User Tools

Site Tools


timeline

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
timeline [2014/01/21 21:58]
stamang removed
timeline [2020/03/05 11:39] (current)
nigam created
Line 1: Line 1:
-====== Denmark Project Timeline ====== +{{ ::patient-timeline.png?nolink&900 |}}
-Back to the main [[denmark|Denmark project]] page +
- +
-=====Fall quarter 2012===== +
-1 - implemented the trajectory stitching using simulated cost data +
-2 - implemented 3 different matching mechanisms for trajectory stitching +
-3 - presented results at end of quarter (Dec 12th) +
- +
-=====Winter quarter 2013===== +
-1 - Engaged with Tommy (first meeting Jan 15th) +
-2 - decision to proceed with the "shopping cart" method using only the 8 years of (unstitched data) +
-3 - started quest to get actual cost data, and data dictionaries from Aarhus (Feb 5th) +
-      * requested 1-year bins on demographics file instead of current 3-year bins +
-      * started getting cost information for drugs and admissions (possibly DRG-classified costs) +
-4 - visited Denmark / Lars to actually get the data we wanted (we hoped to do it remotely, but nothing beats sitting side by side) +
- +
-=====Spring quarter 2013===== +
-1 - started on shopping cart models +
-2 - Both patient-level and population-level validation runs for a test set of 1000 patients (validation, 5/6) +
-3 - Also tried, probabilistic approach to trajectory stitching (didn't work). +
-4 - decided to focus on a certain set of diseases going forward (diabetes / CHF / CAD / COPD) +
-5 - started clustering and template matching as other alternative methods (Ken Jung's work) +
-6 - focused all methods on one goal: find "bifurcations" in the 8 cost trajectories. +
- +
-As of May 15th, we have narrowed the scope to: +
-1.  simply identifying patients with some positive control conditions .. where we believe that bifurcations should exist. +
-2.  finding patients of interest (with respect to their cost trajectory) +
-3.  given 2, determining to first order (i.e., without trying to find intervention points) what drives higher costs in patients with breakpoints in their trajectories +
- +
-=====Summer quarter 2013=====   +
-1.    Characterized cost patterns and concluded that outside of end of life, high costs are concentrated in acute episodes rather than chronic elevated outpatient costs.   +
-2.    Identified high cost subset of patients (defined as >= 90th percentile of total expense) and high cost episodes (defined as >= 90th percentile of annual costs).   +
-3.    Ran association rule mining (apriori) to identify simple combinations of codes that predict high costs (either definition).  +
- +
- +
-//Stanford-Aarhus meeting in late August 2013//+
timeline.txt · Last modified: 2020/03/05 11:39 by nigam