Data-driven healthcare: from patterns to actions

The era of big data opens up new opportunities in personalised medicine, preventive care, chronic disease management and in telemonitoring and managing of patients with implanted devices. The rich data accumulating within online services and internet companies provide a microscope to study human behaviour at scale, and to ask completely new questions about the interplay between behavioural patterns and health. In this paper, we shed light on a particular aspect of data-driven healthcare: autonomous decision-making. We first look at three examples where we can expect data-driven decisions to be taken autonomously by technology, with no or limited human intervention. We then discuss some of the technical and practical challenges that can be expected, and sketch the research agenda to address them.

Nonlinear behaviour of conduction and block in cardiac tissue with heterogeneous expression of connexin 43

Altered gap junctional coupling potentiates slow conduction and arrhythmias. To better understand how heterogeneous connexin expression affects conduction at the cellular scale, we investigated conduction in tissue consisting of two cardiomyocyte populations expressing different connexin levels. Conduction was mapped using microelectrode arrays in cultured strands of foetal murine ventricular myocytes with prede fi ned contents of connexin 43 knockout (Cx43KO) cells. Corresponding computer simulations were run in randomly generated two-dimensional tissues mimicking the cellular architecture of the strands. In the cultures, the relationship between conduction velocity (CV) and Cx43KO cell content was nonlinear. CV fi rst decreased signi fi cantly when Cx43KO content was increased from 0 to 50%. When the Cx43KO content was ≥ 60%, CV became comparabletothatin100%Cx43KOstrands.Co-culturingCx43KOandwild-typecellsalsoresultedinsigni fi cantly more heterogeneous conduction patterns and in frequent conduction blocks. The simulations replicated this behaviour of conduction. For Cx43KO contents of 10 – 50%, conduction was slowed due to wavefront meandering between Cx43KO cells. For Cx43KO contents ≥ 60%, clusters of remaining wild-type cells acted as electrical loads thatimpairedconduction.ForCx43KOcontentsof40 – 60%,conductionexhibitedfractal characteristics,wasprone to block, and was more sensitive to changes in ion currents compared to homogeneous tissue. In conclusion, conduction velocity and stability behave in a nonline ar manner when cardiomyocytes expressing different connexin amounts are combined. This behaviour results from heterogeneous current-to-load relationships at the cellular level. Such behaviour is likely to be arrhythmogenic in various clinical contexts in which gap junctional coupling is heterogeneous.

Scent-marking behaviour and social dynamics in a wild population of Eurasian lynx Lynx lynx.

Scent-marking is widespread among mammals and has been observed in many felid species. Although the behaviour is well-described, little is known about its function in wild felid populations. We investigated patterns of scent-marking and its role in intra- and intersexual communication among resident and non-resident Eurasian lynx Lynx lynx by observing interactions among wild lynx at natural marking sites by means of infrared camera traps. Marking activity of resident animals showed a peak during the mating season and was lowest during the time when females gave birth and lactated. Both sexes scent-marked, but male lynx visited marking sites much more often than females and marked relatively more often when visiting a site. Most visits to marking sites were by residents but we also observed scent-marking by non-residents. Juveniles were never observed marking. We found no evidence of lynx regularly renewing scent-marks after a certain 'expiry date' but the presence of a strange scent-mark triggered over-marking. Males responded similarly to the presence of another individual's scent-mark, irrespective of whether it was the top- or the underlying scent-mark in a mixture of scent-marks they encountered. Our results suggest that marking sites could serve as 'chemical bulletin boards', where male lynx advertise their presence and gain information on ownership relationships in a given area. Females placed their urine marks on top of the ones left by resident males, but further studies are needed to explain the functions of over-marking in females.

Holistic patterns as an instrument to predict promising young football players

Multidimensional talent models represent the current state of the art. However, it remains unclear how these different dimensions interact. Based on current theories of human development, person-oriented approaches seem to be particularly appropriate for talent research. The present study adopts this approach by looking at how a holistic system consisting of the dimensions motivation, motor behaviour and the stage of development goes along with athletic performance. For this purpose, it has to be examined which patterns were formed by the constructs net hope, motor abilities, technical skills and the so far achieved percentage of the predicted adult height and how these patterns are related to subsequent sporting success. 119 young elite football players were questioned and tested three times at intervals of one year, beginning at the age of 12. At the age of 15, the performance level the players had reached was examined. At all three measuring points, four patterns were identified which displayed partial structural and high individual stability. As expected, the players showing values above average in all factors were significantly more likely to advance to the highest performance level. Physically strong, precocious developed players though having some technical weaknesses, have good chances to reach the middle performance level. The results point to the importance of holistic approaches for the prediction of performance among promising football talents in the medium-term and thus provide valuable clues for their selection and promotion.

The importance of socio-cultural factors on sports behaviour of adolescents and young adults in Switzerland – a conceptualisation on the basis of the habitus-approach

Sport participation has often been the topic in sports science and it could be shown that in Europe the population of northern and western countries are more often physically active than southern and eastern countries (European Commission, 2014). In Switzerland the physical activity of the Swiss population also differs between the linguistic regions. The German speaking population is more often physically active than the French or Italian speaking part (Stamm & Lamprecht, 2008). To explain the differences in sport participation structural and cultural factors have been discussed. Because within a country homogenous structural conditions can be assumed, the aim of this study is to analyse how socio-cultural factors correlate with sport participation of adolescents and young adults. In order to analyse this research question, Bourdieu’s concept of habitus (1984) has been used as theoretical background. This sport-related concept of habitus considers cultural determined values, the attribution of meaning and patterns of action which is socially determined and have an influence on individual actions and therefore also on the sport practise. On this basis, a qualitative study including guideline-based interviews with German (n=5) and French (n=3) speaking adolescents and young adults at the age of 16 to 24 (M=21.4) were held in two different linguistic regions of Switzerland. To analyse the interviews the documentary method was applied (Bohnsack, 2010). Initial findings reveal that there are different sport related values, attributions of meanings and patterns of action also called framework of orientations concerning topics like body, health and leisure which correlate with the habitual sports practise in the two different linguistic regions. This study illustrates that the habitus is culturally shaped and that it could help to understand the meaning of socio-cultural factors for sport participation.

Holistic patterns as an instrument to predict performance of promising young football players

Multidimensional talent models represent the current state of the art. However, it remains unclear how these different dimensions interact. Based on current theories of human development, person-oriented approaches seem to be particularly appropriate for talent research. The present study adopts this approach by looking at how a holistic system consisting of the different dimensions motivation, motor behaviour and the stage of development goes along with athletic performance. For this purpose, it has to be examined which patterns were formed by the constructs net hope (Elbe et al., 2003), motor abilities (3 motor tests; Höner et al., 2014), technical skills (3 motor tests; Höner et al., 2014) and the so far achieved percentage of the predicted adult height (Mirwald et al, 2002) and how these patterns are related to subsequent sporting success. 119 young elite football players were questioned and tested three times at intervals of one year, beginning at the age of 12. At the age of 15, the performance level the players had reached was examined (national, regional or no talent card). At all three measuring points, four patterns were identified which displayed partial structural and high individual stability. As expected, the players showing values above average in all factors were significantly more likely to advance to the highest performance level (Odds ratio = 2.2, p < .01). Physically strong, precocious developed players though having some technical weaknesses, have good chances to reach the middle performance level (OR = 1.6, p = .01). Players showing values under average, have an one and a half times higher probability to advance to the lowest performance level (p < .01). The results point to the importance of holistic approaches for the prediction of performance among promising football talents in the medium-term and thus provide valuable clues for their selection and promotion.

Evolution of colour patterns in East African cichlid fish

African cichlid fishes have undergone outbursts of explosive speciation in several lakes, accompanied by rapid radiations in coloration and ecology. Little is known about the evolutionary forces that triggered these events but a hypothesis, published by Wallace Dominey in 1984, has figured prominently. It states that the evolution of colour patterns is driven by sexual selection and that these colour patterns are important in interspecific mate choice, a combination which holds the potential for rapid speciation. Here we present phylogenetic analyses that describe major events in colour evolution and test predictions yielded by Dominey's hypothesis. We assembled information on stripe patterns and the presence or absence of nuptial coloration from more than 700 cichlid species representing more than 90 taxa for which molecular phylogenetic hypotheses were available. We show that sexual selection is most likely the selection force that made male nuptial coloration arise and evolve quickly. In contrast, stripe patterns, though phylogenetically not conserved either, are constrained ecologically. The evolution of vertical bar patterns is associated with structurally complex habitats, such as rocky substrates or vegetation. The evolution of a horizontal stripe is associated with a piscivorous feeding mode. Horizontal stripes are also associated with shoaling behaviour. Strength of sexual selection, measured in terms of the mating system (weak in monogamous, strong in promiscuous species), has no detectable effects on stripe pattern evolution. In promiscuous species the frequency of difference between sister species in nuptial hue is higher than in pair bonding and harem forming species, but the frequency of difference in stripe pattern is lower. We argue that differences between the two components of coloration in their exposure to natural selection explain their very different evolutionary behaviour. Finally, we suggest that habitat-mediated selection upon chromomotor flexibility, a special form of phenotypic plasticity found in the river-dwelling outgroups of the lake-dwelling cichlids, explains the rapid and recurrent ecology-associated radiation of stripe patterns in lake environments, a new hypothesis that yields experimentally testable predictions.

Unusual Southern Hemisphere tree growth patterns induced by changes in the Southern Annular Mode

Recent changes in the summer climate of the Southern Hemisphere extra-tropics are primarily related to the dominance of the positive phase of the Southern Annular Mode1, 2. This shift in the behaviour of the Southern Annular Mode—essentially a measure of the pressure gradient between Southern Hemisphere mid and high latitudes—has been predominantly induced by polar stratospheric ozone depletion2, 3, 4. The concomitant southward expansion of the dry subtropical belts5, 6 could have consequences for forest growth. Here, we use tree-ring records from over 3,000 trees in South America, Tasmania and New Zealand to identify dominant patterns of tree growth in recent centuries. We show that the foremost patterns of growth between 1950 and 2000 differed significantly from those in the previous 250 years. Specifically, growth was higher than the long-term average in the subalpine forests of Tasmania and New Zealand, but lower in the dry-mesic forests of Patagonia. We further demonstrate that variations in the Southern Annular Mode can explain 12–48% of the tree growth anomalies in the latter half of the twentieth century. Tree-ring-based reconstructions of summer Southern Annular Mode indices suggest that the high frequency of the positive phase since the 1950s is unprecedented in the past 600 years. We propose that changes in the Southern Annular Mode have significantly altered tree growth patterns in the Southern Hemisphere.

Palynostratigraphy of the last centuries in Switzerland based on 23 lake and mire deposits: chronostratigraphic pollen markers, regional patterns, and local histories

A total of 23 pollen diagrams [stored in the Alpine Palynological Data-Base (ALPADABA), Geobotanical Institute, Bern] cover the last 100 to over 1000 years. The sites include 15 lakes, seven mires, and one soil profile distributed in the Jura Mts (three sites), Swiss Plateau (two sites), northern Pre-Alps and Alps (six sites), central Alps (five sites), southern Alps (three sites), and southern Pre-Alps (four sites) in the western and southern part of Switzerland or just outside the national borders. The pollen diagrams have both a high taxonomic resolution and a high temporal resolution, with sampling distances of 0.5–3 cm, equivalent to 1 to 11 years for the last 100 years and 8 to 130 years for earlier periods. The chronology is based on absolute dating (14 sites: 210Pb 11 sites; 14C six sites; varve counting two sites) or on biostratigraphic correlation among pollen diagrams. The latter relies mainly on trends in Cannabis sativa, Ambrosia, Mercurialis annua, and Ostrya-type pollen. Individual pollen stratigraphies are discussed and sites are compared within each region. The principle of designating local, extra-local, and regional pollen signals and vegetation is exemplified by two pairs of sites lying close together. Trends in biostratigraphies shared by a major part of the pollen diagrams allow the following generalisations. Forest declined in phases since medieval times up to the late 19th century. Abies and Fagus declined consistently, whereas the behaviour of short-lived trees and trees of moist habitats differed among sites (Alnus glutinosa-type, Alnus viridis, Betula, Corylus avellana). In the present century, however, Picea and Pinus increased, followed by Fraxinus excelsior in the second half of this century. Grassland (traced by Gramineae and Plantago lanceolata-type pollen) increased, replacing much of the forest, and declined again in the second half of this century. Nitrate enrichment of the vegetation (traced by Urtica) took place in the first half of this century. These trends reflect the intensification of forest use and the expansion of grassland from medieval times up to the end of the last century, whereas subsequently parts of the grassland became used more intensively and the marginal parts were abandoned for forest regrowth. In most pollen diagrams human impact is the dominant factor in explaining inferred changes in vegetation, but climatic change plays a role at three sites.

Does the use of mobile devices (tablets and smartphones) affect survey quality and choice behaviour in web surveys?

Web surveys are becoming increasingly popular in survey research including stated preference surveys. Compared with face-to-face, telephone and mail surveys, web surveys may contain a different and new source of measurement error and bias: the type of device that respondents use to answer the survey questions. This is the first study that tests whether the use of mobile devices, tablets or smartphones, affects survey characteristics and stated preferences in a web-based choice experiment. The web survey on expanding renewable energy production in Germany was carried out with 3182 respondents, of which 12% used a mobile device. Propensity score matching is used to account for selection bias in the use of mobile devices for survey completion. We find that mobile device users spent more time than desktop/laptop users to answer the survey. Yet, desktop/laptop users and mobile device users do not differ in acquiescence tendency as an indicator of extreme response patterns. For mobile device users only, we find a negative correlation between screen size and interview length and a positive correlation between screen size and acquiescence tendency. In the choice experiment data, we do not find significant differences in the tendency to choose the status quo option and scale between both subsamples. However, some of the estimates of implicit prices differ, albeit not in a unidirectional fashion. Model results for mobile device users indicate a U-shaped relationship between error variance and screen size. Together, the results suggest that using mobile devices is not detrimental to survey quality.

Thyroid disruption in zebrafish (Danio rerio) larvae: Different molecular response patterns lead to impaired eye development and visual functions

The vertebrate thyroid system is important for multiple developmental processes, including eye development. Thus, its environmentally induced disruption may impact important fitness-related parameters like visual capacities and behaviour. The present study investigated the relation between molecular effects of thyroid disruption and morphological and physiological changes of eye development in zebrafish (Danio rerio). Two test compounds representing different molecular modes of thyroid disruption were used: propylthiouracil (PTU), which is an enzyme-inhibitor of thyroid hormone synthesis, and tetrabromobisphenol A (TBBPA), which interacts with the thyroid hormone receptors. Both chemicals significantly altered transcript levels of thyroid system-related genes (TRα, TRβ, TPO, TSH, DIO1, DIO2 and DIO3) in a compound-specific way. Despite these different molecular response patterns, both treatments resulted in similar pathological alterations of the eyes such as reduced size, RPE cell diameter and pigmentation, which were concentration-dependent. The morphological changes translated into impaired visual performance of the larvae: the optokinetic response was significantly and concentration-dependently decreased in both treatments, together with a significant increase of light preference of PTU-treated larvae. In addition, swimming activity was impacted. This study provides first evidence that different modes of molecular action of the thyroid disruptors can be associated with uniform apical responses. Furthermore, this study is the first to show that pathological eye development, as it can be induced by exposure to thyroid disruptors, indeed translates into impaired visual capacities of zebrafish early life stages.