Background Evidence based health care planning is gaining importance in rheumatology due to the introduction of effective but expensive biological therapies. SLE is among the conditions where knowledge on epidemiological situation may be vital in order to achieve the just distribution of the available treatment resources. Considering the remarkable variance in SLE prevalence it may be misleading to extrapolate the estimations from other populations, and conduction of original studies is preferred. Limitation of the resources in small countries like Estonia can be seen as an obstacle for launching large scale prevalence studies. Calculation of epidemiological estimates based solely on centrally collected statistical data may misjudge the real situation.
Objectives To estimate the period prevalence of SLE in Estonia based on the Estonian Health Insurance Fund (EHIF) database. Health insurance in Estonia is funded through a compulsory scheme; the insurance coverage of population by the EHIF is higher than 95%.
Methods The data was extracted from the EHIF database 2006-2010. Information on all treatment billing episodes for the condition coded M32 in subjects older than 20 years was sought. The requested variables included: billing date, treated person's identification number, sex, birth date, the health care unit, specialty of the care provider. Using the identification number, the data was transformed to person-based form. The persons were grouped by the characteristics of billed treatment activities (number of episodes with M32 coding, specialty of the provider). The upper and lower limits of 5-year period prevalence estimates were calculated using number of patients with different characteristics of treatment activities and period's average size of Estonia's adult population.
Results Totally, 9342 billing episodes with code M32 belonging to 678 persons were identified. The persons were divided to following groups: code M32 recorded by general practitioner only (group N 127; women 71.7%; mean age 56.2 years); by any other specialist but rheumatologist (57; 86.0; 52.4); by rheumatologist 1 or 2 times (143; 91.0; 50.8); by rheumatologist 3 times or more (351; 92.3; 50.0) during the study period.
The total number of identified persons was used for calculation of the upper, and the number of persons treated by rheumatologist more than 3 times was used for the lower limit of prevalence estimation. It was estimated that SLE period prevalence 2006-2010 lay between 34 and 67 per 100 000.
Conclusions The limits found for SLE prevalence from EHIF database accord well with the published estimates from neighboring countries. While the upper limit was calculated using the number of all persons with M32 code in the database, the lower limit accounts solely for individuals for whom having erroneous M32 coding is very unlikely. For more accurate estimation of SLE prevalence, the next step will be verification of SLE diagnosis in the other defined groups using records of the health care providing units. The estimation will be validated against the EHIF database 2011-2013.
Disclosure of Interest None declared