Introduction

The aim of this project was to explore how physicians and dentists in Germany evaluate biostatistics in general and their education in this subject. Furthermore, the importance of the subject for the professional practice of physicians and dentists was determined and insights for teaching during and after medical school were gained.

Method

A total of 2700 physicians and dentists from Schleswig-Holstein were invited by mail to participate in an online survey and to provide sociodemographic data and information on their perception towards the subject of biostatistics in general, in relation to work and to teaching . Data were analyzed with descriptive methods.

Results

Response rate was 13.67%. 50.14% of participants received biostatistical training in medical school. 43.40% and 38.79% of participants reported that biostatistics was useful at study time and for their later career, respectively. Biostatistics was rated as difficult by 58.76%.

93.48% stated biostatistics being a necessary skill for a clinician involved in research, and 93.79% rated it as important for evidence-based medicine. 81.07% agreed that evidence-based medicine is important for clinical practice. Biostatistics was indicated as useful in their own work for evaluating marketing materials from pharmaceutical industry (88.64%), interpreting screening tests (87.88%), reading research publications for general professional interest (85.17%), using research publications to consider new therapeutic options (86.09%), in analyzing data from one's own research (91.89%), but only for 30.18% in clinical contact with patients.

20.00% rated the teaching in biostatistics from their own studies as still useful today. 65.22% would like to understand more about the subject while 86.96% received no further training after graduation. 53.04% of participants said they could do better if they understood more about biostatistics.

Discussion

Our study indicates that the majority of human and dental physicians in Schleswig-Holstein consider biostatistics to be difficult. However, they recognize the value of the subject for evidence-based medicine and research. More than 90%, or nearly one-third, expressed that biostatistics was a necessary skill for clinician scientists or practicing physicians, respectively. Biostatistics was indicated as helpful for a surprisingly large number of physician tasks, so that 13% got training in biostatistics even after graduation.

A large proportion of participants expressed dissatisfaction with biostatistics teaching from undergraduate days. Only one-fifth rated biostatistics teaching from those days as still useful today. Many reported uncertainty about their own biostatistics skills.

Conclusion

The results of this survey show that biostatistics is considered relevant in many areas of physician practice, but that many of the practitioners do not consider themselves well prepared and would like to understand more about biostatistics. Therefore, it seems reasonable to develop further training courses. These should focus precisely on the biostatistical concepts relevant to the medical/dental activities mentioned in the phase of early practical work. We therefore suggest to link the content within the framework of evidence-based medicine with clinically relevant medical topics, thus also increasing the motivation to learn biostatistics.

Here, the issues related to the estimation of parameters in the model of spring and chemical balance weighing designs are presented. These problems are discussed from the point of view of different optimality criteria and different assumptions regard to the errors. We consider the models with uncorrelated errors with different variances and equally correlated errors. The forms of estimators and the necesseary and sufficient conditions determining them are given.

Efficient use of resources is an important aspect of sustainable agriculture systems in future. In agronomy, research is performed to increase water and nutrient use efficiency. There is a range of different definitions of use efficiency that vary in calculation and interpretation in detail. These definitions have in common that use efficiency is defined as the relationship between an investigated input, e.g. water or nutrient, and the output reached by application of the input. The relationship can be calculated in two different ways: The most common approach is to calculate the ratio prior to analysis and submit the ratio to some sort of analysis. The alternative is to perform a regression analysis and estimate the slope in the regression of the output on the input. If use efficiency is calculated relative to a control, both approaches result in identical estimates of use efficiency. However, they vary in the estimated standard error and therefore in the results from significant tests if more than one treatment was compared to the control. The current work showed that both approaches are identical for a single treatment. For more than a single treatment, the ratio approach is to liberal and the regression approach is still fine.

Population-attributable fraction (PAF) is used in epidemiology, health services and public health research to prioritize targets for intervention programs. It allows to quantify, from the population perspective, the overall effect of a set of risk factors on an outcome of interest. It takes the risk factor's effect size as well as its prevalence into account. The average sequential PAF (asPAF), a partialization approach carrying desirable properties, is increasingly used in practical applications recently. The presented simulation study characterizes non-model-based and model-based estimators of the asPAF and Wald and Monte Carlo type confidence intervals.

A good performance of point estimators depends on (a) completeness of variables in the model, and (b) the correct specification of the regression formula with respect to interaction terms in case of model-based estimation. The model-based estimator outperforms the non-model-based estimator and is unbiased even in small samples and in situations with small outcome prevalence. However, computational time of the model-based estimator increases rapidly with increasing sample size and number of variables considered. Resampling-based confidence estimators such as Bootstrap with normality assumption and percentile as well as Jackknife are suited for confidence interval estimation. Here, the computational time especially in conjunction with the model-based estimator increases super-linear with increasing sample size.

Sufficient sample sizes assuring a desirable performance of asPAF exceed those for relative effect measures such as relative risk and odds ratio noticeably. Further, the required sample size increases with a higher number of variables and a lower prevalence of outcome. While sample sizes for PAF are known to decrease with increasing prevalence and effect size of a risk factor, a reverse relation can be observed for asPAF. Sample size estimation can be conducted using Monte Carlo simulation for every conceivable scenario.

Spatio-temporal interactive processes, such as alien species invasions, play a key role in ecology. Existing methods studying such processes often simplify the dynamic structure or the complex interactions of the ecological drivers. In this talk we show how to use relational event modelling (REM) for analysing patterns of ecological interaction processes at large spatial scales including time-varying variables that drive these dynamics. REM relies on temporal interaction dynamics, that encode sequences of relational events connecting a sender node to a recipient node at a specific point in time. We apply REM to the spread of alien species around the globe between 1880 and 2005, following accidental or deliberate introductions into geographical regions outside of their native range. In this context, a relational event represents the new occurrence of an alien species given its former distribution. The application of relational event models to the first reported invasions of 4835 established alien species outside of their native ranges from four major taxonomic groups enables us to unravel the main drivers of the dynamics of the spread of invasive alien species. Combining the alien species' first records data with other spatio-temporal information enables us to discover which factors have been responsible for the spread of species across the globe. Besides the usual drivers of species invasions, such as trade, land use, and climatic conditions, we also find evidence for species-interconnectedness in alien species spread. Relational event models offer the capacity to account for the temporal sequences of ecological events such as biological invasions and to investigate how relationships between these events and potential drivers change over time.