Mapping the Bulgarian Diabetes Register to OMOP CDM: Application Results.

BACKGROUND: The Bulgaria Diabetes Register (BDR) contains more than 380 millions of pseudonymized outpatient records with proprietary data structures and format. OBJECTIVES: This paper presents the application results and experience acquired during the process of mapping such observational health data to OMOP CDM with the objective of publishing it in the European Health Data and Evidence Network (EHDEN) Portal. METHODS: The data mapping follows the activities of the well-structured Extract-Transform-Load process. Unlike other publications, we focus on the need for preprocessing the data structures of raw data, cleaning data and procedures for assuring quality of data. RESULTS: This paper provides quantitative and statistical measures for the records in the CDM database as published in the EHDEN Portal. CONCLUSION: The mapping of data from the BDR to OMOP CDM provides the EHDEN community with opportunities for including these data in large-scale project for evidence generation by applying standard analytical tools.

Application of OMOP Common Data Model for Data Integration: The Bulgarian Diabetes Register.

This paper considers mapping of the Bulgarian Diabetes Register(BDR) onto OMOP Common Data Model (CDM). Research results are referenced and plans for analysis of drug shortages from federated data sources are outlined.

Occupational health assessment summary designed for semantic interoperability.

BACKGROUND: Geopolitical and economic crises force a growing number of people to leave their countries and search better employment opportunities abroad. Meanwhile, the highly competitive labor market provides opportunities for employees to change workplaces and job positions. Health assessment data collected during the occupational history is an essential resource for developing efficient occupational disease prevention strategies as well as for ensuring the physical and psychological well-being of newly appointed workers. The diversity in data representation is source for interoperability problems that are insufficiently explored in the existing literature. OBJECTIVES: This research aims to design a worker's occupational health assessment summary (OHAS) dataset that satisfies the requirements of an international standard for semantic interoperability in the use case for exchanging extracts of such data. The focus is on the need for a common OHAS standard at EU level allowing seamless exchange of OHAS at both cross-border and at the worker's country of origin level. RESULTS: This paper proposes a novelty systematic approach ensuring semantic interoperability in the exchange of OHAS. Two use cases are explored in terms of UML sequence diagram. The OHAS dataset reflects common data requirements established in the national legislation of EU countries. Finally, an EN 13606 archetype of OHAS is designed by satisfying the requirements for semantic interoperability in the exchange of clinical data. Semantic interoperability of OHAS is demonstrated with realistic use case data. CONCLUSIONS: The designed static, non-volatile and reusable information model of OHAS developed in this paper allows to create EN 13606 archetype instances that are valid with respect to the Reference model and the datatypes of this standard. Thus, basic activities in the OHAS use case can be implemented in software, for example, by means of a native XML database as well as integrated into existing information systems.

Diabetes Prevalence and Duration Data Extracted from Outpatient Records Representative for the Bulgarian Population.

This paper presents new data about diabetes prevalence and illness duration from a population of patients with diabetes Type 1(43,818) and Type 2 (457,247). Unlike the usual approach that employs adjusted estimates in similar prevalence reports, this study extracts data from a large number of original clinical documents such as all the outpatient records (6,887,876) issued in Bulgaria to all the 501,065 patients with diabetics during 2018 (9.77% of all the 5,128,172 patients recorded in 2018, 4.43% male and 5.35% female). Diabetes prevalence data are described in terms of distributions of Type 1 and Type 2 diabetes over age and gender. It is mapped to an Observational Medical Outcomes Partnership Common Data Model that is publicly available. The distribution of Type 2 diabetics agrees with peak values of BMI established in related research. A major novelty in this research are the data about the diabetes illness duration. It is an essential metric for evaluating the quality of processes evolving over time. Accurate estimates are obtained for the duration measured in years of Type 1 (95% CI 10.92, 11.08) and Type 2 (95% CI 7.97, 8.02) diabetics from the Bulgarian population. The patients with diabetes Type 1 have longer diabetes duration relatively to those with Type 2. It is recommended to include this metric in official diabetes prevalence reports.

Diabetes and Obesity in Bulgaria. Study of a Large Number of Outpatient Records from 2018.

This paper considers the association between diabetes and obesity by examining body mass index (BMI) values and ICD-10 codes for obesity illnesses. The BMI values are extracted from 6,887,876 anonymized outpatient records describing all the visits of diabetics to general practitioners and specialists in ambulatory care from the latest Bulgarian nationally representative data. The number of adults in this sample having BMI ≥ 25 is 253,841 i.e. 84.121% of the adult diabetics with BMI records are overweight or obese. The objective of the study is to reveal how the BMI recorded values in outpatient records relate diabetics with overweight or obesity illness. In the existing literature sources there is scant empirical data of this subject where the conclusions are founded on a nationally representative sample. A secondary objective is to obtain the distribution of BMI values of adults with respect to their age and gender. The initial computer experiments prove that there is no immediate and unconditional relation between BMI and the obesity illnesses. These results underpin the role of BMI as a risk factor that should be observed regularly as an important part of proactive public health policies.

Using QR Code for Uniform Representation of Content in Cross-Border Exchange of ePrescriptions in the EU.

This paper proposes an approach and demonstrates its application for cross-border exchange of ePrescriptions in the European Union. A business process model of the main use case for exchange of prescription content in the eHealth Digital Service Infrastructure is created and analyzed. The novelty in this approach is the proposed encoding of the basic dataset in a Quick Response (QR) code in terms of an XML scheme that is independent of clinical models or proprietary database structures. It allows to inverse the dataflow control in the chain of message exchanges between Dispenser and National Contact Points. The proposed inversion of control positions the citizen with the QR code of the prescription in the center of that chain of message exchanges between the main actors of the business process. The independent format of content representation in the QR code allows the actors in the message exchange to auto-populate data in their registers when the medicine is dispensed. Initial results are reported and reveal the advantages of embedding prescription details in QR code employing a common independent XML scheme.

Cross-Border Exchange of Clinical Data Using Archetype Concepts Compatible with the International Patient Summary.

This paper proposes an approach and demonstrates its application for cross-border exchange of clinical documents oriented towards the use of archetype concepts and international patient summary standards adopted in the European Union. A novelty in this approach is the management of native XML instances of an archetype concept in the CEN 13606 standard by means of a native XML database and XML technologies. The computer experiments demonstrate that it is suitable for representing relatively small clinical datasets such as those describing rare diseases like the Acromegaly illness, where the semantic context in the relatively small number of symptoms is practicable to tag in terms of SNOMED-CT terminology codes. Additionally, we demonstrate that the semantically enriched information model can facilitate secondary use of clinical data by visualizing the execution of queries based on standard terminologies. Finally, the compatibility of the information model with the IPS standard enables sharing of clinical data among different information models.