Predicting epidemics using search engine data: a comparative study on measles in the largest countries of Europe

Background

In recent years new forms of syndromic surveillance that use data from the Internet have been proposed. These have been developed to assist the early prediction of epidemics in various cases and diseases. It has been found that these systems are accurate in monitoring and predicting outbreaks before these are observed in population and, therefore, they can be used as a complement to other methods. In this research, our aim is to examine a highly infectious disease, measles, as there is no extensive literature on forecasting measles using Internet data,

Methods

This research has been conducted with official data on measles for 5 years (2013–2018) from the competent authority of the European Union (European Center of Disease and Prevention - ECDC) and data obtained from Google Trends by using scripts coded in Python. We compared regression models forecasting the development of measles in the five countries.

Results

Results show that measles can be estimated and predicted through Google Trends in terms of time, volume and the overall spread. The combined results reveal a strong relationship of measles cases with the predicted cases (correlation coefficient R= 0.779 in two-tailed significance p< 0.01). The mean standard error was relatively low 45.2 (12.19%) for the combined results. However, major differences and deviations were observed for countries with a relatively low impact of measles, such as the United Kingdom and Spain. For these countries, alternative models were tested in an attempt to improve the results.

Conclusions

The estimation of measles cases from Google Trends produces acceptable results and can help predict outbreaks in a robust and sound manner, at least 2 months in advance. Python scripts can be used individually or within the framework of an integrated Internet surveillance system for tracking epidemics as the one addressed here.

The C++ programming language in cheminformatics and computational chemistry

This paper describes salient features of the C++ programming language and its programming ecosystem, with emphasis on how the language affects scientific software development. Brief history of C++ and its predecessor the C language is provided. Most important aspects of the language that define models of programming are described in greater detail and illustrated with code examples. Special attention is paid to the interoperability between C++ and other high-level languages commonly used in cheminformatics, machine learning, data processing and statistical computing.

Use of content management systems to address nursing workflow

Nurses are at the forefront of providing healthcare services to individuals of all age groups and with varying medical conditions. Aside from the critical knowledge and technical skills from nursing science, advancement in technology has assisted nurses in delivering quality nursing care by streamlining workflow processes and ensuring that data can easily be retrieved or modified. Electronic health records dramatically changed the landscape of the healthcare practice by providing an electronic means to store data and for healthcare professionals to retrieve and manipulate health information in a secured and collaborative environment. But with the nature of data being stored in the electronic health records, nurses still need to organize and process these data into relevant information, knowledge or wisdom so they can provide better holistic care to patients. This discussion paper details the role of content management systems in addressing nursing workflow by providing a mechanism for nurses to be developers themselves, and not just users or consumers of health innovative technologies. By using content management systems as platform for application development, nurses or other healthcare professionals, may be able to address problems with internal workflow without having to incur huge amounts in software development, or having to extensively learn programming languages.

Machine learning in medicine: a practical introduction

BACKGROUND

Following visible successes on a wide range of predictive tasks, machine learning techniques are attracting substantial interest from medical researchers and clinicians. We address the need for capacity development in this area by providing a conceptual introduction to machine learning alongside a practical guide to developing and evaluating predictive algorithms using freely-available open source software and public domain data.

METHODS

We demonstrate the use of machine learning techniques by developing three predictive models for cancer diagnosis using descriptions of nuclei sampled from breast masses. These algorithms include regularized General Linear Model regression (GLMs), Support Vector Machines (SVMs) with a radial basis function kernel, and single-layer Artificial Neural Networks. The publicly-available dataset describing the breast mass samples (N=683) was randomly split into evaluation (n=456) and validation (n=227) samples. We trained algorithms on data from the evaluation sample before they were used to predict the diagnostic outcome in the validation dataset. We compared the predictions made on the validation datasets with the real-world diagnostic decisions to calculate the accuracy, sensitivity, and specificity of the three models. We explored the use of averaging and voting ensembles to improve predictive performance. We provide a step-by-step guide to developing algorithms using the open-source R statistical programming environment.

RESULTS

The trained algorithms were able to classify cell nuclei with high accuracy (.94 -.96), sensitivity (.97 -.99), and specificity (.85 -.94). Maximum accuracy (.96) and area under the curve (.97) was achieved using the SVM algorithm. Prediction performance increased marginally (accuracy =.97, sensitivity =.99, specificity =.95) when algorithms were arranged into a voting ensemble.

CONCLUSIONS

We use a straightforward example to demonstrate the theory and practice of machine learning for clinicians and medical researchers. The principals which we demonstrate here can be readily applied to other complex tasks including natural language processing and image recognition.

Online Radiology Reporting with Peer Review as a Learning and Feedback Tool in Radiology; Implementation, Validity, and Student Impressions

Communicating radiological reports to peers has pedagogical value. Students may be uneasy with the process due to a lack of communication and peer review skills or to their failure to see value in the process. We describe a communication exercise with peer review in an undergraduate veterinary radiology course. The computer code used to manage the course and deliver images online is reported, and we provide links to the executable files. We tested to see if undergraduate peer review of radiological reports has validity and describe student impressions of the learning process. Peer review scores for student-generated radiological reports were compared to scores obtained in the summative multiple choice (MCQ) examination for the course. Student satisfaction was measured using a bespoke questionnaire. There was a weak positive correlation (Pearson correlation coefficient = 0.32, p < 0.01) between peer review scores students received and the student scores obtained in the MCQ examination. The difference in peer review scores received by students grouped according to their level of course performance (high vs. low) was statistically significant (p < 0.05). No correlation was found between peer review scores awarded by the students and the scores they obtained in the MCQ examination (Pearson correlation coefficient = 0.17, p = 0.14). In conclusion, we have created a realistic radiology imaging exercise with readily available software. The peer review scores are valid in that to a limited degree they reflect student future performance in an examination. Students valued the process of learning to communicate radiological findings but do not fully appreciated the value of peer review.

User-defined functions in the Arden Syntax: An extension proposal

BACKGROUND

The Arden Syntax is a knowledge-encoding standard, started in 1989, and now in its 10th revision, maintained by the health level seven (HL7) organization. It has constructs borrowed from several language concepts that were available at that time (mainly the HELP hospital information system and the Regenstrief medical record system (RMRS), but also the Pascal language, functional languages and the data structure of frames, used in artificial intelligence). The syntax has a rationale for its constructs, and has restrictions that follow this rationale. The main goal of the Standard is to promote knowledge sharing, by avoiding the complexity of traditional programs, so that a medical logic module (MLM) written in the Arden Syntax can remain shareable and understandable across institutions.

OBJECTIVES

One of the restrictions of the syntax is that you cannot define your own functions and subroutines inside an MLM. An MLM can, however, call another MLM, where this MLM will serve as a function. This will add an additional dependency between MLMs, a known criticism of the Arden Syntax knowledge model. This article explains why we believe the Arden Syntax would benefit from a construct for user-defined functions, discusses the need, the benefits and the limitations of such a construct.

METHODS AND MATERIALS

We used the recent grammar of the Arden Syntax v.2.10, and both the Arden Syntax standard document and the Arden Syntax Rationale article as guidelines. We gradually introduced production rules to the grammar. We used the CUP parsing tool to verify that no ambiguities were detected.

RESULTS

A new grammar was produced, that supports user-defined functions. 22 production rules were added to the grammar. A parser was built using the CUP parsing tool. A few examples are given to illustrate the concepts. All examples were parsed correctly.

CONCLUSIONS

It is possible to add user-defined functions to the Arden Syntax in a way that remains coherent with the standard. We believe that this enhances the readability and the robustness of MLMs. A detailed proposal will be submitted by the end of the year to the HL7 workgroup on Arden Syntax.

SU-E-J-114: Web-Browser Medical Physics Applications Using HTML5 and Javascript

PURPOSE

Since 2010, there has been a great attention about HTML5. Application developers and browser makers fully embrace and support the web of the future. Consumers have started to embrace HTML5, especially as more users understand the benefits and potential that HTML5 can mean for the future.Modern browsers such as Firefox, Google Chrome, and Safari are offering better and more robust support for HTML5, CSS3, and JavaScript. The idea is to introduce the HTML5 to medical physics community for open source software developments. The benefit of using HTML5 is developing portable software systems.

METHODS

The HTML5, CSS, and JavaScript programming languages were used to develop several applications for Quality Assurance in radiation therapy. The canvas element of HTML5 was used for handling and displaying the images, and JavaScript was used to manipulate the data. Sample application were developed to: 1. analyze the flatness and symmetry of the radiotherapy fields in a web browser, 2.analyze the Dynalog files from Varian machines, 3. visualize the animated Dynamic MLC files, 4. Simulation via Monte Carlo, and 5. interactive image manipulation.

RESULTS

The programs showed great performance and speed in uploading the data and displaying the results. The flatness and symmetry program and Dynalog file analyzer ran in a fraction of second. The reason behind this performance is using JavaScript language which is a lower level programming language in comparison to the most of the scientific programming packages such as Matlab. The second reason is that JavaScript runs locally on client side computers not on the web-servers.

CONCLUSIONS

HTML5 and JavaScript can be used to develop useful applications that can be run online or offline on different modern web-browsers. The programming platform can be also one of the modern web-browsers which are mostly open source (such as Firefox).