[Patient Level Prediction Workgroup] Meetings

Researchers
1

Dear All,

The Agenda for the bi-weekly meeting tomorrow April 6th:

  1. Revisit the selection of outcomes and prediction problems. Any
    updates from the team? see also this post (link)

  2. Presentation of Narges of a prediction problem on a
    general cohort of newly clinically diagnosed diabetes patients.
    https://www.ispor.org/awards/20Meet/prediction_posterID_PDB5.pdf

  3. Predicting outcomes for new users of celecoxib proposal study (link)

  4. Funding opportunities. Any suggestions from the team?

Meeting details are on the wiki: link

2

Team,

The Agenda for the bi-weekly meeting tomorrow April 20th:

  1. How to run the “Predicting outcomes for new users of celecoxib study” in R

  2. Discusssion on best practices wiki page, e.g.,
    a. How to divide in train test sets
    b. How many folds for cross-validation? Is this dependent on the data size?
    c. How to implement inclusion/exclusion correctly?
    d. How to test sensitivity of model performance on user choices like cohort definition and exclusion criteria
    e. Your input…

  3. AOB

Meeting details are on the wiki: link

3

Team,

In the TC today we will continue our discussion on Best Practices for PLP that we started in the last TC.

Meeting details are on the wiki.

4

Team,

In the TC tomorrow Hamed (@abedtash) will give a presentation on PMML:

“The Potential of PMML for Sharing OMOP-based Predictive Models”

If time allows, we will continue our PLP best practices discussion.

Meeting details are on the wiki page

Looking forward to this.

Peter

5

Team,

In the TC tomorrow Jenna will give a presentation about the Prognosis Research Strategy (PROGRESS) group and the Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD statement)

links:

http://progress-partnership.org

http://www.equator-network.org/reporting-guidelines/tripod-statement/

6

The recording and powerpoint are on the wiki page as always.

Minutes will be created later.

Thanks all for the contribution today.

7

Dear all,

In the TC today Jenna will continue her presentation from last meeting about PROGRESS and the TRIPOD statement.

We hope the round up the best-practices discussion today. I have prepared an abstract for the OHDSI symposium during my flight to US to describe the work done so far on this. I will add any remaining points discussed today and will distribute it for your feedback.

Thanks

8

Dear all,

The agenda of today’s TC:

  1. First results of the “Predicting outcomes for new users of
    celecoxib” study!
  2. Learning Curves
  3. Temporal Modeling in EHR Data
  4. AOB
9

The slides and the recording are available on the wiki page

The minutes will follow later.

Thanks all for the nice TC!

10

Team,

Unfortunately, we have to cancel the PLP meeting tomorrow due to other obligations.

Peter and Jenna

11

Team,

Tomorrow we have the following items on the agenda:

  • Update on the current status of the PLP package
  • Some-by-Some study proposal for OHDSI meeting
  • A.O.B

Hope to talk to you all tomorrow.

Peter

link is on the wiki page as always.

12

Dear all,

We have decided to have one meeting per month in the summer period.

The next meeting will be on Aug 24th.

Peter and Jenna

13

Dear All,

The meeting of Aug 24th is cancelled.

Peter

14

Dear All,

Hope you all had a great Holiday!

In the PLP meeting today we have to following on the agenda:

  • High-level overview of the Patient-Level Prediction presentation at the OHDSI Symposium
  • Sharing of some of the initial results of the Depression experiment

Looking forward to the session today!

Peter

15

Dear All,

Because of the face-to-face OHDSI meeting this Friday we will cancel the meeting of today.
Many team members are either travelling or under high workload at the moment and cannot join.

We look forward to see you all in Washington.

Peter and Jenna

16

Dear All,

Since I am in Korea today and the TC will be in the middle of the night, Jenna will lead a discussion on possible research directions for the PLP workgroup.

Hope you will all join this.

Peter

17

Team,

Tomorrow Kenney Ng will talk about his group at IBM Research and will present an overview of some of the work related to predictive modelling.

Hope you will all join to learn more about the interesting work that is done at IBM.

Peter

1 Like
18

Team,

Tomorrow Kenney Ng will finish his interesting talk about the PLP work done by IBM Research and Mandev Gill will present his work on Sparse Coding.

If you have missed the first part of the presentation of Kenney you can find it on the wiki page

Note also that the PLP Call will be at 12:00 EST which means it will be an hour earlier in Europe (17:00 CET)!

Talk to you tomorrow.

Peter

19

Dear All,

Tomorrow we have the PLP WG TC on the agenda. Unfortunately, I am not available at that time due to a meeting in which I have to present on Patient-Level Prediction.

If there is someone in the group that likes to present their work or like to discuss research opportunities tomorrow let us know and Jenna can start the TC, otherwise we cancel the meeting at the end of today.

Peter

20

Dear All,

Tomorrow in our bi-weekly TC we like to have an open discussion with the team on future research directions for the PLP workgroup.

I hope we can have an interactive brainstorm with you all.

Peter and Jenna.