Data collectors collect health data using Mobile Electronic Data Collection Forms (MEDCFs) particularly in hard to reach areas. However, the usability and user acceptance of these forms by the data collectors are seldom considered, and yet these have an implication on the quality of the data collected and on the health decisions thereof. In this study we aimed at collecting the design preferences the data collectors felt would improve their data collection experience. For that purpose, a mid-fidelity prototype was used to accomplish six tasks and a semi-structured usability questionnaire was given. Forty eight data collectors from Uganda participated in the study between December 2017 and January 2018, indicating their preferences after interacting with the prototype. The results included a detailed feedback regarding the presentation of the forms content, form navigation, error handling, data input, and visualization of progress status. Involvement of users and other stakeholders in the design of MEDCFs using the User Centred Design (UCD) will presumably enhance usability of the data collection forms.

Digitalisation of health care for the purpose of medical documentation lead to huge amounts of data, hence having an opportunity to derive knowledge and associations of different attributes recorded. Many health care events can be prevented when identified. Machine learning algorithms could identify such events but there is ambiguity in understanding the suggestions especially in clinical setup. In this paper we are presenting how we explain the decision based on random forest to health care professionals in the course of the project predicting delirium during hospitalisation on the day of admission.

Comorbidities are multiple co-occurring disorders associated with a primary diagnosis and affect health outcomes and cost of care. Using home health medical claims data, a comorbidity database with frequent item-sets, and an exemplar dashboard were created. The dashboard extends the decision making capacity of clinicians by providing data-driven information about (i) the frequency of comorbidities for any primary diagnosis, and (ii) primary diagnoses sufficiently exclusive to a given comorbidity. Regression models estimate total charges, for any underlying patient comorbidity profile. Using the exemplar dashboard, a panel of healthcare researchers recommended appropriate system parameters to adjust system sensitivity and improve construct validity. The comorbidity database will be useful in future research efforts to study comorbidity, while the exemplar dashboard can provide the foundation for integrated home healthcare decision support systems.

A hot point of phthisiology in Russia during the last decade is the rising morbidity of tuberculosis associated with the drug resistant Micobacteria tuberculosis. The causes of MBT resistance are often related with human factor. Reducing the errors rate in complex Clinical Decision-making in phthisiology requires adequate software support and treatment monitoring. The methodology for structuring of the diagnosis characteristics at the moments of Clinical Decisions in phthisiology is represented. The Quality models and real Clinical Situations have the digital codes. The ontological modeling of the real Clinical Situations is used for analysis, construction the inference rules and improvement of Clinical Decisions and treatment results.

Background: The vast amount of data generated in healthcare can be reused to support decision-making by developing clinical decision-support systems. Since evidence is lacking in Pediatrics, it seems to be beneficial to design future systems towards the vision of generating evidence through cross-institutional data analysis and continuous learning cycles.

Objectives: Presentation of an approach for cross-institutional and data-driven decision support in pediatric intensive care units (PICU), and the long-term vision of Learning Healthcare Systems in Pediatrics.

Methods: Using a four-step approach, including the design of interoperable decision-support systems and data-driven algorithms, for establishing a Learning Health Cycle.

Results: We developed and started to follow that approach on exemplary of systemic inflammatory response syndrome (SIRS) detection in PICU.

Conclusions: Our approach has great potential to establish our vision of learning systems, which support decision-making in PICU by analyzing cross-institutional data and giving insights back to both, their own knowledge base and clinical care, to continuously learn about practices and evidence in Pediatrics.

Data quality (DQ) assessment is advisable before (re)using datasets. Besides supporting DQ-assessment, DQ-tools can indicate data integration issues. The objective of this contribution is to put up for discussion the identified current state of scientific knowledge in DQ-assessment for health data and the planned work resulting from that state of knowledge. The state of scientific knowledge bases on a continuous literature survey and tracking of related working groups' activities. 95 full text publications constitute the considered state of scientific knowledge of which a representative selection of six DQ-tools and -frameworks is presented. The delineated future work explores multi-institutional machine learning on the DQ-measurement results of an interoperable DQ-tool, with the goal to optimize DQ-measurement method combinations and reference values for DQ-issue recognition.

Data informed decision-making (DIDM) plays a fundamental role in the fight against HIV and AIDS in Low-and-Middle-Income Countries (LMICs). Despite the scale-up of HIV care services, supported by health information systems (HIS), cases of under-utilization of data in decision-making are still being reported at different levels of the health system. Literature revealing the process involved in data users' decision to adopt DIDM in LMICs is meager. To fill this gap, we employ the innovation-decision model to describe the stages of adoption of DIDM by data users. Thus, we extract reports on efforts to promote DIDM in HIV from existing literature, and map this to the model. We then identify important stages, which require emphasis in the adoption process. Hence, implementers could benefit from use of the innovation-decision model in understanding adoption process of DIDM.

Nowadays Brazil has a complex cancer care scenario. There are nearly 600.000 new cancer cases each year in Brazil, and the huge majority of patients have some contact with hospital services. However, long waiting queues for diagnostics and treatments have become common. One of the critical success factors in a cancer treatment is early diagnosis. The reduction of waiting time to start therapeutic procedures is one of the main issues for improvement of patient's quality of life and possibilities of cure. The objective of this work is to describe the development of a decision support system that improves the identification of access alternatives, appointment scheduling and employment of available resources. The Theory of Constraints was used to identify bottlenecks in patient treatment flow and a Discrete Events Simulation model was used to reduce patients' waiting time to start cancer treatment.

Physical fitness is important in view of reducing risks for a number of non-communicable diseases, both for individuals and policy-makers. In this paper, we present a prototype tool that combines forecasting of individual fitness parameters of schoolchildren to early adulthood with estimation of relative risk for all-cause early mortality in adulthood based on the forecasted fitness. This tool is a first step in the development of a platform that will show age, gender, and geographical distributions of risk and suggest potential interventions.

Patient specific forecasting tools are an area of active research and very much seen as a necessary tool for future improvements in healthcare. In order to succeed with decision making tools, fine-grained data are required to build models relevant and valid at an individual level. Location and assemble of data to build such tools is not trivial. Even then the ability to perform accurate predictions is not guaranteed. This study outlines a method to integrate existing data sources to base predictions on. A key benefit of the method is the minimal extra burden on the patient and the healthcare system. A pilot study is performed to implement the system architecture on data from total knee arthroplasty. Output from the system is presented using web technologies. In doing so, the viability of the method to implement a tool for the prediction of pre-operative and post-operative follow-up is demonstrated. Future steps will include testing and deployment of the system.

As technology immerses in daily life, all types of clinical practices migrate more on storing medical records on electronically media. The issue is up for debate on several fronts, as both paper and electronic records offer strengths and weaknesses. This paper presents the process of changing from paper medical records to a digital environment in the case of an alternative medicine clinic, having years of stored paper records, with its benefits and challenges. Focusing on quality criteria, the current study shows how beneficial an electronic medical record could be, while arguing how the diagnosis coding from the paper-based patient record resulted in major qualitative disadvantages.

Health analytics is a business-driven term that encompasses a wide spectrum of aspects and dimensions of business intelligence applications and big data analysis. Healthcare organizations recently are eager to know whether they are getting the full value from the massive amounts of data and information they already have, to achieve their strategic effectiveness goals and operational efficiency objectives. It is very crucial to learn more about the diverse functions, types and levels through which health analytics can support such tasks. A careful review of literature was conducted, and a qualitative analysis was used to classify health analytics. Five main types of analytics could be identified; these are descriptive, diagnostic, predictive, prescriptive and discovery analytics, each has its own distinct role in improving healthcare. In addition to the five types, health analytics could also be classified into three levels of performance and engagement, these are the operational, tactical and strategic health analytics.

Bacterial meningitis is a dangerous infectious disease that the entire community can be influenced by its epidemics. The objective of this study is to develop a predictive model as a screening tool to accelerate distinguishing between patients with acute bacterial meningitis and non-bacterial ones to prevent bacterial meningitis epidemics in Iran. This study was conducted on Iranian meningitis registry, which consists of 7,945 suspected cases of the disease between 2009 and 2011. Each sample has 8 predictive and a target variables. The predictive model was developed by decision tree algorithm and, the overall accuracy was 78%, with a sensitivity of 87%, and a specificity of 70%, respectively. This model can help health policymakers and epidemiologists to identify bacterial meningitis outbreaks and support them to make a decision in infection dynamics. In conclusion, we developed and validated a predictive model that can be used in meningitis surveillance system in Iran. However, further research is needed to use the model in practice with different pathogen types of bacterial meningitis in order to proper antimicrobial therapy planning.

Accurate outcome prediction by the means of available clinical contributing factors will support researchers and administrators in realistic planning, workload determination, resource optimization, and evidence-based quality control process. This study is aimed to evaluate APACHE II and SAPS II prediction models in an Iranian population. A a prospective cross-sectional study was conducted in four tertiary care referral centers located in the top two most populated cities in Iran, from August 2013 to August 2015. The Brier score, Area under the Receiver Operating Characteristics Curve (AUC), and Hosmer-Lemeshow (H-L) goodness-of-fit test were employed to quantify models' performance. A total of 1799 patients (58.5% males and 41.5% females) were included for further score calculation. The overall observed mortality (24.4%) was more than international rates due to APACHE II categories. The Brier score for APACHE II and SAPS II were 0.17 and 0.196, respectively. Both scoring systems were associated with acceptable AUCs (APACHE II = 0.745 and SAPS II = 0.751). However, none of prediction models were fitted to dataset (H-L ρ value < 0.01). With regards to poor performance measures of APACHE II and SAPS II in this study, finding recalibrated version of current prediction models is considered as an obligatory research question before applying it as a clinical prioritization or quality control instrument.

Our Oncology Information Systems (OIS) with local User Documentation manages the medical data explored in this 26 item report of a minimal medical dataset. Over 10 years to 2016, 12906 diagnoses were registered (ICD10: C00-C80), with 18.84% of cases, and 63.9% of data points complete. Two sites were quality assured with high completion rates (H&N – 97.4% [4.26% of total cases], RECTUM – 88.74% [4.06%]). Sites lacking clinician attention varied from poor (eg, LUNG – 23.23% [13.24%]) to largely incomplete (eg, BRAIN – 2.01% [0.38%]).This disappointing medical data completion rate makes its use in a ‘Big Data’ effort suspect. Data extrapolation is compromised by variable natural history. Extrapolation techniques are unlikely to cope with only 18.84% complete data. Data mining requires input from domain experts. The 4 requirements of Big Data are not evident in oncological data.

The field of medicine still reports errors because of insufficient knowledge or resources, work load or data not available at the right time and place, and this may be fatal for a patient. To improve the healthcare quality, a doctor needs accurate and complex information processing when accessing drug information. Our work builds on improvement of accessing drug information for a better treatment through homogenization of sections in a prospectus. The sections names in a prospectus may be different for one source to another, and in this article, we propose a method to homogenize the content of all drug prospectuses. Once a correct homogenization of the sections has been established, the prospectuses can be used in clinical decision applications to provide the necessary data for physicians. Classification of the section names is using the Cousine similarity method and the Scikit-learn machine learning software. The best results were obtained with the Scikit-learn software.

This paper presents a method for exploring structural risk of any artificial intelligence-based method in bioinformatics, the A-Test method. This method provides a way to not only quantitate the structural risk associated with a classification method, but provides a graphical representation to compare the learning capacity of different classification methods. Two different methods, Deep Time Growing Neural Network (DTGNN) and Hidden Markov Model (HMM), are selected as two classification methods for comparison. Time series of heart sound signals are employed as the case study where the classifiers are trained to learn the disease-related changes. Results showed that the DTGNN offers a superior performance both in terms of the capacity and the structural risk. The A-Test method can be especially employed in comparing the learning methods with small data size.

Patient data privacy is emphasized due to increase in electronic health data processing. This paper highlights the importance of data privacy in healthcare and its connection to the European Union's (EU) General Data Protection Regulation, which aims to protect all EU citizens from privacy and data violations. The clinical environment is prone to hazards in information management, especially in care coordination, the management of patient data, and verbal communication. The Swiss cheese model was used to discuss the importance of patient data privacy based on statistics from a national patient safety incident registry. Patient identification proved an important factor leading to hazards that can cause severe harm to patients.

Medical registries are in a need of a data set that is based on clinical evidence. In 2014, the Saudi Food and Drug Administration (SFDA) launched a plan to develop the national Comprehensive Implantable Medical Device Registry (CIMDR). One of the primary goals of the CIMDR is to develop a clinical -and population- based data set. The aim of this study is to report on the process of developing the data elements for the CIMDR. We used an iterative process of multi-stakeholder consultation over a two year period (2014–2016). The goal of the multi-stakeholder consultations was to build a dataset to address the need for device traceability, effectiveness, safety, and the recall of implantable medical devices. We investigated international and local standards for implantable medical device information capture, conducted a review of the literature, and consulted expert opinions in the development of the CIMDR dataset. The CIMDR data framework includes demographics, patient history, diagnosis, procedure information, and follow-up details for orthopedic and cardiac related implantable medical devices. Most of the dataset elements are logically validated with minimal free text entry to avoid human error and facilitate ease of entry. We use the International Classification for Diseases-Australian Modification as the standardized nomenclature for the CIMDR.

The proposed application is based on Geographic Information Systems (GIS) and it's used for medical devices spatial mapping. The application aims at improving the efficiency of hospital units. By using appropriate software, digital maps are created. The technical and financial data of the respirators and ultrasound systems are displayed on the maps. The respirators are installed in the Intensive Care Units (ICU) of public hospitals in Peloponnese. The ultrasound scan systems are also installed in the Radiology Departments of the same units. The user is enabled to create interactive maps and visualize a large amount of real-time information. The overarching aim is to develop an integrated surveillance system of medical equipment. This process will promote the loss-making machinery replacement. Furthermore, the efficient management of revenue and expenditure operations of hospital units will be promoted. The reliable decision-making by the Ministry of Health or the Health and Social Welfare Departments will also be enhanced.

Greece is one of the European countries with the largest number of CT and MRI scans. Geographic Information Systems (GIS) are a useful tool for efficient cost reduction in the health sector. ArcGIS software is used to create interpretable maps. Data represents the number of imaging examinations per type and population size. All available data is acquired in 2012–2015. By using special symbols, pop up windows emerge. The windows contain a large number of data. The results are filtered and uploaded to a digital control panel. This central information system combined with the digital maps form a digital platform for estimating the exact number of diagnostic examinations. The survey confirms that the General Hospital of Tripolis performed 27.000 CT examinations in 2012–2015, corresponding to approximately 57% of the population of Tripolis. The approximate percentage of CT scans of the Hospital Unit of Pyrgos was found 55%. The implementation of this platform confirms the large number of diagnostic examinations in most areas of Peloponnese.

The role of today's insurer is changing towards a more preventive and digital or connected approach. In this context, connected health insurance has the potential to contribute towards the general wellbeing of the population. New technologies put to use by the insurance industry might even help deal with major issues related to the rising number of people globally, of chronic disease patients and that of elders while keeping them healthier and the same time protected by insurance.

Registries are increasingly implemented to record the practice of health care. Within a national funding scheme for registries, an accompanying project was launched to support the design of the registries' IT infrastructure amongst other tasks for 16 projects. A challenge of data management systems was organized by the accompanying project in order to enable the projects to define realistic expectations towards IT support in their research protocols. Twelve vendors participated in the challenge. They presented their solutions for selected use cases. In advance, the projects considered a sufficient authorization concept and the possibility to export data to be of highest importance. However, the systems covered mainly core processes of electronic data capture. The accompanying project will continue its support for the next stage of the funding scheme, which will be the implementation of the registries that win a competitive review of their research protocols prepared in the concept development stage.

Electronic Medical Records (EMR) systems show promise for facilitating health care improvement in quality patient care, patient safety and cost reduction. Nevertheless, their adoption requires careful planning and execution for successful implementation and optimal benefits. The main objective of this review was to identify, analyse and categorize facilitators and barriers to the implementation of EMRs in resource constrained settings in order to gain insight for successful EMR implementation. A literature review on papers from 2007 to 2017 concerning facilitators and barriers to EMRs implementation was conducted. The study included 18 articles that met selection criteria. Four categories of facilitators and barriers including a total of 28 sub-categories were identified from content analysis. These are technical, human, processes and organizational. EMR implementers should pay attention to these issues and adopt a change management strategy for sustainable EMR use in resource-constrained settings.