Drug Utilization Visualization experiments
Need to write this. Hope to have a demo up soon.
Need help with this
- in cohort definition we might say we want all exposures to x with a 180-day washout period or, alternatively, all first exposures to x with a 180-day washout period. but we seldom have a clear sense of how different those two cohorts will be. here we can see, for instance, by setting era maxgap to 180, in synpuf5%, half of patients with ANTIHYPERTENSIVES exposure go on to have another exposure after a 180-day gap and we can see the distribution of the actual gaps before the next exposure, and can play with different washout periods (well, I'm not including obs period yet), to (very quickly) see how it affects cohort selection
Need to write this
Things to work on
- stagger overlapping exposures
- load more patients
- provide means to select patients by useful attributes
- put patients in datatable, improve formatting
- show collapsed view of patient with just timeline
cumulative distribution function histogram
- hover tooltip
- hover affects sample patients
- add KDE, maybe do brushing on the KDE
- Carsten's idea: break up gaps by exposure duration blocks
- full height on hover
- legend text for charts
- hover over exposure duration
- tooltip for all recs with same duration
- filter gaps (and maybe all other distributions) to records tied to hover region
- gap duration
- brush gap region to select subsets of patients and show sample patients accordingly
- fix y axis
- make height consistent
- fix choppy gap distributions (or are those meaningful?)
- more distribution charts for overall drug characteristics:
single exposure days (called this All Exposures Together) single gap days (with overlap)
- total exposure days per patient
- total gap days (with overlap) per patient
total exposure days per era gap days between eras
- gap, overlap days within eras
- connect to other data
- overall patient timelines for individual patients
- conditions, concomitant medications, measurements, demographics, etc.
- data composition
- get help validating
- what to do with not enough recs for ntiles?
- next projects:
- treatment pathways
- lifeflow/event flow
- vocab/population navigation
- treatment pathways
This work is funded by The University of Colorado, Health Data Compass under the direction of Michael Kahn, technical guidance and support from Michael Ames and Hajar Homayouni, with subject matter expertise and guidance from Marsha Raebel, Christina Aquilante, and Katy Trinkley.
I'd like to express my great gratitude to Michael Kahn for giving me the chance to work on this project and the creative freedom to experiment and build what we hope will be genuinely innovative and useful tools for health data research.
Contact me (sigfried at sigfried dot org) with any technical questions.
Copyright 2016 Sigfried Gold, Licensed under the Apache License, Version 2.0. You may not use this file except in compliance with the License.
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.