---
title: ""
output: word_document
---
# Variables definition
- ‘DATA_LINK’ column: connects stats and bias tables
- 'review name' column: connects basic wiht the rest
- QUADAS BIAS dataset
- ‘RESULTS’: exclude incomplete items you could filter the spreadsheet
- STATS DATA/SUBGROUPS
- 'SUBGROUP':and it can be NONE, or COVARIATE, and maybe there are othere values. And sometimes there is the TYPE variable just underneath the text I highlighted in the picture, and it can list the IDs of multiple tests. I am not sure at all what any of this means, but maybe it is of interest to extract?
# TO DO list: LENA
- Quadas2 integrated: 7 domains
- Bias results:
- High/Low/Unclear
- Yes/No/Unclear.
- Done: ~~To know if each subgroup produce an estimation in each MA~~
- Which type of subgroup is this? Patient or test group (both)?
- Done: ~~Connection (or clear instruction for the merger ID) of the 3 datasets. The output could be the original 3 datasets and an additional one that unifies them.~~
- Produce two scenario for all the output: with Q2 (only the studies that have Q2) and without it (the whole datasets). Also could be one output with missing data for the not available information.
- Finally, do you see a way to recognize the n more distinct sub-groups in each MA? (You can take more time for this point). Some ideas: does this subgroup have similar studies between of them?
# TO DO list: Philippe
- Lena inclusion as co-authorship: done.
- Re-fit the models with HAS_RESULTS and QUADAS2.
GAM1#: $$ g(CMRI_{score})=\sqrt{CMRI_{score}+11}= \beta_0 + \beta_1 x_i+ \Sigma_{k=1}^K u_k z_k(x_i) + \epsilon_i + \text{Parametric predictors}$$
GAM1#2: $$ g(CMRI_{score})=\sqrt{CMRI_{score}+11}= \beta_0+\beta_1 x_i+ \Sigma_{k=1}^K u_k z_k(x_i) + \epsilon_i + \text{Parametric predictors}$$