Skip to content

Changes

Page history
Second Round Harmonisation - more FAQs added, plus formatting updated authored by Andrei Morgan's avatar Andrei Morgan
Hide whitespace changes
Inline Side-by-side
Second-Round-Harmonisation.md
View page @ 54ec25bf
......@@ -24,20 +24,25 @@
### Dictionary questions
**Q:** Which version of the dictionaries should I use?
| **Question** | **Answer** |
| ---- | :--------- |
| Which version of the dictionaries should I use? | Please use the most recent version of the dictionaries, which can be downloaded above. |
| But I've already started doing the harmonisation with an older version of the dictionaries - what should I do? Do I have to start all over again? | No, do not worry! Just download the dictionary of the view where the harmonisation is being done; make the necessary changes in the dictionary file (as indicated in the changelog above); then, remove the view and create it again, this time using the new dictionary. |
### Data questions - variables
**A:** Please use the most recent version of the dictionaries, which can be downloaded above.
| **Question** | **Answer** |
| ---- | :--------- |
| Do you recommend a specific tool to calculate the requested z-scores for harmonization of birth weights, heights etc as batch? | No. The z-scores only need to be harmonised if they already exist in your data set and/or if there are country-specific norms that are in general usage (e.g. both UK and France have specific growth curves that have been created for their populations). |
| Our follow-up sweep covers a period of several years (i.e. different ages). How do we handle this given that the variables specify outcomes at specific ages? | We suggest that you assume the follow-up occurred at the *average age* of the follow-up - so for 2-3 years, you could use 2.5 years, or for 11-15 years you could use 13 years. Alternatively, you could work out the average age of the children that were seen, e.g. for 11-15 year sweep it may be that the majority of children were seen at 11 years, but a small percentage were difficult to find and only followed-up later on; in this case 11 or maybe 12 years could be a more appropriate age to use. Remember, also, that there variables (in the 'child_vital_status' dictionary) where the exact age of the child at the follow-up should be recorded, so it will be possible for people to use this information if required. |
**Q:** But I've already started doing the harmonisation with an older version of the dictionaries - what should I do? Do I have to start all over again?
**A:** No, do not worry! Just download the dictionary of the view where the harmonisation is being done; make the necessary changes in the dictionary file (as indicated in the changelog above); then, remove the view and create it again, this time using the new dictionary.
### Data questions
**Q:** Do you recommend a specific tool to calculate the requested z-scores for harmonization of birth weights, heights etc as batch?
**A:** No. The z-scores only need to be harmonised if they already exist in your data set and/or if there are country-specific norms that are in general usage (e.g. both UK and France have specific growth curves that have been created for their populations).
### Data questions - subjects
| **Question** | **Answer** |
| ---- | :--------- |
| Should we be including terminations of pregnancy (TOPs)? | We did not initially ask for TOPs to be included in the data sets loaded onto the platform, but we *have* included variables about them. **We now recommend that TOPs *are* included in the data loaded onto the platform.** This is because future studies using the RECAP Preterm platform may wish to include TOPs hence it is preferable (and easier for data managers) if the data are already loaded now. However, the most important thing, whether TOPs are included or not, is that **the dataset description should clearly indicate whether TOPs are included or not.** |