Band 47, Heft 4 - Special Issue: Applied Biostatistics in Laboratory Medicine; Issue Editors: Georg Hoffmann and Frank Klawonn

Objectives Even though most physicians and professionals in laboratory medicine have received basic training in statistics, experience shows that a general understanding of data analysis is not yet available on a broad scale. Therefore, data literacy, data-driven decision making, and computational thinking should be implemented in future educational training. To evaluate the state of digital competence among young scientists (YS) in laboratory medicine, we launched a worldwide online survey. Methods A global online survey was conducted from 25/05/2022 to 26/06/2022 and was disseminated to YS who are listed in three large networks: YS of the DGKL, the EFLM Task Group-YS, and IFCC Task Force-YS and its corresponding members, covering a base of 53 countries. Results A total of 119 young scientists from 40 countries participated in this survey. 80 % did not learn digital skills in their academic education but 96 % felt they needed to. Digital literacy was associated with terms such as programming, artificial intelligence and machine learning, statistics, communication, Big Data and data analytics. Conclusions The results of our survey show that more knowledge and training in the area of digital skills is not just necessary, but also wanted by young scientists. A varied learning environment consisting of tutorial articles, videos, exercises, technical articles, collection of helpful links, online meetings and in person bootcamps is crucial to meet the challenges of an international project with different languages, health systems and time zones.

In many areas of healthcare, digitization has progressed only slowly so far. The SARS-CoV-2 pandemic in particular has shown how valuable software solutions that are deployable at short notice, can be. In this review we present some selected possibilities of the easy-to-learn programming language R and demonstrate potential applications of the package Shiny in the fields of statistical analysis and laboratory medicine. In addition to a brief tabular overview of published applications, we present two examples of their use in routine laboratory workflows. The first example demonstrates how a Shiny app can be used to estimate the minimal difference (MD) of laboratory analytes, while the second example illustrates how pre- and post-analytical processing steps can be integrated into a fully automated workflow using R and the Shiny package.

Objectives Indirect methods for the indirect estimation of reference intervals are increasingly being used, especially for validation of reference intervals, as they can be applied to routine patient data. In this study, we compare three statistically different indirect methods for the verification and validation of reference intervals in eight laboratories distributed throughout Europe. Methods The RefLim method is a fast and simple approach which calculates the reference intervals by extrapolating the theoretical 95 % of non-pathological values from the central linear part of a quantile-quantile plot. The Truncated Maximum Likelihood (TML) method estimates a smoothed kernel density function for the distribution of the mixed data, for which it is assumed that the ‘‘central’’ part of the distribution represents the healthy population. The refineR utilizes an inverse modelling approach. This algorithm identifies a model that best explains the observed data before transforming the data with the Box-Cox transformation. Results We show that the different indirect methods each have their advantages but can also lead to inaccurate or ambiguous results depending on the approximation of the mathematical model to real-world data. A combination of different methodologies can improve the informative value and thus the reliability of results. Conclusions Based on routine measurements of four enzymes alkaline phosphatase (ALP), total amylase (AMY), cholinesterase (CHE) and gamma-glutamyl transferase (GGT) in adult women and men, we demonstrate that some reference limits taken from the literature need to be adapted to the laboratory’s particular local and population characteristics.

Reference intervals are an important component in the interpretation of medical laboratory findings. Especially in children and adolescents, their limits sometimes can change very rapidly with age. We suggest continuous methods to better represent the age-dependent progression of reference intervals. The generalized additive models for location, scale, and shape parameters (GAMLSS) from the R package gamlss generates continuous percentile plots of laboratory values. A user-friendly Shiny application called AdRI_GAMLSS (Age-dependent Reference Intervals), available at github.com/SandraKla/AdRI_GAMLSS, has been developed for this purpose. Using alkaline phosphatase (ALP) as an example, we obtain different smoothed percentile curves depending on the model used. We demonstrate the superiority of continuously modeled reference intervals compared to fixed age groups and provide the Shiny application AdRI_GAMLSS to make the technique easily accessible to clinicians and other experts.

Despite substantial gains facilitated by Artificial Intelligence (AI) in recent years, it has to be applied very cautiously in sensitive domains like medicine due to the lack of explainability of many methods in this field. We aim to provide a system to overcome these issues of medical AI applications by means of our concept of medical operational AI detailed in this paper. We make use of various methods of AI and utilize knowledge graphs in particular. The latter is continuously updated by medical experts based on medical literature such as peer-reviewed papers and standard online sources such as UpToDate. We thoroughly derive a multi-level system tackling the corresponding challenges. In particular, its design encompasses (i) holistic diagnostic assistance on a macro level, (ii) predicitions and detailed suggestions for specific medical domains on a micro level, as well as (iii) AI-based optimizations of the overall system on a meta level. We detail practical merits of medical operational AI and discuss the state of the art beyond our solution.

The amount of data generated in the field of laboratory medicine has grown to an extent that conventional laboratory information systems (LISs) are struggling to manage and analyze this complex, entangled information (“Big Data”). Statistical learning, a generalized framework from machine learning (ML) and artificial intelligence (AI) is predestined for processing “Big Data” and holds the potential to revolutionize the field of laboratory medicine. Personalized medicine may in particular benefit from AI-based systems, especially when coupled with readily available wearables and smartphones which can collect health data from individual patients and offer new, cost-effective access routes to healthcare for patients worldwide. The amount of personal data collected, however, also raises concerns about patient-privacy and calls for clear ethical guidelines for “Big Data” research, including rigorous quality checks of data and algorithms to eliminate underlying bias and enable transparency. Likewise, novel federated privacy-preserving data processing approaches may reduce the need for centralized data storage. Generative AI-systems including large language models such as ChatGPT currently enter the stage to reshape clinical research, clinical decision-support systems, and healthcare delivery. In our opinion, AI-based systems have a tremendous potential to transform laboratory medicine, however, their opportunities should be weighed against the risks carefully. Despite all enthusiasm, we advocate for stringent added-value assessments, just as for any new drug or treatment. Human experts should carefully validate AI-based systems, including patient-privacy protection, to ensure quality, transparency, and public acceptance. In this opinion paper, data prerequisites, recent developments, chances, and limitations of statistical learning approaches are highlighted.