Abstract

Background: The prevalence of chronic kidney disease (CKD) in Estonia is unknown. Since the average prevalence of CKD in the
world is 9.1%, the number of CKD patients in Estonia should be approximately 118,300. However, according to the 2017 global report, the number of CKD patients in Estonia is approximately 258,859, which is probably overestimated. Estonian e-health is the foundation for registration and collection of healthcare data. At the same time, data analysis is often problematic, because the output documents are in different formats, which complicates greatly their analysis. In addition, relevant diagnosis codes and/or severity category of CKD are often not specified in medical records.
Aim: In order to find out the prevalence and management of CKD patients in Estonia, the aim of this study was to retrospectively
investigate the number of adult CKD patients and their distribution according to the CKD risk profile, analyzing the data of the e health information system and for the first time using the aid of artificial intelligence.
Methods: The study was based on a combined dataset from the databases of health insurance claims and drug prescriptions, and from electronic health records in the Health Information System (including specialist care, primary care, purchased drugs, laboratory measurement data, hereinafter e-health data) for a random sample of 10% of the Estonian population. The study population was defined as all patients at least 18 years of age who had been diagnosed with at least one disease that is a risk factor for CKD between 2016 and 2019 and/or had at least one recorded estimate of glomerular filtration rate (eGFR) and/or albumin-creatinine ratio (U-Alb/UCrea) value during 2019. The number of medicines purchased by the patients, as well as the number hospitalisations and/or emergency care cases were evaluated. The aid of artificial intelligence was used to analyse the output documents in different formats, where information in the epicrisis text file was transformed into the form used in the analysis.
Results: Based on the e-health data, we identified 5% of the population with an existing CKD diagnosis and in addition 2.4%
potential G3-G5 CKD patients. According to these data, there may be a total of 83,710 patients with chronic kidney disease in Estonia. Thus, according to e-health data, the prevalence of CKD in the adult population is 7.4%. The eGFR values are available in for 52% of at-risk patients, while U-Alb/U-Crea values are available for only 12%. During multi-morbidity assessment, we found that hypertension (79%), cardiovascular disease (CVD) (63%) and diabetes (28%) occur most often in patients diagnosed with CKD. Almost half of the patients diagnosed with CKD have been hospitalised or referred to the emergency department within one year. The main reason for this has been CVD (11%). Patients with the diagnosis of CKD differ from patients in the CKD risk group (diabetes, hypertension, CVD) in terms of higher rate of hospitalisation and the need for emergency care.
Summary: According to the e-health data, the prevalence of CKD in the adult population in Estonia is 7.4%. Despite the existence
of a national CKD treatment guideline and a CKD screening system for diabetes and hypertension, there are gaps in patient testing, kidney disease risk assessment, and evidence-based patient management. In order to characterize the risk of CVD and to ensure appropriate treatment of patients in the CKD risk groups, it is necessary to follow the treatment guidelines regarding measurement of the U-Alb/U-Crea value and specification of severity of CKD involving also the albuminuria category. The disease
burden of CKD is high and requires close cooperation between primary health care and specialists. Therefore, prevention and treatment of CKD requires even more attention and resources.