If it's not in my outlook calendar, it doesn't exist.

Thanks for the explanation!

news.ki.se/digital-twin...

There has been a suggestion of a shared genetic risk between some cancers and sarcoidosis. This might explain the slightly increased risk of cancer associated with sarcoidosis that persists for years.

We need a better evidence base for choosing treatments. If someone has a history of cancer, what treatment for sarcoidosis is best? Do sarcoidosis treatments increase cancer risk? We show that sarcoidosis, regardless of treatment is associated with cancer. How should people be screened?

How to predict prolonged pregnancy outcomes, to identify who should be induced at week 41? When you introduce a new intervention, planning ahead and making time/space and having funding should be considered

We still need more data to build good prediction models - they are still not good enough (e.g. predicting vaginal birth after a first birth via cesarean )

In 2017 results from a randomized trial showed aspirin reduced preeclampsia risk in high risk pregnancies. First preventative medicine for PE
Now women in Sweden at high rusk are receiving aspirin during pregnancy www.nejm.org/doi/full/10....

A lot of effort has been put into predicting preeclampsia - using all the information we already collect, we can improve prediction.

Next: NIPT test - a blood sample which detects cell-free fetal DNA if KUB shows high risk. Now only 1% of pregnancies have amniocentesis.

We need to be more rigorous - need to test and validate in multiple cohorts. How do we take what we learn in academia into clinical care?

He gave really good examples of identifying preclinical biomarkers in IBD. Signatures for treatment response did not pan out. Future: get longitudinal samples to capture dynamics. The heterogeneity of patients is a challenge.

www.sciencedirect.com/science/arti...

An example: using unsupervised clustering to identify novel subgroups of diabetes. However the clusters identified were not clinically useful.