading. METHODS: One hundred and twenty-two participants (experienced dermatologists and researchers) rated 20 digital images of different PT reactions twice, first blinded, then unblinded regarding substance, concentration and vehicle. Agreement between both ratings was quantified with Cohen's kappa, and systematic differences statistically examined with tests for marginal homogeneity. RESULTS: Mostly, ratings remained stable, e.g. in > 90% of cases of images showing typical strong or extreme positive PT reactions. Reclassification was comparatively often observed in images depicting irritant reactions. Conversely, 16 of 122 participants re-rated doubtful reactions to thiuram mix and dichromate, respectively, as 'allergic' (weak or strong positive) after knowing the substance. CONCLUSIONS: The considerable proportion of participants who choose an 'allergic' rating, despite a morphological presentation definitely not justifying this, points to a conceptual problem of the PT reading scale: the mixing up of morphological classification and interpretation. We therefore suggest amending the scale. Moreover, standardization of PT reading can be improved by continual PT training sessions.

Outdoor work and skin cancer incidence: a registry-based study in Bavaria.

OBJECTIVE: To analyse the association between occupational ultraviolet (UV) light exposure and skin cancer (basal cell carcinoma, BCC; squamous cell carcinoma, SCC; cutaneous malignant melanoma, CMM) based on data from the Bavarian population-based cancer registry. METHODS: The population-based cancer registry of Bavaria (Germany) provided data on incident cases of BCC, SCC, and CMM, respectively, during the period 2001 until 2005. Eleven Bavarian districts with complete skin cancer registration were included in this analysis based on 2,156,336 person years. Cases were assigned to "indoor", "mixed indoor/outdoor", and "outdoor" exposure categories according to their job title. We computed age-specific and age-adjusted incidence rates of BCC (n = 1,641), SCC (n = 499), and CMM (n = 454) by work type, and the relative risk (RR) of skin cancer occurrence for "outdoor" and "mixed indoor/outdoor" workers, respectively, compared to "indoor" workers. RESULTS: The risk of BCC was substantially elevated in male (RR, 2.9; 95% CI, 2.2-3.9) and female (RR, 2.7; 95% CI, 1.8-4.1) outdoor workers compared to male and female indoor workers, respectively. We also found an elevated risk of similar magnitude for SCC in male (RR, 2.5; 95% CI, 1.4-4.7) and female (RR, 3.6; 95% CI, 1.6-8.1) outdoor workers compared to male and female indoor workers, respectively. CMM risk was not significantly associated with outdoor work. CONCLUSION: Our study confirms previous reports on the increased risk of BCC and SCC in outdoor workers compared to indoor workers.

Point and interval estimation of partial attributable risks from case-control data using the R-package 'pARccs'.

The attributable risk (AR) is an epidemiologic measure quantifying the relationship between an exposure factor and a disease at the population level. In addition to its original use as a one-dimensional parameter the AR is increasingly applied in multifactorial epidemiologic situations when the combined impact of multiple exposure factors has to be partitioned into factor-specific components. We discuss the point and interval estimation of the resulting multidimensional parameter termed partial attributable risk (PAR) and introduce the R-package 'pARccs', a comprehensive software enabling the application of the methods. 'pARccs' allows for point and interval estimation of PAR from case-control data utilizing the non-parametric bootstrap with stratified resampling in combination with the percentile or BC(a) method to compute confidence intervals. We illustrate the concept of partial attributable risks and the application of the software by an example from a recent case-control study on risk factors for melanoma. We also discuss practical aspects of the software application for epidemiologic purposes and its limitations.

Double-smoothing in kernel hazard rate estimation.

OBJECTIVES: In oncological studies, the hazard rate can be used to differentiate subgroups of the study population according to their patterns of survival risk over time. Nonparametric curve estimation has been suggested as an exploratory means of revealing such patterns. The decision about the type of smoothing parameter is critical for performance in practice. In this paper, we study data-adaptive smoothing. METHODS: A decade ago, the nearest-neighbor bandwidth was introduced for censored data in survival analysis. It is specified by one parameter, namely the number of nearest neighbors. Bandwidth selection in this setting has rarely been investigated, although the heuristical advantages over the frequently-studied fixed bandwidth are quite obvious. The asymptotical relationship between the fixed and the nearest-neighbor bandwidth can be used to generate novel approaches. RESULTS: We develop a new selection algorithm termed double-smoothing for the nearest-neighbor bandwidth in hazard rate estimation. Our approach uses a finite sample approximation of the asymptotical relationship between the fixed and nearest-neighbor bandwidth. By so doing, we identify the nearest-neighbor bandwidth as an additional smoothing step and achieve further data-adaption after fixed bandwidth smoothing. We illustrate the application of the new algorithm in a clinical study and compare the outcome to the traditional fixed bandwidth result, thus demonstrating the practical performance of the technique. CONCLUSION: The double-smoothing approach enlarges the methodological repertoire for selecting smoothing parameters in nonparametric hazard rate estimation. The slight increase in computational effort is rewarded with a substantial amount of estimation stability, thus demonstrating the benefit of the technique for biostatistical applications.

The impact of meteorological conditions on patch test results with 12 standard series allergens (fragrances, biocides, topical ingredients).

BACKGROUND: Fluctuating irritability of the skin induced by low ambient temperature and humidity may compromise the reproducibility of patch testing. OBJECTIVES: To assess the impact of temperature and absolute humidity at the time of patch testing on the occurrence of irritant or doubtful (IR/?), weak positive (+) and (strong) positive (++/+++) reactions, respectively, among 12 allergens included in the German Standard Series. METHODS: Analysis of clinical data collected in the surveillance network IVDK (http://www.ivdk.org) between January 1993 and December 2001 (n = 73 691 patients) combined with meteorological data obtained by the national services in Germany and Austria. Multinomial logistic regression analysis was used to estimate the risk associated with temperature, absolute humidity and vapour pressure, respectively, adjusted for sex, age, atopic dermatitis and duration of patch test application. RESULTS: For low temperature and humidity, a relevant increase of IR/? reaction frequency was observed in the cases of paraben mix and (chloro-) methylisothiazolinone. Both IR/? and + reactions were significantly increased with respect to the allergens fragrance mix, oil of turpentine, methyldibromo glutaronitrile + phenoxyethanol and particularly formaldehyde, while ++/+++ reactions were hardly affected by weather conditions. CONCLUSIONS: The observed increase of IR/? reactions may be due to epidermal barrier function impairment. The impact of dry/cold weather on + reactions in terms of possibly false-positive reactions is restricted to few allergens. In the case of + reactions of unknown relevance, a re-test under warm conditions or verification tests such as the repeated open application test or the provocative use test may be recommendable.

Flexible modeling of malignant melanoma survival.

OBJECTIVES: This paper compares the diagnostic capabilities of flexible ensemble methods modeling the survival time of melanoma patients in comparison to the well established proportional hazards model. Both a random forest type algorithm for censored data as well as a model combination of the proportional hazards model with recursive partitioning are investigated. METHODS: Benchmark experiments utilizing the integrated Brier score as a measure for goodness of prediction are the basis of the performance assessment for all competing algorithms. For the purpose of comparing regression relationships represented by the models under test, we describe fitted conditional survival functions by a univariate measure derived from the area under the curve. Based on this measure, we adapt a visualization technique useful for the inspection and comparison of model fits. RESULTS: For the data of malignant melanoma patients the predictive performance of the competing models is on par, allowing for a fair comparison of the fitted relationships. Newly introduced MODplots visualize differences in the fitting structure of the underlying models. CONCLUSION: The paper provides a framework for comparing the predictive and diagnostic performance of a parametric, a non-parametric and a combined approach.

Association between split selection instability and predictive error in survival trees.

OBJECTIVES: To evaluate split selection instability in six survival tree algorithms and its relationship with predictive error by means of a bootstrap study. METHODS: We study the following algorithms: logrank statistic with multivariate p-value adjustment without pruning (LR), Kaplan-Meier distance of survival curves (KM), martingale residuals (MR), Poisson regression for censored data (PR), within-node impurity (WI), and exponential log-likelihood loss (XL). With the exception of LR, initial trees are pruned by using split-complexity, and final trees are selected by means of cross-validation. We employ a real dataset from a clinical study of patients with gallbladder stones. The predictive error is evaluated using the integrated Brier score for censored data. The relationship between split selection instability and predictive error is evaluated by means of box-percentile plots, covariate and cutpoint selection entropy, and cutpoint selection coefficients of variation, respectively, in the root node. RESULTS: We found a positive association between covariate selection instability and predictive error in the root node. LR yields the lowest predictive error, while KM and MR yield the highest predictive error. CONCLUSIONS: The predictive error of survival trees is related to split selection instability. Based on the low predictive error of LR, we recommend the use of this algorithm for the construction of survival trees. Unpruned survival trees with multivariate p-value adjustment can perform equally well compared to pruned trees. The analysis of split selection instability can be used to communicate the results of tree-based analyses to clinicians and to support the application of survival trees.