YouEDC Biostat

Cost estimate by tariffs

The healthcare resource use the sites record, the tariff table of the study and what the recorded care would cost by it, per patient and per patient-year.

Health economics needs to know what the care of the patients would cost. Cost estimate takes the healthcare resource use the sites record in the study (hospital stays, day hospital, visits, procedures and tests, emergency calls) and prices it by the tariff table the study uploads: the tariffs of the compulsory medical insurance (ОМС) of the region and year, or of another DRG system. It is an estimate by tariffs, not the actual costs, and every page and file says so.

Resource use Kind of careHospital stay Admission – discharge12.03 – 15.03 Tariff codeST-D01 3 bed-days Tariff table ST-D0128 000 × 1.12 AMB-D011 450 PR-D02640 Region, 2026, RUBuploaded in Excel/CSV Estimate Arm A 21 300 Placebo 35 900 per patient-year, ₽ An estimate by the tariffs of the compulsory medical insurance, not the actual costs

Recording the resource use

Add the form Healthcare resource use from the form library (category Health economics) to an unscheduled log, or use a form of your own with the same variables (CDASH domain HO). One record is one hospital stay, or one kind of visit, procedure or emergency call with their number:

  • HOCAT, the kind of care: HOSP (a round-the-clock stay), DAYHOSP, OUTPT, PROC or EMERG; HOSTDAT, the date (of admission); HOENDAT, the date of discharge;
  • HONUM, how many (empty counts as 1); HOCODE, the tariff code: the КСГ (DRG) code of a stay, the service code of a visit or procedure.

Bed-days of a stay are discharge minus admission (at least 1); of a day hospital, the days from admission to discharge inclusive. A stay without a discharge date has no bed-days yet.

The tariff table

The portal ships no real tariffs: they are regional and change every year. In Cost estimate, Tariff table lets the project administrator, a data manager or the statistician upload the table in Excel or CSV. The columns may be named in Russian or English: code, kind (КСГ or DRG, visit, procedure, emergency call), tariff, or base rate and coefficient (the price is their product); name, region, year, valid from and valid to are optional. The page offers a sample file. A file with an error saves nothing and lists the lines to fix. A new table replaces the one in use; the old one stays in the history, and every upload goes to the audit trail with the file's SHA-256.

How the estimate is counted

  • A record costs the price of the tariff row of its kind and code valid on its date, times its number. A record without a code, with a code not in the table or not valid on that date costs nothing and is listed under Not priced with the reason, to correct the form or the table.
  • Per patient: mean (SD) and median, patients without any record counting as zero. Per patient-year: the sum of the group divided by its years of follow-up, from enrolment to leaving the study or to today.
  • By arm for readers who may see the arms; in a blinded study the others see all patients together. Patients who withdrew the use of their data are left out. Only numbers are shown, never patient codes.
  • Every number has a stable key, for example cost.cost_total.A.per_py: the same key is on the page and in the Excel file, which goes to the audit trail.

In the demo

The diabetes and hypertension demos have the resource-use form and an invented tariff table marked as such: visits to the specialist about every three months with blood tests, a stay in hospital for each serious adverse event and a few others (fewer on the active arm), now and then a day hospital or an emergency call. A few visits lack the code or carry an unknown one, so the list of records not priced shows how it works.