Comparison of Data from the Health Information System and the Health Insurance Fund in the Period of 2012-2019 by Patients, Diagnoses, Types of Treatment and Health Service Providers
Name
Kristiina Miller
Abstract
Examining insurance claims and electronic health records is critical to improving the efficiency, quality, and effectiveness of health care delivery, while helping audits to simplify the work of health care professionals and prevent fraud. It is known that health care providers send fewer electronic health records to the Health Information System than they send insurance claims to the Health Insurance Fund, but the current researchers have looked into this issue so far based on aggregated health data. In this work, health data between 2012-2019 from the Health Insurance Fund's insurance claims and electronic health records from Health Information System are used to compare two data sources by patients, by diagnoses, types of treatment and service providers. Since those are two separate data sets, personalized pseudonymized data were used and transferred to the OMOP CDM data model. The purpose of the work, apart from comparing the data, is to find how much of the data overlaps between two data sources, whether the goal of one-time data entry and reuse is fulfilled, how big is the problem of data continuity in Estonian health data and how much has the obligation of health care providers to send data in the health information system improved. As a result of the work, it was found that even if the healthcare service provider has an obligation to send health documents to the Health Information System, this obligation was not fully fulfilled by the end of 2019. A large part of the data was available in the data of the Health Insurance Fund, but not in the Health Information System. In addition, it was found that for inpatient health records, the amount of health data is almost the same between the two data sources, and the difference comes in with outpatient health records. It was also found that most of the data related to diagnoses is duplicated between two data sources, which does not meet the goal of single data entry.
Graduation Thesis language
Estonian
Graduation Thesis type
Master - Data Science
Supervisor(s)
Marek Oja, Sirli Tamm
Defence year
2024