Dear all,

We just had the Skype call today with Nino, Maia and colleagues and discussed the first results of the data we were provided last week.

I would like thank you (and Triin joins me) for the work you have done to derive the data from your database and prepare them for first check.

The general picture looks good. Data includes 108808 unique rows. 

Nevertheless, there are some things which need to be fixed before we can proceed with the grouping exercise.

In current stage we focus only for data needed for groping, and only after that we start to deal with different tariffs and cost related data.

Here are the main findings followed by actions to be taken. The findings follow the order of data-fields needed for grouping:

1. Case ID (column A).
   1. CaseDistplayNumber is used. It’s a unique number and can be used later for linking additional data if needed.
      No action needed
2. Main diagnosis (column AA)
   
   1. in 248 cases the main diagnosis is missing
      Action: to find the main diagnosis for those cases.
   2. in 28 cases instead of diagnosis code “hemotransfusion” is used.
      Action: to find correct main diagnosis code for those cases and use NCSP code for hemotransfusion (if there exist one in GEO version)
3. Secondary diagnosis (columns AB-AJ)
   1. in some cases instead of diagnosis code “hemotransfusion” and “haemodialysis” is used.
      Action: to find correct secondary diagnosis code for those cases (if coded) and use NCSP codes for those procedures (if there exist one in GEO version)
4. Procedure codes
   1. There are 16 columns for procedure codes
      1. NCSP_ძირითადი_2- NCSP_ძირითადი_8 (#7)
      2. NCSP_თანმხლები_1- NCSP_თანმხლები_9 (#9)
         Action: The names of the columns need to be clarified. In addition, it should be checked whether or not some columns are duplicating each other. If so, the duplicates need to be removed.
5. Gender (column C)
   1. Information is there but it needs formatting
      Action: If male -> use 1, if female -> use 2, if unknown –> use 3
6. Age (column D)
   1. AgeInDays, the age ranges from 0 to 35272 days (0-99 years) and should be ok
      No action needed
7. Discharge
   1. For this the HospitalizationOutcome (column N) can be used. This information should be mapped to four different characters needed for grouping:
      1. 1=any other including normal discharge to home or unknown
      2. 2=transferred to another hospital
      3. 3=left against medical advice
      4. 4=dead
         Action: map every HospitalizationOutcome characteristic (9 different in the table) to one of the above (1-4). Add respective column "Discharge" in the table. Keep column "HospitalizationOutcome".
8. Length of stay (LOS)
   1. Information is correct (minus one case).
      Action: Check the case 3310497435, were LOS is -5540.

1. Birthweight (column S)
   1. 1170 cases with birthweight information in the data.
   2. Birthweight ranges from 300g to 7000g? Is 300 and 7000g correct?
   3. In 25 cases the birthweight information seems not to be relevant because the age of those cases ranges from 18 to 39 years
      Action: Check the age of those patients. If incorrect, correct the age. If correct, there is no need for birthweight information.

There is one more comment regarding the OrganizationType (column Y). There are 62 cases where the organization type is Family Medicine Center. Most of the cases (#49) are related to cataract issues, some of them have procedure codes reflecting the cataract surgery. Are those cases relevant to Family Medicine Centre? Can there be the coding issues or do the ophthalmologist work in Family Medicine Centres and provide respective services?

We agreed that the corrected data will be provided by Friday 22^nd Sept, if possible. If not, then by Monday 25^th Sept, as latest.

The next steps (given that the data are ok):

• Final preparation for grouping (week 39).
  • general check
  • replacing the NCSP+ code with Estonian and Finnish NCSP codes
  • replacing the birthweight information with respective ICD10 codes which are currently missing in ICD10 version used in Georgia
• Grouping the data with NordDRG grouper (EST/FIN) (week 39).
• Providing the preliminary results of the grouping to working group (via Skype, preliminary meeting agreed on 2^nd Oct)
• Providing feedback to SSA for additional check of certain cases with suspicious coding quality (50-100 cases), based on the grouping results (week 40).
• Agree on next steps regarding cost analyses by using current SSA and providers’ tariffs (tbc).

Thank you once again and looking forward to getting the corrected data so that we can proceed with action our plan.

Regards,

Kristiina