4.7.2. Issues specific to Estonia¶
The sample in Estonia is restricted to the users of health care, and therefore excludes healthy people; people with undiagnosed chronic conditions; and patients with diagnosed conditions that did not use any healthcare (e.g. not even visiting a doctor for a brief consultation). Below is excerpt from an Estonian report [Thi{\'{e}}baut, 2017 [58]].
‘In Estonia, 95% of the population…is covered by a mandatory public insurance while private insurance is almost non-existent. The public health insurance is financed through a solidarity-based mandatory contributions in the form of an earmarked social payroll tax collected by the Estonian Health Insurance Fund (EHIF), an independent public institution. EHIF is the main purchaser of health care (HC) in Estonia… The dataset is an extract from EHIF discharge database, it contains all reimbursement costs for primary care, specialists, hospital stays and rehabilitation claimed to EHIF by each Estonian people who have used the HC system during the year 2013.’
After removing missing cases and people under 18 years old, 817 522 individuals were available for analysis.
EHIF reimbursement dataset was combined with the Estonia Census to make it more representative of the whole population. The predicted costs \(E(C|C>0, disease=1)\) and \(E(C|C>0, disease=0)\) were representatives of the users of healthcare. To make them representative of the whole population, such costs were multiplied by \(P(C>0)\) using logit parameters estimated on the combined EHIF and Census sample.
Primary care in Estonia is financed by a capitation system: family doctors are practicing on the basis of a list of enrolled patients for whom they receive an annual capitation payment from EHIF. As a result, reimbursement costs for primary care reported into EHIF claims file are very small, and often equal to zero (i.e.: GP’s visit are “free” for every Estonian person enrolled). A high proportion of patients in EHIF sample had zero reimbursed costs but were nevertheless users of healthcare. In such cases, individual cost has been corrected by computing an implicit cost of GP’s visit valued at 9.14 Euros, that was added to each bill for family medicine visits. The same procedure was followed for injury data from Estonia.
The claim file does not contain individual medication consumption. However EHIF accounts provide annual prescribed medicine cost compensated by EHIF: 112 793 thousands Euros in 2013. Knowing total EHIF claimed reimbursement from our dataset is equal to 490 740 thousands Euros, we compute a ratio of prescribed medicine as share of EHIF reimbursed cost : 112 793 / 490 740 = 0.23, or pharmaceuticals cost = 0.23 times observed reimbursed cost. We apply this factor to each bill amount to get individual prescribed medicine cost.
Table 3 in a report by [Thi{\'{e}}baut, 2017 [58]] presents distribution of OOP expenditure highlighting the main cost components, namely: prescribed and OTC medicines, dental care, specialists outpatient, LTC, glasses, and outpatient rehabilitation. OOP for prescribed medicines was differentiated by disease group according to table 11 (in the above cited report): previously estimated prescribed medicine cost is multiplied by OOP/EHIF ratio according to NCD identified as main diagnosis for visit/stay using diagnosis sequence variable. Other OOP components were not included (OTC, LTC, glasses and dental care).
Since the time since diagnosis is not defined in the EHIF dataset, we assume that the average extra disease costs apply at all years since diagnosis. Also, because of the data limitations, the disease groups are different in Estonia compared to France. Specifically, the marginal treatment costs were estimated for the following disease categories:
AMI (Acute Myocardial Infarction)
Alcohol use
Cancers (combined)
Strokes (combined)
Cirrhosis
Chronic Kidney disease
Chornic Obstructive Pulmonary Disease
Depression
Diabetes
Neurological disorders
Residual costs.